Josef Rudinger Memorial Lecture: Use of peptides to modulate protein‐protein interactions

Peptides are destined to play a major role as therapeutic agents. My laboratory is contributing to speeding up this process. On the one hand, we devote efforts to studying the molecular details and dynamics of the events that occur during molecular recognition at protein surfaces. We succeeded to design and synthesize peptides able to modulate these recognition events either permanently or in response to light. On the other hand, we are discovering and designing peptides able to cross biological barriers. Our aim is to use these peptides as shuttles for targeting therapeutic agents to organs, tissues, or cells, with a special emphasis on drug delivery to the brain. Copyright © 2015 European Peptide Society and John Wiley & Sons, Ltd.     

Peptidic modulators of protein‐protein interactions: Progress and challenges in computational design

With the decline in productivity of drug‐development efforts, novel approaches to rational drug design are being introduced and developed. Naturally occurring and synthetic peptides are emerging as novel promising compounds that can specifically and efficiently modulate signaling pathways in vitro and in vivo. We describe sequence‐based approaches that use peptides to mimic proteins in order to inhibit the interaction of the mimicked protein with its partners. We then discuss a structure‐based approach, in which protein‐peptide complex structures are used to rationally design and optimize peptidic inhibitors. We survey flexible peptide docking techniques and discuss current challenges and future directions in the rational design of peptidic inhibitors. © 2009 Wiley Periodicals, Inc. Biopolymers 91: 505–513, 2009. This article was originally published online as an accepted preprint. The “Published Online”date corresponds to the preprint version. You can request a copy of the preprint by emailing the Biopolymers editorial office at biopolymers@wiley.com     

Short PlGF‐derived peptides bind VEGFR‐1 and VEGFR‐2 in vitro and on the surface of endothelial cells

The placental growth factor (PlGF), a member of VEGF family, plays a crucial role in pathological angiogenesis, especially ischemia, inflammation, and cancer. This activity is mediated by its selective binding to VEGF receptor 1 (VEGFR‐1), which occurs predominantly through receptor domains 2 and 3. The PlGF β‐hairpin region spanning residues Q87 to V100 is one of the key binding elements on the protein side. We have undertaken a study on the design, preparation, and functional characterization of the peptide reproducing this region and of a set of analogues where glycine 94, occurring at the corner of the hairpin in the native protein, is replaced by charged as well as hydrophobic residues. Also, some peptides with arginine 96 replaced by other residues have been studied. We find that the parent peptide weakly binds VEGFR‐1, but replacement of G94 with residues bearing H‐bond donating residues significantly improves the affinity. Replacement of R96 instead blocks the interaction between the peptide and the domain. The strongest affinity is observed with the G94H (peptide PlGF‐2) and G94W (peptide PlGF‐10) mutants, while the peptide PlGF‐8, bearing the R96G mutation, is totally inactive. The PlGF‐1 and PlGF‐2 peptides also bind the VEGFR‐2 receptors, though with a reduced affinity, and are able to interfere with the VEGF‐induced receptor signaling on endothelial cells. The peptides also bind VEGFR‐2 on the surface of cells, while PlGF‐8 is inactive. Data suggest that these peptides have potential applications as PlGF/VEGF mimic in various experimental settings.     

Synthesis,characterization, and biological activity of poly(arginine)‐derived cancer‐targeting peptides in HepG2 liver cancer cells

The solid‐phase synthesis, structural characterization, and biological evaluation of a small library of cancer‐targeting peptides have been determined in HepG2 hepatoblastoma cells. These peptides are based on the highly specific Pep42 motif, which has been shown to target the glucose‐regulated protein 78 receptors overexpressed and exclusively localized on the cell surface of tumors. In this study, Pep42 was designed to contain varying lengths (3–12) of poly(arginine) sequences to assess their influence on peptide structure and biology. Peptides were effectively synthesized by 9‐fluorenylmethoxycarbonyl‐based solid‐phase peptide synthesis, in which the use of a poly(ethylene glycol) resin provided good yields (14–46%) and crude purities >95% as analyzed by liquid chromatography–mass spectrometry. Peptide structure and biophysical properties were investigated using circular dichroism spectroscopy. Interestingly, peptides displayed secondary structures that were contingent on solvent and length of the poly(arginine) sequences. Peptides exhibited helical and turn conformations, while retaining significant thermal stability. Structure–activity relationship studies conducted by flow cytometry and confocal microscopy revealed that the poly(arginine) derived Pep42 sequences maintained glucose‐regulated protein 78 binding on HepG2 cells while exhibiting cell translocation activity that was contingent on the length of the poly(arginine) strand. In single dose (0.15 mM) and dose‐response (0–1.5 mM) cell viability assays, peptides were found to be nontoxic in human HepG2 liver cancer cells, illustrating their potential as safe cancer‐targeting delivery agents. Copyright © 2014 European Peptide Society and John Wiley & Sons, Ltd.     

Virtual screening on an α‐helix to β‐strand switchable region of the FGFR2 extracellular domain revealed positive and negative modulators

The secondary structure of some protein segments may vary between α‐helix and β‐strand. To predict these switchable segments, we have developed an algorithm, Switch‐P, based solely on the protein sequence. This algorithm was used on the extracellular parts of FGF receptors. For FGFR2, it predicted that β4 and β5 strands of the third Ig‐like domain were highly switchable. These two strands possess a high number of somatic mutations associated with cancer. Analysis of PDB structures of FGF receptors confirmed the switchability prediction for β5. We thus evaluated if compound‐driven α‐helix/β‐strand switching of β5 could modulate FGFR2 signaling. We performed the virtual screening of a library containing 1.4 million of chemical compounds with two models of the third Ig‐like domain of FGFR2 showing different secondary structures for β5, and we selected 32 compounds. Experimental testing using proliferation assays with FGF7‐stimulated SNU‐16 cells and a FGFR2‐dependent Erk1/2 phosphorylation assay with FGFR2‐transfected L6 cells, revealed activators and inhibitors of FGFR2. Our method for the identification of switchable proteinic regions, associated with our virtual screening approach, provides an opportunity to discover new generation of drugs with under‐explored mechanism of action. Proteins 2014; 82:2982–2997. © 2014 Wiley Periodicals, Inc.     

Investigation of α/γ hybrid peptide self‐assembled structures with antimicrobial and antibiofilm properties

The present work describes the synthesis and characterization of α/γ hybrid peptides, Boc‐Phe‐γ⁴‐Phe‐Val‐OMe, P1 ; Boc‐Ala‐γ⁴‐Phe‐Val‐OMe, P2 ; and Boc‐Leu‐γ⁴‐Phe‐Val‐OMe, P3 together with the formation of self‐assembled structures formed by these hybrid peptides in dimethyl sulfoxide (DMSO)/water (1:1). The self‐assembled structures were characterized by infrared (IR) spectroscopy, circular dichroism (CD), and scanning electron microscopy (SEM). Further, α/γ hybrid peptide self‐assembled structures were evaluated for antibacterial properties. Among all, the self‐assembled peptide P1 exhibited the antimicrobial activity against Escherichia coli and Klebsiella pneumoniae, while self‐assembled peptide P3 inhibited the biofilms of Salmonella typhimurium and Pseudomonas aeruginosa. In this study, we have shown the significance of self‐assembled structures formed from completely hydrophobic α/γ hybrid peptides in exploring the antibacterial properties together with biofilm inhibition.     

Gαs protein C‐terminal α‐helix at the interface: does the plasma membrane play a critical role in the Gαs protein functionality?

The heterotrimeric guanine nucleotide-binding regulatory proteins (G proteins, Galphabetagamma) mediate the signalling process of a large number of receptors, known as G protein-coupled receptors. The C-terminal domain of the heterotrimeric G protein alpha-subunit plays a key role in the selective activation of G proteins by their cognate receptors. The interaction of this domain can take place at the end of a cascade including several successive conformational modifications. Galpha(s)(350-394) is the 45-mer peptide corresponding to the C-terminal region of the Galpha(s) subunit. In the crystal structure of the Galpha(s) subunit it encompasses the alpha4/beta6 loop, the beta6 beta-sheet segment and the alpha5 helix region. Following a previous study based on the synthesis, biological activity and conformational analysis of shorter peptides belonging to the same Galpha(s) region, Galpha(s)(350-394) was synthesized and investigated. The present study outlines the central role played by the residues involved in the alpha4/beta6 loop and beta6/alpha5 loops in the stabilization of the C-terminal Galpha(s)alpha-helix. H(2)O/(2)H(2)O exchange experiments, and NMR diffusion experiments show interesting evidence concerning the interaction between the SDS micelles and the polypeptide. These data prompt intriguing speculations on the role of the intracellular environment/cellular membrane interface in the stabilization and functionality of the C-terminal Galpha(s) region.     

β‐Amino acids containing peptides and click‐cyclized peptide as β‐turn mimics: a comparative study with ‘conventional’ lactam‐ and disulfide‐bridged hexapeptides

The increasing interest in click chemistry and its use to stabilize turn structures led us to compare the propensity for β‐turn stabilization of different analogs designed as mimics of the β‐turn structure found in tendamistat. The β‐turn conformation of linear β‐amino acid‐containing peptides and triazole‐cyclized analogs were compared to ‘conventional’ lactam‐ and disulfide‐bridged hexapeptide analogs. Their 3D structures and their propensity to fold in β‐turns in solution, and for those not structured in solution in the presence of α‐amylase, were analyzed by NMR spectroscopy and by restrained molecular dynamics with energy minimization. The linear tetrapeptide Ac‐Ser‐Trp‐Arg‐Tyr‐NH₂ and both the amide bond‐cyclized, c[Pro‐Ser‐Trp‐Arg‐Tyr‐D ‐Ala] and the disulfide‐bridged, Ac‐c[Cys‐Ser‐Trp‐Arg‐Tyr‐Cys]‐NH₂ hexapeptides adopt dominantly in solution a β‐turn conformation closely related to the one observed in tendamistat. On the contrary, the β‐amino acid‐containing peptides such as Ac‐(R)‐β³‐hSer‐(S)‐Trp‐(S)‐β³‐hArg‐(S)‐β³‐hTyr‐NH₂, and the triazole cyclic peptide, c[Lys‐Ser‐Trp‐Arg‐Tyr‐βtA]‐NH₂, both specifically designed to mimic this β‐turn, do not adopt stable structures in solution and do not show any characteristics of β‐turn conformation. However, these unstructured peptides specifically interact in the active site of α‐amylase, as shown by TrNOESY and saturation transfer difference NMR experiments performed in the presence of the enzyme, and are displaced by acarbose, a specific α‐amylase inhibitor. Thus, in contrast to amide‐cyclized or disulfide‐bridged hexapeptides, β‐amino acid‐containing peptides and click‐cyclized peptides may not be regarded as β‐turn stabilizers, but can be considered as potential β‐turn inducers. Copyright © 2011 European Peptide Society and John Wiley & Sons, Ltd.     

A synthetic peptide corresponding to a region of the human pericentriolar material 1 (PCM‐1) protein binds β‐amyloid (Aβ1‐42) oligomers

We have recently reported that a ~19‐kDa polypeptide, rPK‐4, is a protein kinase Cs inhibitor that is 89% homologous to the 1171–1323 amino acid region of the 228‐kDa human pericentriolar material‐1 (PCM‐1) protein (Chakravarthy et al. 2012). We have now discovered that rPK‐4 binds oligomeric amyloid‐β peptide (Aβ)_1‐42 with high affinity. Most importantly, a PCM‐1‐selective antibody co‐precipitated Aβ and amyloid β precursor protein (AβPP) from cerebral cortices and hippocampi from AD (Alzheimer's disease) transgenic mice that produce human AβPP and Aβ_1‐42, suggesting that PCM‐1 may interact with amyloid precursor protein/Aβ in vivo. We have identified rPK‐4′s Aβ‐binding domain using a set of overlapping synthetic peptides. We have found with ELISA, dot‐blot, and polyacrylamide gel electrophoresis techniques that a ~ 5 kDa synthetic peptide, amyloid binding peptide (ABP)‐p4‐5 binds Aβ_1‐42 at nM levels. Most importantly, ABP‐p4‐5, like rPK‐4, appears to preferentially bind Aβ_1‐42 oligomers, believed to be the toxic AD‐drivers. As expected from these observations, ABP‐p4‐5 prevented Aβ_1‐42 from killing human SH‐SY5Y neuroblastoma cells via apoptosis. These findings indicate that ABP‐p4‐5 is a possible candidate therapeutic for AD.     

N‐terminal diproline and charge group effects on the stabilization of helical conformation in alanine‐based short peptides: CD studies with water and methanol as solvent

Protein folding problem remains a formidable challenge as main chain, side chain and solvent interactions remain entangled and have been difficult to resolve. Alanine‐based short peptides are promising models to dissect protein folding initiation and propagation structurally as well as energetically. The effect of N‐terminal diproline and charged side chains is assessed on the stabilization of helical conformation in alanine‐based short peptides using circular dichroism (CD) with water and methanol as solvent. A1 (Ac–Pro–Pro–Ala–Lys–Ala–Lys–Ala–Lys–Ala–NH₂) is designed to assess the effect of N‐terminal homochiral diproline and lysine side chains to induce helical conformation. A2 (Ac–Pro–Pro–Glu–Glu–Ala–Ala–Lys–Lys–Ala–NH₂) and A3 (Ac–d Pro–Pro–Glu–Glu–Ala–Ala–Lys–Lys–Ala–NH₂) with N‐terminal homochiral and heterochiral diproline, respectively, are designed to assess the effect of Glu...Lys (i , i  + 4) salt bridge interactions on the stabilization of helical conformation. The CD spectra of A1 , A2 and A3 in water manifest different amplitudes of the observed polyproline II (PPII) signals, which indicate different conformational distributions of the polypeptide structure. The strong effect of solvent substitution from water to methanol is observed for the peptides, and CD spectra in methanol evidence A2 and A3 as helical folds. Temperature‐dependent CD spectra of A1 and A2 in water depict an isodichroic point reflecting coexistence of two conformations, PPII and β‐strand conformation, which is consistent with the previous studies. The results illuminate the effect of N‐terminal diproline and charged side chains in dictating the preferences for extended‐β, semi‐extended PPII and helical conformation in alanine‐based short peptides. The results of the present study will enhance our understanding on stabilization of helical conformation in short peptides and hence aid in the design of novel peptides with helical structures. Copyright © 2017 European Peptide Society and John Wiley & Sons, Ltd.     

Rational design of amyloid β peptide–binding proteins: Pseudo‐Aβ β‐sheet surface presented in green fluorescent protein binds tightly and preferentially to structured Aβ

Some neurodegenerative diseases such as Alzheimer disease (AD) and Parkinson disease are caused by protein misfolding. In AD, amyloid β‐peptide (Aβ) is thought to be a toxic agent by self‐assembling into a variety of aggregates involving soluble oligomeric intermediates and amyloid fibrils. Here, we have designed several green fluorescent protein (GFP) variants that contain pseudo‐Aβ β‐sheet surfaces and evaluated their abilities to bind to Aβ and inhibit Aβ oligomerization. Two GFP variants P13H and AP93Q bound tightly to Aβ, K_d = 260 nM and K_d = 420 nM, respectively. Moreover, P13H and AP93Q were capable of efficiently suppressing the generation of toxic Aβ oligomers as shown by a cell viability assay. By combining the P13H and AP93Q mutations, a super variant SFAB4 comprising four strands of Aβ‐derived sequences was designed and bound more tightly to Aβ (K_d = 100 nM) than those having only two pseudo‐Aβ strands. The SFAB4 protein preferentially recognized the soluble oligomeric intermediates of Aβ more than both unstructured monomer and mature amyloid fibrils. Thus, the design strategy for embedding pseudo‐Aβ β‐sheet structures onto a protein surface arranged in the β‐barrel structure is useful to construct molecules capable of binding tightly to Aβ and inhibiting its aggregation. This strategy may provide implication for the diagnostic and therapeutic development in the treatment of AD. Proteins 2010. © 2009 Wiley‐Liss, Inc.     

Structural refinement of the hERG1 pore and voltage‐sensing domains with ROSETTA‐membrane and molecular dynamics simulations

The hERG1 gene (Kv11.1) encodes a voltage‐gated potassium channel. Mutations in this gene lead to one form of the Long QT Syndrome (LQTS) in humans. Promiscuous binding of drugs to hERG1 is known to alter the structure/function of the channel leading to an acquired form of the LQTS. Expectably, creation and validation of reliable 3D model of the channel have been a key target in molecular cardiology and pharmacology for the last decade. Although many models were built, they all were limited to pore domain. In this work, a full model of the hERG1 channel is developed which includes all transmembrane segments. We tested a template‐driven de‐novo design with ROSETTA‐membrane modeling using side‐chain placements optimized by subsequent molecular dynamics (MD) simulations. Although backbone templates for the homology modeled parts of the pore and voltage sensors were based on the available structures of KvAP, Kv1.2 and Kv1.2‐Kv2.1 chimera channels, the missing parts are modeled de‐novo. The impact of several alignments on the structure of the S4 helix in the voltage‐sensing domain was also tested. Herein, final models are evaluated for consistency to the reported structural elements discovered mainly on the basis of mutagenesis and electrophysiology. These structural elements include salt bridges and close contacts in the voltage‐sensor domain; and the topology of the extracellular S5‐pore linker compared with that established by toxin foot‐printing and nuclear magnetic resonance studies. Implications of the refined hERG1 model to binding of blockers and channels activators (potent new ligands for channel activations) are discussed. Proteins 2010. © 2010 Wiley‐Liss, Inc.     

Molecular design and optimization of hepatic cancer SLP76‐derived PLCγ1 SH3‐binding peptide with the systematic N‐substitution of peptide PXXP motif

The phospholipase Cγ1 (PLCγ1) is essential for T‐cell signaling and activation in hepatic cancer immune response, which has a regulatory Src homology 3 (SH3) domain that can specifically recognize and interact with the PXXP‐containing decapeptide segment (¹⁸⁵QP P VP P QRPM¹⁹⁴, termed as SLP76^185–194 peptide) of adaptor protein SLP76 following T‐cell receptor ligation. The isolated peptide can only bind to the PLCγ1 SH3 domain with a moderate affinity due to lack of protein context support. Instead of the traditional natural residue mutagenesis that is limited by low structural diversity and shifted target specificity, we herein attempt to improve the peptide affinity by replacing the two key proline residues Pro187 and Pro190 of SLP76^185–194 PXXP motif with nonnatural N‐substituted amino acids, as the proline is the only endogenous N‐substituted amino acid. The replacement would increase peptide flexibility but can restore peptide activity by establishing additional interactions with the domain. Structural analysis reveals that the domain pocket can be divided into a large amphipathic region and a small negatively charged region; they accommodate hydrophobic, aromatic, polar, and moderate‐sized N‐substituted amino acid types. A systematic replacement combination profile between the peptide residues Pro187 and Pro190 is created by structural modeling, dynamics simulation, and energetics analysis, from which six improved and two reduced N‐substituted peptides as well as native SLP76^185–194 peptide are identified and tested for their binding affinity to the recombinant protein of the human PLCγ1 SH3 domain using fluorescence‐based assays. Two N‐substituted peptides, SLP76^185–194(N‐Leu187/N‐Gln190) and SLP76^185–194(N‐Thr187/N‐Gln190), are designed to have high potency (K_d = 0.67 ± 0.18 and 1.7 ± 0.3 μM, respectively), with affinity improvement by, respectively, 8.5‐fold and 3.4‐fold relative to native peptide (K_d = 5.7 ± 1.2 μM).     

Binding modes of Bcl-2 homology 3 (BH3) peptides with anti-apoptotic protein A1 and redesign of peptide inhibitors: a computational study

The interaction between protein and peptide ligand is a challenging problem in molecular biology and drug design. The binding of the Bcl-2 homology 3 (BH3) peptide to the anti-apoptotic protein A1 was revealed as a critical step in the regulation of apoptosis. These BH3 peptides hold high structural similarity, but are diverse in their regulation abilities. Based on molecular simulations and MM-P(G)BSA methods, this work presented a detailed analysis on binding mechanism of the BH3 peptides derived from PUMA and BMF. Residue-level energy decomposition showed that the core regions of BH3 peptides maintain in stable helical conformations and the four conserved hydrophobic residues together with an invariant aspartic acid contribute the major driving force for binding, whereas their two terminal segments exhibit obvious flexibility and various binding modes. Such kind of behavior was suggested as the reason for binding diversity and selectivity of BH3 peptides. As a further step, several BH3-mimetic peptides have been redesigned by computational mutation. Those new peptides showed not only stronger affinities when binding to protein A1, as well transferable binding patterns at some specific positions. A long-range coupling effect was disclosed for BH3 peptides, side-chain orientation and binding contribution of terminal residues were even affected by mutations at large sequence interval. Overall, this work reports that the binding modes of BH3 peptides are primarily dependent on its two terminal segments. The computational methods applied herein are also demonstrated to be of great assistance in the rational design of peptide inhibitors.     

Comparing the binding properties of peptides mimicking the Envelope protein of SARS‐CoV and SARS‐CoV‐2 to the PDZ domain of the tight junction‐associated PALS1 protein

The Envelope protein (E) is one of the four structural proteins encoded by the genome of SARS‐CoV and SARS‐CoV‐2 Coronaviruses. It is an integral membrane protein, highly expressed in the host cell, which is known to have an important role in Coronaviruses maturation, assembly and virulence. The E protein presents a PDZ‐binding motif at its C‐terminus. One of the key interactors of the E protein in the intracellular environment is the PDZ containing protein PALS1. This interaction is known to play a key role in the SARS‐CoV pathology and suspected to affect the integrity of the lung epithelia. In this paper we measured and compared the affinity of peptides mimicking the E protein from SARS‐CoV and SARS‐CoV‐2 for the PDZ domain of PALS1, through equilibrium and kinetic binding experiments. Our results support the hypothesis that the increased virulence of SARS‐CoV‐2 compared to SARS‐CoV may rely on the increased affinity of its Envelope protein for PALS1.     

Five-stranded β-sheet sandwiched with two α-helices: A structural link between restriction endonucleases EcoRI and EcoRV

Examination of crystal structures of restriction endonucleases EcoRI and EcoRV complexes with their cognate DNA revealed a common structural element, which forms the core of both proteins. This element consists of a five-stranded β-sheet and two α-helices packed against it and could be described as α–β sandwich in which helices and β-strands lie in two stacked layers. While the spatial structure of this α–β sandwich is conserved in both enzymes, there are no detectable similarities between amino acid sequences except of a few residues involved in active site formation. Probably, other restriction endonucleases which have similar organization of the active site might possess similar structural element regardless of DNA sequence recognized and recognition elements in the enzyme used. © 1994 Wiley-Liss, Inc.     

Dead‐end elimination with perturbations (DEEPer): A provable protein design algorithm with continuous sidechain and backbone flexibility

Computational protein and drug design generally require accurate modeling of protein conformations. This modeling typically starts with an experimentally determined protein structure and considers possible conformational changes due to mutations or new ligands. The DEE/A* algorithm provably finds the global minimum‐energy conformation (GMEC) of a protein assuming that the backbone does not move and the sidechains take on conformations from a set of discrete, experimentally observed conformations called rotamers. DEE/A* can efficiently find the overall GMEC for exponentially many mutant sequences. Previous improvements to DEE/A* include modeling ensembles of sidechain conformations and either continuous sidechain or backbone flexibility. We present a new algorithm, DEEPer (D ead‐E nd E limination with Per turbations), that combines these advantages and can also handle much more extensive backbone flexibility and backbone ensembles. DEEPer provably finds the GMEC or, if desired by the user, all conformations and sequences within a specified energy window of the GMEC. It includes the new abilities to handle arbitrarily large backbone perturbations and to generate ensembles of backbone conformations. It also incorporates the shear, an experimentally observed local backbone motion never before used in design. Additionally, we derive a new method to accelerate DEE/A*‐based calculations, indirect pruning, that is particularly useful for DEEPer. In 67 benchmark tests on 64 proteins, DEEPer consistently identified lower‐energy conformations than previous methods did, indicating more accurate modeling. Additional tests demonstrated its ability to incorporate larger, experimentally observed backbone conformational changes and to model realistic conformational ensembles. These capabilities provide significant advantages for modeling protein mutations and protein–ligand interactions. Proteins 2013. © 2012 Wiley Periodicals, Inc.     

β‐amyloid model core peptides: Effects of hydrophobes and disulfides

The mechanism by which a disordered peptide nucleates and forms amyloid is incompletely understood. A central domain of β‐amyloid (Aβ21–30) has been proposed to have intrinsic structural propensities that guide the limited formation of structure in the process of fibrillization. In order to test this hypothesis, we examine several internal fragments of Aβ, and variants of these either cyclized or with an N‐terminal Cys. While Aβ21–30 and variants were always monomeric and unstructured (circular dichroism (CD) and nuclear magnetic resonance spectroscopy (NMRS)), we found that the addition of flanking hydrophobic residues in Aβ16–34 led to formation of typical amyloid fibrils. NMR showed no long‐range nuclear overhauser effect (nOes) in Aβ21–30, Aβ16–34, or their variants, however. Serial ¹H‐¹⁵N‐heteronuclear single quantum coherence spectroscopy, ¹H‐¹H nuclear overhauser effect spectroscopy, and ¹H‐¹H total correlational spectroscopy spectra were used to follow aggregation of Aβ16–34 and Cys‐Aβ16–34 at a site‐specific level. The addition of an N‐terminal Cys residue (in Cys‐Aβ16–34) increased the rate of fibrillization which was attributable to disulfide bond formation. We propose a scheme comparing the aggregation pathways for Aβ16–34 and Cys‐Aβ16–34, according to which Cys‐Aβ16–34 dimerizes, which accelerates fibril formation. In this context, cysteine residues form a focal point that guides fibrillization, a role which, in native peptides, can be assumed by heterogeneous nucleators of aggregation.