Breaking the Chain: 'Molecular Cap' Blocks Processes That Lead to Alzheimer's, HIV - 3 articles below
"The devastating and incurable dementia known as Alzheimer's disease affects the thinking, memory, and behavior of dozens of millions of people worldwide. Although amyloid fibers and oligomers of two proteins, tau and amyloid-ß, have been identified in association with this disease, the development of diagnostics and therapeutics has proceeded to date in a near vacuum of information about their structures. Here we report the first atomic structures of small molecules bound to amyloid.......Our results indicate that peptides from structure-based designs can disrupt the fibril formation of full-length proteins, including those, such as tau protein, that lack fully ordered native structures.......The complexes presented here suggest routes for structure-based design of combinations of compounds that can bind to a spectrum of polymorphic aggregates, to be used as markers of fibers and as inhibitors of aggregation."|
ScienceDaily (June 22, 2011) - A new advance by UCLA biochemists has brought scientists one step closer to developing treatments that could delay the onset of Alzheimer's disease and prevent the sexual transmission of HIV.
The researchers report that they have designed molecular inhibitors that target specific proteins associated with Alzheimer's disease and HIV to prevent them from forming amyloid fibers, the elongated chains of interlocking proteins that play a key role in more than two dozen degenerative and often fatal diseases.
"By studying the structures of two key proteins that form amyloids, we were able to identify the small chain of amino acids responsible for amyloid fiber formation and engineer a 'molecular cap' that attaches to the end of the fibers to inhibit their growth," said research leader David Eisenberg, director of the UCLA-Department of Energy Institute of Genomics and Proteomics and a Howard Hughes Medical Institute investigator.
The study was published online June 15 in the journal Nature and will be available in an upcoming print edition.
"This research is an important first step toward the development of structure-based drugs designed against amyloid diseases," said Eisenberg, who is a UCLA professor of chemistry, biochemistry and biological chemistry and a member of the California NanoSystems Institute at UCLA. "Our results have opened up an avenue so that universities and industry can start creating therapeutics that could not have been produced 10 years ago."
Toward delaying Alzheimer's disease
Amyloid fibers are elongated, water-tight structures formed from two linked protein sheets. Proteins from each sheet contribute side chains, causing them to interlock like the teeth of a zipper, Eisenberg said.
The fibers are found not only Alzheimer's disease but in a variety of conditions, including Lou Gehrig's disease, Parkinson's disease, type II diabetes and a family of disorders related to mad cow disease, among others. In Alzheimer's and other neurodegenerative diseases, the tau protein forms amyloid fibers inside brain cells, destroying them through a mechanism that is still being investigated.
Though many serious diseases are characterized by amyloid fibers, Alzheimer's is the most prevalent, Eisenberg said. Today there are 5 million patients in the U.S. who suffer from Alzheimer's, with 500,000 new cases every year. Alzheimer's health care cost this year alone have been estimated at $178 billion, including the value of unpaid care for Alzheimer's patients provided by nearly 10 million family members and friends.
"By the year 2050, it is projected that there will be 19 million Alzheimer's patients," Eisenberg said. "The care of so many patients with this debilitating illness could be a substantial fraction of the gross domestic product of the United States."
Eisenberg and his research team found that of the entire tau protein, a small chain of just six amino acids -- abbreviated VQIVYK -- was responsible for the formation of amyloid fibers. By studying the structure of the fibers using microcrystallography, a method developed at UCLA for this research, the team was able to use the fibers as a template to design an inhibitor that could 'cap' the fiber and stop it from growing.
The results were dramatic. The introduction of the inhibitor into a tau protein solution completely prevented amyloid fiber formation, validating the idea that the structure-based design of therapeutics for amyloid diseases is a plausible option.
Despite this success, there is still a long road ahead before a viable therapeutic can be developed to combat the onset of Alzheimer's in human patients, Eisenberg said. The inhibitor, a chain of amino acids, is far too large to penetrate deep into the brain where the tau proteins form amyloid fibers.
"This research is an important step toward identifying smaller molecules that can be utilized to develop a therapeutic," Eisenberg said. "Our goal is to be able to delay the onset of Alzheimer's disease."
Preventing the transmission of HIV
Unlike the tau protein, the SEVI (semen-derived enhancer of viral infection) protein is a far more accessible target for a molecular blocker because it builds amyloid fibers in a vaginal environment, a key process in the sexual transmission of HIV, Eisenberg said.
"The presence of SEVI makes the rate of HIV infection through sexual transmission up to 100,000 times more likely," he said. "By blocking SEVI, we have a method for inhibiting the sexual transmission of HIV."
Though the tau and SEVI proteins have different structures and unrelated functions, they both form amyloid fibers with similar morphology, making it possible to design two separate inhibitors using the same process, according to Eisenberg.
The SEVI blocker proved to be equally effective in preventing fiber growth, bolstering the idea that blockers can be designed for other diseases associated with amyloid fibers as well.
"Though many tests remain, it seems we could be on the way to developing a therapeutic," Eisenberg said. "Our hope is that we could make a blocker that could be applied with a vaginal gel or spray that would help to prevent HIV infection."
The tau and SEVI protein inhibitors were designed using synthetic amino acids, similar to the standard protein building blocks of the human body. But these synthetic amino acids were flipped, as if viewed in a mirror, or had added side chains not normally found in nature. Enzymes in the human body that are programmed to break apart protein-like chains are, in principle, unable to recognize the non-natural amino acids, keeping the blockers safe to latch on to the target proteins.
This research was federally funded by the National Institutes of Health, the National Science Foundation and the U.S. Department of Energy, as well as by the Howard Hughes Medical Institute and the Joint Center for Translational Medicine.
Other co-authors of this study included UCLA postdoctoral scholars Stuart Sievers and Lin Jiang; UCLA graduate students Howard Chang and Anni Zhao; John Karanicolas, an assistant professor at the University of Kansas; Jason Stevens, an undergraduate at the University of Kansas; David Baker, a professor at the University of Washington; and professor Jan Munch and researcher Onofrio Zirafi, of the University of Ulm in Germany.
Small molecules, big job
A second research team also led by Eisenberg recently announced that it had identified four small molecules that bind to amyloid fibers, including a promising candidate called 'orange-G' that wedges into the zipper-like fiber and may be able to break it apart.
This study was published June 14 in PLoS Biology, an online journal of the Public Library of Science. (See paper below)
"These are the first small molecules visualized as they bind to amyloid-like fibers," Eisenberg said. "These small molecules are less likely to be broken up in the body and can potentially be modified to force apart amyloid fibers or serve as diagnostic tools to identify infected areas of the body."
Eisenberg and his research team found that orange-G was uniquely able to pierce the impenetrable "steric zippers" that seal the water-tight amyloid fibers of the amyloid-beta protein that is responsible for forming senile plaques in Alzheimer's disease.
"In 10 years we have gotten to the point where we are starting to understand the structural biology of amyloid fibers and how to inhibit them and how to interfere with them," Eisenberg said. "The next step is to make practical molecules that inhibit and break amyloid fibers -- that is the ultimate goal."
Co-authors on this UCLA research included Kym Faull, professor of psychiatry and biobehavioral sciences; Jorge Barrio, professor of molecular and medical pharmacology; researchers Michael Sawaya and Jie Liu; postdoctoral scholars Meytal Landau, Lin Jiang and Stuart Sievers; and graduate student Arthur Laganowsky.
Structure-based design of non-natural amino-acid inhibitors of amyloid fibril formation - pdf attached
Nature | Letter
15 June 2011
Stuart A. Sievers1*, John Karanicolas2,3*, Howard W. Chang1*, Anni Zhao1*, Lin Jiang1*, Onofrio Zirafi4, Jason T. Stevens3, Jan Mu nch4, David Baker2 & David Eisenberg1
1Departments of Biological Chemistry and Chemistry and Biochemistry, Howard Hughes Medical Institute, UCLA, Box 951970, Los Angeles, California 90095-1570, USA. 2Department of Biochemistry and Howard Hughes Medical Institute, University of Washington, Seattle, Washington 98195, USA. 3Center for Bioinformatics and Department of Molecular Biosciences, University of Kansas, 1200 Sunnyside Avenue, Lawrence, Kansas 66045-7534, USA. 4Institute of Molecular Virology, University Hospital Ulm, Meyerhofstrasse 1, 89081 Ulm, Germany.
Many globular and natively disordered proteins can convert into amyloid fibrils. These fibrils are associated with numerous pathologies1 as well as with normal cellular functions2, 3, and frequently form during protein denaturation4, 5. Inhibitors of pathological amyloid fibril formation could be useful in the development of therapeutics, provided that the inhibitors were specific enough to avoid interfering with normal processes. Here we show that computer-aided, structure-based design can yield highly specific peptide inhibitors of amyloid formation. Using known atomic structures of segments of amyloid fibrils as templates, we have designed and characterized an all-d-amino-acid inhibitor of the fibril formation of the tau protein associated with Alzheimer's disease, and a non-natural l-amino-acid inhibitor of an amyloid fibril that enhances sexual transmission of human immunodeficiency virus. Our results indicate that peptides from structure-based designs can disrupt the fibril formation of full-length proteins, including those, such as tau protein, that lack fully ordered native structures. Because the inhibiting peptides have been designed on structures of dual-ß-sheet 'steric zippers', the successful inhibition of amyloid fibril formation strengthens the hypothesis that amyloid spines contain steric zippers.
The finding that dozens of pathologies, including Alzheimer's disease, are associated with amyloid fibrils has stimulated research on fibril inhibition. One approach uses the self-associating property of proteins that form fibrils to poison fibril formation with short peptide segments6, 7, 8, 9, 10, 11. A second approach is based on screening for molecules that can disrupt fibril formation12, 13. Here we take a third approach to fibril inhibition: structure-based design of non-natural peptides targeted to block the ends of fibrils. With advanced sampling techniques and by minimizing an appropriate energy function, we identify novel candidate inhibitors computationally from a large peptide space that interact favourably with our template structure. This approach has been made possible by the determination of several dozen fibril-like atomic structures of segments from amyloid-forming proteins14, 15, 16.
These structures reveal a common motif called a steric zipper, in which a pair of ß-sheets is held together by the interdigitation of their side chains14. Using as templates the steric-zipper structures formed by segments of two pathological proteins, we have designed inhibitors that cap fibril ends. As we show, the inhibitors greatly slow the fibril formation of the parent proteins of the segments, offering a route to designed chemical interventions and supporting the hypothesis that steric zippers are the principal structural elements of these fibrils.
One of the two fibril-like steric zippers that we have chosen as a target for inhibitor design is the hexapeptide VQIVYK, residues 306-311 of the tau protein, which forms intracellular amyloid fibrils in Alzheimer's disease17. This segment has been shown to be important for fibril formation of the full-length protein and itself forms fibrils with biophysical properties similar to full-length tau fibrils15, 18, 19. Our second template for inhibitor design, identified by the '3D profile' algorithm20, 21, is the steric-zipper structure of the peptide segment GGVLVN from the amyloid fibril formed by 248PAP286, a proteolytic fragment containing residues 248-286 of prostatic acid phosphatase, a protein abundant in semen. 248PAP286 fibrils, also known as semen-derived enhancer of virus infection (SEVI), enhance human immunodeficiency virus (HIV) infection by orders of magnitude in cell culture studies, whereas the monomeric peptide is inactive22.
Our computational approach to designing non-natural peptides that inhibit fibril formation is summarized in Fig. 1 for the VQIVYK segment of tau protein; the same general strategy is used for the GGVLVN segment of 248PAP286. In both systems, we design a tight interface between the inhibiting peptide and the end of the steric zipper to block additional segments from joining the fibril. By sampling l or d amino acids, or commercially available non-natural amino acids, we can design candidate inhibitors with side chains that maximize hydrogen bonding and hydrophobic interactions across the interface.
We propose that the steric-zipper structures of the VQIVYK and GGVLVN segments represent the spines of the fibrils formed by the parent proteins containing these segments. Supporting our hypothesis are our results that d-amino-acid inhibitors designed on the VQIVYK steric-zipper template inhibit fibril formation not only of the VQIVYK segment, but also of two tau constructs, K12 and K1923, 24 (Fig. 2a). Similarly, the peptide composed of non-natural amino acids designed on the GGVLVN template inhibits the fibril formation of 248PAP286 and greatly inhibits the HIV infectivity of human cells in culture.
To design a d-amino-acid hexapeptide sequence that interacts favourably with the VQIVYK steric zipper15, and prevents further addition of tau molecules to the fibril, we used the Rosetta software25. This led to the identification of four d-amino-acid peptides: d-TLKIVW, d-TWKLVL, d-DYYFEF and d-YVIIER, in which the prefix signifies that all α-carbon atoms are in the d configuration (Fig. 2b, c, Supplementary Figs 1 and 2 and Supplementary Table 1). In the d-TLKIVW design model (Fig. 2b, c and Supplementary Fig. 1), the inhibitor packs tightly across the top of the VQIYVK steric-zipper structure, maintaining all main-chain hydrogen bonds. The side-chain hydrogen bonding between layers of stacked Gln 307 residues is replaced in the designed interface by an interaction with d-Lys 3. Several hydrophobic interactions between d-TLKIVW and the two VQIVYK ß-strands contribute to the favourable binding energy (Supplementary Table 1). In the design, the d-peptide blocks the addition of another layer of VQIVYK, both above the d-peptide and across on the mating ß-sheet (Supplementary Fig. 3). d-Leu 2 of the designed inhibitor prevents the addition of a VQIVYK molecule above it through a steric clash with Ile 308 of VQIVYK and on the mating sheet through a clash with Val 306 and Ile 308 (Supplementary Fig. 3). These steric clashes involving d-Leu 2 are intended to block fibril growth.
We used fluorescence spectroscopy and electron microscopy to assess whether the designed d-peptides inhibit the fibril formation of the tau segment VQIVYK and of the tau constructs K12 and K19. Of our designed inhibitors, d-TLKIVW is the most effective (Supplementary Fig. 4). Electron microscopy, performed after three days, verified that incubation with equimolar d-TLKIVW prevents K19 fibril formation, which would otherwise have occurred within the elapsed time (Fig. 1, upper right). d-TLKIVW delays fibril formation of VQIVYK, K12 and K19 even when present in sub-equimolar concentration (Supplementary Fig. 5). A fivefold molar excess of d-TLKIVW delays K12 fibril formation for more than two weeks in some experimental replicates (Supplementary Fig. 5c, d). In tenfold molar excess, d-TLKIVW prevents the fibril formation of K12 for more than 60 hours in the presence of preformed K12 fibril seeds, suggesting that the peptide interacts with fibrils (Fig. 2d). Also, kinetic analysis shows that the fibril elongation rate decreases in the presence of increasing concentrations of inhibitor peptide (Supplementary Fig. 6). The large increase in lag time in unseeded reactions may be due to interactions with small aggregates formed during the process of fibril formation.
To investigate the specificity of the designed inhibitor, we tested scrambled sequence variants of d-TLKIVW that have poor (that is, high) calculated energies and unfavourable packing (Supplementary Table 1). The scrambled peptides d-TIKWVL, d-TIWKVL and d-LKTWIV have little inhibitory effect when present at an equimolar ratio with VQIVYK, K12 and K19 (Fig. 2e and Supplementary Fig. 7), showing that the inhibition is sequence specific. Also, the diastereomer, l-TLKIVW, is less effective than d-TLKIVW (Supplementary Fig. 8). As a further test of the specificity of our design, we confirmed that d-TLKIVW is unable to block the fibril formation of amyloid-ß, which also is associated with Alzheimer's disease (Supplementary Fig. 9). This suggests that the d-peptide inhibitor is not general to amyloid systems, but is specific to the VQIVYK interface in tau protein. Such specificity is essential for designed inhibitors if they are not to interfere with proteins that natively function in an amyloid state3.
To confirm that the designed d-peptide inhibits in accordance with the design model (Fig. 2b, c and Supplementary Fig. 1), we performed several additional tests. First we visualized the position of the inhibitor d-TLKIVW relative to fibrils of the tau construct K19 using electron microscopy. We covalently linked Monomaleimido Nanogold particles both to the inhibitor and, separately, to a scrambled hexapeptide, d-LKTWIV. We used a blind counting assay and found that, relative to Nanogold alone, d-TLKIVW shows a significant binding preference for the end of fibrils, in contrast to the scrambled control peptide, d-LKTWIV (Fig. 3a and Supplementary Fig. 10).
As a further test of the model, we used NMR to characterize the binding affinity of d-TLKIVW for tau fibrils. The 1H NMR spectra for d-TLKIVW were collected in the presence of increasing concentrations of VQIVYK or K19 fibrils. Because neither K19 nor VQIVYK contains tryptophan, we were able to monitor the 1H resonance of the indole proton of the tryptophan in our inhibitor. When bound to a fibril, the inhibitor, d-TLKIVW, is removed from the soluble phase and the 1H resonance is diminished26 (Fig. 3b and Supplementary Fig. 11). As a control, we also measured spectra for the non-inhibiting peptide d-LKTWIV present with d-TLKIVW in the same reaction mixture. As shown in Fig. 3b, the presence of VQIVYK fibrils at a given concentration reduces the d-TLKIVW indole resonance much more than it does the d-LKTWIV indole resonance. Spectra of the two peptides are shown in Supplementary Fig. 12. By monitoring the d-TLKIVW indole resonance over a range of VQIVYK fibril concentrations, we estimate the apparent dissociation constant of the interaction between d-TLKIVW and VQIVYK fibrils to be ~2 μM (Supplementary Fig. 11a and Methods). This value corresponds to a standard free binding energy of ~7.4 kcal mol-1, with ~2.5 kcal mol-1 from non-polar interactions and ~4.9 kcal mol-1 from six hydrogen bonds (Methods). Repeating the NMR binding experiment with K19 fibrils yields a similar trend (Supplementary Fig. 11b). To determine whether d-TLKIVW has affinity for soluble VQIVYK, we measured 1H NMR spectra of d-TLKIVW and d-LKTWIV in the presence of increasing amounts of soluble VQIVYK. Only a slight change in the respective chemical shifts of the indole proton peaks of d-TLKIVW and d-LKTWIV is observed, even at a 70-fold molar excess of VQIVYK (Supplementary Fig. 13). This, together with the ability of the peptide to prevent seeded fibril formation, suggests that d-TLKIVW does not interact with monomers but rather with a structured, fibril-like species.
As another test of our design model, we replaced the d-Leu residue with d-Ala in d-TLKIVW. Our structural model suggests that d-Leu 2 of d-TLKIVW is important for preventing tau fibril formation because of its favourable interaction with the Ile residue of the VQIVYK molecule below and with Ile and the first Val of VQIVYK across the steric zipper (Fig. 2b, c and Supplementary Fig. 1). The d-Ala replacement eliminates these interactions and, furthermore, removes a steric clash that would occur were another VQIVYK molecule placed across from the inhibitor (Supplementary Fig. 3 and Supplementary Table 1). When the d-Ala variant is incubated with VQIVYK and the tau constructs, it has no inhibitory effect on fibril formation (Fig. 2f and Supplementary Fig. 14). This confirms that d-Leu 2 is critical for the efficacy of d-TLKIVW, consistent with our model.
In summary, although our electron microscopy, NMR and d-Ala replacement results support a model in which the designed peptide d-TLKIVW binds to the ends of tau fibrils, they do not constitute proof that the inhibitors bind exactly as anticipated in the designs (Supplementary Fig. 15).
To expand on our design methodology, we computationally designed an inhibitor of 248PAP286 fibril formation containing non-natural l-amino acids (Fig. 4b and Supplementary Fig. 16), using the GGVLVN structure as a template (Fig. 4a and Supplementary Table 2). This peptide, Trp-His-Lys-chAla-Trp-hydroxyTic (WW61), contains an Ala derivative, ß-cyclohexyl-l-alanine (chAla) and a Tyr/Pro derivative, 7-hydroxy-(S)-1,2,3,4-tetrahydroisoquinoline-3-carboxylic acid (hydroxyTic), both of which increase contact area with the GGVLVN template. The non-natural chAla forms hydrophobic interactions with the Leu residue in the steric-zipper interface, and hydroxyTic supports the favourable placement of chAla through hydrophobic packing (Fig. 4b and Supplementary Fig. 16b). Moreover, we propose that the bulky side chains and steric constraints of hydroxyTic provide hindrance to further fibril growth.
This designed peptide, WW61, effectively delays both seeded and unseeded fibril formation of 248PAP286 in vitro (Fig. 4c and Supplementary Figs 17 and 18). In the presence of a twofold molar excess of this inhibitor, seeded fibril formation is efficiently blocked for more than two days (Fig. 4c). Furthermore, we see that increasing the concentration of this inhibitor extends the fibril formation lag time (Supplementary Fig. 19). These inhibition assay results were further confirmed by electron microscopy (Supplementary Fig. 20). As a control for specificity, we tested the effect of GIHKQK, from the amino terminus of 248PAP286, and PYKLWN, a peptide with the same charge as WW61. Neither peptide affected fibril formation kinetics, indicating that the inhibitory activity of the designed peptide is sequence specific (Supplementary Fig. 21).
Because 248PAP286 fibrils (SEVI) have been shown to enhance HIV infection22, using a functional assay we investigated whether WW61 is able to prevent this enhancement. In this experiment, we treated HIV particles with 248PAP286 solutions that had been agitated for 20 hours (to allow fibril formation) in the presence or absence of WW61, and infected TZM-bl indicator cells. As has been previously observed, SEVI efficiently enhanced HIV infection22. However, 248PAP286 incubated with the designed inhibitor prevented HIV infection (Fig. 4d).
We performed several control experiments to verify that the lack of infectivity observed in the assay is indeed due to the inhibition of SEVI formation. First we confirmed that in the absence of SEVI the designed inhibitor WW61 does not affect HIV infectivity (Supplementary Fig. 22a). We also found that the control peptides GIHKQK and PYKLWN, which do not inhibit 248PAP286 fibril formation, fail to decrease HIV infectivity (Supplementary Fig. 22b). Additionally, we observed that WW61 has no inhibitory effect on polylysine-mediated HIV infectivity27, further ruling out a non-specific electrostatic interaction mechanism (Supplementary Fig. 22a). Together, these results demonstrate that a peptide capable of preventing 248PAP286 fibril formation also inhibits the generation of virus-enhancing material.
Structure-based design of inhibitors of amyloid fibril formation has been challenging in the absence of detailed information about the atomic-level interactions that form the fibril spine. So far, one of the most successful structure-based approaches to preventing fibril formation has been to stabilize the native tetrameric structure of transthyretin28. That approach is well suited to the prevention of fibril formation of proteins with known native structures, but other proteins involved in amyloid-related diseases, such as tau protein, amyloid-ß and 248PAP286, lack fully ordered native structures29. Our structure-based approach makes it possible to design inhibitors independent of native structure. Instead, the templates are atomic-level structures of short, fibril-forming segments14, 15. By using these fibril-like templates, and adopting computational methods successful in designing novel proteins and protein-protein interfaces25, 30, we have created specific inhibitors of proteins that normally form fibrils. These results support the hypothesis that the steric zipper is a principal feature of tau-related and SEVI fibrils, and suggest that, with current computational methods and steric-zipper structures, we have the tools to design specific inhibitors to prevent the formation of other amyloid fibrils.
Towards a Pharmacophore for Amyloid
Towards a Pharmacophore for Amyloid
Meytal Landau, Michael R. Sawaya, Kym F. Faull, Arthur Laganowsky, Lin Jiang, Stuart A. Sievers, Jie Liu, Jorge R. Barrio, David Eisenberg Research Article, published 14 Jun 2011
Diagnosing and treating Alzheimer's and other diseases associated with amyloid fibers remains a great challenge despite intensive research. To aid in this effort, we present atomic structures of fiber-forming segments of proteins involved in Alzheimer's disease in complex with small molecule binders, determined by X-ray microcrystallography. The fiber-like complexes consist of pairs of ß-sheets, with small molecules binding between the sheets, roughly parallel to the fiber axis. The structures suggest that apolar molecules drift along the fiber, consistent with the observation of nonspecific binding to a variety of amyloid proteins. In contrast, negatively charged orange-G binds specifically to lysine side chains of adjacent sheets. These structures provide molecular frameworks for the design of diagnostics and drugs for protein aggregation diseases.
The devastating and incurable dementia known as Alzheimer's disease affects the thinking, memory, and behavior of dozens of millions of people worldwide. Although amyloid fibers and oligomers of two proteins, tau and amyloid-ß, have been identified in association with this disease, the development of diagnostics and therapeutics has proceeded to date in a near vacuum of information about their structures. Here we report the first atomic structures of small molecules bound to amyloid. These are of the dye orange-G, the natural compound curcumin, and the Alzheimer's diagnostic compound DDNP bound to amyloid-like segments of tau and amyloid-ß. The structures reveal the molecular framework of small-molecule binding, within cylindrical cavities running along the ß-spines of the fibers. Negatively charged orange-G wedges into a specific binding site between two sheets of the fiber, combining apolar binding with electrostatic interactions, whereas uncharged compounds slide along the cavity. We observed that different amyloid polymorphs bind different small molecules, revealing that a cocktail of compounds may be required for future amyloid therapies. The structures described here start to define the amyloid pharmacophore, opening the way to structure-based design of improved diagnostics and therapeutics.
The challenge of developing chemical interventions for Alzheimer's disease has proceeded in a virtual vacuum of information about the three-dimensional structures of the two proteins most widely accepted as being involved in the etiology. These are amyloid-beta (Aß) and tau ,. Both convert from largely natively disordered, soluble forms to toxic oligomers and fibers , that may be related in structure . Indeed, analogs of the well-established ligands to amyloid fibers, congo-red and thioflavin T, also bind Aß oligomers labeling them in vitro and in vivo . Screens of chemical libraries have uncovered dozens of small molecules that interact with amyloid -. Curcumin and various antibiotics are a few of many fiber inhibitors that also inhibit oligomer formation ,,, supporting a common underlying structure in fibers and oligomers. Despite this progress, until now there have been no atomic-level structures showing how small molecules bind to amyloid and, consequently, no means for structure-based design of specific binders.
More is known about the molecular structure of amyloid fibers, both those associated with Alzheimer's disease and with the numerous other amyloid conditions -. Common to all amyloid fibers is their X-ray fiber-diffraction pattern, with two orthogonal reflections at about 4.8 Å and 10 Å spacing suggesting a "cross-ß structure" ,. The determination of the first amyloid-like atomic structures revealed a motif consisting of a pair of tightly mated ß-sheets, called a "steric zipper," which is formed from a short self-complementary segment of the amyloid-forming protein ,,. The steric zipper structures elucidate the atomic features that give rise to the common cross-ß diffraction pattern, corresponding to the 4.8 Å spacing between strands forming ß-sheets and the ~10 Å spacing between two mating ß-sheets. The structures imply that stacks of identical short segments form the "cross-ß spine" of the protofilament, the basic unit of the mature fiber, while the rest of the protein adopts either native-like or unfolded conformations peripheral to the spine ,.
The short segments forming steric zippers, when isolated from the rest of the protein, form well-ordered fibers on their own, with essentially all properties of the fibers of their full-length parent proteins ,. These properties include similar fiber diameters and helical pitch, similar cross-ß diffraction patterns, similar fiber-seeding capacities, similar stability, and similar dye binding. That stacked short amyloidogenic segments can constitute the entire spine of an amyloid-like fiber has been demonstrated with the enzyme RNase A, containing an insert of a short amyloidogenic segment ,. These RNase A fibers retain enzymatic activity, showing that native-like structure remains intact with only the stacked segments forming the spine. Thus while short amyloidogenic segments cannot recapitulate the entire complexity of their parent proteins, they nonetheless serve as good models for full amyloid fibers  and offer the informational advantage that they often grow into microcrystals whose atomic structures can be determined . To date, structures for over 50 such steric zippers have been determined from a variety of disease-associated proteins (,,- and Colletier et al. unpublished results).
Here we use one such amyloid-forming segment from Aß and one from tau to form co-crystals with low molecular weight compounds, with the aim of illuminating the nature of interactions of small molecules with amyloid. These complexes reveal a molecular framework which partially defines the amyloid pharmacophore, the structural features responsible for the binding of small molecules to amyloid aggregates.
Our crystal structures of small molecules bound within amyloid-like steric zippers begin to define the molecular frameworks, or pharmacophores, for the design of diagnostics and drugs for Alzheimer's and other aggregation diseases. The amyloid components in our structures are steric zippers formed by stacks of six-residue segments from Alzheimer-related proteins. Although these steric zippers cannot represent all aspects of the full-length amyloid parent proteins, they share many properties and are commonly used as models of the amyloid ß-spine and of aggregation ,. The small molecules in our structures bind along the ß-spine, and because the parent amyloids contain the same segments, we expect a similar mode of binding along the spine of the full-length parent amyloid fibers. Moreover, we anticipate the steric zipper spine of the parent fibers to be flanked with the rest of the protein residues in a native-like or unfolded conformation , and therefore to contain more solvent channels, or accessible sites for the binding of the small molecules, compared to the very compact packing of the steric zipper segments. Consistently, orange-G, curcumin, and DDNP all bind to, or affect fibrillation of, full-length fibers ,,.
Molecular Frameworks of Amyloid Binders
Overall, the complexes presented here suggest the nature of two molecular frameworks for the binding of small molecules to amyloid fibers. The first molecular framework pertains to site-specific binders, such as charged compounds that form networks of interactions with sequence motifs, and is relatively well defined. The second molecular framework, far less well defined at this point, pertains to broad-spectrum binders, such as uncharged aromatic compounds that bind to tube-like cavities between ß-sheets. For binding amyloid deposits in the brain, uncharged molecules could be more effective because of superior blood-brain-barrier penetrability. The same frameworks, offering cavities along ß-sheets, might also exist in amyloid oligomers known to be rich in ß-sheets and possibly fiber-like , similar to the observed binding of amyloid markers to ß-sheets in non-fibrillar structures ,. Consistent with this, both oligomers and fibers are inhibited by similar compounds, including curcumin ,.
The specific binding of orange-G allows definition of the chemical properties of a specific molecular framework. The prominent feature of amyloid structures is the separation of ß-strands (forming a ß-sheet) by ~4.8 Å. In structures with strands packed in an antiparallel orientation, as observed for the KLVFFA fibers and for a rare mutation in Aß that is associated with massive depositions of the mutant protein and early onset of the disease ,, the separation of repeating units (2 strands) is twice as great, ~9.6 Å. Orange-G contains two negatively charged sulfonic acid groups facing the same direction, with the sulfur atoms spaced ~5 Å apart and the oxygen atoms separated by 4.5-7.5 Å. This framework allows the formation of salt links between the sulfonic acid groups and lysine ammonium ions from every repeating strand in both KLVFFA (anti-parallel orientation) and VQIVYK (parallel orientation) fibers (Figures 1 and 4). This shows that a specific framework includes two charged moieties spaced either ~4.8 Å or ~9.6 Å apart. The specific sequence motif of the spine of the fiber and the separation of the ß-strands dictates the signs of the necessary charges in the small molecule and their separation.
Within our framework, an apolar aromatic spine is another essential moiety . The largely apolar KLVFFA segment attracts the apolar surface of orange-G, stabilizing the binding (Figure 2). In the complex with VQIVYK, the aromatic rings of orange-G are also packed against apolar side chains, but the binding is largely mediated via polar interactions with glutamine and lysine side-chains at the edges of two steric zippers (Figure 5). The differences in the binding cavities between the KLVFFA and VQIVYK fibers may account for the higher molecular stoichiometry within the KLVFFA-orange-G crystals observed by mass spectrometric analyses, and the correspondingly greater order of this complex (Figures S5-S6).
Despite the lack of atomized electron density for the binding of curcumin and DDNP in VQIVYK fibers, the location of the binding cavity is clear. It is narrow, restricting rotation of the small molecule (Figures 6-7). The atomic groups lining the binding cavity are about half apolar and half polar (Figure 7). The tube-like shape favors the binding of uncharged molecules, such as DDNP and curcumin. The binding site is, however, insufficiently site-specific to allow for high occupancy and ordered interactions and is not yet well defined in atomic detail.
Our structures show that different small molecules bind along the ß-spine of amyloid-like fibers. In case fibers contain more than a single spine, the molecules might bind to multiple sites. This is more likely for the broad-spectrum hydrophobic compounds but can also apply for charged compounds. For example, we observed orange-G to bind to two different steric zippers, of KLVFFA and VQIVYK, with the commonality of binding to lysine side chains protruding from the ß-sheets.
Congo-red, a known amyloid marker, contains two sulfonic acid groups, similar to orange-G, but they are spaced ~19 Å apart, which might account for its lack of specificity . In a recent model, built using NMR constrains, congo-red was computationally docked to the fungal prion domain HET-s(218-289), suggesting that the sulfonic acid groups interact with lysine residues protruding from the sheets , similar to orange-G in our structures. However, in the model, the strands of HET-s are arranged in an anti-parallel orientation and the sulfonic acid groups of congo-red interact with every other lysine along the fiber , while orange-G interacts with every single lysine in both the KLVFFA and VQIVYK complexes (Figures 1 and 4). Both congo-red and thioflavin T, another known marker, bind to numerous different ß-structures, even in a non-fibrillar form ,. Despite their limited specificity and low affinity ,, these dyes play a major role in amyloid research because their binding is detectible via birefringence or fluorescence ,. An important application of our structures is for the design of new markers for aggregation that will be more potent and can also be used in vivo.
The Two Molecular Frameworks and Function
Defining these two molecular frameworks illuminates functional attributes of specific and broad-spectrum amyloid binders. This distinction is consistent with competitive kinetic experiments demonstrating that the binding of FDDNP (the fluoridated analog of DDNP) to Aß fibrils is displaceable by the uncharged non-steroidal anti-inflammatory naproxen, but not by the common charged dyes congo-red and thioflavin T . Moreover, in vitro FDDNP labels amyloid-like structures in a fashion similar to congo-red and thioflavin T, providing further evidence for the broad-spectrum type of binding . Knowledge of both frameworks can lead to the design of more potent and specific compounds. These molecules can act as binders and be used as diagnostics, or serve as inhibitors of aggregation by either destabilizing steric zippers by wedge action (Figure 1) or binding between steric zippers preventing higher-order ß-sheet interactions (Figure 4).
In the case of the complexed curcumin and DDNP structures, we hypothesize that the tube-like cavity along the ß-sheets provides an adequate site for the binding of many compounds of similar properties. However, the lack of specific interactions allows the small molecule to drift along the fiber axis, leading to lower occupancy and a degree of fluidity in the structure. Extrapolating from our structures, we expect that various aromatic compounds, such as polyphenols , would bind to a variety of amyloid-forming sequences because of a cylindrical, partially apolar cavity that forms between the pairs of ß-sheets forming the fibers. These cavities might also provide binding sites for various kinds of apolar drugs, such as benzodiazepines and anesthetics, explaining some of the altered pharmacokinetic properties and increased sensitivity detected in elderly .
A subtle implication of our structures for the design of effective therapeutic treatments is the specificity they reveal of ligand binding to particular fiber polymorphs (Figure 3). Various amyloid proteins show diverse fiber morphologies that are correlated with different patterns of pathology and toxicity ,. In earlier work, we have suggested that fiber polymorphism has its molecular basis in different steric zippers (ß-sheet packing) formed by the same sequence . Our new findings show that different compounds bind to different fiber polymorphs formed by the same sequence. For example, orange-G displaces one VQIVYK zipper relative to its mate; that is, wedges between protofilaments (Figures 3F and 4). In contrast, both DDNP and curcumin opportunistically bind to cylindrical cavities at the edges of VQIVYK zippers, in a void formed within a different VQIVYK ß-sheet packing (Figures 3E and 6). This suggests that each compound binds to only a sub-population of fibers. Thus, just as cocktails of anti-HIV drugs are necessary to inhibit different viral strains, a combination of compounds may be necessary to bind to the several amyloid polymorphs present.
Four crystal structures of small molecules bound to fiber-forming segments of the two main Alzheimer's disease proteins show common features. The small molecules bind with their long axes parallel to the fiber axis. The structures reveal a sequence-specific binder which forms salt links with side-chains of the steric zipper spines of the fibers and non-specific binders which lie in cylindrical cavities formed at the edges of several steric zippers. Small-molecule binding is specific to particular steric-zipper polymorphs, suggesting that effective Alzheimer's diagnostics and therapeutics may have to be mixtures of various compounds to bind to all polymorphs present. The complexes presented here suggest routes for structure-based design of combinations of compounds that can bind to a spectrum of polymorphic aggregates, to be used as markers of fibers and as inhibitors of aggregation.