Sequence-Dependent Heterotypic Assemblies of Intrinsically Disordered Peptides

Abstract

Intrinsically disordered regions are ubiquitous within proteins and critical for protein functions. A recent Cryo-EM study revealed that a pyrene conjugated hexalysine (Pyn-KKKKKK) a pyrene-tyrosines (EY) motif peptide (Pyn-EYEYEY) co-assemble into intrinsically disordered nanofibers through aromatic-aromatic and electrostatic interactions. To probe the sequence effects, we generated a series of peptides by mutating amino acid residues in them and examined their assemblies using negative-stained TEM, fluorescence, and CD spectroscopy. Homochiral and heterochiral mixtures displayed similar morphologies, indicating a role of chirality in their structures. Sequence modifications revealed that aromatic residues patterning governs self-assembly: “XZXZXZ” or “XXXXXZ” (X = charged residue, Z = aromatic residue) favored fiber formation, while “XXXZXZ” promoted pyrene excimer generation. Fiber assembly consistently correlated with an intermediate emission peak at 420–430 nm, suggesting a specific pyrene stacking geometry, particularly in sequences containing aromatic residues and in heterochiral mixtures of oppositely charged peptides. These findings highlights how sequence and chirality modulate aromatic-aromatic and electrostatic interactions in IDPs, providing insights of heterotypic nanofiber formation and confirming the utility of lysine-tyrosine pair in rational design of supramolecular peptide materials that maintain both disorder and self-assembly.

Graphical Abstract

Two oppositely charged intrinsically disordered peptides (IDPs) co-assemble into heterotypic nanofibers. Homochiral mixtures (L/L IDPs) form homogeneous, aligned bundles with more positive IDP. Heterochiral mixtures (L/D IDPs) yield heterogeneous, twisted bundles as positive IDP content increases. The insights on morphologies dictated by sequence and chirality provide a useful guide to design peptide assemblies as biomaterials.

Introduction

This article reports on how sequence composition influences the heterotypic assemblies of inherently disordered peptides (IDPs), a subset of intrinsically disordered regions (IDRs) that lack stable secondary structure. IDRs are critical for diverse protein functions, including tethering at membrane contact sites (MCSs),^[1] acting as entropic barriers to regulate protein orientation and dynamics, facilitating amyloid fiber formation,^[2] driving RNP granule assembly,^[3] and mediating protein condensation and partner recruitment.^[4] Often flanking folded cores, IDRs contain small linear motifs (SLiMs)^[5] that enable higher-order assemblies^[6] through self-oligomerization and liquid–liquid phase separation.^[7] They also underlie transcription factor binding specificity and transcriptional regulation.^[8] Notably, several Csg (curli-specific genes) proteins required for curli amyloid assembly are predicted to be intrinsically disordered.^[9] Furthermore, charged disordered proteins can form hydrogel networks,^[10] highlighting the functional diversity of IDRs in biomolecular assembly.

While considerable progress has been made in understanding IDRs, studies on IDPs remain limited. Experimental studies show IDPs can adopt secondary structures in less polar environments.^[11] drive hierarchical organization as antimicrobial domains,^[12] and form fibers, as in glycine-rich FUS IDPs.^[13] IDPs attached to alkyl chains^[14] enable self-assembly,^[15] and small aggregating fragments of tau are intrinsically disordered.^[16] IDP function roles span the C-terminal linker of Escherichia coli FtsZ,^[17] mineral stabilization,^[18] bone regeneration,^[19] non-endosomal cell entry,^[20] and miRNA complexation,^[21] and attenuation of oncogenic signatures.^[22] Other examples include membrane-binding IDPs,^[23] the Ebola virus VP35 IDP,^[24] antibacterial Pln149,^[25] and redox-sensitive lunasin.^[26] Incorporating non-natural amino acids into eIF4G-derived IDPs promotes disorder-to-order transitions,^[27] while a redesigned IDP disrupts MYC-mediated responses in Arabidopsis.^[28] Hierarchical assembly of short IDPs under ionic or pH changes provides new insights,^[29] and cryo-EM has revealed atomistic insights into heterotypic IDP assemblies for the first time.^[30]

Beyond experimental studies, molecular dynamic (MD) simulations have provided valuable insights into IDPs. They revealed the low critical solution temperature (LCST) behavior of elastin-like IDPs,^[31] trimethylamine N-oxide (TMAO)-induced shifts in monomer populations,^[32] and sequence-dependent effects in retro IDPs^[33]. Simulations have characterized disordered sequences from a DEAD-box protein,^[34] probed dynamic heterogeneity at IDP interfaces,^[35] and examine amyloid-beta (Aβ42) conformations in solution.^[36] Recent work showed IDPs can form complex coacervates or nanovesicles,^[37] with membrane remodeling driven by polyampholyte-lipid and counterion interactions.^[38] Advances in force fields, such as Amber ff24EXP-GA, improved modeling of GxG peptide ensembles in agreement with spectroscopic data.^[39] Recently, Shea et al. highlighted the context dependency of hydrophobicity in IDPs,^[40] underscoring its role in self-assembly and providing a useful guide for IDP exploration.

The above advances in IDP research offer new opportunities to explore their assemblies,^{[15b, 20, 41]} utilizing their dynamic features to expand the functions and applications of peptide assemblies.^[42] While short peptides are typically disordered in solution, cryo-EM has shown that they can adopt ordered conformations in the solid state^[43]. We recently reported context-dependent heterotypic assemblies of IDPs^[30], where some short peptides form β-sheets while others remain disordered within the same fibril, providing structural evidence for short-peptide IDPs. Building on this, we expanded the IDP library to probe sequence-dependent interactions. In addition to previously synthesized L-peptides, we synthesized their D-enantiomers (Pyn-eyeyey, Pyn-eeeyey, Pyn-eeeeey, and Pyn-eeeeee) to examine the impact of chirality on assembly. Although opposite charges in peptides are known to drive self-assembly^{[43b, 44]}, their role in the context of IDPs remain underexplored.

Here, we generated homochiral and heterochiral mixtures of oppositely charged IDPs and observed nanoparticle-to-fiber transitioneds with increasing positively charged peptides, with excess leading to fiber bundling. Heterochiral mixtures formed less uniform fiber networks, exhibiting helical bundling with varied diameters at equimolar ratios and greater disorder with higher tyrosine content (Scheme 1). CD and fluorescence spectroscopy confirmed peptide interactions: mixtures displayed excimer formation (increased E/M ratios), mirrored CD signals, and a consistent 430 nm emission peak correlating with fiber formation (and suggesting close pyrene–pyrene proximity within the assembled structures^[45]), where the 480 nm excimer peak was not essential, as confirmed by TEM.

Scheme 1.

Two oppositely charged IDPs self-assemble to form heterotypic nanofibers or bundles of nanofibers according to their chirality.

To probe assembly rules, we varied lysine content, aromatic motifs, and pH. Both charge density and the presence of aromatic motifs proved critical for generating ordered assemblies. Substitution studies revealed that “XZXZXZ” or “XXXXXZ” (X = charged residue, Z = aromatic residue) arrangements favored ordered fibers, while “XXXZXZ” promoted excimer emission but less structural order. These results highlight how sequence, charge, and π–π interactions govern IDP self-assembly, providing insights for rational design of supramolecular peptide biomaterials.

Results and Discussion

Molecular Design.

As shown in Figure 1, the parent negatively charged peptide sequence for this study is EEEEEE, a disordered region that appears in multiple human proteins.^[46] The corresponding positively charged peptide sequence is KKKKKK, which is also exist as a disordered region in many human proteins.^[47] We previously discussed the presence of EYEYEY, EEEYEY, and EEEEEY sequences in human proteins^[30]. BLAST analysis further revealed that EEEEEF is an IDR in human proteins,^[48] while sequences KYKYKY, KKKYKY, EEEFEF, and EFEFEF lack any experimentally verified structure. These BLAST results confirm the central rationale for the sequence selection: the peptides studied here are derived from IDRs of human proteins. As disordered, yet naturally occurring segments, they may be functionally relevant while offering a versatile platform for probing assembly behavior.

Figure 1.

Molecular structures of the negatively and positively charged IDPs and relevant peptides with aromatic amino acid residues (colors represent pKa of the residues).

To further expand this design, we replaced L-amino acid residues in KKKKKK, EEEEEE, EEEEEY, EEEYEY and EYEYEY by the corresponding D-amino acid residues, which leads to kkkkkk, eeeeee, eeeeey eeeyey and eyeyey, respectively.

Pyrene promotes self-assembly by forming a hydrophobic core within fibers, as supported by previous structural models^[30]. This core organizes otherwise disordered peptide sequences, facilitating intermolecular interactions that drive fiber formation. Conjugating these peptides to pyrene would generate peptides 1-19. Additionally, we reduce the length of the peptide to generate Pyn-KKK (9) or omit the pyrene group to use Ac-KKKKKK (13) and KKK (10) that lack aromatic motifs for testing heterotypical assemblies with negatively charged IDPs.

Homotypic Assemblies.

CAC measurement.

We synthesized the designed peptides using solid-phase peptide synthesis^[49] (Figure S27–S44) and used the pyrene motif fluorescence (Figure S2) to measure their critical aggregation concentrations (CACs)^[51]. All CACs ranged from 300 to 500 μM with minimal variation (Figure 2A and Figure S3), suggesting that these charged IDPs remain soluble in water. Pyn-EYEYEY (2), Pyn-EFEFEF (16),and Pyn-KYKYKY (17), the analogs richest in aromatic residues, exhibited the lowest CACs. In the lysine-tyrosine series, CAC increased as tyrosine content decreased: Pyn-KYKYKY (17) < Pyn-KKKYKY (18) < Pyn-KKKKKY (19). The phenylalanine-glutamic acid sequences showed a similar trend, with Pyn-EFEFEF (14) having the lowest CAC and reduced phenylalanine content raising CACs. These trends highlight the role of hydrophobic residues in promoting aggregation at lower concentrations. However, tyrosine and glutamic acid sequences broke this pattern: Pyn-EEEYEY (7) had the highest CAC among the L-glutamic acid-tyrosine IDPs, while Pyn-eeeyey (11) had the lowest CAC among its D-counterparts. Difference may arise from log-based fitting evaluations or polymorphism of in IDP assemblies.

Figure 2.

(A) The critical aggregation concentrations (CACs) of all IDPs in water at pH = 7. (B) The ratios of pyrene excimer/monomer of multiple concentrations.(C) The ratios of pyrene excimer/monomer for all IDPs at 500 μM^[50]. The pHs of the solutions are all at 7.

E/M Quantification.

We quantified the correlation between pyrene monomer and excimer by calculating the area under the curve (AUC) for monomer (360 nm to 430 nm) and excimer (430 nm to 600 nm), then derived the excimer to monomer ratio (E/M) to generate Figure 2B. As concentration increased, E/M initially decrease to a minimum near 500 μM, then rose, indicating concentration-dependent aggregation that alters the pyrene proximity. At low concentrations (10 to 50 μM), weak fluorescence produced E/M values > 1 because the excimer region spans a broader range. Above 2 mM, particularly in peptides with alternating tyrosine and charged side chains such as Pyn-EYEYEY (2), Pyn-eyeyey (5), and Pyn-KYKYKY (17), excimer emission dominated, showing that the aromatic tyrosine residues promote close packing of pyrene groups at high concentration.

E/M comparison and TEM correlation.

At 500 μM, near their CACs (Figure 2C), Pyn-KYKYKY (17) stands out with an E/M of 1.10, indicating nearly comparable amounts of monomers and excimers. Other IDPs display E/M ratios ≤ 0.50, dominated by monomers. At this concentration, 17 forms fiber bundles visible by TEM (Figure S4), consistent with its high E/M and hydrogen bonding between tyrosine and lysine side chains revealed in cryo-EM.^[30] Most other IDPs form nanoparticles or droplet morphologies (Figure S4), except Pyn-EFEFEF (14), which also forms fibers. Its fluorescence peak at ~430 nm intensifies with concentration (Figure S2L), suggesting a distinct π–π stacking mode, unlike Pyn-EYEYEY (2), which favors 480 nm excimer emission. This packing mode may contribute to the unique fibrillization behavior of 14. Aromatic free analogs, such as Pyn-EEEEEE (1), yield featureless nanoparticle aggregates, while pyrene-conjugated lysines produce droplet morphologies in TEM.

Overall, homotypic assemblies showed stronger pyrene excimer emission in peptides with higher aromatic content. Most IDPs near their CAC adopted disordered morphologies, except for peptides containing multiple KY or EF motifs, suggesting distinct intrinsic interactions.

Homochiral Heterotypic Assemblies.

Charge-Driven Excimer Formation.

Using pyrene fluorescence, we measured the fluorescent spectra of heterotypic IDP mixtures (Figure S5). For negatively charged Pyn-EEEEEE (1), adding 0.1 equivalent of the positively charged Pyn-KKKKKK (3) to a 500 μM solution of 1 decreased in the pyrene monomer peak (360–430 nm), and slightly increased the excimer peak (430–600 nm). The E/M ratio rose from 0.1 (control) to 0.3 (Figure 3B), indicating that charge interaction promotes the excimer formation (Figure 3A). Increasing 3 further enhanced the effect: at a 1:1 ratio, the excimer peak overtook the monomer peak, with E/M rising 10 times to 6.7. At double the concentration of 3, monomer peak nearly vanished, while the excimer dominated with the E/M reaching 42.1, confirming that stronger positive charge enhances the conversion of monomers into excimers.

Figure 3.

(A) Fluorescent spectra, (B) the ratios of pyrene excimer/monomer and (C) TEM images of 500 μM Pyn-EEEEEE (1) mixed with 0.1, 0.2, 0.5, 1, and 2 equivalent of Pyn-KKKKKK (3). The pH of the mixture is 7.

Fiber Formation and Phase Separation.

TEM images confirm the fluorescence and E/M trends (Figure 3C). Pyn-EEEEEE (1) forms nanoparticles at 500 μM, and the morphology remains largely unchanged with the addition of 0.1 equivalence of Pyn-KKKKKK (3). At 0.2 and 0.5 equivalence, single fibers emerge from the nanoparticle clusters. An equimolar mixture produced more fibers, consistent with the stronger excimer peak in fluorescence. With a twofold excess of 3, fibers give way to large bundled aggregates, cause immediate turbidity. After standing overnight, a clear supernatant separates from an opaque bulk phase, indicating condensed assemblies rather than amorphous precipitates. This phase separation arises from electrostatic interactions that promote fiber bundling. A similar pattern occurs with the Pyn-eeeeee (4) and Pyn-kkkkkk (6) (Figure S8), where increasing 6 drives to progressive fiber bundling.

Tyrosine Suppresses Pyrene Fluorescence.

Because the phenolic group in tyrosine residue quenches pyrene fluorescence, peptides containing tyrosine showed markedly lower fluorescent intensity than those without (Figure S2). As a result, mixtures of tyrosine-containing peptides with Pyn-KKKKKK (3) only weakly follow the excimer/monomer trends observed for mixtures of Pyn-EEEEEE (1) and Pyn-KKKKKK (3). Consistent with prior reports^[30], L-homochiral mixtures of Pyn-EEEEEE (1), Pyn-EEEEEY (8), Pyn-EEEYEY (7), and Pyn-EYEYEY (2) with 3 show characteristic morphologies, fluorescence, and CD features driven by pyrene stacking and electrostatic interactions. The corresponding D-homochiral mixtures exhibited the same overall trends (Figure S7–S14), reflecting mirrored assembly behavior.

CD corresponds with TEM.

CD spectra (Figure 4A) showed that neither Pyn-eyeyey (5) nor Pyn-EYEYEY (2) alone displayed ordered features. Adding positively charged IDPs in homochiral mixtures induced peaks at 270–290 nm, with negative signals for L-tyrosine and positive for D-tyrosine^[52], along with a ~280 nm band partly attributed to pyrene stacking^[53]. Equimolar mixtures produced the strongest signals, which decreased at 0.5 and 2 equivalents positively charged IDP partners, consistent with similar assembly processes across enantiomers.

Figure 4.

(A) CD spectra and (B) TEM images of 500 μM Pyn-EYEYEY (2) mixed with 0.5, 1, and 2 equivalent of Pyn-KKKKKK (3) (solid lines), and 500 μM Pyn-eyeyey (5) mixed with 0.5, 1, and 2 equivalent of Pyn-kkkkkk (6) (dashed lines).

The spectra symmetry corresponded with the TEM morphological (Figure 4B). Both Pyn-EYEYEY (2) and its D-enantiomer Pyn-eyeyey (5) formed nanoparticles alone, but 0.5 equivalents of the positively charged IDP (2 + 3 or 5 + 6) promoted fiber formation. At equimolar ratios, these fibers coalesced, and at two equivalents of Pyn-KKKKKK (3) or Pyn-kkkkkk (6), they aligned into densely packed fiber bundles.

The morphological similarities persisted when one or two tyrosines were replaced with glutamic acids, yielding Pyn-EEEYEY (7) and Pyn-EEEEEY (8) and their D-enantiomers Pyn-eeeyey (11) and Pyn-eeeeey (12). In both L-homochiral mixtures (7 + 3, and 8 + 3) and D-homochiral mixtures (11 + 6, and 12 + 6), excess positively charged IDP (3 or 6) consistently promoted ordered fibers that bundled further (Figure S11–S14).

Homochiral mixtures involving 7, 8, 11 or 12 showed weak CD signals (<5 mdeg) and lacked the pronounced spectral symmetry observed in Figure 4A (Figure S6). Without atomically resolved structures, the precise molecular arrangement remains undetermined.

Fluorescence of D-enantiomers.

Fluorescence spectra of the homochiral D-enantiomer mixtures closely mirrored those of their L counterparts, with only minor intensity differences across stoichiometric ratios, indicating broadly similar assembly behavior (Figure S4). In mixtures containing three tyrosines (2+3 or 5+6), fluorescence was largely quenched, yet the E/M ratio increased (Table S1), suggesting monomer depletion and separation of higher-order assemblies from the solution. In contrast, Pyn-EEEYEY (7) + Pyn-KKKKKK (3) and Pyn-eeeyey (11) + Pyn-kkkkkk (6) exhibited stronger excimer emission (Figure S4E and S4H), likely due to glutamic acid substitution for tyrosine, which increased pyrene separation and favored excimer formation. Mixing Pyn-EEEEEY (8) or Pyn-eeeeey (12) with their homochiral positively charged IDP partners produced a distinct 430 nm peak (Figure S4I and S4L), intermediate between the monomer (374 nm) and excimer (480 nm), possibly representing an excited-state intermediate in which pyrene chromophores are positioned too far for true excimer formation yet closer than isolated monomers.

Effects of Lysine Content and Aromatic Stacking.

To assess the effect of lysine content and electrostatic density, we synthesized Pyn-KKK (9) with fewer lysine residues. To evaluate the role of aromatic stacking, we prepared a non-pyrene analog, Ac-KKKKKK (13), and used commercially available KKK (10) as a shorter sequence analog of Pyn-KKKKKK (3). These analogs, when mixed with negatively charged IDPs, are unable to produce the fiber formation observed with 3 or its enantiomer, yet the appearance of bulky aggregates upon their addition indicates some degree of interaction. (Figures S18–S20).

Fluorescence spectra revealed distinct assembly behavior of peptide mixtures (Figure 5). For Pyn-EEEEEE (1), adding Ac-KKKKKK (13) or KKK (10) produced little change, while Pyn-KKK (9) slightly increased the excimer peak, raising the E/M ratio from 0.1 (control) to 0.3–0.4. In contrast,, mixing 1 with Pyn-KKK (10) resulted in a pronounced, concentration-dependent increase in the excimer peak: E/M of 3.7 at one equivalent and 10.4 at two equivalents. For the aromatic-rich sequences Pyn-EYEYEY (2), all three positively charged analogs promoted significant excimer formation through aromatic–aromatic and lysine-tyrosine interactions. Among the three, Pyn-KKK (9) induced the strongest excimer signal, followed by Ac-KKKKKK (13), and then KKK (10), indicating the combined importance of the aromatic motif and charge density. Notably, 9 not only enhanced the excimer peak but also reduced the monomer peak, suggesting closer pyrene–pyrene interactions. In contrast, mixtures with non-pyrene analogs increased the excimer signal in proportion to the added equivalents, but the monomer peak remained largely unchanged.

Figure 5.

Fluorescent spectra of 500 μM (A) Pyn-EEEEE (1) and (B) Pyn-EYEYEY (2) mixed with 1 and 2 equivalence of Pyn-KKK (9), Ac-KKKKKK (13) and KKK (10) in water. (C) Ratios of pyrene excimer to monomer emission (E/M) for each sample.

These observations suggest that direct pyrene–pyrene interaction is essential for excimer formation. While non-pyrene analogs can promote molecular crowding and support aggregation, they appear insufficient to bring pyrene moieties into the proximity required for efficient excimer emission. Furthermore, the presence of aromatic residues and adequate peptide length appears critical for fiber formation. Positively charged analogs that lack aromatic groups or are too short likely cannot engage in the π–π or electrostatic interactions necessary for ordered assembly with negatively charged partners.

pH modulation.

pH strongly influences electrostatically driven self-assembly. To assess this, we examined the heterotypic fiber-forming pair, Pyn-EYEYEY (2) and Pyn-KKKKKK (3) at pH 4.0, 6.4, 7.6, and 8.5 (Figure S22–S23). These values span the pKa of glutamic acid while keeping lysine and tyrosine protonated. Pyn-EYEYEY (2) displayed pH-sensitive assembly, forming fibers with monomer-to-excimer fluorescence transition and β-sheet characteristics at pH 4.0 (below the pKa of glutamic acid). A positive peak near 280 nm suggests changes in tyrosine-pyrene environment or peptide conformation under this protonation state. At higher pH, the morphology became disordered, fluorescence shifted to monomer emission, and CD spectra lost secondary structure. In contrast, Pyn-KKKKKK (3) remained positively charged from pH 4.0–7.6, forming micelles with monomer fluorescence and random coil CD signatures. At pH 8.5, nanosheets appeared with slight excimer induction in fluorescence and a small CD shift toward β-sheet character, consistent with partial lysine deprotonation (pKa ≈ 9).

Mixing Pyn-EYEYEY (2) and Pyn-KKKKKK (3) under different pH conditions resulted in clear morphological transitions (Figure S24). At pH 4.0, 0.5 equiv. of 3 induced fibers similar to those of 2 alone but slightly more bundled; higher ratios of 3 had little effect. At pH 6.4, thinner fibers appeared alongside thicker fiber bundles, with bundling increasing at higher concentrations of 3. At pH 7.6 and 8.5, bundling diminished and individual fibers became more distinct, consistent with earlier findings that higher pH favors separated fibers^[30].

Fluorescence spectroscopy supported these observations (Figure S25). At pH 4.0, the spectral shape remained unchanged as Pyn-KKKKKK (3) increased, but overall fluorescence rose due to the intrinsic fluorescence of 3, without the tyrosine-induced quenching seen in Pyn-EYEYEY (2). At pH 6.4, excimer fluorescence grew relative to monomer fluorescence as 3 increased, though the intensity stayed below 40 A.U., likely due to significant precipitation from self-assembly. At pH 7.6 and 8.5, stronger self-assembly further reduced fluorescence intensity below 30 A.U, eliminating clear pyrene signals.

CD spectroscopy provided further insights (Figure S26). At pH 4.0, the CD signal closely resembled that of Pyn-EYEYEY (2) alone, with only a slight decrease in intensity and a red shift as Pyn-KKKKKK (3) was added, indicating subtle changes in pyrene electronic transitions and molecular geometry. At pH 6.4, a characteristic tyrosine-pyrene peak at 280 nm became evident and intensified with increasing 3, consistent with positively charged lysines attracting hydrophobic motif via cation–π interactions^[54], positioning pyrene near tyrosine residues and promoting strong exciton coupling. At pH 7.6, a distinct negative peak at 250 nm emerged, enhancing high-energy π–π* transitions and indicating reduced pyrene–tyrosine stacking due to decreased lysine charge and an altered local environment. At pH 8.5, the 250 nm negative peak diminished while the 280 nm peak re-emerged, suggesting re-alignment of tyrosine and pyrene dipoles toward a more favorable coupling geometry^{[53, 55]}.

Overall, pH critically modulates the self-assembly of Pyn-EYEYEY (2) and Pyn-KKKKKK (3), influencing both morphology and spectroscopic properties. At low pH, fiber formation relies on hydrogen bonding and charge interactions with minimal structural changes, while higher pH enhances electrostatic interactions, leading to more distinct, well-separated fibers. Fluorescence and CD spectra data confirm these trends, highlighting the role of aromatic interactions and secondary structure in driving assembly.

Lysine-Tyrosine Interactions.

Previous atomic structures of fibers from Pyn-EYEYEY (2) and Pyn-KKKKKK (3) suggest that lysine–tyrosine interactions significantly contribute to fiber formation^[30]. To further probe this and the role of glutamic acid–lysine electrostatics, we examined lysine–tyrosine sequence analogs Pyn-KYKYKY (17), Pyn-KKKYKY (18), and Pyn-KKKKKY (19), counterparts to glutamic acid–tyrosine sequences (2, 7, and 8). Replacing one terminal lysine with tyrosine, 19 behaved like 3, with similar CAC, E/M values, and micellar morphologies. At 500 μM, 18 formed both nanoparticles and nanosheets, indicating polymorphic packing. With three tyrosine residues, 17 self-assembled readily without oppositely charged IDPs, likely through its intrinsic K-Y interactions. Mixing with Pyn-EEEEEE (1), 17 formed bundled, tangled fibers versus the straight fibers of 17 alone. Reducing the tyrosine content altered this morphology: 18 with 1 produced clustered nanospheres, whereas 19 with 1 formed fibers (Figure S20).

Fluorescence spectroscopy further confirmed these differences (Figure S21A). Tyrosine-rich peptides showed quenched signals, but mixing with Pyn-EEEEEE (1) reduced monomer peak and induced excimer signals. Pyn-KKKYKY (18) showed the strongest excimer peak upon mixing, correlating with bead-like nanostructures in TEM. In contrast, 19 mixtures showed reduced signals, likely from precipitation, while 17 mixtures exhibited a slight increase due to 1’s lack of tyrosine quenching.

CD spectroscopy further clarified secondary structures (Figure S21B). Pyn-KYKYKY (17) exhibited β-sheet features, which weakened upon mixing with Pyn-EEEEEE (1), accompanied by a tyrosine-pyrene associated negative peak at 280 nm. This suggests that introducing electrostatic interactions with 1 altered the spatial arrangement of pyrene and tyrosine groups. Pyn-KKKYKY (18) showed no defined structural features, with only a weak signal at 210 nm in equimolar mixtures with 1. Pyn-KKKKKY (19) exhibited increased β-sheet signals and a 280 nm tyrosine-pyrene peak when mixed with 1, though both remained weak (< 5 mdeg). This indicates that introducing electrostatic counterparts induced a more ordered arrangement during assembly.

Heterochiral Heterotypic Assemblies.

Heterochiral mixtures of heterotypic assemblies showed morphological and spectral patterns similar to their enantiomeric counterparts, consistent with the trends in homochiral heterotypic assemblies. Equimolar heterochiral pairs generally exhibited mirrored CD spectral patterns, with slight wavelength shifts indicating minor deviations from perfectly mirrored assembly (Figure 6A and S6). The peak near 280 nm showed opposite signs compared to those associated with L- or D-tyrosines in Figure 4A, suggesting reversed aromatic stacking orientation driven by the partner’s opposite chirality. A weak additional peak near 255 nm likely arose from pyrene stacking with contributions from aromatic amino acid residues.

Figure 6.

(A) CD spectra and (B) TEM images of 500 μM Pyn-EEEEEE (1), Pyn-EEEEEY (8), Pyn-EEEYEY (7), Pyn-EYEYEY (2) or Pyn-eeeeee (4), Pyn-eeeeey (12), Pyn-eeeyey (11), Pyn-eyeyey (5) mixed with equal molar of heterochiral Pyn-KKKKKK (3) or Pyn-kkkkkk (6).

TEM imaging (Figure 6B) showed that L-peptide Pyn-EEEEEE (1) with the D-peptide Pyn-kkkkkk (6) formed helical twisted fiber bundles, closely resembling those from the D-peptide Pyn-eeeeee (4) with the L-peptide Pyn-KKKKKK (3) (Figures S7, S8). Substituting one glutamic acid with tyrosine (Pyn-EEEEEY, 8) and pairing with 6 yielded more prevalent helical and tangled structures. Varying the ratio of 6 preserved the overall morphology but reduced nanoparticle content relative to nanofibers (Figure S13). Its D-enantiomer (Pyn-eeeeey, 12) with 3 produced similarly tangled bundles (Figure S14). Replacing two glutamic acids with tyrosine (Pyn-EEEYEY, 7; Pyn-eeeyey, 11) and mixing with 6 or 3 generated larger helical bundles entangled into dense networks (Figures S11, S12). With three tyrosines, Pyn-EYEYEY (2) mixed with 6 produced less uniform fibers, with more broken strands and unordered aggregates (Figure S9). The enantiomeric pair Pyn-eyeyey (5) with 3 showed even stronger effect, forming disorganized chunks rather than elongated fibers (Figure S10).

Fluorescence spectra support these findings (Figure S5). In heterochiral mixtures, the intermediate emission peak at 430 nm, between the pyrene monomer and excimer peaks, was strongly enhanced, indicating molecular orientations where pyrene units adopt intermediate proximities. Heterochiral mixtures of the positively charged IDP with three-tyrosine sequences (2 or 5) also showed pronounced excimer formation, consistent with stronger aromatic interactions from increased tyrosine content. While homochiral mixtures (e.g., Pyn-EEEEEE+Pyn-KKKKKK, 1+3) induced excimer emission at 480 nm, heterochiral pairs (e.g., Pyn-eeeeee+Pyn-KKKKKK, 4+3) favored arrangements that enhanced 430 nm emission instead. These results suggests that heterochirality alters molecular packing, shifting pyrene stacking and local polarity.

Phenylalanine Analogs.

To further probe electrostatic and aromatic interactions in the IDPs, we synthesized phenylalanine-containing analogs. Pyn-EFEFEF (14) alone formed fibers, exhibited quenched fluorescence, and showed an induced 420 nm peak (Figure S2L), indicating strong intrinsic aromatic interactions. Adding positively charged IDPs, either homochirally with Pyn-KKKKKK (3) or heterochirally with Pyn-kkkkkk (6), decreased monomer emission and increased excimer intensity (Figure S5U–V). TEM confirmed fibrous morphologies, with increased aggregation at higher equivalents of positively charged IDPs (Figure S17). CD spectra showed minimal differences between homochiral and heterochiral mixtures (Figure S6U–V). Notably, 14 alone exhibited two distinct peaks at 240 nm and 250 nm; with increasing positively charged IDPs, the 240 nm band shifted to 250 nm, suggesting that electrostatic effects altered pyrene–phenylalanine interactions by reposition or partially bury certain aromatic residues.

Replacing tyrosine with phenylalanine in Pyn-EEEYEY (7) yielded Pyn-EEEFEF (15). In homochiral mixtures with 3, the monomer peak dropped sharply while a strong 480 nm excimer peak emerged, raising the E/M ratio to 33 at 2 equivalents, compared to 0.08 without 3 (Figure S5S and Table S1). In heterochiral mixtures with Pyn-kkkkkk (6), the monomer peak decreased but no clear 480 nm peak appeared; instead, a 430 nm emission emerged (Figure S5T). CD spectra of the homochiral mixture showed minimal signals (< 10 mdeg) (Figure S6S), while the heterochiral mixture produced a strong positive peak near 280 nm (41 mdeg; Figure S6T). Overall, 15 displayed spectral behavior similar to 7 when mixed with positively charged IDPs, but its fluorescence intensity was higher. TEM revealed distinct morphologies (Figure S16): Pyn-EEEFEF (15) alone formed nanoparticles (Figure S4); with 3, nanoparticles aggregated into larger, interconnected clusters; with 6, bundles and nanoparticles co-existed, with bundles thickening at higher cationic IDP levels. These differences from tyrosine analog 7 likely reflect the packing preferences of phenylalanine and the absence of tyrosine-lysine hydrogen bonding.

Fluorescence spectra of Pyn-EEEEEF (16) mixed with 3 reduced monomer and slightly increased excimer emission at 480 nm relative to 16 alone (Figure S5Q). In heterochiral mixtures with Pyn-kkkkkk (6), 16 exhibited a distinct 430 nm intermediate peak, indicative of specific pyrene interactions (Figure S5R). This behavior closely resembled that observed for the heterochiral mixture of 6 with its tyrosine-containing counterpart Pyn-EEEEEY (8) (Figure S5), suggesting structural similarities in the underlying interactions. CD spectra of homochiral mixtures of Pyn-KKKKKK (3) to Pyn-EEEEEF (16) revealed negative signals at 220–230 nm upon addition of 0.5 and 1.0 equivalents of 3 (Figure S6Q). In heterochiral mixtures, a positive peak emerged at 190 nm with 0.5 equivalents of Pyn-kkkkkk (6), shifting toward 200 nm with increasing 6’s concentration (Figure S6R). These observations suggest a conformational transition from a random coil to a β-sheet structure upon binding of the positively charged IDP to 16. TEM results revealed that 16 formed both nanoparticles and nanofibers when mixed with 0.5 equivalents of Pyn-KKKKKK (3). At equimolar concentrations, the fibers exhibited increased entanglement, with short fiber protrusions visible along fiber edges and in the background. At 2.0 equivalents, these short fibers assembled into thicker aggregates. Adding Pyn-kkkkkk (6) to 16 produced fibers of varying diameters, with increased bundling observed at higher concentrations of the positively charged IDP (Figure S15).

Discussion.

Mixing Pyn-EEEFEF (15) with Pyn-KKKKKK (3), and similarly mixing Pyn-KKKYKY (18) with Pyn-EEEEEE (1), produced strong pyrene excimer emission, whereas the individual IDPs displayed only monomer fluorescence. However, as discussed earlier, excimer peaks do not necessarily indicate ordered fiber formation. Among the phenylalanine analogs 14-16, only 15 failed to form fibers when homochirally mixed with oppositely charged IDPs, instead producing aggregates of nanoparticles. 18 showed a similar pattern when compared to Pyn-KYKYKY (17) and Pyn-KKKKKY (19). These results suggest that sequence patterning influences self-assembly, with an “XXXZXZ” arrangement being less favorable for ordered fiber assembly than “XZXZXZ” or “XXXXXZ.”

Pyn-EEEYEY (7) and Pyn-eeeyey (11) represents an exception, as they formed fibers when homochirally mixed with positively charged IDPs despite following the “XXXZXZ” pattern. Differences in Y–K interactions between the two components likely explain this behavior.

In peptide mixture systems containing aromatic residues (Y or F), we consistently observed an intermediate fluorescence peak at ~420–430 nm, except for Pyn-EEEFEF (15) and Pyn-KKKYKY (18) with their oppositely charged homochiral partners. In contrast, heterochiral mixtures always exhibited this intermediate peak. Notably, mixtures that displayed this intermediate peak always formed ordered fiber morphologies in TEM, suggesting that the intermediate peak, which indicates specific pyrene proximity, may be a prerequisite for fiber formation. This hypothesis, however, requires further validation through more cryo-EM structures and detailed molecular modeling.

Overall, CD and fluorescence data aligned with the TEM morphologies, though interpretation requires careful consideration due to solubility differences, aggregate exclusion, intrinsic fluorescence variations, and complex peptide interactions, particularly in mixtures where tyrosine and phenylalanine CD peaks may overlap with pyrene signals in the 250–300 nm region.

Conclusion

In summary, this study extensively examined the role of sequence patterning, aromatic residue content, and peptide chirality collectively govern the self-assembly of IDPs through the interplay of electrostatic and aromatic–aromatic interactions. Unlike the alternating charged–aromatic peptides such as (FKFE)n reported before^[56], this study places opposite charges on separate peptides and organizes polar and nonpolar residues in blocks rather than in an alternating pattern, leading to distinct sequence-dependent heterotypic assemblies. Our findings, complementing the reported cryo-EM structure of such assemblies^[30], emphasize the critical role of chirality, lysine–tyrosine interactions, charge distribution, aromatic sidechain arrangement, and pH in determining morphology and stability. Notably, the intermediate fluorescence peak at 420–430 nm consistently correlated with fiber formation, suggesting a specific pyrene proximity essential for π–π stacking. Although, in some cases, the D-enantiomers of IDPs exhibit minor differences in behavior compared to their corresponding L-enantiomers, we speculate that these differences likely result from the slightly lower enantiomeric purity of D-amino acids in the starting materials or the kinetics of the assembling process. Overall, this work contributes to the molecular-level understanding of how IDP sequences and chirality dictate supramolecular architecture and may provide certain designing principles for the rational design of functional peptide-based biomaterials.

Supplementary Material

Supporting information for this article is given via a link at the end of the document.

Experimental procedures, peptide synthesis and LC/MS characterization data, additional TEM images, fluorescence and CD spectra, CAC determinations, and a summary table of morphological measurements and spectral analyses for individual peptides and their mixtures.