A Peptide-Based Strategy to Disrupt Oxidation-Induced Aggregation of γC-Crystallin Missense Mutant

Abstract

Cataracts, a leading cause of blindness worldwide, are frequently associated with mutations in Crystallin genes. The G129C mutation in human γC-Crystallin (γC) leads to severe early onset congenital cataracts by introducing a solvent-exposed cysteine residue, though its pathogenic mechanism remains incompletely understood. This study demonstrates that the mutation renders the protein highly susceptible to oxidative stress, lowering the threshold for a deleterious phase transition that culminates in abnormal aggregation. Although the G129C mutant retains its native folded state, it undergoes rapid, concentration-dependent aggregation under mild oxidative stress, forming solid-like aggregates with distinctive structural properties. To counteract this pathology, we further identified an affinity peptide, Cand.2. Cand.2 selectively binds to G129C-induced aggregates and dissolves preformed aggregates in vitro. More importantly, it can alleviate oxidative stress-induced intracellular aggregation in cell models. This study elucidates the molecular pathway by which the G129C mutation leads to cataract formation and establishes a peptide-based targeted therapeutic strategy, offering a promising nonsurgical alternative to conventional broad-spectrum antioxidants for the treatment of hereditary cataracts.

Introduction

Cataracts represent a major global cause of vision loss, exhibiting distinct age-dependent etiologies. In the elderly population, age-related cataracts predominate, with multifactorial pathogenesis involving ultraviolet exposure, antioxidant deficiency, drug reactions, smoking, metabolic disorders (e.g., diabetes, gout), alongside genetic predisposition. In contrast, congenital cataracts primarily result from genetic abnormalities, with Crystallin gene mutations responsible for >50% of inherited forms, highlighting the crucial role of these proteins in maintaining lens homeostasis. , Although surgery is currently the only clinically proven and effective treatment, the vision of developing nonsurgical intervention strategies continues to drive researchers to explore the molecular mechanisms of cataracts, aiming to develop preventive measures or pharmacological interventions.

Crystallins, constituting ∼90% of lens soluble proteins (400–1000 mg/mL solubility), maintain optical clarity through tightly regulated interactions. , Disruption of this equilibrium via abnormal protein aggregation directly initiates cataractogenesis. , The Crystallin superfamily includes molecular chaperone α-crystallins (belonging to the small heat shock protein family) and structural β/γ-crystallins, the latter comprising seven isoforms (γA, γB, γC, γD, γE, γF, and γS). , In congenital cataracts, β/γ-Crystallin mutations typically cause misfolding and precipitation, triggering aggregation cascades that ultimately lead to lens opacity. , For example, γD-Crystallin P23T markedly reduces protein solubility, whereas γD-Crystallin R36S increases crystallization propensity. Beyond these mechanisms, redox-related cysteine mutations have been reported in other crystallins, including γD-Crystallin R14C, γA-Crystallin R48C, γS-Crystallin S39C. γC is a 174-residue structural protein essential for lens refraction. , Mutations in γC can lead to its aberrant aggregation. Among reported cataract-associated CRYGC missense mutations, G129C is a representative cysteine-substitution variant, identified in a Chinese pedigree and associated with severe early onset congenital cataracts. , Previous studies have shown that this mutation perturbs the stability and unfolding behavior of γC-Crystallin and promotes the formation of aggregation-prone intermediates. However, whether the acquired cysteine also lowers the oxidative threshold for aggregation has remained unclear.

Among emerging strategies to counteract Crystallin aggregation, small-molecule therapeutics, particularly those amenable to topical ocular delivery, represent a promising direction due to their favorable ocular permeability and noninvasive administration profile, advantages over surgical intervention which carries risks such as intraocular inflammation. Current efforts to identify effective small-molecule agents have largely centered on antioxidants, including aspirin, pantethine, flavonoids, and lanosterol: , aspirin and pantethine lack efficacy; flavonoids are only effective against selenite-induced cataracts; , and lanosterol has not demonstrated a curative effect despite advancing to Phase II trials. , These failures underscore the necessity for therapies that target specific pathological processes, such as protein aggregation, rather than employing broad-spectrum antioxidant approaches.

In this work, we investigate the molecular pathology of the G129C mutation in γC, a variant linked to hereditary cataracts. We show that the mutation-introduced cysteine residue acts as a primary driver of pathogenesis by initiating a deleterious phase transition. This G129C substitution primes γC for oxidative damage, predisposing it to aberrant phase separation that drives protein aggregation (Figure A). Furthermore, we identified Cand.2, an affinity peptide that selectively binds to G129C-induced aggregates and effectively restores their solubility. This work establishes a novel therapeutic strategy for counteracting cataracts by targeting mutant-specific aggregation, with Cand.2 representing a potential nonsurgical candidate.

1.

G129C mutant exhibits a higher propensity to aggregate compared to the WT γC under oxidative stress. (A) Structure of WT γC (PDB: 2NBR) and a computational model of the G129C mutation generated using PyMol. Blue, glycine (Gly129); red, cysteine (Cys129). At low concentrations of H₂O₂, WT γC remains clear in 1× PBS, pH 7.4, whereas the G129C mutant forms aggregates. (B and D) Aggregation kinetics of 100 μM G129C mutant (B) and WT γC (D) in 1× PBS (pH 7.4) under a range of H₂O₂ concentrations or with 5 mM DTT, by measuring the optical density at 350 nm (N = 3). Data show mean ± s.e.m. (C,E) Time-lapse DIC microscopy images of 100 μM G129C mutant (C) and WT γC (E) over time under 0.1% or 1% H₂O₂. Scale bars, 10 μm.

Results

The G129C Mutant Exhibits a Higher Aggregation Propensity under Oxidative Stress

γC is composed of two globular Greek-key motifs, which form a stable hydrophobic core and a well-packed interdomain interface, conferring high solubility to the protein. Gly129 is situated within a solvent-exposed loop of the C-terminal domain, adjacent to this critical interface (Figure A). We introduced a cysteine substitution at position 129 of γC (G129C) and verified using solution nuclear magnetic resonance spectroscopy (NMR) that the resulting mutant retains a native-like fold, consistent with previous literature reports (Figure S1). We evaluated the thermal stability of the G129C mutant via turbidity measurements (OD₃₅₀) from 37 to 60 °C (Figure S2). The mutant protein remains soluble across tested concentrations at physiological temperatures. By contrast, aggregation initiates at 55 °C and is markedly enhanced at 60 °C in a concentration-dependent manner. Since these temperatures exceed the physiological range of the lens, heat-induced denaturation of the G129C mutant does not likely represent the principal mechanism driving γC aggregation in vivo.

The G129C mutation significantly increases the oxidative sensitivity of γC. Additional aggregation kinetics assays reveal that the G129C mutant exhibits concentration-dependent aggregation under a gradient of H₂O₂ (0.01–1% v/v, ≈3–300 mM), reaching OD₃₅₀ values of 0.4 and 1.4 at 0.01% and 1% H₂O₂, respectively, within 180 min (Figure B). In contrast, the WT protein shows no significant aggregation below 1% H₂O₂ (Figure D), indicating a markedly lower oxidative threshold for the mutant (Figure S3A). Although the H₂O₂ concentrations used in these assays are higher than reported physiological levels in the lens, they can be used as an accelerator in vitro oxidative-stress condition to induce measurable aggregation. Differential interference contrast (DIC) microscopy confirms that 0.1% H₂O₂ induces micron-scale aggregates (average diameter: 1.7 μm) in G129C whereas the WT remains soluble (Figure C,E). Increasing oxidant concentrations (0.1–1% H₂O₂) promote progressive aggregate growth in the mutant, yielding larger spherical structures (Figure S3B,C). Fourier transform infrared (FTIR) spectroscopy reveals concomitant secondary structural alterations in G129C aggregates, notably a reduction in random coil content, as evidenced by decreased amide I band intensity at 1637–1648 cm^–1 , (Figure S3D). Collectively, these data establish that the acquired thiol group in G129C γC substantially lowers the oxidative threshold for aggregation and promotes concentration-dependent aggregate formation.

G129C Mutant Aggregates Display Partial Solid-State Characteristics

Under 1% H₂O₂ treatment, comparative aggregation kinetics between G129C and WT reveal distinct characteristics (Figure A). Both proteins exhibit sigmoidal aggregation curves, though the mutant demonstrates accelerated kinetics. At 100 μM protein concentration, G129C displays significantly shorter onset time (16 ± 2 min) versus WT (31 ± 2 min), indicating faster nucleation (Figure B). The OD₃₅₀ absorbance values during the plateau phase correlate positively with protein concentration, suggesting that higher concentrations facilitate the formation of larger or more numerous aggregates. At each concentration tested, the plateau OD₃₅₀ of the G129C mutant is higher than that of WT γC; for example, 1.03 ± 0.07 (G129C) versus 0.66 ± 0.06 (WT) at 50 μM (Figure C). To quantitatively characterize the kinetics, we determined the aggregation half-time (t _1/2), defined as the time to reach 50% of the maximum OD₃₅₀, and analyzed its logarithm, log­(t _1/2), across protein concentrations (Figure D). The G129C mutant consistently shows lower log­(t _1/2) values than WT γC at all concentrations, confirming its faster aggregation. Notably, whereas WT exhibits a monotonic decrease in log­(t _1/2) with increasing concentration, the mutant displays a nonmonotonic dependence, implying a concentration-dependent shift in its aggregation mechanism.

2.

G129C mutant aggregates display partial solid-state characteristics. (A) Aggregation kinetics of G129C mutant and WT γC in 1 × PBS (pH 7.4) under 1% H₂O₂ were determined by measuring the optical density at 350 nm across a range of protein concentrations (N = 3). (B) Comparison of the aggregation onset time for 100 μM G129C mutant versus WT γC under 1% H₂O₂ treatment (N = 3; ***P < 0.001). (C,D) Comparison of plateau absorbance at 350 nm (C) and aggregation half-time (D) for the G129C mutant versus WT γC under 1% H₂O₂ at varying protein concentrations. (E) Time-lapse DIC microscopy images of 100 μM G129C mutant and WT γC after incubation with 1% H₂O₂ at 37 °C for 60 min. Images were acquired immediately within 10 min. No fusion events were observed in aggregates over time. Scale bars, 5 μm. (F) Full-FRAP analysis of aggregates from 100 μM G129C mutant (left) and WT γC (right) after incubation with 1% H₂O₂ at 37 °C for 60 min (N = 3). Measurements were initiated 10 min after aggregate formation. Dark green lines, bleached regions; light green lines, unbleached regions; gray lines, background. Representative images show the time points before bleaching, immediately after bleaching (0 s), and postbleaching at 180 s of recovery. Scale bars, 1 μm. (G) Recovery rates of G129C mutant aggregates compared with those of WT γC (N = 3). Quantified data in this figure are presented as mean ± s.e.m.

For the introduced Cys129 is adjacent to the native Cys130, we further examined whether the G129C aggregation depends on formation of a local Cys129-Cys130 intramolecular disulfide bond. We constructed a G129C/C130A double mutant, in which Cys130 was removed while Cys129 was retained. Under oxidative conditions, removal of Cys130 did not suppress G129C-associated aggregation. Instead, the G129C/C130A double mutant exhibited even stronger aggregation than G129C. At 100 μM protein concentration, the G129C/C130A double mutant showed detectable aggregation even in the presence of 0.001% H₂O₂ (Figure S4A). Under 1% H₂O₂, the G129C/C130A double mutant displayed concentration-dependent aggregation across the tested protein concentrations, with earlier aggregation onset and faster approach to the plateau phase than the G129C mutant (Figure S4B). These results indicate that Cys130 is not required for the aggregation and illustrate the local Cys129-Cys130 intramolecular disulfide bond not as the primary driver of G129C-induced aggregation.

Characterization of aggregates formed by the G129C mutant and WT γC by DIC microscopy shows no fusion events over a 10 min imaging period (Figure E), suggesting that the aggregates possess solid-state properties. Quantitative fluorescence recovery after photobleaching (FRAP) measurements, performed using a 2 μm diameter bleach region with full laser depletion, reveal significantly restricted fluorescence recovery at 180 s postbleach: 26.7 ± 13.1% for G129C versus 38.9 ± 11.5% for WT (Figure F,G). Although the difference between G129C and WT did not reach statistical significance, both aggregates showed restricted fluorescence recovery, indicating limited molecular mobility and partial solid-like behavior. The substantial standard deviations (>10%) in recovery rates point to significant structural heterogeneity within the aggregates, likely arising from variations in packing density and intermolecular interactions.

Identification of Peptide Cand.2 Targeting G129C Mutant Aggregates

To identify peptides capable of disrupting G129C mutant aggregation, we screened a custom peptide microarray library comprising 264 unique peptides for high-affinity binders to the G129C mutant. The peptide library, comprising 264 unique sequences, was designed to encompass a broad spectrum of biophysical properties, including sequence length, charge state, hydrophobicity, and amino acid composition. Sequences varied in length from 4 to 39 amino acids, with a median of 15 residues (Figure S4A). In terms of electrostatic properties, the predicted net charge at pH 7.0 ranged from −8.99 to +13.99 (median = 1.00), covering a wide span from highly anionic to strongly cationic peptides. Correspondingly, the predicted isoelectric points (pI) spanned from 0.0 to 14.0 (median = 9.2), representing acidic, near neutral, and basic peptides. Hydrophobicity, assessed by the grand average of hydropathicity (GRAVY) based on the Kyte–Doolittle scale ^, ranged from −4.50 to 4.50 (median = −0.97). This overall negative median indicates a bias toward hydrophilic character in the library (Figure S4B). Analysis of amino acid composition confirmed substantial residue-level diversity. At the residue level across the entire library, Arg, Gly, and Lys were the most abundant amino acids, whereas His, Trp, and Met were relatively underrepresented (Figure S5A). At the peptide level, the median fractions were 7.7% for aromatic residues, 33.6% for hydrophobic residues, 66.5% for hydrophilic residues, 35.7% for charged residues, 11.1% for acidic residues, and 24.7% for basic residues (Figure S5B). Collectively, these data demonstrate that the library possesses considerable diversity in residue composition, polarity, and electrostatic characteristics.

The microarray was incubated with 0.5 mg/mL WT γC or G129C mutant labeled with Alexa Fluor 488 (see Materials and Methods for microarray preparation details), followed by fluorescence scanning (Figure A). Peptides exhibiting higher fluorescence intensity with the G129C mutant than with WT were selected as candidates for targeted recognition validation. Six candidate polypeptides (designated Cand.1–6) were identified; their sequences are provided in Table .

3.

Identification of peptide Cand.2 targeting G129C γC aggregates. (A) Schematic diagram of the peptide microarray. The left panel shows the binding results of candidate peptides to wild-type (WT) γC and the G129C mutant, with spot fluorescence intensity representing relative binding strength (light green: bound; dark green: unbound). The right panel shows representative fluorescence images of candidate peptides Cand.1–Cand.6 incubated with WT γC and the G129C mutant. (B) Changes in the absorbance at 350 nm of preformed G129C mutant aggregates upon treatment with 1 mM candidate peptides (Cand.1–Cand.6) or 5 mM DTT. (C) Bright-field microscopy images of G129C mutant aggregates after 3 h of incubation with 1 mM candidate peptides (Cand.1–Cand.6) or 5 mM DTT in 1× PBS. Scale bars, 20 μm. (D) Absorbance changes at 350 nm before and after incubation with varying concentrations of peptide Cand.2 or with 5 mM DTT (N = 3; **P < 0.01). Data show mean ± s.e.m. (E) Representative diagram of the BLI assay showing sequential steps: peptide loading, association of the G129C mutant, and dissociation. (F-G) Association and dissociation kinetics of the G129C mutant (F) and WT γC (G) on Cand.2-immobilized BLI sensors. (H) Comparison of BLI responses for WT γC versus the G129C mutant upon binding to the Cand.2-immobilized surface.

1. Amino Acid Sequences and Structural Modifications of the Selected Peptides.

peptide	sequence
Cand.1	Biotin-PESFDGDPASNTAPLQPEQLQ
Cand.2	Biotin-YGRKKRRQRRR
Cand.3	Biotin-FIHHIIGGLFSIGKIIHRLIRRRRR
Cand.4	EEGFFSA[citrulline]GHRPLDKK-Biotin
Cand.5	Biotin-CPRGNYVNEKPY
Cand.6	Biotin-GGRSFILLRIIQGCRRRNTVDD

To evaluate the disaggregation potential of Cand.1–6 peptides, preformed G129C aggregates (generated under 0.1% H₂O₂ at 37 °C) were incubated with 1 mM of each peptide at 37 °C. Aggregate dissolution was monitored by turbidimetry at 350 nm (Figure B) and confirmed by optical microscopy after 3 h incubation (Figure C). In control experiments with either no peptide or 5 mM dithiothreitol (DTT), aggregates remained stable. While peptides Cand.1, 3, 4, 5, and 6 show minimal activity, Cand.2 treatment results in a progressive decrease in OD₃₅₀ (Figure B) and corresponding microscopically visible clarification (Figure C). Transmission electron microscopy (TEM) imaging further verified the capacity of the Cand.2-mediated reduction in aggregate size (Figure S6). Cand.2 demonstrates concentration-dependent disaggregation activity against preformed G129C aggregates (Figure D). Comparative analysis of OD₃₅₀ values before and after 180 min incubation shows that Cand.2, at concentrations ≥10 μM, exhibits significantly greater efficacy in reversing G129C aggregation than other candidate peptides. Also, we tested its disaggregation potential against assemblies formed by the cataract-associated γD-Crystallin R36S mutant. Cand.2 did not show disaggregation activity to γD R36S crystalline assemblies, supporting that, Cand.2 is not a broad-spectrum reversing peptide (Figure S10).

Comparative analysis of the six candidate peptides (Cand.1-Cand.6) revealed that, Cand.2 possessed distinct properties. As summarized in Table S2, Cand.2 exhibited the most negative grand average of hydropathicity (GRAVY) score (−3.64), indicating substantially stronger hydrophilicity compared to the other candidates. Furthermore, it displayed the highest proportions of basic residues (72.7%), charged residues (72.7%), and hydrophilic residues (90.9%), while having the lowest hydrophobic-residue fraction (9.1%) and a strongly positive predicted net charge at pH 7.0 (+7.00). In contrast, the other, less active peptides were generally characterized by higher hydrophobicity (Cand.3 and Cand.6), anionic character (Cand.1), or a less cationic nature coupled with elevated fractions of aromatic or hydrophobic residues (Cand.4 and Cand.5). Collectively, these features distinguish Cand.2 from the other candidates and may contribute to its superior disaggregation activity. To further explore the molecular basis of Cand.2 activity, we performed AlphaFold3-based complex prediction using the γC-Crystallin solution structure as a reference (Figure S7A). The alteration of hydrophobic and charged residues suggests Cand.2 engages G129C through sequence-dependent interactions (Figure S7B,C). We also characterized Cand.2 alone by static light scattering under the same conditions. The intensity of light scattered from Cand.2 remained low and stable over 180 min, indicating Cand.2 did not undergo self-aggregation (Figure S8).

To investigate the interaction between peptide Cand.2 and the G129C mutant, we biotinylated the N-terminus of Cand.2 and immobilized it on a streptavidin-coated biolayer interferometry (BLI) sensor (Figure E). Real-time BLI measurements monitored the association and dissociation kinetics through optical interference shifts resulting from G129C mutant or WT γC binding at the sensor interface. The G129C mutant bound with an equilibrium dissociation constant (K _D) of 1.7 ± 0.7 × 10^–7 M, corresponding to an approximately 2-fold tighter binding than WT γC (K _D = 4.0 ± 1.9 × 10^–7 M) (Figure F,G). This difference in affinity is attributed to favorable binding kinetics: the mutant exhibits both a higher association rate constant (k _a = 3.3 ± 1.3 × 10³ M^–1 s^–1 versus 1.6 ± 0.8 × 10³ M^–1 s^–1 for WT) and a lower dissociation rate constant (k _d = 5.4 ± 0.5 × 10^–4 s^–1 versus 6.3 ± 1.1 × 10^–4 s^–1 for WT). The preferential affinity for the mutant is further corroborated by BLI, where Cand.2-functionalized sensors yielded consistently stronger responses to G129C across all concentrations tested (Figure H).

Peptide Cand.2 Reverses G129C Mutant Aggregates

To evaluate the therapeutic potential of peptide Cand.2 against pathological Crystallin aggregation, N-terminally eGFP-tagged versions of WT γC and the G129C mutant were expressed in HeLa cells and human lens epithelial cells (HLEB3). Expression of eGFP-tagged WT γC results in uniform cytoplasmic fluorescence in both cell lines (Figures A and S7). In contrast, expression of the eGFP-tagged G129C mutant yields two distinct cellular phenotypes: (i) an aggregation-negative phenotype with diffuse fluorescence (Figure B, left), and (ii) an aggregation-positive phenotype characterized by discrete intracellular fluorescent clusters (Figure B, right). Based on this dichotomy, we established a quantitative metric, i.e., the aggregation-positive rate (calculated as the number of aggregation-positive cells divided by the total number of eGFP-positive cells × 100%), to provide an objective standard for subsequent compound evaluation.

4.

Peptide Cand.2 reduces G129C mutant aggregation in cells. (A) Fluorescence image of HeLa cells transfected with eGFP-tagged WT γC. (B) HeLa cells transfected with eGFP-tagged G129C mutant, showing an aggregation-negative (left) and an aggregation-positive (right) phenotype. (C) HeLa cells transfected with eGFP-tagged WT γC or G129C mutant and treated with exogenous H₂O₂ or ASA. Scale bars, 50 μm. (D) Percentage of HeLa cells containing aggregates after transfection with eGFP-tagged G129C mutant and treatment with indicated concentrations of H₂O₂ or ASA (N = 6; ***P < 0.001). Data show mean ± s.e.m. (E) Percentage of HeLa cells containing aggregates after transfection with eGFP-tagged G129C mutant and treatment with peptide Cand.2 (N = 6; **P < 0.01). Data show mean ± s.e.m. (F) Representative fluorescence images of HeLa cells transfected with eGFP-tagged G129C mutant and treated with Cand.2 for 24 h. Scale bars, 50 μm. (G) Time-lapse live-cell imaging of Cand.2-mediated dissolution of eGFP-tagged G129C mutant aggregates in HeLa cells. For each condition, representative cells from different regions are shown in the upper and lower panels. Scale bars, 10 μm.

Given preliminary in vitro evidence suggesting oxidative stress as a key driver of G129C mutant aggregation, we employed exogenous H₂O₂ treatment to simulate intracellular oxidative stress. The experimental data reveals that WT γC maintains a stable cytoplasmic distribution even under 500 μM H₂O₂ treatment (Figure C). In contrast, the G129C mutant shows 1.32-fold and 1.21-fold increases in its aggregation-positive rate at 100 μM and 500 μM H₂O₂, respectively (P < 0.01; Figure D). Notably, the broad-spectrum antioxidant ascorbic acid (ASA, 200 μM) fails to significantly reverse this oxidative stress-induced aggregation (P > 0.05; Figure C,D). This result underscores that conventional antioxidants have limited utility in preventing the aggregation of the G129C mutant.

Our study identified peptide Cand.2 as possessing unique antiaggregation properties. In the eGFP-tagged G129C mutant transfection model, 24 h Cand.2 treatment dose-dependently reduces aggregation-positive rates, with the 500 μM treatment group showing 18% reduction compared to controls (P < 0.001) (Figure E). Fluorescence images corroborate this reduction (Figures F and S8). Live-cell imaging further demonstrates that the control cells expressing the eGFP-tagged G129C mutant progressively form aggregates within 24 h, whereas those treated with 500 μM Cand.2 show active aggregate dissolution within 6 h (Figure G). Western blot showed approximate same level of total eGFP-tagged G129C mutant levels with or without Cand.2 treatment, supporting Cand.2 dissolve aggregates without G129C degradation (Figure S13). Safety assessments confirm Cand.2’s biocompatibility, maintaining >95% cell viability at effective concentrations (≤1000 μM in HeLa cells, ≤300 μM in HLEB3 cells) after 48 h treatment (CCK-8 assay, Figure S9). These results collectively demonstrate Cand.2’s capacity to specifically reverse pathological aggregation of the G129C mutant.

Conclusion

In summary, this study delineates a pathological pathway for the G129C cataractogenic mutation: the acquired cysteine residue substantially lowers the oxidative threshold of γC, initiating a deleterious phase transition that culminates in protein aggregation. Beyond elucidating this mechanism, we have identified an affinity peptide, Cand.2, which specifically targets and dissolves G129C-induced aggregates. For its highly cationic, arginine-rich, and hydrophilic composition, Cand.2 may modulate pathological protein aggregates through noncovalent interaction, , extending peptide-based antiaggregation strategies from amyloid aggregates toward amorphous aggregates. , Our work establishes a proof-of-concept for a targeted therapeutic strategy against hereditary cataracts, moving beyond broad-spectrum antioxidant approaches to directly counter the protein aggregation caused by specific mutations. The mutant-specific affinity peptide described here presents a promising lead for the development of nonsurgical interventions, offering a potential paradigm shift in the management of Crystallin aggregation-related eye diseases. It is noteworthy that, although numerous cataract-associated Crystallin mutations have been identified, their precise pathogenic mechanisms remain largely elusive. , Future delineation of these molecular mechanisms promises to catalyze the development of diversified precision interventions. , Beyond peptide-based strategies, recent studies indicate that small molecules such as celastrol can also exert protective effects by suppressing the aggregation of specific mutants and potentiating chaperone activity, underscoring the vast potential of multitarget therapeutic strategies for the definitive treatment of hereditary cataracts.

Materials and Methods

Protein Expression and Purification

The DNA sequences of γC WT and mutated γC G129C, γC G129C/C130A or γD R36S were subcloned into a pET28a­(+) vector and expressed in Escherichia coli BL21­(DE3) pLysS. Cells were grown in LB medium to OD₆₀₀ of 0.8–1.2 at 37 °C and induced with 0.5 mM IPTG at 37 °C for 4–6 h. Cells were harvested by centrifugation at 6000g for 10 min at 4 °C, resuspended in lysis buffer (20 mM MES, 1 mM EDTA, 1 mM DTT, pH 6.0), and then lysed by sonication (supplemented with 1 mM PMSF). Lysates were centrifuged at 12,000g for 15 min at 4 °C. The supernatant was loaded onto a cation exchange column (HiTrap SP Sepharose, Cytiva) pre-equilibrated with the lysis buffer. Target proteins were eluted using extra 100 mM NaCl in lysis buffer. Target protein fractions identified by SDS-PAGE were pooled, concentrated by ultrafiltration, and further purified by size-exclusion chromatography. The protein solution was loaded onto a pre-equilibrated (equilibrium buffer: 1 × PBS, 5 mM EDTA, 5 mM DTT, pH 7.4) column (Superdex 75 10/30 GL, Cytiva). The elution was monitored by UV absorbance at 280 nm, and fractions were collected based on the expected molecular weight range of the target protein.

The mutated γC G129C was cloned into a pET28a­(+) vector with a N-terminal His₆-tag and expressed in E. coli BL21­(DE3) pLysS. For preparation of ¹⁵N-labeled protein, ¹⁵NH₄Cl was used in the M9 minimal medium. The γC protein was purified by Ni-followed by gel filtration (Superdex 75 10/30 GL, Cytiva) and NTA affinity chromatography (Ni-NTA Beads 6FF, LABLEAD).

NMR Spectroscopy

Samples for NMR spectroscopy were prepared in a 90% H₂O/10% D₂O solvent with 20 mM potassium phosphate buffer (pH 7.0) containing 20 mM NaCl, 5 mM DTT. The protein concentration was 0.5 mM. ¹H and ¹⁵N-Heteronuclear Single Quantum Coherence (HSQC) spectra were performed at 298 K on Bruker Avance 600 MHz spectrometers equipped with a cyroprobe. All spectra were processed with NMRPipe and analyzed with NMRView.

Turbidimetric Analysis of Protein Aggregation Kinetics

Solution turbidity was monitored by Synergy H1 (Biotek) at 350 nm wavelength. The protein, in the presence or absence of peptides, was transferred to a 96-well plate with a final volume of 150 μL per well. The 96-well plate was placed in a plate reader and incubated at 37 °C (or as otherwise specified) with continuous shaking at 220 rpm. The OD₃₅₀ was recorded at 3 min intervals for 3 h. The resulting kinetic profiles were obtained by plotting OD₃₅₀ against time.

Reconstitution of Peptide Assemblies in Solution

The peptides (purity >98%) were obtained from Anhui Guoping Pharmaceutical Co. Ltd. The lyophilized powders were initially dissolved in 1,1,1,3,3,3-hexafluoro-2-propanol (HFIP; Innochem) to promote peptide unfolding. Following evaporation under a nitrogen gas stream, the residual peptides were resuspended in 1 × PBS (pH 7.4).

Peptide Concentration Determination

For peptides containing tyrosine (Y) or tryptophan (W), the sample was dissolved in 6 M guanidinium chloride aqueous solution to ensure complete denaturation. Peptide concentrations were determined using a UV–vis spectrophotometer (PerkinElmer Lambda 450, USA). The absorbance (A) at 280 nm was measured and correlated with the molar extinction coefficient (ε), optical path length (b), and peptide concentration (c) according to the Beer–Lambert Law ,

$$A=\epsilon bc$$ 1

where ε­(W) = 5690 M^–1 cm^–1 and ε­(Y) = 1280 M^–1 cm^–1 at 280 nm. If the peptide contains more than one W or Y, or a mixture, then the extinction coefficients are added. If the peptide lacks W or Y, its concentration was determined by weighing out.

DIC Microscopy Imaging

A droplet of 5 μL sample was placed onto a coverslip (2% w/v BSA coating). The sample was immediately imaged using a DIC-equipped inverted microscope (Leica DMi8) with a 63× Oil immersion lens. Images were captured using a high-resolution digital camera. All observations were conducted at room temperature.

TEM Imaging

A 10 μL sample aliquot was deposited onto a 200-mesh Formvar/carbon-coated copper grid and allowed to adsorb for 2 min at room temperature. Excess solution was removed by blotting with filter paper, and the grid was air-dried overnight. Subsequently, the grid was stained with 1% phosphotungstic acid for 30 s and air-dried again overnight. Samples were examined using a Hitachi H-7650 transmission electron microscope (Hitachi, Tokyo, Japan).

FTIR Spectroscopy

A 100 μL aliquot of the sample was deposited onto a CaF₂ window and dried completely. FTIR spectroscopy was performed on a PerkinElmer Spectrum One spectrometer. Spectra were recorded from 4000 to 450 cm^–1 at a resolution of 2 cm^–1, with 3 accumulations per scan. The amide I band (1600–1700 cm^–1) was examined to assess protein secondary structure. Peak deconvolution and second-derivative analysis were conducted using PeakFit (v4.12) to resolve overlapping peaks and quantify the relative percentages of structural components.

FRAP Analysis

FRAP was performed on a Leica Stellaris 5 confocal microscope. The proteins (WT γC and the G129C mutant) were fluorescently labeled using an Alexa Fluor 488 protein labeling kit (Invitrogen, A10235). For each aggregate, a small region of interest (ROI) was selected, and baseline fluorescence was recorded under low laser intensity. The selected ROI was then photobleached with a high-intensity laser pulse. Fluorescence recovery was monitored by time-lapse imaging until the signal reached a plateau. Recovery curves were generated by normalizing the fluorescence intensity in the bleached ROI to its prebleach value, with correction for background photobleaching using an unbleached reference region. All experiments were performed in at least three independent replicates, and data were analyzed using GraphPad Prism software.

Peptide Microarray Screening for Affinity Peptide Identification

A custom peptide microarray comprises 264 peptides, each with a biotin moiety at either the N- or C-terminus (peptide sequences are provided in Table S1). Using an Arrayjet microarrayer, all peptides were printed in parallel and in duplicate onto a 3D-functionalized slide surface (Capital Biochip Corp). The array included negative controls (PBS, 100 μg/mL BSA (Sigma-Aldrich)). The prepared microarrays were stored at −20 °C until use.

WT γC and G129C γC were conjugated with fluorescence by Protein Labeling Kits (Alexa Fluor 488, A10235 Invitrogen). Labeled proteins were diluted to 0.5 mg/mL in 1× PBS (pH 7.4). Peptide microarrays were incubated with labeled protein separately (200 μL) for 1 h at room temperature in the dark, followed by three-times TBST buffer washes. Fluorescence images were acquired on an imager equipped with a 488 nm channel at 10 μm pixel size (Typhoon, FLA9500). Images were quantified in ImageQuant, extracting the mean fluorescence intensity of two circle areas per peptide. Peptides were selected based on the criterion that the mean fluorescence intensity for its binding to the G129C mutant exceeded that for WT γC by a difference higher than 5.

Light Scattering Measurements

Static light scattering measurements were conducted using a Zetasizer Nano ZS90 instrument (Malvern) at room temperature (25 °C). A 4 mW, 633 nm laser was used to illuminate the sample cell, and the scattering angle was set at 90°. Samples were equilibrated for 30 s before measurement, and each sample was measured in at least three replicates.

BLI Analysis

The peptide N-terminus was biotinylated for the immobilization onto the sensor surfaces. Biotinylated peptides were dissolved to 2 μM in 1× PBS for the immobilization onto Super Streptavidin Octet biosensors (ForteBio). Aliquots (200 μL) of 1× PBS buffer, peptide solution (in 1× PBS), or target protein solution (in 1× PBS) were dispensed into designated wells of a 96-well plate (ForteBio). Prior to binding measurements, BLI sensors were equilibrated by incubating in 1× PBS for 10 min. All affinity measurements were performed at 30 °C according to the following steps: (1) a 120 s baseline was established in 1× PBS; (2) sensors were loaded with 2 μM biotinylated peptide for 180 s; (3) a second baseline was recorded in 1× PBS for 180 s; (4) association with the target protein was monitored for 180 s; (5) dissociation was carried out in 1× PBS for 600 s. Data were analyzed using the ForteBio Data Analysis software.

BLI response R(t) was modeled by a Langmuir binding scheme. In experiments, peptides (L) were immobilized on the sensor surfaces. Proteins (A) in the liquid phase were transported to bind to L. The stoichiometric ration of A/L is marked as n. The association and dissociation reactions can be represented as

$$nA+L\stackrel{k_{\mathrm{a}}}{\to }{A}_{n}L{A}_{n}L\stackrel{k_{\mathrm{d}}}{\to }nA+L$$

For such reactions, the rate equation can be written as

$$\frac{\mathrm{d}R}{\mathrm{d}t}=k_{\mathrm{a}}{C}_{\mathrm{a}}(n{R}_{\max }-R)-k_{\mathrm{d}}R$$ 2

Herein, this rate equation is represented based on the percent sensor surface coverage of the adsorbate A, marked as R. C _a is the concentration of A. The k _a is referred to the association-rate constant, and k _d is referred to the dissociation-rate constant. The R _max is maximum percent coverage at saturated analyte binding. The t is the time.

When reactions reach an equilibrium state, here comes the assumption that

$$\frac{\mathrm{d}R}{\mathrm{d}t}=0$$ 3

It can be rewritten as

$$k_{\mathrm{a}}{C}_{\mathrm{a}}(n{R}_{\max }-R)-k_{\mathrm{d}}R=0$$ 4

Denote the static association rate constant as K _A, and the static dissociation rate constant as K _D

$${K_{\mathrm{A}}}_{=}\frac{k_{\mathrm{a}}}{k_{\mathrm{d}}}$$ 5

$$K_{\mathrm{A}}{C}_{\mathrm{a}}(n{R}_{\max }-R)-R=0$$ 6

The eq can be linearized as

$$\frac{1}{R}=\frac{1}{n{R}_{\max }{K}_{\mathrm{A}}{C}_{\mathrm{A}}}+\frac{1}{{nR}_{\max }}$$ 7

For K _D and K _A with relationship as

$$K_D=\frac{1}{K_{\mathrm{A}}}$$ 8

The eq can be linearized as

$$\frac{1}{R}=\frac{K_{\mathrm{D}}}{n{R}_{\max }{C}_{\mathrm{A}}}+\frac{1}{{nR}_{\max }}$$ 9

For each analyte concentration C _A, an apparent equilibrium response R was obtained by averaging the response of associating phase at the end of the 2 s.

Dissociation kinetics were analyzed assuming a monoexponential decay of the sensor surface response

$$\mathrm{\Delta }R(t)=\mathrm{\Delta }{R}_0{e}^{-k_{\mathrm{d}}t}$$ 10

where ΔR(t) is the baseline-corrected response at time t and ΔR ₀ is the response at the start of dissociation. Taking natural logarithms gives a linear relationship

$$\mathrm{I}\mathrm{n}\mathrm{\Delta }R(t)=\mathrm{I}\mathrm{n}\mathrm{\Delta }{R}_0-k_{\mathrm{d}}t$$ 11

For each of the analyte concentrations, selected dissociation time in which the response followed an exponential decay (450–520 s in this study).

Cell Culture and Transfection

The HeLa cells were obtained from the Peking Union Medical College Cell Culture Center. The HeLa cells were cultured in Dulbecco’s modified Eagle’s medium (DMEM; Gibco) containing 10% fetal bovine serum (FBS; Gibco) and 1% Pen/Strep (VivaCell) at 37 °C in 5% CO₂ incubator. The HLEB3 cells were purchased from the BeNa Culture Collection (BNCC). The HLEB3 cells were cultured in Minimum Essential Medium (MEM; Gibco) containing 10% FBS and 1% Pen/Strep at 37 °C in 5% CO₂ incubator.

The genes of the WT γC and G129C mutant were inserted into the eukaryotic expression vector pEGFP-N1 (XhhoI-BamHI). Plasmids were dissolved in 1× Tris–EDTA buffer. To assess the effect of γC WT and G129C intracellular aggregation, HeLa or HLEB3 cells were transfected with of pEGFP-N1-γC WT or pEGFP-N1-γC G129C. Cells were grown to 80–90% confluence before transfection with plasmids using Lipofectamine 3000 transfection reagents (Thermo Fisher Scientific) at a 1:1.5 ratio (100 ng vector/1.5 μL reagent), according to the manufacturer’s instructions. Before transfection, cells were serum-starved by Opti-MEM (Gibco) for 2 h. After transfection for 4 h, the cells were cultured in fresh media for another 24 h, and then analyzed by fluorescence microscopy.

To assess the effect of peptide Cand.2 on the G129C mutant aggregation, HeLa cells were transfected with the pEGFP-N1-γC-G129C plasmid. Cells were grown to 80–90% confluence before transfection with plasmids using Lipofectamine 3000 transfection reagents (Thermo Fisher Scientific) at a 1:1.5 ratio (100 ng vector/1.5 μL reagent), according to the manufacturer’s instructions. Before transfection, cells were serum-starved by Opti-MEM (Gibco) for 2 h. After transfection for 4 h, the cells were cultured in fresh media for another 12 h. The rescue experiments were performed by addition of peptide Cand.2 to fresh culture medium and cultured for 24 h.

Fluorescence Microscopy

Cells were washed three times with 1× PBS and fixed with 4% paraformaldehyde for 40 min at room temperature. After three additional PBS washes, nuclei were counterstained with Hoechst 33342 (LABLEAD). Samples were mounted and imaged using an EVOS FL Auto 2 microscope (Thermo Fisher Scientific).

SDS-PAGE

Protein samples were mixed with SDS-PAGE loading buffer (LABLEAD, G2526) and denatured. Denatured samples (10 μL) and a molecular weight marker (LABLEAD, P1018) were loaded onto a 4–20% precast polyacrylamide gel (LABLEAD, P42015). Electrophoresis was performed at 120 V in MOPS running buffer. Subsequently, the gel was stained with Coomassie Brilliant Blue R-250 for 2 min and destained in H₂O for 30 min until clear bands were visible. The gel was visualized using a gel documentation system, and protein bands were analyzed by comparing their migration distances to the molecular weight marker.

Western Blot Analysis

Proteins were extracted by using RIPA Lysis Buffer (Cocktail) from HeLa cells. Proteins were separated by SDS-PAGE and then transmembrane to PVDF. The membranes were incubated with primary antibodies against EGFP (Bioss, bs-2194R) or GAPDH (Cell Signaling, 2118T), followed by detection with secondary antibodies conjugating HRP.

Cell Viability Assay

The cytotoxicity of peptide Cand.2 was evaluated using a Cell Counting Kit-8 (CCK-8) assay. Cells were seeded in a 96-well plate at a density of 5000 cells per well and allowed to adhere for 12 h. The cells were then treated with various concentrations of the peptide in fresh medium for 48 h. Subsequently, 10 μL of CCK-8 solution was added to each well, followed by a 3 h incubation. Absorbance at 450 nm (A ₄₅₀) was measured using a Synergy H4 microplate reader (BioTek, USA). Cell viability was calculated relative to untreated control cells (set to 100%).

Live-Cell Imaging

Live-cell imaging was performed using an Incucyte S3 Live-Cell Analysis System (Sartorius). Cells were seeded at an appropriate density in 48-well plates (Corning, no. 3548) and cultured/transfected according to standard protocols. Following treatment with peptide Cand.2, time-lapse imaging was conducted over 24 at 1 h intervals, using a 20× objective to capture both GFP and bright-field channels.

Visualization

Structure visualization was carried out using PyMol. The protein structure file (2NBR, PDB format) was obtained from the Protein Data Bank. Mutating residue was performed using the software’s command line.

Statistical Analysis

All experiments were repeated at least three times. Divergences between the experimental and control groups were analyzed using Student’s t-test. Statistical significance was set at *P < 0.05, **P < 0.01, and ***P < 0.001. GraphPad Prism 9.0 (GraphPad Software, Inc., San Diego, CA, USA) was used for the statistical analysis.

The Supporting Information is available free of charge at https://pubs.acs.org/doi/10.1021/jacsau.6c00727.

• Method for peptide property calculation. ¹H–¹⁵N HSQC of the G129C mutant (Figure S1); Aggregation kinetics of the G129C mutant in 1 × PBS (pH 7.4) (Figure S2); The impact of exogenous hydrogen peroxide on the G129C mutant (Figure S3); Aggregation kinetics of the G129C/C130A double mutant under oxidative conditions (Figure S4); Sequence length and properties of the 264-member custom peptide microarray library (Figure S5); Amino acid composition and residue-class distribution of the custom peptide microarray library (Figure S6); AlphaFold3 prediction of the interaction between G129C γC-Crystallin and Cand.2 or Cand.2 variants (Figure S7); Dynamic light scattering count-rate analysis of Cand.2 (Figure S8); TEM images of G129C mutant aggregates with or without the incubation with peptide Cand.2 (Figure S9); In situ bright-field microscopy images of γD-Crystallin R36S crystalline assemblies treated with Cand.2 (Figure S10); Image of HLEB3 cells transfected with eGFP-tagged WT γC or the G129C mutant (Figure S11); Images of HeLa cells transfected with eGFP-tagged G129C mutant and treated with Cand.2 (Figure S12); Western blotting analysis of eGFP-tagged G129C in HeLa cells (Figure S13); The cell cytotoxicity of the peptide Cand.2 evaluated by the CCK-8 assay for HLEB3 and HeLa cells (Figure S14); Peptide amino acid sequences and structural modifications of peptide microarray (Table S1); Comparison of sequence and properties of Cand.1-Cand.6 (Table S2) (PDF) (PDF)

C.W., W.Z. and L.Y. conceived the study and designed the research. M.C., Z.J., Z.D., S.M., Zhenyan Li., Y.W., H.L. and Zhenlin Liu. conducted the experiments. All the authors analyzed the data. C.W., W.Z., M.C. and L.Y. wrote the paper.

The authors declare the following competing financial interest(s): W.Z., M.C., C.W. and L.Y. are inventors on a pending Chinese patent related to this work filed by Institute of Basic Medical Sciences & School of Basic Medicine, Chinese Academy of Medical Sciences & Peking Union Medical College (application no. CN202510620221.7). The authors declare no other compet-ing interests.