A novel amyloid designable scaffold and potential inhibitor inspired by GAIIG of amyloid beta and the HIV-1 V3 loop

Abstract

The GAIIG sequence, common to the amyloid beta peptide (residues 29–33) and to the HIV-1 gp120 (residues 24–28 in a typical V3 loop), self-assembles into amyloid fibrils, as suggested by theory and the experiments presented here. The longer YATGAIIGNII sequence from the V3 loop also self-assembles into amyloid fibrils, of which the first three and the last two residues are outside the amyloid GAIIG core. We postulate that this sequence, with suitably selected modifications at the flexible positions, can serve as a designable scaffold for novel amyloid-based materials. Moreover, we report the single crystal X-ray structure of the beta-breaker peptide GAIPIG at 1.05 Å resolution. The structural information provided in this study could serve as the basis for structure-based design of potential inhibitors of amyloid formation.

Abbreviations

MD, molecular dynamics

TEM, Transmission electron microscopy

Naturally occurring peptide sequences extracted from amyloid proteins or β-sheet protein regions can self-assemble outside the context of the entire sequence into amyloid β-sheets and can serve as scaffolds for novel materials [1-12]. GAIIGL [13] and NSGAITIG [14] are two peptide sequences similar in sequence which are part of the amyloid-β (Aβ) peptide, linked to Alzheimer's disease, and the adenovirus fiber shaft [15, 16], respectively. Both GAIIGL and NSGAITIG form amyloid β-sheets outside the context of the entire peptide or protein. According to experimental X-ray and computational molecular dynamics (MD) simulation studies, in both peptides residues outside the GAIIG or GAITIG sequences are not part of the amyloidogenic β-sheet core: the C-terminal leucine, and the N-terminal asparagine and serine residues in the two peptides, respectively, are exposed. The latter provided impetus [14] for the discovery of a series of amyloid materials with several applications [1-4] by modifying the NS-residues.

While aromatic residues are key components of amyloid self-assembly (as pioneered by Gazit and colleagues [17-19]), patterns of aliphatic residues are also key self-assembly components [20-22], contributing to the amyloid properties of GAIIGL [13]. In this study, we carried out experimental and computational studies that show that the shorter GAIIG peptide (amidated at the C-terminus) self-assembles into amyloid β-sheets. This sequence is common to the amyloid-β (Aβ) peptide (residues 29–33) and to HIV gp120 (residues 24–28 in a typical 35-residue long V3 loop), according to the HIV sequence database (https://www.hiv.lanl.gov/). We also show that the longer YATGAIIGNII sequence from the V3 loop self-assembles into amyloid fibrils of which the first and last three residues are outside the amyloid GAIIG core. Finally, we report on the single crystal structure of the beta-breaker peptide GAIPIG and the computational investigation of its binding properties to Aβ and the V3 loop. We discuss potential implications for material design and structure-based inhibitor design.

Materials and methods

Materials and experimental methods

Peptides and chemicals

The following peptides were studied: NH₃⁺-GAIIG-CONH₂, NH₃⁺-GAIPIG-CONH₂, and NH₃⁺-YATGAIIGNII-COO⁻. The peptides were purchased from Genecust (Luxemburg) and possessed a degree of purity higher than 95%. The Aβ_1–40 peptide was purchased from Bachem (Switzerland). Phosphate-buffered saline and Thioflavin T were purchased from Sigma, and sodium azide from Serva.

Transmission electron microscopy (TEM)

Samples for TEM analysis were prepared by depositing 8 μL of the sample on carbon-coated formvar copper grids (Agar Scientific, Stansted, UK), left aside for 2 min, dried with a filter paper, and then the same procedure was repeated with the stain. The samples were negatively stained with 8 μL 1% (w/v) phosphotungstic acid for 2 min for GAIIG and with 8 μL 2% (w/v) uranyl acetate for YATGAIIGNII. The TEM experiments were performed using a JEOL JEM 2100 High Resolution microscope, operating at 80 kV (University of Crete, Biology Department).

X-ray fiber diffraction

A droplet of a peptide fibril solution was placed between two glass rods that were supported by two plasticine balls and allowed to dry while pulling to induce shear alignment as previously described [23]. The X-rays were focused on the aligned fibers at right angles and the diffraction patterns were recorded.

For the GAIIG peptide fibrils, the diffraction patterns were recorded with a SIEMENS M18XHF Rotating anode generator equipped with a MarResearch 345 image detector system, at a wavelength of 1.541 Å (Cu Kα edge). The exposure times were 30 min per image.

For the YATGAIIGNII peptide fibrils, X-ray fiber diffraction experiments were carried out at the European Synchrotron Radiation Facility (ESRF in Grenoble) on the MASSIF-1 beamline [24] with a wavelength of 0.966Å and a beam size of 15 μm. The exposure time was of the order of 3 s. The sample–detector distance was set to 435.14 mm for giving a resolution of 3.5 Å at the edge of the detector.

Single-crystal X-ray diffraction

Crystallization trials were carried out by sitting-drop vapor diffusion in MRC crystallization plates (Molecular Dimensions, Newmarket, Suffolk, England). Crystals were obtained in wells with different concentrations of ammonium sulfate and pH values. Optimization was performed using 50 μL of reservoir solution and drops consisting of 2 μL of GAIPIG peptide aqueous solution at 38 mg·mL⁻¹ mixed with 2 μL of reservoir solution. Crystals appeared after a period of more than 3 months, only after removing the covering tape in five occasions for 3–5 min during the month previous to appearance of the needle-shaped crystals. The dataset used for structure solution was collected from a crystal grown when 0.1 m Tris-HCl pH 7.5, 1 m ammonium sulfate was used as a reservoir solution. A crystal was mounted in a LithoLoop (Molecular Dimensions) and vitrified in liquid nitrogen for native data collection at 100 K.

The X-ray diffraction dataset was collected at ESRF beamline ID23-2 [25] with a Dectris Pilatus3 X 2M detector using the MXCuBE control software [26]. The wavelength used was 0.8729 Å and a transmittance of 26%, oscillation range of 0.1 deg, and exposures of 0.04 s were employed. Indexing, data integration, and reduction were performed using CrysAlisPro 1.171.39.32a (Rigaku OD, 2017). Absorption correction was performed by the multiscan method implemented in CrysAlisPro. The space group was determined after detwinning the rhombohedral obverse/reverse twinned data using XPREP software (Bruker AXS Inc., (2008) Madison, Wisconsin, USA) and confirmed after solving the structure.

The structure was solved using SHELXD-2013/2 [27] and refined using SHELXL-2018/1 [28, 29] (full-matrix least-squares on F²) through SHELXLe graphical interface program [30]. All non-hydrogen atoms were refined anisotropically. Hydrogen atoms were placed in geometrically calculated positions and were included in the refinement process using a riding model with isotropic thermal parameters. Diffraction data were observed only up to 1.05 Å resolution approximately; therefore, all atoms were refined with geometrical and ADP restraints. The model was refined in space group R3 as a two-component obverse/reverse twin using the HKLF5 format in SHELXL. Crystal data and refinement conditions are shown in Table S1.

Thioflavin T assay

Synthetic lyophilized peptide Aβ_1–40 (Bachem) was dissolved in DMSO to a concentration of 100 μm, aliquoted and stored at −20 °C. Aβ_1–40 aliquots were thawed, sonicated in ice cold water for 20 s to prevent preaggregation, and immediately diluted with 10 mm phosphate-buffered saline containing 10% (v/v) DMSO and 0.05% (w/v) sodium azide [150 mm NaCl (pH 7.4)] to a final concentration of 10 μm. The Aβ_1–40 solution was immediately mixed with the GAIPIG stock solution (1 mm) to a final Aβ_1–40 concentration of 5 μm and GAIPIG concentration of 50 μm. A 10 μm Aβ_1–40 control solution was mixed with the equivalent amount of 10 mm phosphate-buffered saline to a final concentration of 5 μm. A solution of the GAIPIG peptide at the same final concentration (50 μm) without Aβ_1–40 was also prepared_. The samples were incubated without agitation at 37 °C, and the fibrillogenesis rate was monitored by using ThT fluorescence analysis. A quantity of 40 μL of each sample was taken and mixed with 360 μL of 4 μm ThT. The respective excitation and emission wavelengths were 450 nm (5 nm slit) and 480 nm (5 nm slit). The fluorescence of ThT was measured using a SPEX FluoroMax fluorimeter.

Computational methods

Computational methods are included in the Supporting Information.

Results and Discussion

Self-assembly of the GAIIG and YATGAIIGNII sequences

According to 16 μs replica exchange MD simulations and a subsequent computational analysis performed analogously to previous studies [14, 31-35], GAIIG self-assembles into primarily antiparallel off-register β-sheets (Fig. S1A,D,G,J and S2A,D) which possess a high degree of order and alignment of peptides (Fig. S3A). The β-sheet core of the peptide is predominantly composed of the two isoleucine residues Ile3 and Ile4. A representative structure of a highly ordered and well-aligned β-sheet composed by four peptide strands is presented in Fig. 1A. Compared to the antiparallel β-sheets formed by GAIIGL [13], our analysis suggests that the presence of an additional leucine at the end of GAIIG is not necessary for self-assembly. The complete methods and analysis are included in the Supporting Information. Experiments validated the amyloidogenic properties of GAIIG, revealing that the peptide forms non-branched fibrils with diameters of around 10 nm and lengths reaching the order of microns as revealed by transmission electron microscopy (Fig. 2A). Fibrous rods of the peptide display the characteristic cross-beta signature in X-ray fiber diffraction: a 4.6 Å meridional reflection that corresponds to the distance between β-strands and a 10.1 Å equatorial reflection that corresponds to the distance between β-sheets (Fig. 2B).

Figure 1

Molecular graphics images of representative highly ordered and well aligned 4-stranded β-sheet structures of (A) GAIIG and (B) YATGAIIGNII peptides and an example of the disordered structure of (C) GAIPIG peptides. (A) Residues Ile3 and Ile4, shown in licorice representation, form β-bridge interactions, indicated with black dotted lines, and make up the β-sheet core, indicated using transparent surface representation. (B) β-sheet interactions within the GAIIG motif of the YATGAIIGNII peptides, indicated using dotted black lines, are preserved while the N- and C-terminal ends are exposed and flexible. The three N-terminal residues, Tyr1, Ala2, and Thr3 as well as the two C-terminal residues Ile10 and Ile11 are indicated with maroon spheres. The amyloid zipper-like region formed within the GAIIG domain and Asn9 is indicated using transparent surface representation. (C) Pro4 of the GAIPIG peptides prevents the formation of ordered β-sheet structures and disrupts the formation of extended β-sheet conformations (e.g., see the left and middle pairs of peptides forming single β-bridge interactions in contrast to the right pair of peptides where an extended β-sheet conformation is formed).

Figure 2

(A) TEM micrograph of a 20 mg·mL⁻¹ solution of the GAIIG peptide following 62 days of incubation in phosphate buffer pH 7, negatively stained with phosphotungstic acid 1%. (B) X-ray fiber diffraction pattern of rods formed from a 7 mg·mL⁻¹ solution in phosphate buffer following aging for 40 days.

The recent experimentally resolved structures of entire Aβ fibrils [36-38] show that the last glycine of GAIIG introduces a turn into the amyloid β-sheets. The aforementioned information suggests that GAIIG can be a sufficiently short amyloidogenic core of larger amyloid forming peptides containing additional (e.g., 2–3) residues at both termini, which can be outside the amyloid β-sheet, as glycine residues can act as β-turn promoters halting β-sheet elongation. The discovery of such amyloid peptide scaffolds can serve as a source of inspiration for the discovery of amyloid materials with advanced properties, as the exposed residues can be modified accordingly depending on the desired application.

Our previous computational studies [39, 40] showed that the HIV-1 gp120 V3 loop adopts a structure in which the opposite stems of the loop form a β-sheet in its interaction with chemokine receptors CXCR4 and CCR5. Interestingly, we observed that a GAIIG sequence fragment (or other homologous fragments including GQIIG, GQIVG, etc.) is part of one of the two stems, comprising residues 24–28 in a typical 35-residue long V3 loop, according to the HIV sequence database (https://www.hiv.lanl.gov/). This observation further supported our aforementioned suggestion and led us to postulate that such variable sequences derived from the HIV-1 gp120 V3 loop containing GAIIG as an amyloid core and additional (e.g., 2–3) residues at both termini could serve as a source of inspiration for novel amyloid material scaffolds. Additionally, this led us to a second postulation that sequence fragments including GAIPIG, which are homologous to GAIIG but are not amyloidogenic due to the presence of a proline beta-breaker [22], could potentially bind to GAIIG fragments of HIV-1 gp120 V3 loop and to Αβ.

As for the first postulation, in this study, we focused on one such sequence: YATGAIIGNII derived from a V3 loop [41] (without any modifications at the termini). Similar to GAIIG, according to 16 μs replica exchange MD simulations and a subsequent computational analysis, which was performed analogously to previous studies [14, 31-35], the YATGAIIGNII peptide self-assembles into primarily antiparallel off-register β-sheets (Fig. S1C,F,I,L,N and S2C,F,H), which possess a high degree of order and alignment of peptides (Fig. S3C). The β-sheet core of the peptide is predominantly composed of residues Gly4 to Asn9, encompassing the GAIIG domain. A representative structure of a highly ordered and well-aligned β-sheet composed by four peptide strands is presented in Fig. 1B. Interestingly, we observe that both glycine residues act as β-turn promoters, which halt the elongation of the β-sheet core outside the domain. Thus, N-terminal residues Tyr1, Ala2, Thr3, as well as C-terminal residues Ile10 and Ile11 are rarely involved in β-sheet formation and are outside the amyloid zipper-like region formed within the GAIIG domain and Asn9 (Fig. 1B and Fig. S2C,F,H). As a result, the amyloid scaffolds formed by YATGAIIGNII can be considered as excellent designable scaffolds for the synthesis of functional amyloid materials. This can be achieved by introducing suitably selected mutations at the non-β-sheet forming terminal residue positions 1, 2, 3, 10, and 11, which would not disrupt the amyloid self-assembly properties and at the same time would allow the newly designed amyloid fibrils to bind to ions, molecules, or surfaces. The complete methods and analysis are included in the Supporting Information. Experiments validated the amyloidogenic properties of YATGAIIGNII. TEM micrographs of YATGAIIGNII fibrils reveal a typical amyloid-type morphology (Fig. 3A) and X-ray fiber diffraction of rods display the characteristic cross-β signature with a 4.67 Å reflection at the meridian and 9.92 Å at the equator (Fig. 3B). Additional replica exchange MD simulation runs with a longer 13-residue peptide with sequence AFYATGAIIGNII extracted from the same HIV-1 gp120 V3 loop show that the inclusion of additional residues result in the formation of U-shaped β-sheets (preliminary results not shown) similarly to β-sheets formed by LSFDNSGAITIG [22, 35]. Thus, 11-residue peptides containing three residues before and after the GAIIG domain can be optimal designable amyloid scaffolds containing the maximum number of mutable positions and at the same time comprising linear shaped peptides in which the non-β-sheet residues are exposed for functionalization purposes.

Figure 3

(A) TEM micrograph of a 5 mg·mL⁻¹ solution of the YATGAIIGNII peptide following 5 days of incubation in water, negatively stained with uranyl acetate 2%. (B) X-ray fiber diffraction pattern of rods formed from a 3 mg·mL⁻¹ solution in water after 2 h of incubation.

Additional peptide sequences containing the GAIIG amyloid core plus 2–3 residues at both termini, and inspired by either the HIV-1 V3 loop sequence variability or other naturally occurring proteins encompassing the amyloid GAIIG, may also be designable amyloid scaffolds that could be further investigated. Interestingly, this can be supported by the fact that GAIIG domain can be found in amyloid or β-sheet rich regions of proteins of known structure, including Αβ (₂₉GAIIG₃₃ in PDB ID: 5OQV [37]), α-tubulin acetyl transferase (₉₁GAIIG₉₅ in PDB ID: 4PK2 [42] Chain A), Mycobacterium smegmatis alpha-ketoglutarate decarboxylase homodimers (₂₇₅GAIIG₂₇₉ in PDB ID: 2XT6 [43] Chain A), and Tryparedoxin from Trypanosoma brucei (₁₂₀GAIIG₁₂₄ in PDB ID: 1073) [44].

X-ray and computational analysis of the GAIPIG peptide

While GAIIG is an amyloid peptide and as a domain it comprises the amyloid core of the larger peptide YATGAIIGNII, our previous experiments showed that the insertion of a proline between the two isoleucine residues (GAIPIG peptide) results in the disruption of amyloid formation [22]. Here, we performed additional X-ray and computational analyses on the structures formed by GAIPIG. A crystallographic structure reveals that two extended molecules of the GAIPIG peptide interact with each other to form an antiparallel dimer, reminiscent of a small beta-sheet. In the crystal, these dimers associate in helical fashion around a 3₁ symmetry axis. The center of the helix is hydrophobic, while the outside is hydrophilic (Fig. 4A,B,C,D).

Figure 4

(A) Hydrogen-bonded dimer formed by the two peptide molecules in the crystallographic asymmetric unit (AU). Single-letter amino acid code is displayed for one peptide. (B) Same as (A) including two additional symmetry related molecules (cyan and green) that extend the hydrogen bonded helical assembly. Hydrogen bonds for symmetry related molecules in orange. Water molecules omitted for clarity. (C) Ribbon representation of the helical hydrogen-bonded assembly of GAIPG molecules displayed along crystallographic a axis and (D) along c axis. Side chains are showed in wireframe representation and solvent water molecules as red dots. The antiparallel β-strand corresponding with the crystallographic AU is colored in green; in cyan and purple are the 3-fold screw symmetry related ones. Figures were generated with MERCURY [67] (A) and (B) and PyMOL (C) and (D) (PyMOL 1.9: The PyMOL Molecular Graphics System, Schrodinger, LLC.).

The simulation-based analysis depicts that the peptide forms β-sheet rich structures and the interactions occur frequently within the AIPI moiety of the peptide (Fig. S2B,E,G) which primarily involve β-isolated bridges formed between Ile5 residues, in line with our X-ray studies. Yet, the presence of proline at the middle of the core disallows the formation of highly ordered and well-aligned β-sheets (Fig. 1C and S3B). This in contrast to GAIIG, providing evidence for the inability of the GAIPIG peptide to self-assemble into amyloid cross-β-sheets. Additionally, in the simulations, we observed the infrequent formation of complex conformations in which a peptide on one side forms nearly in-register antiparallel β-sheets while, on the opposite side, its proline side chain disallows the formation of an in-register β-sheet. Instead off-register antiparallel β-sheet interactions are formed, which result in breaking of the symmetry required for cross-β-sheet interactions. These conformations, which are part of the complex conformations in our analysis [34] (Fig. S4A), are reminiscent of the elementary structural units observed in the X-ray studies (Fig. S4B).

Computational investigation of binding properties of GAIPIG to Aβ and the V3 loop

As for the second postulation, driven by the non-amyloid character of GAIPIG and our present results, we computationally investigated the binding properties to the sequentially similar GAIIG domains of Aβ and the HIV-1 gp120 V3 loop, and its ability to potentially inhibit Aβ fibril elongation or prevent the binding of the HIV-1 gp120 V3 loop to chemokine receptors. Of note, according to experiments, in the presence of elevated concentrations of GAIPIG, an increase in the length of the lag phase as well as a reduction in the quantity of fibril formation by Aβ was observed; a 10-fold molar excess of GAIPIG led to about 25–30% reduction in fibril formation after 168 h (7 days) of incubation (Fig. 5).

Figure 5

Kinetics of the Aβ_1–40 peptide fibril formation (5 μm) as assessed by the Thioflavin-T binding assay in the absence (full circles) or in the presence of 50 μm (open diamonds) GAIPIG peptide. Full squares, 50 μm of GAIPIG peptide alone. One representative experiment from four independent ones is shown.

Simulations of Aβ fibrils in the presence of GAIPIG peptides provide insights for the experimentally observed inhibition, depicting that GAIPIG peptides bind to and form β-sheet interactions with ₁₉FFA₂₁ and ₃₀AII₃₂ of the VFFA and GAIIG motifs of Aβ respectively in such a way that the proline side chain is facing outwards from the fibril (Figs 6 and S5B,C). Thus, GAIPIG likely inhibits the elongation through binding Aβ recognition motifs at the edge of the fibrils in such a way that the proline side chains prevent the formation of β-sheet interactions. Additional details are provided in Supporting Information. In addition, according to simulations of a specific HIV-1 gp120 V3 loop in the presence of GAIPIG peptides, GAIPIG peptides predominantly bind and form β-sheet interactions with residues within the GAIIG motif of loop and, to a lesser extent, residues on the opposite site of the loop comprising the GIHIG motif (Fig. S5A and S6). Furthermore, as GAIPIG peptides favor interactions to the II motif of GAIIG and the GXIIG motif is abundant (e.g. GAIIG, GQIIG, GEIIG, GKIIG, etc.) despite the high variability of the HIV-1 gp120 V3 loop [45, 46] the future study and design of peptides having a GAIPIG- based core and designed extensions (e.g., analogously to ref [47].) targeting V3 loop sequences can be of interest. GAIPIG-based peptide analogs could constitute seeds for potential HIV-1 gp120 V3 loop entry inhibitors, preventing its binding to chemokine receptors, similarly to polyanionic HIV-inhibitors [48, 49].

Figure 6

Molecular graphics images of representative structures of GAIPIG peptides, shown in blue cartoon representation, binding to Aβ fibrils, shown in red cartoon representation. GAIPIG peptides form β-sheet interactions, indicated with black dotted lines, with residues within the (A) GAIIG and (B) VFFA motifs.

GAIIG as another Aβ core recognition sequence?

Of the residues belonging to the core of the amyloid fibril, the stretch comprising residues 16–21, KLVFFA, comprising the aromatic dipeptide motif FF has been most investigated as a core recognition sequence in Alzheimer's disease [50]. The beta-breaker peptide LPFFD was designed and studied both theoretically and experimentally [51-54]. Moreover, endomorphin analogs comprising aromatic amino acids along with proline residues were recently studied as inhibitors of β-amyloid oligomerization [55]. In addition to the role of the KLVFFA stretch, a number of different approaches were pointing at an important role of the GAIIG stretch in Aβ aggregation and fibril elongation. Residues ₃₀AIIG₃₃ were identified as part of the aggregation-prone region 30–42 of the Aβ peptide using a predictive algorithm of aggregation propensities developed by Dobson and colleagues [56]. Thorough proline-[57] and alanine-scanning [58] mutagenesis studies carried out by Wetzel and colleagues suggested that residues ₃₁IIG₃₃ were highly sensitive to proline and alanine replacements. Both replacements were increasing the ΔG of elongation equilibrium of the fibrils, indicating a destabilization of the fibril (reviewed in ref [59].). The peptide sequence NKGAII comprising residues 27–32 of Aβ, was also recently identified as one of the “hot regions” for self- and cross-interaction between Aβ and IAPP [60]. By using an in vivo reporter system, Ventura and colleagues calculated experimental aggregation propensities of the 20 amino acids with isoleucine ranking highest, followed by phenylalanine [61], and identified the ₃₀AIIGLM₃₅ stretch as a “hot spot” for aggregation [62]. Moreover, according to theoretical studies, Ile is avoided in protein interfaces, has a high propensity to be involved in amyloid formation, and molecular dynamics studies of Nussinov and colleagues pointed to a stable Ile-Ile cluster holding the structure of the Aβ 25–35 region [63]. The Ile-Ile dipeptide motif might constitute a core recognition motif, analogous to the dipeptide motif Phe-Phe extracted from the KLVFFA sequence. Taking this into account, the GAIPIG peptide was studied and the computational results as well as the single crystal structure presented here confirm that proline insertion between the two isoleucines disrupts amyloid formation. Most efforts on discovering peptide inhibitors inspired by the sequence of Aβ focused primarily on modified peptides homologous to the sequence fragment KLVFFA [64, 65]. Interestingly, our computational results indicate that the inhibition of GAIPIG to Aβ can be attributed to its capacity to bind to both sequence fragments KLVFFA and GAIIG of Αβ, suggesting that modified GAIPIG peptides are worthy of further investigation for Αβ inhibition.

Conclusion

The computational and experimental studies presented here (summarized in Fig. 7) point to the GAIIG sequence as an amyloid-forming building block that merits further investigation as a potential core recognition sequence in Aβ. By introducing suitably selected mutations at the non-β-sheet forming terminal residue positions of the longer sequence YATGAIIGNII, novel materials could be designed. Furthermore, our computational results suggest interaction of the beta-breaker GAIPIG peptide with Aβ that could inhibit Aβ fibril elongation and experiments showed that the peptide delays the aggregation of the peptide Aβ_1–40 in vitro. A natural peptide like GAIPIG cannot be envisaged as a potential therapeutic per se, due to degradation and stability issues in vivo; however, non-natural analogs and peptidomimetics structurally related to the parent peptide can be developed [66]. Thus, the single-crystal structural information on the beta-breaker GAIPIG peptide could be exploited as a minimal framework for future structure-based design of Aβ inhibitors. Our computational results also suggested interaction of GAIPIG with the GAIIG sequence of gp120 V3 loop, especially with the II motif common to the GXIIG sequence of the V3 loop. Although we have not performed experimental investigations of inhibition of the V3 loop binding to chemokine receptors, the computational and single-crystal information presented in this paper can be of interest for the future design of V3 loop entry inhibitors in the future.

Figure 7

Illustration of the molecular linkage between an amyloid-forming core, material scaffold, and beta-breaker peptide. The GAIIG amyloid-forming core (center) surrounded by flexible residues (in red color, bottom panel) becomes amenable to modifications that can target a variety of materials. When a proline residue (in red, upper panel) is introduced between the two isoleucines, the resulting GAIPIG peptide does not form amyloids and crystallizes yielding a single crystal structure. Isoleucine residues are in cyan color.

Acknowledgments

This research is supported by startup funding by the Artie McFerrin Department of Chemical Engineering at Texas A&M University, a Seed Fund by the Texas A&M Engineering Experiment Station (PT), and by the Texas A&M University Graduate Diversity Fellowship from the TAMU Office of Graduate and Professional Studies (AAO). All MD simulations and free energy calculations were performed on the Ada supercomputing cluster at the Texas A&M High Performance Research Computing Facility. M.J.v.R. acknowledges funding by the grant BFU2014-53425-P (AEI/FEDER, EU). C.K. acknowledges support from a Manassaki Foundation Fellowship of the University of Crete, and M.S.-B. from an FPI fellowship from the Spanish Ministry of Economy, Industry and Competitiveness. C.P.A. acknowledges the financial support of the Stavros Niarchos Foundation within the framework of the project ARCHERS “Advancing Young Researchers’ Human Capital in Cutting Edge Technologies in the Preservation of Cultural Heritage and the Tackling of Societal Challenges”. We thank Prof. Georgios Chalepakis and Ms. Sevasti Papadogiorgaki for expert technical assistance with Transmission Electron Microscopy experiments.

Data Accessibility

Research data pertaining to this article is located at figshare.com: https://dx.doi.org/10.6084/m9.figshare.6259337.

Author contributions

CK, CPA, MK and AC conducted biochemical studies; SVRJ, AAO, JMJ and PT conducted computational studies; MS-B, ALL-S and MJvR conducted crystallization studies and solved crystal structures; EM, VTF, EPM and MWB conducted X-ray fiber diffraction studies; PT and AM conceived the study, supervised experiments and wrote the manuscript. All authors have seen and approved the manuscript.