Backbone engineering in the hydrophobic core of villin headpiece

Abstract

Changing the backbone connectivity of proteins can impart useful new traits while maintaining essential structural and functional features. In design of artificial proteomimetic agents, backbone modification is usually isolated to sites that are solvent-exposed in the folded state, as similar changes at buried residues can alter the fold. Recent work has shown that core backbone modification without structural perturbation is possible; however, the modifications in that study were consistently destabilizing and made in a prototype of exceptionally high conformational stability. Here, we report efforts to broaden the scope and improve the efficacy of core backbone engineering by applying it to the C-terminal subdomain of villin headpiece. A series of variants are prepared in which different artificial residue types are incorporated at core positions throughout the sequence, including a crucial aromatic triad. Impacts on folding energetics are quantified by biophysical methods, and high-resolution structures of several variants determined by NMR. We go on to construct a variant with ∼40% of its core modified that adopts a fold identical to the prototype while showing enhanced thermodynamic stability.

---

Introduction

Engineering the chemical composition of peptides and proteins can yield biomimetic macromolecules with intricate folded structures and potent biological functions.^1–4 Expanding beyond ribosomally encoded amino acids as the monomer pool for construction of protein mimetics has been used to hone structure–activity relationships as well as facilitate biomedical applications.^5,6 One chemical change commonly applied in such work is the incorporation of amino acids with modified backbones.^7–9 Examples include alteration of stereochemistry (i.e., D-α-residues), addition of new substituents (e.g., N-Me-α-, N-amino-α-, C_α-Me-α-residues), and backbone elongation (e.g., β-residues). Compared to side-chain substitution, altering backbone connectivity can lead to large effects on folded structure, folding energetics, and biological properties from relatively small chemical changes.¹⁰

While modification of backbone composition has proven utility in construction of artificial peptide and protein mimetics, moving from an isolated secondary structure (e.g., a helical peptide) to a higher order tertiary fold (e.g., a small protein) as the target for mimicry poses challenges.¹ In tertiary structures, local conformational preferences must be considered alongside a complex network of weak, long-range interactions critical to the fold. Related to this challenge, backbone modification in the hydrophobic core of sequences with tertiary folding patterns has been shown to lead to a variety of unexpected effects.^11–13

Seeking to broaden the scope of strategies for protein mimetic design to encompass reliable modification within the hydrophobic core, we recently reported efforts to engineer the backbone of the G-related albumin-binding module from bacterial protein PAB (1PRB).¹⁴ In that work, we showed judicious backbone alteration in the core could be structurally tolerated but was consistently destabilizing. While promising, these results left open questions. Would the design strategies prove viable in a protein domain with a more delicate fold than 1PRB, which has a thermal unfolding temperature ∼90 °C? Further, would expanding the pool of artificial monomers used in the construction of core-modified protein mimetics enable the creation of variants with enhanced conformational stability relative to the prototype? Here, we report efforts to address these questions through the application of backbone modification at core positions of the C-terminal subdomain of villin headpiece (HP35, Fig. 1).^15–17

Fig. 1 (A) Sequence of HP35. Residues shaded in gray make up the hydrophobic core, and underlined residues indicate sites selected for backbone modification. (B) Crystal structure of HP35 (PDB 3TRY); core side chains are shown as spheres. (C) Chemical structures of an α-residue and three artificial monomer types used in core-modified HP35 analogues. The R group in each C_α-Me-α and β³ residue matches the side chain of the α-residue in HP35 it replaces.

HP35 is one of the smallest domains that folds autonomously and enjoys a rich history as a model for fundamental studies of protein folding.^18–29 In the realm of protein mimetics, backbone modification in HP35 has been made at solvent-exposed sites, and the structural and energetic effects of β-residue and C_α-Me-α-residue incorporation described by Gellman³⁰ and our lab,³¹ respectively. Backbone modification to the HP35 core has not been reported; however, Berlicki described a computationally designed de novo protein mimetic that contained several carbocyclic β-residues in its core and was intended to mimic HP35.¹³ Interestingly, this sequence was observed to form a domain-swapped dimer fold that deviated significantly from the design.¹³ In the present study, we set out to systematically examine the impacts of altered backbone composition in the hydrophobic core of HP35 on folded structure and stability of the protein. Our results show artificial residues can be accommodated at several core positions in this system—individually or in tandem—without altering folded structure and that some modifications enhance folded stability relative to the canonical backbone.

Results and discussion

Design of core-modified HP35 variants

The hydrophobic core of HP35 (Fig. 1(A) and (B)) is primarily composed of three phenylalanine side chains (F6, F10, F17), which are located within the first and second helices and form an aromatic cluster that is critical to stabilizing the tertiary fold of the domain.³² Solvent-accessible surface area (SASA) analysis shows these three residues are each >90% buried. In addition to the phenylalanine triad, M12 in the loop between helix 1 and 2, L20 in the loop between helix 2 and 3, and L28 in helix 3 are partially buried (74–81%). Finally, polar residues Q25 and K29 in helix 3 occupy hydrophobic core-flanking positions and are also partially buried (67–85%). Based on the above analysis, we selected five sites from the core of HP35 as targets for backbone modification (Fig. 1(A)): the key aromatic cluster F6, F10 and F17, buried aliphatic L28, and interfacial charged K29.

Three different artificial residue types were employed in design of core-modified HP35 variants: β³-residues, chiral C_α-Me-α-residues, and the cyclic β-residue trans-2-aminocyclopentane carboxylic acid (ACPC) (Fig. 1(C)). All three of these classes are known to be well accommodated in solvent-exposed helical contexts in a variety of heterogeneous-backbone tertiary structure mimetics, including HP35.^30,31 Relative to a canonical α-residue, a methylene group is inserted between the carbonyl and C_α in a β³ monomer, which adds a rotatable bond and increases backbone flexibility. In a C_α-Me-α residue, the H_α is replaced by a methyl group, which restricts backbone conformational freedom. In the cyclic β-residue ACPC, the cyclopentane ring constrains the central torsion, which increases rigidity relative to the β³ residue at the expense of a lost side chain. Collectively, the application of these three monomer types allows for a comparative analysis of the importance of rigidity and side chain functionality when modifying the backbone at buried core positions.

For HP35 variants modified at core position L28 and core-flanking position K29, the canonical α-residue was replaced by a β³-residue or ACPC. As a control to understand context-dependent effects of these modifications at a comparable non-core site, we incorporated the same monomer types at neighboring solvent-exposed residue K30. Given the crucial role of side-chain packing of the phenylalanine triad in HP35, backbone substitutions at those positions were limited to β³ and C_α-Me-α monomers that maintain the side chain.

The above design considerations yielded a set of 12 single-site backbone substitution variants of HP35 for analysis. As data for the β³K30 variant have been reported,³⁰ the remaining 11 peptides were prepared by Fmoc solid-phase methods. All syntheses were completed without major issues and yielded good crude purities (Fig. S1). Peptides were purified by preparative reverse phase high-performance liquid chromatography (HPLC). The identity of isolated products was confirmed by electrospray ionization mass spectrometry (ESI-MS) and purity assessed by analytical HPLC (Fig. S2–S13) prior to biophysical and structural characterization. To avoid complications from oxidation during storage and handling, M12 was replaced by norleucine in the HP35 prototype and all variants.

β-Residue substitution at a core, core-flanking, and solvent-exposed site

We first examined folding behavior of the HP35 variants modified in helix 3, comparing effects of β³ or ACPC substitution at three neighboring sites: core L28, core-flanking K29, and solvent-exposed K30. Results for the variants were benchmarked against our recently reported biophysical data set for prototype HP35.³¹ Circular dichroism (CD) scans at 20 °C on samples consisting of 50 µM peptide in 50 mM phosphate pH 7 show minima at ∼208 and ∼222 nm consistent with the predominantly helical fold of the prototype; however, the magnitude of the CD peaks in β³L28 and β³K29 is slightly attenuated (Fig. 2(A)).

Thermal unfolding experiments monitoring CD signal at 222 nm as a function of temperature exhibit cooperative unfolding transitions in all cases (Fig. 2(B)), with temperature midpoint (T_m) values ranging from 47–69 °C (Table 1). Backbone modification at core position L28 results in a consistent reduction in the thermal stability of the fold (ΔT_m −19 °C for β³L28 and −14 °C for ACPC28). At core-flanking position K29, variant β³K29 is comparably destabilized (ΔT_m −15 °C), while ACPC29 shows a small increase in thermal stability relative to prototype (ΔT_m +3 °C). For solvent-exposed position 30, ACPC incorporation is slightly destabilizing (ΔT_m of −2 °C), and prior results indicate β³ substitution is moderately destabilizing (ΔT_m of −12 °C).³⁰

Fig. 2 CD scans at 20 °C (A) and thermal melts (B) for HP35 and indicated variants. Conditions: 50 µM peptide in 50 mM phosphate buffer, pH 7.

Table 1 Biophysical properties of HP35 and variants^a

	Tm (°C)	ΔTm (°C)	(kcal mol^−1)	(kcal mol^−1)
a Temperature midpoint (Tm) of the thermal unfolding transition and change in Tm relative to HP35 (ΔTm) determined by variable temperature CD (conditions: 50 µM peptide in 50 mM phosphate buffer, pH 7). Folding free energy () and change in folding free energy relative to HP35 () determined via chemical denaturation with guanidinium chloride monitored by tryptophan fluorescence (conditions: 10 µM peptide in 50 mM phosphate buffer, pH 7, 25 °C). Uncertainties for Tm and are standard errors from the fits and uncertainties for ΔTm and determined by error propagation.b Data sourced from published report;^30 the sequence (M12/H27) and buffer composition (acetate pH 5) differed slightly from the present work, so the ΔTm value is reported relative to the prototype Tm from that study.
HP35	65.6 ± 0.7	—	−2.14 ± 0.12	0
β^3L28	47.0 ± 2.6	−19 ± 3	—	—
ACPC28	51.2 ± 1.3	−14 ± 1	—	—
β^3K29	50.6 ± 1.7	−15 ± 2	—	—
ACPC29	68.7 ± 0.7	+3 ± 1	−2.38 ± 0.15	−0.2 ± 0.2
β^3K30	57.2 ± 1.7^b	−12 ± 2^b	—	—
ACPC30	63.3 ± 0.8	−2 ± 1	—	—
β^3F6	36.3 ± 3.5	−29 ± 4	—	—
β^3F10	47.3 ± 2.5	−18 ± 3	—	—
β^3F17	—	—	—	—
α^MeF6	71.5 ± 0.7	+6 ± 1	−3.63 ± 0.26	−1.5 ± 0.3
α^MeF10	45.3 ± 5.0	−20 ± 5	—	—
α^MeF17	57.4 ± 1.3	−8 ± 1	−0.90 ± 0.21	+1.2 ± 0.2
α^MeF6/α^MeF17/ACPC29	—	—	−3.00 ± 0.24	−0.9 ± 0.3

---

To determine the effects of backbone modification at core residues on the folded structure of HP35, we subjected the helix-3 variants to analysis by NMR spectroscopy. 1-Dimensional ¹H and 2-dimesional ¹H/¹H COSY, TOCSY, and NOESY spectra were recorded for β³L28, β³K29, ACPC28, ACPC29, and ACPC30 at 1 mM peptide in 9 : 1 H₂O/D₂O at pH 5 (Fig. S15–S20). Assignment of proton resonances followed by simulated annealing with restraints derived from the NMR measurements yielded a 10-model structure ensemble for each peptide (Fig. 3 and Tables S1–S5).

Fig. 3 NMR structure ensembles for HP35 (PDB 9MF7) and variants. Backbone rmsd (average and standard deviation for residue range 1–33) for overlay of each ensemble to the X-ray structure of HP35 (PDB 3TRY) is shown in parentheses.

All the helix-3 variants adopt tertiary folds close to that of HP35 (0.5–0.8 Å backbone rmsd to the X-ray structure). Further support for this finding comes from observed backbone H_α chemical shifts throughout the domain, which are also very similar to the prototype (Fig. S27). Inspection of hydrophobic core packing shows that modification at L28 does not perturb the key aromatic triad; however, the artificial residues pack less efficiently (94%, 89%, and 70% burial for the α, ACPC, and β³ residue at position 28, respectively) (Fig. S28). This reduction in packing efficiency may give rise to the thermal stability loss observed by CD. At core-flanking position 29, both the β³K side chain as well as the ACPC ring engage core residues F10 and F17 in a manner similar to K29 in HP35.

From a structural perspective, all the modifications examined in helix 3 of HP35 are well tolerated. Even when backbone composition is altered at buried aliphatic L28, the variant folded structure is indistinguishable from the prototype. With respect to stability, effects of backbone modification vary with monomer type and sequence context. Results for the β³ variants point to a similar impact on thermal stability from substitution at the core, core-flanking, or solvent-exposed site. This suggests the thermal stability loss from β³ substitution arises primarily from properties inherent to the monomer, such as enhanced flexibility. In contrast, effects of ACPC substitution are context dependent—destabilizing at core site L28 but stabilizing at neighboring core-flanking K29. This finding indicates long-range interactions involving the ACPC side chain may influence fold stability in HP35. Prior studies have shown that methionine or norleucine substitution at K29 stabilizes HP35 due to elimination of unfavorable electrostatic interactions.³³ Incorporation of ACPC at this position may provide a similar benefit. Correlations between the side chain of K29 and the aromatic rings of F10 and F17 are observed in the NOESY spectrum of HP35, and similar interactions are seen for the ACPC residue in variant ACPC29 (Fig. S29). These results support the hypothesis that long-range interactions involving residue 29 are important to the folded stability of HP35.

β-Residue substitution in the HP35 core aromatic triad

Given β-residue incorporation was viable at aliphatic core position L28 in HP35, we hypothesized it might also be tolerated in the crucial aromatic core cluster of F6, F10, and F17. To test this hypothesis, each of these residues was individually replaced by a β³ analogue. CD scans and thermal melts suggest that α → β³ substitution in the aromatic triad is consistently disruptive, though the degree depends on the site modified (Fig. 4(A), (B) and Table 1). β³F6 and β³F10 show evidence for a predominantly helical tertiary fold but are both significantly destabilized relative to the prototype (ΔT_m = −30 °C and −20 °C, respectively). Backbone alteration in β³F17 leads to CD spectral properties characteristic of a random coil and complete loss of a cooperative thermal folding transition, suggesting this variant does not adopt an ordered fold.

Fig. 4 (A), (B) CD scans at 20 °C (A) and thermal melts (B) for HP35 and variants modified with β³ residues in the core aromatic triad. Conditions: 50 µM peptide in 50 mM phosphate buffer, pH 7. (C) NMR structure ensemble for β³F6; backbone rmsd (average and standard deviation; residue range 1–33) for overlay of the ensemble to the X-ray structure of HP35 (PDB 3TRY) is shown in parentheses. (D) Overlay of the lowest energy model from the NMR ensemble of β³F6 with the X-ray structure of HP35 (PDB 3TRY). Alignment is based on helices 2 and 3, and arrows highlight repositioning of helix 1, which contains the F → β³F modification.

Comparison of the NMR spectra of β³F6 and β³F10 to that of HP35 shows reduced peak dispersion and increased peak broadening (Fig. S21 and S22). This effect suggests exchange among multiple ordered conformational states on an intermediate time scale.^34,35 Peak overlap and broadening in β³F10 precluded resonance assignment; however, we were able to generate the NMR structure for β³F6 (Fig. 4(C) and Table S6). β³F6 adopts a predominantly helical fold; however, tertiary packing is altered relative to prototype HP35. A shift in the β³F side chain away from the core is accompanied by an increase in local conformational heterogeneity in helix 1 and a change in the position of helix 1 relative to the remainder of the protein (Fig. 4(D)). Additional experimental support for a greater degree of disorder in this variant comes from the H_α chemical shifts, which show a lower magnitude deviation from random coil values throughout the domain than in any other folded variant (Fig. S27). The structural rearrangement of helices is accompanied by a change in composition of the hydrophobic core. Residues β³F6, F10, and K29 become more exposed in the variant, while residues T13, K24, and K32 become more buried and form new long-range contacts.

Overall, these findings show that β³ residue incorporation in the core aromatic triad of HP35 is significantly disruptive to both folded structure and folded stability. The differences among the three β³F variants reinforce the idea that the position of modification is crucial. Among the three isomers (β³F6, β³F10, β³F17), which differ only in the placement of a single methylene group in the backbone, one adopts an ordered but non-native fold, one a partially ordered molten globule, and the other a disordered random coil.

C_α-Me-α-residue substitution in the HP35 core aromatic triad

Given β³ substitution at core aromatic positions in HP35 was not tolerated, we wondered whether an alternate monomer might be better accommodated at these sites. Thus, we examined a corresponding set of variants in which each of the key phenylalanine residues was individually replaced by a chiral C_α-Me-α analogue bearing the same side chain. Given chiral C_α-methyl residues are strong helix inducers,^31,36–42 we were interested to see how side chain retaining α → C_α-Me substitution would compare to α → β³ replacement when targeted to critical buried core positions in HP35.

CD spectra of variants α^MeF6 and α^MeF17 are similar to HP35 (Fig. 5(A)), while that of α^MeF10 suggests significantly reduced helicity. Variable-temperature CD measurements exhibit cooperative folding transitions; however, the folded baseline for α^MeF10 is poorly defined (Fig. 5(B)). T_m values range from 45–72 °C (Table 1), and relative thermal stabilities follow the trend α^MeF6 > HP35 > α^MeF17 > α^MeF10. Folded structures of α^MeF6 and α^MeF17 determined by NMR show tertiary folds close to HP35 (Fig. 5(C) and Tables S7, S8). Packing of the aromatic triad in these variants is unchanged compared to the prototype, including at the artificial C_α-Me-α residues. The α-methyl group at each of these sites is buried in the core, in close contact with other side chains. In α^MeF6, the organization of the remainder of the core is indistinguishable from HP35. In the case of α^MeF17, the side chains of norleucine 12 and L20 shift slightly to accommodate the new methyl group. While subtle, this change may give rise to the observed decrease in thermal stability. NMR spectra of α^MeF10 show significant peak overlap compared to the other α^MeF variants (Fig. S23–S25), preventing resonance assignment. Taken with a low intensity CD signature and minimal thermal unfolding cooperativity, the small chemical shift dispersion suggests this variant does not adopt an ordered tertiary fold. The origin of the destabilization of α^MeF10 relative to the other two α^MeF variants is unclear.

Fig. 5 (A), (B) CD scans at 20 °C (A) and thermal melts (B) for HP35 and variants modified with C_α-Me-α residues in the core aromatic triad. Conditions: 50 µM peptide in 50 mM phosphate buffer, pH 7. (C) NMR structure ensembles for α^MeF6 and α^MeF17; backbone rmsd (average and standard deviation; residue range 1–33) for overlay of each ensemble to the X-ray structure of HP35 (PDB 3TRY) is shown in parentheses.

Generation of an HP35 variant with a backbone modification throughout the core

Given that several individual core sites in HP35 proved amenable to backbone modification, we next sought to establish whether multiple core residues could be replaced in tandem. Thus, we synthesized and characterized a triple substitution variant α^MeF6/α^MeF17/ACPC29. With one modification in each of the three helices, this analogue has a core that is 38% artificial in backbone composition. As detailed above, each of the corresponding individual substitutions combined in this sequence are innocuous with respect to the folded structure and two are thermally stabilizing.

CD scans suggest that the triple variant adopts a fold very similar to HP35 (Fig. 6(A)). Thermal melts reveal an exceptionally high thermal stability, with the almost complete absence of a defined unfolding transition (Fig. 6(B)). To obtain quantitative insights into thermodynamic effects of the altered backbone composition in this analogue, we compared the free energy of folding (ΔG°) for it, the prototype and each corresponding single-site variant (i.e., α^MeF6, α^MeF17, and ACPC29) through chemical denaturation with guanidinium chloride monitored by tryptophan fluorescence (Fig. 6(C)). All five sequences showed a cooperative unfolding transition. Folding free energies () obtained from the fits follow a similar trend as T_m (Table 1). Relative to HP35, single-site variant α^MeF6 is significantly stabilized (ΔΔG° −1.5 kcal mol⁻¹), ACPC29 is comparable in stability (ΔΔG° −0.2 kcal mol⁻¹), and α^MeF17 is slightly destabilized (ΔΔG° +1.2 kcal mol⁻¹). The triple mutant is also more thermodynamically stable than the prototype (ΔΔG° −0.9 kcal mol⁻¹), and more stable than all but one of the single-site variants. NMR analysis for the α^MeF6/α^MeF17/ACPC29 triple variant (Fig. S26 and Table S9) shows that, despite substantial fraction of the core being backbone modified, it adopts a very similar fold to HP35 (Fig. 6(D)).

Fig. 6 (A), (B) CD scans at 20 °C (A) and thermal melts (B) for HP35 and triple-modified variant α^MeF6/α^MeF17/ACPC29. Conditions: 50 µM peptide in 50 mM phosphate buffer, pH 7. (C) Chemical denaturation of HP35, the triple-modified variant, and the single-site variants α^MeF6, α^MeF17, and ACPC29 by guanidinium chloride (Gnd) monitored by tryptophan fluorescence. Conditions: 10 µM peptide in 50 mM phosphate buffer, pH 7. Data points are average and standard deviation from triplicate measurements. (D) NMR structure ensemble for α^MeF6/α^MeF17/ACPC29; backbone rmsd (average and standard deviation; residue range 1–33) for overlay to the X-ray structure of HP35 (PDB 3TRY) is shown in parentheses.

Conclusions

In summary, we have reported here a systematic examination of the impacts of backbone modification in the hydrophobic core of the C-terminal subdomain of the villin headpiece on the folded structure and conformational stability of the protein. We find that backbone alteration through substitution of α-residues with β³, ACPC, or C_α-Me-α analogues can be made at core or core-flanking positions without perturbing the fold of the protein. The sequence context for a given backbone modification is an important factor in determining its exact effect; however, even residues with side chains that engage in crucial tertiary packing interactions in the hydrophobic core can be replaced. Many of the single-site variants are destabilized relative to the prototype, but some show improved conformational stability. Combining three well tolerated modifications into a single construct yields a variant with 38% of its core residues modified, a fold virtually identical to the prototype, and significantly enhanced thermal stability and folding free energy.

Compared to prior work on the application of backbone modification at core positions in villin and other domains, the fact that many of the variants adopt folds very similar to prototype is noteworthy. Further, the enhanced stability seen for several of the variants is significant. Backbone modification in tertiary structure contexts is often destabilizing—even when made at solvent-exposed sites. Examples of tertiary structure mimetics with multiple backbone modifications and enhanced conformational stability relative to a prototype protein are rare.^43–46 Achieving this outcome when the modification sites are in the core of the domain is noteworthy, considering the unique composition of the hydrophobic core, the small size, and the moderate unfolding free energy of HP35. Collectively, the present results demonstrate that the entirety of a sequence—both buried core residues and solvent-exposed sites—should be considered fair game in the design of heterogeneous-backbone protein mimetics in the context of α-helical secondary structure.

A significant challenge in development of protein mimetics based on artificial backbones is the incorporation of non-proteogenic amino acids. Although methods for ribosomal synthesis with non-canonical backbones continue to advance,⁴⁷ chemical synthesis remains the method of choice for larger sequences with multiple modifications. Reliance on total chemical synthesis poses limitations to the size of entities that can be produced and, thus, the complexity of folds that can be mimicked. Methods such as native chemical ligation and semi-synthesis can expand the chain lengths accessible;⁴⁸ however, spontaneous refolding of large synthetic constructs absent the typical cellular milieu can prove difficult.⁴⁹ The above technical challenges notwithstanding, studies on the folding behavior of artificial protein-like backbones—such as the present work—show what is possible in these entities and provides insights into their design. Improved understanding of these fundamental issues along with continued advances in methods for construction of protein-like artificial macromolecules will help to realize the full potential of proteomimetics as a concept.¹

Conflicts of interest

There are no conflicts to declare.

Abbreviations

1PRB	G-Related albumin-binding module from bacterial protein PAB
ACPC	trans-2-Aminocyclopentane carboxylic acid
BMRB	Biological magnetic resonance bank
CD	Circular dichroism
COSY	Correlated spectroscopy
ESI-MS	Electrospray ionization mass spectrometry
Gnd	Guanidinium chloride
HP35	Villin headpiece
HPLC	High-performance liquid chromatography
NMR	Nuclear magnetic resonance
NOESY	Nuclear Overhauser effect spectroscopy
PDB	Protein data bank
rmsd	Root mean square deviation
SASA	Solvent-accessible surface area
TOCSY	Total correlation spectroscopy

Data availability

Coordinates and experimental data for newly reported NMR structures are deposited in the PDB (9YLZ, 9YM0, 9YM1, 9YM2, 9YM3, 9YM4, 9YM5, 9YM6, 9YM7) and BMRB (31270, 31271, 31272, 31273, 31274, 31275, 31276, 31277, 31278). Other data are available from the corresponding author upon request.

Additional data supporting the findings of this study can be found in the SI for the article. Supplementary information: Fig. S1–S29, Tables S1–S9, materials and methods. See DOI: https://doi.org/10.1039/d5cb00269a.

Acknowledgements

Funding for this work was provided by a grant from the National Institutes of Health (R35GM149220 to W. S. H.). R. M. D. was supported by the National Science Foundation (REU 2244200).

Notes and references

1. W. S. Horne and T. N. Grossmann, Nat. Chem., 2020, 12, 331–337.
2. C. S. Swenson, G. Mandava, D. M. Thomas and R. E. Moellering, Chem. Rev., 2024, 124, 13020–13093.
3. S. H. Hong, T. Nguyen, J. F. Ongkingco, A. Nazzaro and P. S. Arora, Chem. Rev., 2025, 125, 6819–6869.
4. L. Lombardi, V. D. Genio, F. Albericio and D. R. Williams, Chem. Rev., 2025, 125, 7099–7166.
5. E. Lenci and A. Trabocchi, Chem. Soc. Rev., 2020, 49, 3262–3277.
6. J. L. Hickey, D. Sindhikara, S. L. Zultanski and D. M. Schultz, ACS Med. Chem. Lett., 2023, 14, 557–565.
7. Z. E. Reinert and W. S. Horne, Org. Biomol. Chem., 2014, 12, 8796–8802.
8. K. L. George and W. S. Horne, Acc. Chem. Res., 2018, 51, 1220–1228.
9. P. Sang and J. Cai, Chem. Soc. Rev., 2023, 52, 4843–4877.
10. M. M. Müller, Biochemistry, 2018, 57, 177–185.
11. W. S. Horne, J. L. Price and S. H. Gellman, Proc. Natl. Acad. Sci. U. S. A., 2008, 105, 9151–9156.
12. M. W. Giuliano, W. S. Horne and S. H. Gellman, J. Am. Chem. Soc., 2009, 131, 9860–9861.
13. M. Bejger, P. Fortuna, M. Drewniak-Świtalska, J. Plewka, W. Rypniewski and Ł. Berlicki, Chem. Commun., 2021, 57, 6015–6018.
14. Y. Lin and W. S. Horne, Chem. – Eur. J., 2024, 30, e202401890.
15. J. C. McKnight, D. S. Doering, P. T. Matsudaira and P. S. Kim, J. Mol. Biol., 1996, 260, 126–134.
16. C. J. McKnight, P. T. Matsudaira and P. S. Kim, Nat. Struct. Biol., 1997, 4, 180–184.
17. D. E. Mortenson, K. A. Satyshur, I. A. Guzei, K. T. Forest and S. H. Gellman, J. Am. Chem. Soc., 2012, 134, 2473–2476.
18. Y. Duan and P. A. Kollman, Science, 1998, 282, 740–744.
19. C. D. Snow, H. Nguyen, V. S. Pande and M. Gruebele, Nature, 2002, 420, 102–106.
20. J. Kubelka, W. A. Eaton and J. Hofrichter, J. Mol. Biol., 2003, 329, 625–630.
21. M. Wang, Y. Tang, S. Sato, L. Vugmeyster, C. J. McKnight and D. P. Raleigh, J. Am. Chem. Soc., 2003, 125, 6032–6033.
22. A. Fernández, M.-y Shen, A. Colubri, T. R. Sosnick, R. S. Berry and K. F. Freed, Biochemistry, 2003, 42, 664–671.
23. S. H. Brewer, D. M. Vu, Y. Tang, Y. Li, S. Franzen, D. P. Raleigh and R. B. Dyer, Proc. Natl. Acad. Sci. U. S. A., 2005, 102, 16662–16667.
24. H. Lei, C. Wu, H. Liu and Y. Duan, Proc. Natl. Acad. Sci. U. S. A., 2007, 104, 4925–4930.
25. D. L. Ensign, P. M. Kasson and V. S. Pande, J. Mol. Biol., 2007, 374, 806–816.
26. D. E. Shaw, P. Maragakis, K. Lindorff-Larsen, S. Piana, R. O. Dror, M. P. Eastwood, J. A. Bank, J. M. Jumper, J. K. Salmon, Y. Shan and W. Wriggers, Science, 2010, 330, 341–346.
27. K.-N. Hu, W.-M. Yau and R. Tycko, J. Am. Chem. Soc., 2010, 132, 24–25.
28. J. K. Chung, M. C. Thielges and M. D. Fayer, J. Am. Chem. Soc., 2012, 134, 12118–12124.
29. G. Žoldák, J. Stigler, B. Pelz, H. Li and M. Rief, Proc. Natl. Acad. Sci. U. S. A., 2013, 110, 18156–18161.
30. D. F. Kreitler, D. E. Mortenson, K. T. Forest and S. H. Gellman, J. Am. Chem. Soc., 2016, 138, 6498–6505.
31. T. W. Harmon, Y. Lin, R. T. Sutton, S. W. J. Osborne and W. S. Horne, ChemBioChem, 2025, 26, e202401022.
32. B. S. Frank, D. Vardar, D. A. Buckley and C. J. McKnight, Protein Sci., 2002, 11, 680–687.
33. J. Kubelka, T. K. Chiu, D. R. Davies, W. A. Eaton and J. Hofrichter, J. Mol. Biol., 2006, 359, 546–553.
34. H. J. Dyson and P. E. Wright, Methods Enzymol., 2001, 339, 258–270.
35. A. K. Mittermaier and L. E. Kay, Trends Biochem. Sci., 2009, 34, 601–611.
36. C. Zhang, W. Miller, K. J. Valenzano and D. J. Kyle, J. Med. Chem., 2002, 45, 5280–5286.
37. N. A. Tavenor, Z. E. Reinert, G. A. Lengyel, B. D. Griffith and W. S. Horne, Chem. Commun., 2016, 52, 3789–3792.
38. R. Islam, D. O. Sviridov, S. K. Drake, J. Tunyi, G. Abdoulaeva, L. A. Freeman, R. W. Pastor and A. T. Remaley, Biochem. Biophys. Res. Commun., 2020, 526, 349–354.
39. V. Bauer, B. Schmidtgall, G. Gógl, J. Dolenc, J. Osz, Y. Nominé, C. Kostmann, A. Cousido-Siah, A. Mitschler, N. Rochel, G. Travé, B. Kieffer and V. Torbeev, Chem. Sci., 2021, 12, 1080–1089.
40. R. A. D. Bathgate, P. Praveen, A. Sethi, W. I. Furuya, R. R. Dhingra, M. Kocan, Q. Ou, A. L. Valkovic, I. Gil-Miravet, M. Navarro-Sánchez, F. E. Olucha-Bordonau, A. L. Gundlach, K. J. Rosengren, P. R. Gooley, M. Dutschmann and M. A. Hossain, J. Am. Chem. Soc., 2023, 145, 20242–20247.
41. A. Chandramohan, H. Josien, T. Y. Yuen, R. Duggal, D. Spiegelberg, L. Yan, Y.-C. A. Juang, L. Ge, P. G. Aronica, H. Y. K. Kaan, Y. H. Lim, A. Peier, B. Sherborne, J. Hochman, S. Lin, K. Biswas, M. Nestor, C. S. Verma, D. P. Lane, T. K. Sawyer, R. Garbaccio, B. Henry, S. Kannan, C. J. Brown, C. W. Johannes and A. W. Partridge, Nat. Commun., 2024, 15, 489.
42. T. W. Harmon, J. Song, A. J. Gulewicz, Y. P. Di and W. S. Horne, ChemBioChem, 2025, 26, e202400951.
43. K. L. George and W. S. Horne, J. Am. Chem. Soc., 2017, 139, 7931–7938.
44. S. D. Melton, E. A. E. Brackhahn, S. J. Orlin, P. Jin and D. M. Chenoweth, Chem. Sci., 2020, 11, 10638–10646.
45. J. R. Santhouse, J. M. G. Leung, L. T. Chong and W. S. Horne, Chem. Sci., 2022, 13, 11798–11806.
46. M. M. Wright, B. H. Rajewski, T. A. Gerrein, Z. Xu, L. J. Smith, W. S. Horne and J. R. Del Valle, Commun. Chem., 2025, 8, 76.
47. M. Sigal, S. Matsumoto, A. Beattie, T. Katoh and H. Suga, Chem. Rev., 2024, 124, 6444–6500.
48. V. Agouridas, O. El Mahdi, V. Diemer, M. Cargoët, J.-C. M. Monbaliu and O. Melnyk, Chem. Rev., 2019, 119, 7328–7443.
49. A. J. Callahan, A. Rondon, R. M. Reja, L. L. Salazar, S. Gandhesiri, J. Rodriguez, A. Loas and B. L. Pentelute, J. Am. Chem. Soc., 2024, 146, 28696–28706.

---