Catalytic Amyloids: Is Misfolding Folding?

Abstract

Originally regarded as a disease symptom, amyloids have shown a rich diversity of functions, including biologically beneficial ones. As such, the traditional view of polypeptide aggregation into amyloid-like structures being “misfolding” should rather be viewed as “alternative folding”. Various amyloid folds have been recently used to create highly efficient catalysts with specific catalytic efficiencies rivalling those of enzymes. Here we summarize recent developments and applications of catalytic amyloids, derived from both de novo and bioinspired designs, and discuss how progress in the last two years reflects on the field as a whole.

1.1. Introduction.

Traditional thermodynamic diagrams depict the process of protein folding in a funnel-like shape, with the protein minimizing its free energy by folding into its native (deemed to have a particular function) state. More recently, a potential for many proteins to form various distinct low energy aggregated assemblies (often β-sheet rich)[1] has been recognized and the field of intrinsically disordered proteins (i.e., those not having a distinct unique 3-D structure) has grown tremendously.[2] Moreover, advances in structural characterization of amyloid assemblies show high complexity of the folding landscape and the corresponding diversity of stable structures even when formed by the same polypeptide sequence.[3] Originally studied as a symptom of diseases, amyloids have shown a rich diversity of functions beyond aggregation and may have beneficial roles within biological systems, such as when curli amyloids fibrils are produced which act as a protective case for bacterial communities during biofilm formation.[4,5] Taking all this into consideration, the traditional view of polypeptide aggregation in amyloid-like structures being “misfolding” can no longer stand. Instead, we should think of it in terms of “alternative folding” very much in line with multiple conformations adopted by proteins (sometimes irreversibly) in response to external stimuli. Indeed, reproducible production of well-defined 3-D structures with function is the very hallmark of protein folding and the recent discovery of enzymatic activities exhibited by amyloid-like assemblies further supports this notion.[6–8] Here we will review the rapid progress in the design and discovery of catalytic amyloids highlighting approaches, challenges, and possible future directions.

Despite quite a few spectacular examples of highly efficient catalytic amyloids reported, their design approaches are still in their infancy and are largely based on the principles developed in the late 1980’s.[6,7,9,10] All of the examples reported to date can be broadly put into two categories that we define as de novo designed and bioinspired (Fig. 1a,b). In the first case the peptide sequence is rationally designed based on the basic mechanistic requirements for the reaction to be catalyzed and a minimalist self-assembling scaffold. The advantages of this approach include the versatility of the reactions to be catalyzed, a large number of possible structures produced and the flexibility of the peptide sequence space. The very advantages of the de novo design can be its disadvantages: since only an approximate idea of final fold geometry is available a priori, a larger number of possibilities needs to be experimentally tested to identify the optimal sequence, that often is quite different from the predicted one. The bioinspired approach works in the opposite direction. The design is based on the sequence and 3-D structural information available for naturally occurring amyloids and then modifications are made to create an active site capable of catalysis. This approach allows for quite accurate modelling of the possibilities achieved at the expense of sequence flexibility. Naturally, these two different approaches are better suited for the different types of reaction catalyzed (Fig. 1), so we will summarize the progress in the field over the past two years by both the type of catalysis and the design approach.

Figure 1.

Approaches to catalytic amyloid design can be broadly defined as 1) de novo (hydrophobic residues are shown in green and polar residues shown in blue) or 2) bioinspired (the core of Aβ42 is highlighted in blue). Both design strategies have led to an ever increasing number of examples of the chemical reactions catalyzed, detailed as follows a) pNPA hydrolysis, described within references [6,7,11–17] b) CO₂ hydration described within reference [14] c) Paraoxon hydrolysis [18] d) dimethoxyphenol oxidation [19] e) The cascade formation of 2’,2’-dichlorofluoroscein from the sequential hydrolysis and oxidation of 2’,7’-dichloroflurescin acetate [18] f) oxidation of 2,4-dichlorophenol and subsequent trapping of the quinone with 4-aminoantipyrine [20] g) electrochemical pyrrole polymerisation [17] h) hydrolysis of a phosphoester unit from A, G, C or TTP [21,22] g) cyclopropane formation from styrenes in the presence of diazo compounds [23] j) 2-methoxyphenol oxidation [24,25] k) ABTS oxidation [26] l) TMB oxidation [26,27] m) hydrolysis of 4-aminophenyl alanine ester n) catalysis of the methodol retroaldol reaction [28] o) hydrolysis of 4-nitrophenyl 4-oxopentanoate [29] p) hydrolysis of 4-nitrophenyl pentanoate [29]q) hydrolysis of napthalene-2-yl-acetate [29] r) hydrolysis of 6-formyl napthalene-2-yl acetate [29] s) the aldol reaction between 2-acetonapthone and 6-methoxy-2-napthaldehyde yielding 3-hydroxy-3-(6-methoxynaphthalen-2-yl)-1-(naphthalen-2-yl)propan-1- one [30] t) the aldol reaction between cyclohexanone and p-nitrobenzaldehyde yielding 2-[(4-nitrophenyl)hydroxymethyl]cyclohexanone [31] u) imide condensation of 6-amino-2-napthaldehyde [28] v) hydrolysis of 2-methoxy phenyl acetate [24] w) oPD oxidation [26]. The image of nanotube assembly was reproduced from Sarkhel et. al. [29] with permission, copyright American Chemical Society 2019.

1.2. Hydrolytic reactions promoted by catalytic amyloids.

Hydrolytic reactions were the first reported to be catalyzed by catalytic amyloids.[6] Initial efforts in functional amyloid design focused on para-nitrophenyl acetate (pNPA), a common and well-benchmarked substrate used to assay many different types of enzymes and small molecule catalysts.[6,7,13] Despite the seeming simplicity, creating highly efficient catalysts for hydrolyzing pNPA is surprisingly difficult.[32] Given the low molecular weights of the building blocks in the self-assembling systems specific catalytic efficiencies of catalytic amyloids can be remarkably high. Recently, Gazit and co-workers took the minimalist approach to de novo designing catalytic amyloid-like systems to its limit. They showed that a single amino acid residue of phenylalanine can assemble in the presence of zinc into amyloid-like structures that is capable of hydrolyzing pNPA (k_cat/K_M 77 M⁻¹ s⁻¹) and promoting carbon dioxide hydration (k_cat/K_M 962 M⁻¹ s⁻¹).[14] The specific ability of Zn-Phe₂ (k_cat/K_M = 46 × 10⁻² l⁻¹g⁻¹s⁻¹) to promote pNPA hydrolysis is higher than that of carbonic anhydrase (k_cat/K_M = 5.7 × 10⁻² l⁻¹g⁻¹s⁻¹), an inspiration for many designs.

While amyloid-like structures formed by a single amino acid residue are impressive, they fundamentally lack the functional diversity offered by longer peptides. Given their reduced sequence complexity and extreme stability, catalytic amyloids have high tolerance for introduced mutations and offer additional possibilities for functional diversity in forming heteromeric mixtures with improved properties.[15,16] Recently, we demonstrated that synergistic interactions were observed in all heteromeric amyloid assemblies formed by the leucine core 7-residue peptides studied in our previous work (Fig. 2).[11] Peptide sequences of the family LHLHLXL (where X is varied) can be mixed to create heteromeric fibrils with increased catalytic efficiency. Complementary charge mixing (i.e. R and E) as well as R-Q mixtures show an increase in catalytic efficiency greater than the sum of both fractions of individual peptide. This approach allows for easily screening of a number of different arrangements of functional residues by simply mixing different peptides to create more complex sites within the bulk fibril with increased catalytic activity. These systems currently lack high resolution characterization of the amyloid surface, such as solid-state NMR or X-ray crystallography. Key to improving these designs is the ability to define what changes these mixed systems undergo to increase their catalytic efficiency. In their characterizations of the self-assembly of mixtures of CATCH(+) and CATCH(−) peptides into β-sheets, Shao et. al. found that within the same fibril they observed A-B pairing, as would be expected, but also significant mismatched CATCH(+)-CATCH(+) and CATCH(−)-CATCH(−) pairs also.[33] This behavior was also observed within King-Webb peptide fibrils [34]. This supports the idea that co-assembled structures should be treated as statistical mixtures, rather than perfect A-B assemblies. This level of detailed characterization of co-assembled amyloid systems is key to exploring further co-assembled designs.

Figure 2.

Mixing different peptides provides opportunities for rapid and facile identification of productive arrangements of functional groups that can be further improved upon by creating longer and more complex macrocyclic or linked peptide sequences. Rather than interrupting amyloid formation, a variety of linkers were tolerated by the amyloid sequences but had separate effects upon their hydrolysis efficiencies. Figure reproduced from refs. [11] and [12], copyright Wiley 2020.

Interestingly, we observed that mixing catalytic amyloid-forming peptides with their respective enantiomers led to formation of amyloids completely devoid of catalytic properties. Such necessity of homochirality for efficient catalysis is interesting in the context of the Amyloid World hypothesis, stating that amyloids were the first species capable of self-replication, information transmission, and evolution.[35,36] The loss of catalysis in enantiomeric mixtures suggests a possible cause for the single-handedness of life. This is also consistent with other observations where propensity of peptides to form assemblies did not always correlate with their ability to promote catalysis. Importantly, the productive arrangements of functional groups identified in the mixing experiments can also be used to build more complex assemblies that do not have the sequence space restrictions imposed by homomeric assembly.[12]

Introduction of an optimized linker (PSGSGSP) yields amyloid sequences which assemble more rapidly while preserving their increased catalytic efficiency. However, linking two starnds with ornithine residues to create self-assembling macrocyclic peptides yields clear structure-activity relationships at the expense of catalytic activity. Additional tuning of functional properties can be achieved by controlling the nature and arrangement (parallel vs anti-parallel) of the linker(s). In addition to improving the properties of the assemblies, this approach helps in elucidating the mechanistic aspects of the catalytic amyloids’ reactivity, an especially important feature given difficulties in structural characterization of heteromeric peptide assemblies.

Hydrolytic abilities of catalytic amyloids are not limited to model substrates. We have shown that a catalytic amyloid formed by Ac-IHIHIYI-NH₂ in the presence of copper is capable of hydrolyzing paraoxon, an effective pesticide with high toxicity and a low background rate of hydrolysis (7.97 × 10⁻⁷ min⁻¹ at pH 7.8) with a catalytic efficiency of 1.7 M⁻¹ min⁻¹ (pH 8).[18] This level of catalytic efficiency is on par with those of catalytic antibodies and small molecule catalysts, highlighting the ability of a simple 7-residue amyloid to catalyze difficult reactions.

Diaz-Espinoza and co-workers have developed manganese containing amyloids with ATPase-like activity. Their initial sequence (Ac-NADFDGDQMAVHV-NH₂) was inspired by the active site of RNA polymerase.[21] The manganese dependent activity of the catalytic amyloids was low, (k_cat/K_M = 3.3 × 10⁻⁵ M⁻¹ s⁻¹) but highly specific. Subsequently, taking inspiration from a DNA polymerase sequence, they designed a new manganese dependent amyloid (Ac-SDIDVFI-NH₂) which displayed activity for the hydrolysis of the ATP, GTP, CTP and UTP, with clear substrate preference for GTP over the other nucleotides.[22] In this work, changing the design of the active site (utilizing carboxylates rather than histidines) allowed the use of a previously unexplored cofactor, which in turn resulted in new catalytic activity.

1.3. Cofactor dependent redox catalysis.

Given the ability of catalytic amyloids to productively bind several different cofactors, much attention has been recently devoted to their ability to promote a diverse set of redox transformations. In a first example of redox-mediated transformation promoted by catalytic amyloids, we reported the ability of the fibrils formed by Ac-IHIHIQI-NH₂ to productively bind copper (II) ions to efficiently promote a redox-mediated coupling of dimethoxyphenol using dioxygen as an oxidant.[19] This bimolecular reaction results in a formation of a new carbon-carbon bond and a product that is more complex than the starting materials, a feat uncommon in model reactions. In addition to their redox properties, the Ac-IHIHIYI-NH₂/Cu^II fibrils can also promote hydrolysis. Such behavior opens a path for the catalytic amyloids to promote cascade reactions in which consecutive reactions occur only as a result of the preceding reaction.[18] The catalytic amyloids catalyze the cascade formation of 2’,7’-dichlorofluorescein (DCF) from 2’,7’-dichlorofluorescin acetate (DCFH-DA), a pathway which requires hydrolysis of the acetate to form 2’,7’-dichlorofluorescin (DCFH) with a subsequent oxidation step to form the final product DCF. The overall rate of the reaction is higher than that of the slower of the two transformations measured alone. This unique feature of the catalytic amyloids is due to their ability to promote both transformations at the same reaction site without the need for the substrate to diffuse in and out.

As mentioned above, catalytic amyloids can be formed by very short peptide sequences and recently Gazit and co-workers showed that amyloid-like structures can be formed by complexing Cu (II) with phenylalanine. These stable, recyclable supramolecular assemblies formed by [Cu^II-Phe₂] repeats have laccase-like activity that exceeds the activity of the native laccase in the model reaction of 2,4-dichlorophenol oxidation, where the resulting quinone is trapped with 4-aminoantipyrine; they can also promote oxidation of several common neurotransmitters.[20]

The ability of catalytic amyloids to productively associate with metallocofactors is not limited to simple metal ions. By modifying the sequences of the Ac-LHLHLXL-NH₂ family designed to bind zinc to accommodate bulkier, more hydrophobic cofactors,[6] we developed an efficient hemin binding motif of the representative sequence Ac-LILHLFL-NH₂.[23] Ac-LILHLFL-NH₂ self-assembles to form catalytic amyloid fibrils that promote efficient cyclopropanation using ethyl diazoacetate and a variety of aromatic styrenes. Remarkably, enantiomeric excesses of up to 40% e.e. were observed in these reactions, running contrary to the idea that catalytic amyloids are flat and featureless assemblies. This finding is in line with commonly observed β-sheet twists that create grooves on fibril surfaces,[37,38] which in turn leads to more complex arrangements of functional groups reminiscent to those found in the tertiary structures of proteins.

Given the versatility of heme-promoted catalysis it is not surprising that catalytic amyloids facilitate other types of reactions,[39] including peroxidation, a well-established hallmark for heme containing enzymes.[40] We have recently shown that catalytic amyloids formed by Ac-LMLHLFL-NH₂ bind hemin and catalyze the oxidation of a wide range of substrates with different oxidation modes. In the case of ABTS and hydrogen peroxide, the catalytic efficiency of the catalytic amyloids reaches 3 × 10⁵ M⁻¹ s⁻¹ at pH 6.[26] This level of activity is higher than those of all other peptide assemblies previously reported and is within an order of magnitude of the best examples of de novo designed peroxidases.[41]

Using a bioinspired approach Das and co-workers introduced a histidine into the nucleating core of Aβ−42. The resulting peptide, Ac-HLVFFAL-NH₂, assembles into amyloid nanotubes with exposed histidines capable of binding heme and promoting oxidation of 2-methoxy phenol peroxidation using H₂O₂ as the oxidant. This catalyst was also capable of promoting tandem hydrolysis/oxidation when using 2-methoxyphenyl acetate as the substrate.[24]

Catalytic amyloids can also be used in conjunction with naturally occurring enzymes to harness their reactivity and take part in extended cascade reactions.[25] Das and workers found that the binding capabilities of the short nanotube forming sequence Im-KLVFFAL allowed them to load sarcosine oxidase (SOX) and hemin to catalyze multi-step reactions. SOX oxidizes sarcosine, which generates H₂O₂ utilized in guaiacol oxidation catalyzed by the hemin present.

1.4. Cofactor free catalysis.

While metal cofactors allow for great catalytic versatility, the functional groups provided by amino acids alone and tunable environment created in self-assembling supramolecular structures allows for cofactor-free catalysis by catalytic amyloids. Lynn and co-workers modified the nucleating core of the well-known amyloid Aβ−42 into Ac-KLVFFAL-NH₂, yielding a design which assembles at neutral pH to form nanotubes, with the N-terminal lysine residues positioned above the cross-β groove within the nanotubes.[28] This results in the ability of the nanotube to perform multiple types of catalysis. It promotes condensation of 6-amino-2-napthaldehyde into a dimeric colored product, as well as retroaldol catalysis. Interestingly, even subtle changes to the sequence affect the activity of the assembly, e.g., mutations of the terminal leucine cause the nanotube to show no catalytic activity.

Das and co-workers reported an amyloid nanotube, Im-KLVFFAL-NH₂, again derived from the Aβ−42.[29] The terminal lysine and imidazole residues work cooperatively upon assembly through the formation of imine bonds between the lysine and the substrate. This catalytic amyloid hydrolyzes a range of carboxyester substrates, both aromatic and aliphatic, and is reliant on the concerted action of both terminal residues. Replacement of lysine with arginine, incapable of forming imine bonds, showed greatly decreased catalysis. Turning their attention to aldolase reactions, Das and co-workers truncated their design to the central VFF and rendered the system by amphiphilic through the introduction of a C-terminal lysine and N-terminal decanoic acid, which displayed time dependent assembly behavior.[30] Initial assemblies (7–15 days) were fibrillar, but after aging (30 days), formed homogenous nanotubes capable of promoting retroaldol reactions showing preference for methodol over other substrates studied. These fibrils also promote the direct aldol reaction, although with significant substrate selectivity: only 2-acetonapthone and 6-methoxy-2-napthaldehyde were found to form the corresponding aldol product. These reactions can be successfully combined in a cascade fashion, as mixing methodol, acetone, 2-hexanone and 6-methoxy-2-napthaldehyde yields the aldol product of methodol and 6-methoxy-2-napthaldehyde.

Most oxidation catalysts reported rely on a metal cofactor to carry out the oxidation of interest. However, recently catalytic amyloids have been reported to carry out oxidations in the absence of metallocofactors. Liu et. al. reported the cofactor free oxidation of TMB promoted by assemblies formed by histidine oligomer H15, an exciting result which demonstrates how oligomers of a single amino acid can form highly ordered nanostructures capable of catalyzing redox reactions without the use of metal cofactors.[27] Ultrasonic treatment of the crystalline assemblies increased the activity, attributed to an increase in surface area, and conjugation of the oligohistidine with the fibrillating sequence QQKFQFQFEQQ led to fibril formation, and an increase in k_cat. Yet no catalyst turnover was reported (product/catalyst ratio of less than 20%) in the H15 system. We observed TMB, ABTS and oPD oxidation promoted by the heptapeptide Ac-VHVHVQV-NH₂ in the absence of any cofactors.[26] The oxidation of oPD was observed across a range of pH values, with the highest catalytic efficiency of 0.45 M⁻¹ s⁻¹ observed at pH 6.5 with at least two catalyst turnovers observed. While the mechanism of cofactor free oxidation promoted by catalytic amyloids has yet to be elucidated, it is feasible that the selective binding of the substrates by amyloids with a subsequent change in redox potential plays a role in it. While heme containing designs offer far greater efficiencies for redox applications, these results also support a major evolutionary role for catalytic amyloids as early peptide assemblies were not likely to rely on complex cofactors like heme for their redox activity.[36]

1.5. Practical applications of catalytic amyloids.

Simple repeating motifs with minimal residue diversity can be used to mimic natural protein sequences and their activity with a reduction in complexity, and so be used as a starting point for new designs. Rather than considering them model systems as a stepping-stone to more complex proteins, in recent years it has become increasingly clear that short amyloid designs are an effective method to design de novo catalysts in their own right. This has become increasingly evident in exploration of their practical applications. Catalytic amyloids are extremely robust, and their properties are desirable for a variety of nanotechnology purposes.[15,42–44] As a result of this, they have the potential to be immobilized and applied in situations where they can be subject to multiple reaction cycles or utilized under conditions which might permanently denature globular proteins.

Winter and co-workers have explored the behavior of catalytic amyloids under extreme pressures. These high stress environments are prevalent around deep-sea vents proposed to be sites where life on Earth may have originated.[45] FTIR (Fourier-transform infrared) spectra of the representative catalytic amyloids (Ac-LHLHLRL-NH₂ and Ac-IHIHIQI-NH₂) showed no changes in secondary structure even when the pressure reached 2000 atm. Moreover, their catalytic efficiencies in pNPA hydrolysis showed a marked increase of almost one order of magnitude reaching k_cat/K_M of 450 M⁻¹ s⁻¹.[15] The high concentration of osmolytes within deep sea organisms led Winter and co-workers to study the effects of cosolvents and crowding agents on the catalytic efficiency of Ac-IHIHIQI-NH₂.[16] No deleterious effects on catalytic efficiency were observed in the present of cosolvent or in a crowded system, further highlighting the practical robustness and versatility of catalytic amyloids.

The stability and the heterogeneous nature of catalytic amyloids makes them easily compatible with flow chemistry. They can be easily applied to commonly available syringe filters directly from the buffer they were prepared in, with no further processing required, to make hydrolytic or redox-active flow catalysts. [18],[26]

Catalytic amyloids are also compatible with electrochemically driven reactions. Ventura and co-workers reported the ability of copper coordinating catalytic amyloids comprised of an alternating histidine-tyrosine sequence (HY8 or HY9) to catalyze the oxidative polymerization on polypyrrole.[17] Polypyrrole can be synthesized electrochemically, but the reaction is very slow at potentials lower than 0.8 V. The HY peptides promoted near complete polymerization of pyrrole within 10 minutes in the presence of copper when potential was applied.

1.6. Conclusion.

The development of self-assembling catalysts for a variety of different chemical reactions is a rapidly growing field.[42,46,47] While a number of different approaches using various building blocks have been tried, [31,48–51] self-assembling peptides that form catalytic amyloids have been shown to particularly well suited for the task. In less than a decade, there is a rapidly growing number of highly active catalytic amyloids reported developed using some very basic principles (Fig. 3). In several cases, their specific activities are higher than those of enzymes. The resulting materials are extremely robust and can be used in a variety of nanotechnology applications, as their properties (high activity, ease of synthesis, stability and promiscuity) lend them to such applications as biocompatible materials and catalysts.

Figure 3.

Structurally characterized functional amyloid systems A) The solid-state NMR structure of Ac-IHVHLQI-NH₂, an amyloid catalyst for pNPA hydrolysis, reproduced with permission from ref. [52], copyright National Academy of Sciences 2017. The hydrated Zn within the structure can react with pNPA when substrate diffuses to sites in the fibril. B) the reported crystal structure of hydrolase Phe-Zn-Phe, reproduced with permission from ref. [14], copyright Springer Nature 2019. C) Computational models of hydrolytic amyloids Ac-IHIHIXI-NH₂ (X= Y, Q, R) with the structures of X= Q and R validated with solid-state NMR structures, reproduced from ref [53] with permission, copyright American Chemical Society 2019. These are proposed to have a similar reaction mechanism to those in ref. [52]. D) X-ray crystal structure of Phe-Cu-Phe [20], reproduced under Creative Commons CC BY 4.0 license. This laccase-type catalysis displayed by the species stems from a hydrogen atom transfer from the phenolic substrate to a deprotonated carboxylate within the F moiety, with the proton then transferred to an oxygen molecule. The resulting phenolic radical then diffuses from the active site, and forms polymeric species with radicals from other active sites in the reaction mixture. E) A structural model of peroxidase-like H16 nanosheets, derived from X-ray diffraction data and transmission electron microscopy, reproduced with permission from ref. [27], copyright Springer Nature 2020. Non-covalent interactions lead to TMB adsorption followed by diffusion of H₂O₂ generates the active catalytic species, after which proton abstraction and addition of a further equivalent of TMB results in water generation and TMB^.+ release.

The “rules” to effective catalytic amyloid design are still being discovered, rendering the creation of highly active amyloid species more of an art than a science. It is a very active, young field within which it is difficult to establish complex rules as of yet. Additional research into these systems is needed, and the next challenge in the field as it grows is to establish the principles that guide the design. Amyloids have been shown to be highly active hydrolysis catalysts, with a burgeoning ability to catalyze bond forming reactions. Efforts to make the active sites of fibrils more complex would yield greater control over the reactions catalyzed, but this is difficult without detailed structural information. Similar to the breakthroughs in protein science driven by structural information in the 1990’s we need fast and accurate tools to establish the structure-morphology-activity relationships to enable further development. Currently, no single technique can offer sufficient level of structural detail for catalytic amyloids comprised of more than few residues, and long and laborious in-depth studies that combine solid state NMR, microscopy and theoretical experiments are required, although available characterization suggests that structural complexity can arise within even simple systems (Figure 3).[14,52,53] We anticipate that development of cryo-EM tools will allow for fast and accurate structure prediction that will greatly facilitate future designs that will undoubtedly become more complex and functional. Given the ease of designing very active catalysts, it is extremely likely that more efficient catalytic amyloids will be developed and catalytic activities that rival those of enzymes can finally be achieved.

Highlights.

• Catalytic amyloids promote a variety of different chemical transformations with high efficiency.

• Current approaches of catalytic amyloid design can be classified as de novo and bioinspired.

• Catalytic amyloids can rely on cofactors or be cofactorless.

• Traditional view of amyloid formation as misfolding is an oversimplification.