What is biological information?

Biological information is a well-defined concept. It denotes the specific amino acid sequence of a protein that corresponds to a specific base sequence in DNA. This concept can be represented simply as:

Biological information = amino acid sequence = DNA sequence x the genetic code, where 3 specific bases = 1 specific amino acid.

This simple definition of biological information is the core principle at the foundation of modern molecular biology since Francis Crick introduced it in his seminal paper “On Protein Synthesis” published in 1958¹. In a subsection appropriately titled "The essence of the problem”, Crick lays out the definition of biological information: “By information I mean the specification of the amino acid sequence of the protein”. He called this the Sequence Hypothesis which states “that the specificity of a piece of nucleic acid is expressed solely by the sequence of its bases, and that this sequence is a (simple) code for the amino acid sequence of a particular protein.” He added another hypothesis and called it the Central Dogma, which states that “once information has passed into protein it cannot get out again. In more detail, the transfer of information from nucleic acid to nucleic acid, or from nucleic acid to protein may be possible, but transfer from protein to protein, or from protein to nucleic acid is impossible.”

In later years both of his hypotheses were shown to be correct. In the 1960s, the genetic code was deciphered by the work of Marshall Nirenberg and his coworkers, and subsequently, the exquisite and elaborate components of transcription and translation were identified revealing a complex machinery that involves RNA polymerases, mRNAs, tRNAs, and ribosomes. Despite its complexity, the system was shown to work in accordance with Crick’s simple arrow of information transfer: from nucleic acid to nucleic acid (DNA --> RNA, transcription, or, RNA --> DNA, reverse transcription), or from nucleic acid to protein (translation), but never from protein to nucleic acid. The reason for the impossibility of what would have been called “reverse translation” is, as Crick himself pointed out in 1970 in a paper titled Central Dogma of Molecular Biology: “The transfer protein --> RNA (and the analogous protein --> DNA) would have required (back) translation, that is, the transfer from one alphabet to a structurally quite different one. It was realized that forward translation involved very complex machinery. Moreover, it seemed unlikely on general grounds that this machinery could easily work backwards. The only reasonable alternative was that the cell had evolved an entirely separate set of complicated machinery for back translation, and of this there was no trace, and no reason to believe that it might be needed.” This remains true to this day.

Within this elegant edifice of molecular biology, a crucial component seems to have been missing; where is protein conformation? How does the linear sequence information of amino acids become a 3D functional protein? In Crick’s 1958 prescient paper, the answer to that question was to assume that “folding is simply a function of the order of the amino acids". This was the theoretical choice he opted for at the time because, similar to the problem of reverse or back translation, if protein conformation was not simply a function of the amino acid sequence, a separate and parallel system of conformational information would have had to be accounted for. A whole system of protein conformational templates for all the possible proteins in all the possible conformations would have had to be faithfully and independently replicated in parallel with DNA replication and transferred during mitotic and miotic cell divisions according to Mendelian rules. How could the structure of the chromosome possibly carry templates for all the possible protein 3D structures a cell might need in the future? How would such a system be stored and maintained to protect and transmit such information with fidelity? And if such a system of templates did exist, why would the cells bother replicating DNA in the first place and have dedicated complex machinery for replication, transcription, and translation? Obviously, the absence of such a system in addition to its insurmountable theoretical and practical challenges made Crick’s assumption that protein conformation is a function of the sequence of the amino acids the logical choice. In time, Crick’s assumption was experimentally verified by the work of Christian Anfinsen, who showed that enzymes such as ribonuclease fold spontaneously in the test tube after denaturation and regain their enzymatic activity in the absence of any protein template or machinery. This was later called the Thermodynamic Hypothesis of Protein Folding or Anfinsen’s dogma, which states that all the information required by a protein to adopt its final conformation is present in its amino acid sequence and that this process “is driven entirely by the free energy of conformation that is gained in going to the stable, native structure”². In other words, it is a spontaneous process driven by the laws of thermodynamics acting on the primary sequence of the protein.

Anfinsen’s dogma is a pillar of molecular biology, without it the entire structure of genetics built over decades from Mendel to this day falls apart. Without it, we would be required to postulate a parallel system of information transfer to explain 3D protein structure; a system that does not exist and would have been impossible to integrate with everything else we know about genetics. The thermodynamic hypothesis of protein folding complements the fact that the only form of biological information is sequence information, and that “conformational information” does not need accounting for as it is simply a function of the sequence information already coded for with high fidelity and dedicated machinery. Thus, adopting a certain protein conformation does not involve a new layer of information that requires separate code, machinery, or template.

All this seemed to apply universally except for one phenomenon, protein misfolding into amyloids in diseases such as Creutzfeldt Jakob disease (CJD), Alzheimer’s disease, or Parkinson’s disease. As an exception to the rest of biology and genetics, this conformation is postulated to require a conformational template, or a prion, which can move from organism to organism or within a tissue and act as a template to “imprint” its corrupt conformation on similar proteins. The prion hypothesis introduced a new and parallel system of biological information: conformational information transfer through templating. It is explicitly compared to DNA where “just as DNA mediates inheritance by templating its own sequence, these proteins act as genes by templating their conformation”³. This is an extraordinary claim. In science, extraordinary claims require extraordinary evidence. However, in this case, there is an extraordinary amount of evidence against, and not for, the prion hypothesis.

The prion hypothesis in its entirety is based on one experimental phenomenon: seeding. This is when an amyloid fibril fragment, a seed, is added to a concentrated solution of other proteins, they also become amyloids. For decades, however, what exactly is the prion conformation that is being templated and what is the mechanism of templating remained unclear. In most papers, it is usually represented diagrammatically as squares and circles (or squares and triangles), where the square protein with the bad conformation binds to the circle protein with the normal conformation and makes it turn into a square (Figure 1A).  More recently, and thanks to a better understanding of the structure of amyloid fibrils, the term “conformational information” now denotes the specific 2D cross-sectional shape (the specific pattern of folds and turns) of the amyloid protofilament or fibril (Figure 1B). An amyloid fibril usually contains two or more protofilaments. In this framework, a prion, which is an amyloid fibril fragment, templates its 2D cross-sectional shape on incoming protein molecules binding to its tip during the process of elongation, where the incoming molecules must accommodate the 2D cross-sectional shape of the fibril by binding in a parallel, in-register manner (i.e., the N and C termini align in the same direction, and each amino acid of the incoming molecule stack on top of the identical residue at the tip of the fibril). This is also the underlying principle behind the concept of prion strains, where different 2D cross-sectional seed shapes are postulated to imprint their distinctive shape on incoming protein molecules, leading to different disease phenotypes^4–6. Similarly, prion propagation takes place via breakage or fission of a fibril with a particular 2D cross-sectional shape into smaller fragments or seeds, which then template that 2D shape on incoming protein molecules. Thus, in its modern formulation, the prion templating and propagation of conformational information is based on elongation and fission, a mechanism that has been compared to the Watson-Crick base pairing, even using similar language to the famous concluding line of Watson and Crick’s 1953 DNA paper to describe it: “It did not escape our notice that the in-register beta sheet architecture can explain inheritance of prion variants”⁷.

Figure 1. A. Diagrammatic representation of prion conformational templating. B. The modern formulation of the prion hypothesis proposes that different 2D cross-sectional shapes of protofilaments or fibrils represent different prion strains.

However, for the elongation and fission mechanism to preserve and propagate specific 2D cross-sectional information sustainably across multiple generations of fibrils and across different cells, tissues, and hosts, it must fulfill some basic criteria:

1. Incoming proteins must only bind to the fibril tip to accommodate its specific cross-sectional shape. 

2. Incoming proteins must have the same sequence as the protein in the fibril tip to enable parallel in-register stacking.

However, there is an overwhelming amount of robust experimental evidence that shows that both criteria are not fulfilled during amyloid growth:

1. Amyloid growth cannot maintain elongation even for one generation of fibrils because the process is immediately taken over by branching, where new fibrils grow as branches on the surface of the parent fibril (Video 1.); a process termed secondary nucleation^8–18. The fibril surface carries no resemblance to the 2D cross-sectional shape of the fibril (the conformational information) and cannot engage in parrel, in-register protein binding.

2. Seeds of one protein can induce amyloid aggregation of another protein without the sequence homology needed for parallel-in register elongation, a process termed cross-seeding^19–29. In this case, heterologous seeds act mainly as catalytic surfaces that do not relay any conformational information.

Based on these two points alone, the prion hypothesis as it pertains to sustained conformational templating & propagation via elongation and fission contradicts experimental evidence. Furthermore, these two points are among many other fundamental problems with the hypothesis including:

3. There is no energetic driver of a protein molecule to exit its thermodynamically stable native conformation to mold itself on top of a fibril. Also, seeds cannot just go around in solution fishing for a protein to mold it into their shape.

4. No machinery has ever been found that could restrict amyloid growth to tip elongation and prevent secondary nucleation on the surface to preserve the 2D template information for multiple generations of fibrils.

5. While parallel in-register is the most common β-sheet stacking architecture in amyloids, amyloids with anti-parallel and out-of-register architecture have been experimentally found³⁰.

6. In a process termed heterogeneous nucleation, amyloid growth can be initiated by any surface including lipid membranes^31–35, polysaccharides^36,37, nucleic acids³⁸, nanoparticles^39–41, and viruses ⁴², which lack any templating information.

7. In a process termed homogenous nucleation, amyloids form spontaneously at high protein concentrations in the absence of any templating information⁴³.

8. The final shape of the protofilament or fibril depends on environmental conditions and not on the 2D cross-sectional shape of the seed ^44–51.  

Video 1. Compiled, 2X sped supplementary videos from the paper titled: Branching in Amyloid Fibril Growth, published by Christian B. Andersen et al., Biophys J. 2009 Feb 18; 96(4): 1529–1536.

Taken together, the prion hypothesis is simply wrong, and it can’t be saved. Built entirely on one phenomenon, seeding, it completely mischaracterized the process, mislabeling a simple phenomenon of protein phase transition as replication. Seeding is only one pathway of amyloid formation, where templating cannot be preserved or maintained. Furthermore, proteins can form amyloids spontaneously without any template, and amyloid formation can be catalyzed by non-proteinaceous surfaces. In all cases, what drives amyloid formation, similar to any other phase transition (crystallization for example), is the increased probability of forming intermolecular bonds at higher protein concentrations (supersaturation). In this case, no template is needed at all, just a nucleation trigger, which can be any surface. 

In fact, biological information is lost during the process of amyloid formation. All proteins, irrespective of their sequence, adopt the same cross-β architecture when they become amyloids, where layers and layers of protein monomer units align on top of each other in long, interdigitating, β-sheet ladder pairs. The specific sequence information that was preserved in DNA for generations and elegantly replicated, transcribed, and translated to a specific amino acid sequence via highly sophisticated machinery is lost to such a generic architecture. Specific domains that have evolved to do specific functions by adopting intricate sequence-dependent conformations are lost to aimless, sequence-independent intermolecular β-sheet ladders, where such domains are mostly consumed in self-interactions and sequestered out of function in insoluble, inert precipitates.  

In conclusion, sequence information is the only form of biological information and no parallel system for conformational information exists. Protein conformation for both native and amyloid folding is simply a function of thermodynamics acting on protein chemistry. If this was not the case, and a protein required another protein template to adopt a certain conformation, the entire structure of molecular biology would crumble. Anfinsen’s dogma saved molecular biology from becoming a problem of infinite turtles (where did the first template come from?), or an impossible storage problem, where cells had to carry templates of all possible protein conformations all the time. Conformational templating according to the prion hypothesis cannot be maintained even for one generation of fibrils and is neither necessary nor sufficient to explain the experimental results of amyloid formation, which is a simple process of phase transition.

The famous economist John Maynard Keynes once said about the power of theories as frameworks of thinking: “The ideas of economists and political philosophers, both when they are right and when they are wrong, are more powerful than is commonly understood. Indeed, the world is ruled by little else.” This applies to scientific theories too. The prion hypothesis is the wrong framework of thinking about the amyloid phenomenon. It is not incomplete, it is not inaccurate, it is just flat wrong. Its core assumptions about sustained conformational information templating contradict endless lines of experimental evidence. There are no two ways about it. Unfortunately, the prion hypothesis has been incredibly powerful in framing research and therapeutic development for neurodegenerative diseases for over four decades leading the field to focus almost exclusively on one mechanism of amyloid formation (seeding) and one mechanism of toxicity (gain-of-function), ignoring other equally if not more important pathways of amyloid formation and pathogenesis. The result is zero therapies. 

Real progress can only happen when the fatal flaws of such an influential theory are acknowledged and replaced by the proper physical understanding of the phenomenon as a spontaneous folding event under the conditions of supersaturation and phase transition. Prions (and their cousin oligomers) are indeed toxic to the brain, not as real entities, but as flawed concepts that prevent us from properly understanding the phenomenon in order to develop suitable interventions. We must cure our intellectual framework first to be able to cure these diseases. 

References

1. Crick F. H. On Protein Synthesis. Symp Soc Exp Biol 12, 138–163 (1958).
2. Anfinsen, C. B. Principles that govern the folding of protein chains. Science (1979) 181, 223–230 (1973).
3. Wickner, R. B. et al. Amyloid diseases of yeast: Prions are proteins acting as genes. Essays Biochem 56, 193–205 (2014).
4. Collinge, J. Mammalian prions and their wider relevance in neurodegenerative diseases. Nature 539, 217–226 (2016).
5. Manka, S. W. et al. A structural basis for prion strain diversity. Nat Chem Biol (2022) doi:10.1038/s41589-022-01229-7.
6. Scheres, S. H., Zhang, W., Falcon, B. & Goedert, M. Cryo-EM structures of tau filaments. Curr Opin Struct Biol 64, 17–25 (2020).
7. Reed Brendon Wickner, M.D. | Principal Investigators | NIH Intramural Research Program. https://irp.nih.gov/pi/reed-wickner.
8. Ferrone, F. A., Hofrichter, J. & Eaton, W. A. Kinetics of sickle hemoglobin polymerization. II. A double nucleation mechanism. J Mol Biol 183, 611–631 (1985).
9. Andersen, C. B. et al. Branching in amyloid fibril growth. Biophys J 96, 1529–1536 (2009).
10. Mukhopadhyay, S., Bera, S. C. & Ramola, K. Observation of two-step aggregation kinetics of amyloid-β proteins from fractal analysis. Phys Biol 19, 046001 (2022).
11. Törnquist, M. et al. Secondary nucleation in amyloid formation. Chemical Communications 54, 8667–8684 (2018).
12. Cohen, S. I. A. et al. Proliferation of amyloid-β42 aggregates occurs through a secondary nucleation mechanism. Proc Natl Acad Sci U S A 110, 9758–63 (2013).
13. Thacker, D. et al. Direct observation of secondary nucleation along the fibril surface of the amyloid 42 peptide. 120, 1–9 (2023).
14. Gaspar, R. et al. Secondary nucleation of monomers on fibril surface dominates α-synuclein aggregation and provides autocatalytic amyloid amplification. Q Rev Biophys 50, (2017).
15. Padrick, S. B. & Miranker, A. D. Islet amyloid: Phase partitioning and secondary nucleation are central to the mechanism of fibrillogenesis. Biochemistry 41, 4694–4703 (2002).
16. Ruschak, A. M. & Miranker, A. D. Fiber-dependent amyloid formation as catalysis of an existing reaction pathway. Proc Natl Acad Sci U S A 104, 12341–12346 (2007).
17. Foderà, V., Librizzi, F., Groenning, M., Van De Weert, M. & Leone, M. Secondary Nucleation and Accessible Surface in Insulin Amyloid Fibril Formation. Journal of Physical Chemistry B 112, 3853–3858 (2008).
18. Gaspar, R. et al. Secondary nucleation of monomers on fibril surface dominates α-synuclein aggregation and provides autocatalytic amyloid amplification. Q Rev Biophys 50, e6 (2017).
19. Subedi, S., Sasidharan, S., Nag, N., Saudagar, P. & Tripathi, T. Amyloid Cross-Seeding: Mechanism, Implication, and Inhibition. (2022).
20. Motifs, A. S. et al. Partial Prion Cross-Seeding between Fungal and Mammalian. 12, 1–23 (2021).
21. Polido, S. A. et al. Cross-seeding by prion protein inactivates TDP-43. Brain (2023) doi:10.1093/brain/awad289.
22. Lucas, M. J. et al. Cross-Seeding Controls Aβ Fibril Populations and Resulting Functions. J Phys Chem B (2022) doi:10.1021/acs.jpcb.1c09995.
23. Ren, B. et al. Fundamentals and Introductory of Cross-seeding of Amyloid Proteins. J Mater Chem B (2019) doi:10.1039/c9tb01871a.
24. Koloteva-Levine, N. et al. Amyloid particles facilitate surface-catalyzed cross-seeding by acting as promiscuous nanoparticles. Proc Natl Acad Sci U S A 118, 1–12 (2021).
25. Vaneyck, J., Segers-Nolten, I., Broersen, K. & Claessens, M. M. A. E. Cross-seeding of alpha-synuclein aggregation by amyloid fibrils of food proteins. Journal of Biological Chemistry 296, 100358 (2021).
26. Nirwal, S., Bharathi, V. & Patel, B. K. Amyloid-like aggregation of bovine serum albumin at physiological temperature induced by cross-seeding effect of HEWL amyloid aggregates. Biophys Chem 278, 106678 (2021).
27. Anand, B. G., Prajapati, K. P. & Kar, K. Aβ 1-40 mediated aggregation of proteins and metabolites unveils the relevance of amyloid cross-seeding in amyloidogenesis. Biochem Biophys Res Commun 1–7 (2018) doi:10.1016/j.bbrc.2018.04.198.
28. Hartman, K. et al. Bacterial curli protein promotes the conversion of PAP248-286 into the amyloid SEVI: cross-seeding of dissimilar amyloid sequences. PeerJ 1, e5 (2013).
29. Morales, R., Moreno-Gonzalez, I. & Soto, C. Cross-Seeding of Misfolded Proteins: Implications for Etiology and Pathogenesis of Protein Misfolding Diseases. PLoS Pathog 9, 1–4 (2013).
30. Tycko, R. & Wickner, R. B. Molecular structures of amyloid and prion fibrils: Consensus versus controversy. Acc Chem Res 46, 1487–1496 (2013).
31. Knight, J. D. & Miranker, A. D. Phospholipid Catalysis of Diabetic Amyloid Assembly. J Mol Biol 341, 1175–1187 (2004).
32. Terakawa, M. S., Yagi, H., Adachi, M., Lee, Y. H. & Goto, Y. Small liposomes accelerate the fibrillation of amyloid beta (1- 40). Journal of Biological Chemistry 290, 815–826 (2015).
33. Terakawa, M. S. et al. Impact of membrane curvature on amyloid aggregation. Biochimica et Biophysica Acta (BBA) - Biomembranes 0–1 (2018) doi:10.1016/j.bbamem.2018.04.012.
34. Habchi, J. et al. Cholesterol catalyses Aβ42 aggregation through a heterogeneous nucleation pathway in the presence of lipid membranes. Nat Chem (2018) doi:10.1038/s41557-018-0031-x.
35. Galvagnion, C. et al. Lipid vesicles trigger α-synuclein aggregation by stimulating primary nucleation. Nat Chem Biol 11, 229–234 (2015).
36. Cohlberg, J. A., Li, J., Uversky, V. N. & Fink, A. L. Heparin and other glycosaminoglycans stimulate the formation of amyloid fibrils from α-synuclein in vitro. Biochemistry 41, 1502–1511 (2002).
37. Iannuzzi, C., Irace, G. & Sirangelo, I. The effect of glycosaminoglycans (GAGs) on amyloid aggregation and toxicity. Molecules 20, 2510–2528 (2015).
38. Liu, C. & Zhang, Y. Nucleic acid-mediated protein aggregation and assembly. Adv Protein Chem Struct Biol 84, 1–40 (2011).
39. Gladytz, A., Abel, B. & Risselada, H. J. Gold-Induced Fibril Growth: The Mechanism of Surface-Facilitated Amyloid Aggregation. Angewandte Chemie International Edition 1–6 (2016) doi:10.1002/anie.201605151.
40. John, T. et al. Impact of nanoparticles on amyloid peptide and protein aggregation: A review with a focus on gold nanoparticles. Nanoscale 10, 20894–20913 (2018).
41. Linse, S. et al. Nucleation of protein fibrillation by nanoparticles. Proc Natl Acad Sci U S A 104, 8691–8696 (2007).
42. Ezzat, K. et al. The viral protein corona directs viral pathogenesis and amyloid aggregation. Nature Communications 2019 10:1 10, 1–16 (2019).
43. Srivastava, A. K. et al. β‐Amyloid Aggregation and Heterogeneous Nucleation. Protein Science pro.3674 (2019) doi:10.1002/pro.3674.
44. Goldsbury, C. S. et al. Polymorphic Fibrillar Assembly of Human Amylin. J Struct Biol 119, 17–27 (1997).
45. Meinhardt, J., Sachse, C., Hortschansky, P., Grigorieff, N. & Fändrich, M. Aβ(1-40) Fibril Polymorphism Implies Diverse Interaction Patterns in Amyloid Fibrils. J Mol Biol 386, 869–877 (2009).
46. Lutter, L., Aubrey, L. D. & Xue, W. F. On the Structural Diversity and Individuality of Polymorphic Amyloid Protein Assemblies. J Mol Biol 433, 167124 (2021).
47. Annamalai, K. et al. Polymorphism of Amyloid Fibrils In Vivo. Angewandte Chemie International Edition 55, 4822–4825 (2016).
48. Wilkinson, M. et al. Structural evolution of fibril polymorphs during amyloid assembly. Cell 186, 5798-5811.e26 (2023).
49. Peduzzo, A., Linse, S. & Buell, A. The Properties of α-Synuclein Secondary Nuclei are Dominated by the Solution Conditions Rather than the Seed Fibril Strain. 1–22 (2019) doi:10.26434/CHEMRXIV.9757778.V1.
50. Lövestam, S. et al. Seeded assembly in vitro does not replicate the structures of α-synuclein filaments from multiple system atrophy. FEBS Open Bio 11, 999–1013 (2021).
51. Li, X. et al. Subtle change of fibrillation condition leads to substantial alteration of recombinant Tau fibril structure. iScience 25, 105645 (2022).