Sequence‐Defined Short Peptide‐Derived Coacervate Vesicles for Targeted Therapeutics

ABSTRACT

Compartmentalization underpins biological organization and the design of programmable therapeutic carriers.  Here, we report a sequence‐defined short peptide, Biotin‐Ser(‐OPO₃ ^{2 −})‐Phe‐Phe‐Arg (Biotin‐SR), that undergoes physiological liquid–liquid phase separation to form lipid‐free coacervate vesicles (CVs) with enhanced structural persistence.  The designer peptide integrates oppositely charged functionalities that drive pH‐responsive coacervation, incorporating an alkaline phosphatase (ALP)‐cleavable anionic phosphoester, a diphenylalanine motif that promotes hydrophobic packing, and a cationic guanidinium group that stabilizes the coacervate architecture while serving as a precursor for nitric oxide (NO) generation. These CVs selectively target cancer cells via biotin‐mediated recognition. CVs of Biotin‐SR with pronounced morphological persistence, as confirmed by restricted molecular exchange in FRAP analysis, and encapsulate glucose oxidase (GOx) within this stabilized vesicular architecture, undergo ALP‐induced enzymatic dephosphorylation in ALP‐rich tumor environments to trigger an irreversible loss of charge complementarity, GOx release and a synergistic therapeutic cascade integrating enzymatic starvation, oxidative stress amplification, and NO‐mediated mitochondrial dysfunction. The resulting glucose depletion couples with NO generation to elevate reactive oxygen species (ROS), inducing mitochondrial dysfunction and suppressing tumor growth in both spheroids and xenograft models. This work demonstrates how minimal peptide sequences can translate liquid‐liquid phase separation into stable coacervate vesicles with enzyme‐responsive supramolecular function for targeted therapeutics.

Alkaline phosphatase‐triggered peptide coacervate vesicles integrate metabolic starvation with nitric oxide gas therapy, enabling synergistic tumor‐selective cancer therapy through enzyme‐responsive supramolecular transformation.

1. Introduction

Biological organization relies on compartmentalization, often mediated by lipid membranes [1]. While weak multivalent interactions can form dynamic, membrane‐less condensates [2, 3, 4, 5, 6, 7], these structures can suffer from coalescence and limited boundary definition [8]. Recent efforts have focused on coacervate vesicles (CVs)—lipid‐free condensates with membrane‐mimetic boundaries providing enhanced structural persistence and encapsulation capabilities [9, 10, 11]. CVs have emerged as vesicle‐like condensates formed through programmed electrostatic complementarity, hydrophobic interactions, and charge patterning [12, 13].

Unlike simple droplets, CVs exhibit membrane‐mimetic boundaries that confer enhanced structural persistence and encapsulation capability [9, 10, 11]. Diphenylalanine‐based peptide assemblies are known to promote LLPS, yielding metastable disordered condensates [14]. Spruijt and co‐workers demonstrated that a disulfide‐linked hydrophobic diphenylalanine derivative undergoes self‐coacervation at sub‐millimolar concentrations, forming functional microcompartments [15]. The same group further showed that tyrosine‐rich short peptide derivatives assemble into coacervates with semi‐permeable, membrane‐like boundaries arising from zwitterionic organization, imparting enhanced stability and size‐selective permeability (molecular weight cut‐off ≈2.5 kDa) [10]. More recently, Liu, Li, Gu, and co‐workers reported that CVs formed via LLPS of DNA and nucleoproteins enhance cellular uptake of oncolytic viruses through a receptor‐independent, energy‐dependent macropinocytosis pathway [9]. In a related context, Vicentini et al. demonstrated enzyme loading within nanocoacervates generated by polyanion–polycation complexation [11]. These studies establish that LLPS can generate vesicle‐like coacervate condensates with membrane‐mimetic boundaries and appreciable structural integrity. However, whether short‐peptide–derived CVs can be engineered to simultaneously retain condensate stability while enabling stimulus‐responsive and target‐specific therapeutic delivery remains unresolved [9]. In most reported CV systems, vesicular integrity arises primarily from electrostatic complexation between oppositely charged components, which governs interfacial enrichment and boundary stabilization [9, 10, 11]. We hypothesized that embedding electrostatic complementarity directly within a short peptide, while preserving its intrinsic aromatic sticker motif and avoiding flexible spacers, would sustain LLPS and enhance interfacial cohesion, thereby yielding mechanically persistent yet responsive vesicle‐like condensates. Functionalization with a targeting ligand was further anticipated to bias cellular engagement and reduce nonspecific interactions [16, 17]. To test this hypothesis, we designed a short peptide, Biotin‐Ser(‐OPO₃ ^{2 −})‐Phe‐Phe‐Arg (Biotin‐SR), integrating tumor targeting, enzymatic responsiveness, and therapeutic functionality within a single molecular framework. Biotin‐SR comprises a biotin ligand, an ALP‐responsive phosphoester, and an arginine residue that provides electrostatic complementarity and serves as a latent NO source. At physiological pH, multivalent interactions between the anionic phosphoester, phenyl ring and the delocalized guanidinium group of arginine (Figure S44) drive LLPS and the formation of vesicle‐like coacervate condensates (Figure 1a). Recently, Cheng et al. demonstrated that guanidinium‐phosphate interactions can effectively drive LLPS [18]. Arginine is particularly effective owing to its charge delocalization and interaction multiplicity [12]. The diphenylalanine motif further stabilizes the assembly through hydrophobic packing, enhancing interfacial cohesion [6, 19, 20]. These multivalent interactions kinetically arrest vesicle coalescence by increasing the fusion energy barrier and restricting membrane fusion [21]. Together, these interactions enable Biotin‐SR to form stable CVs under physiological conditions while encapsulating GOx. Minimal Fluorescence Recovery after Photo bleaching (FRAP) confirms a membrane‐like coacervate vesicle architecture, imparting structural stability essential for sustained cargo loading and therapeutic functionality. Biotin receptors that mediate ligand‐induced endocytosis are overexpressed in many aggressive cancers [16], while ALP activity is elevated in tumor microenvironments [22]. ALP‐catalysed dephosphorylation disrupts local charge complementarity at the vesicular interface, destabilizing the CVs (Figure 1b) and triggering localized release of encapsulated GOx [23]. The released GOx catalyses intracellular glucose oxidation with concomitant hydrogen peroxide (H₂O₂) generation, inducing metabolic stress associated with starvation therapy [24, 25, 26]. The locally generated H₂O₂ further reacts with arginine to yield NO, initiating gas‐mediated cytotoxicity and apoptotic signalling [27]. Consistent with this design logic, control peptides establish that arginine's delocalized guanidinium charge is indispensable for LLPS‐driven CV formation, yet it is distinguished by the encapsulation of GOx within the core while maintaining structural integrity that can be selectively destabilized by an enzymatic (ALP) trigger and coupled therapeutic function [28, 29]. Together, these finding demonstrate structurally stable CVs that integrate stimulus responsiveness, target specificity, and controlled therapeutic release, addressing a largely unexplored design space in the contemporary literature (Figure 2).

FIGURE 1.

(a) Molecular structures of the short peptide system, Biotin–SR and control molecules (Biotin–SK and Biotin–SE), along with a schematic illustration of coacervate vesicle (CV) formation driven by multivalent intermolecular interactions; (b) Schematic illustration of the liquid–liquid phase separation (LLPS) of Biotin–SR into coacervate vesicles, followed by ALP‐mediated dephosphorylation‐driven transformation into fibrous assemblies.

FIGURE 2.

Schematic illustration of the synergistic metabolic starvation and gas therapy mediated by GOx@Biotin‐SR coacervate vesicles in biotin receptor and ALP overexpressing cancer cells. Upon ALP‐triggered activation, GOx is released and catalyzes cellular glucose to form H₂O₂, which further reacts with Arg to form NO. This induces the MCF7 cancer cell death through ROS generation and mitochondrial dysfunction.

2. Results and Discussion

Biotin‐SR, Biotin‐SK, and Biotin‐SE were synthesized by Fmoc solid‐phase peptide synthesis (SPPS) on Rink amide resin using HBTU/DIPEA coupling. Biotin was conjugated at the N‐terminus, and peptides were cleaved, lyophilized, and purified by RP‐HPLC (C18, water‐acetonitrile gradient). Product identity was confirmed by ESI‐MS and ¹H NMR and ³¹P NMR (Figures S1–S7). For self‐assembly visualization, FITC‐SR (λ_ex/λ_em = 491/516 nm) and RhB‐SR (λ_ex/λ_em = 546/567 nm) were prepared by replacing biotin with 5‐carboxyfluorescein and rhodamine B respectively, characterized by ESI‐MS mass (Figures S8 and S9), forming a FRET pair [30]. Following successful synthesis, lyophilized Biotin‐SR was dissolved in 1X PBS buffer (pH 7.4) to prepare a homogeneous 1 mm solution, to examine its coacervate‐forming ability. Upon dissolution, the solution immediately turned milky and turbid, indicative of coacervate formation [31]. The turbid solution was taken into a confocal disc, and bright‐field images were acquired using a Leica TCS SP8 fluorescence confocal microscope. Microscopy revealed the formation of nano‐ to micro‐metre‐sized CVs (Figure 3a).

FIGURE 3.

Characterization and morphological stability of Biotin–SR coacervate vesicles. (a) Bright‐field microscopy image of Biotin–SR (1 mm in 1× PBS, pH 7.4; stock: 5 mm in 99.9:0.1 water: DMSO, v/v) showing spontaneous coacervate droplet formation. (b) Macroscopic view illustrating liquid–liquid phase separation (LLPS) of Biotin–SR (1 mm, PBS). (c) Post‐centrifugation image (4,000 rpm, 3 min) revealing a peptide‐rich opaque coacervate phase. (d) Turbidity analysis (A₆₀₀) of Biotin–SR (0.005–5 mm) demonstrating concentration‐dependent coacervation. (e) Field‐emission scanning electron microscopy (FE‐SEM) image (750 µm, PBS, pH 7.4) displaying spherical vesicular morphology. (f, g) Cryogenic transmission electron microscopy (cryo‐TEM) and three‐dimensional atomic force microscopy (AFM) images of Biotin–SR (750 µm) confirming vesicle‐like droplet structures. (h) TEM image of Biotin–SK (750 µm) showing micelle‐like assemblies (∼500 nm). (i) TEM image of Biotin–SE (500 µm) exhibiting fibrillar nanostructures. (j) Raman spectra of Biotin–SR and Biotin–SE highlighting distinct conformational signatures between vesicular and fibrillar morphologies.

To examine the phase separation behavior of Biotin‐SR, a turbid solution in aqueous 1X PBS (pH 7.4) buffer was incubated overnight at room temperature, resulting in a biphasic system with a peptide‐rich condensed phase and a peptide‐depleted supernatant (Figure 3b) [32]. Upon centrifugation at 4000 rpm for 3 min, the coacervate‐rich phase exhibited distinct separation into two layers: a supernatant upper layer and a viscous lower layer, confirming LLPS (Figure 3c). A concentration‐dependent study using turbidity measurements at 600 nm revealed a sharp increase in turbidity at 750 µm, identifying the critical coacervate concentration (CCC) required for phase separation (Figure 3d) [14]. The self‐assembly behavior of Biotin‐SR was further investigated using bright field, cryogenic transmission electron microscopy (cryo‐TEM), field‐emission scanning electron microscopy (FE‐SEM), and atomic force microscopy (AFM) to confirm vesicle formation (Figure 3a,e–g) [33]. Cryo‐TEM imaging together with line‐profile intensity analysis revealed enhanced electron density at the vesicle periphery and a low‐density interior, confirming the hollow architecture of the coacervate vesicles (Figure S16). Consistent observations were further obtained from DIC microscopy, and optimized FE‐SEM imaging clearly revealed lumen‐containing vesicular structures, while partially broken coacervate vesicles (CVs) exhibited distinct hollow cavities, further supporting their vesicular architecture (Figures S17 and S18) [34].

The oppositely charged phosphate and guanidinium groups in Biotin‐SR promote CV formation through electrostatic interactions, presumably adopting an intermolecular arrangement that favors multivalent interactions. The resulting vesicles have an average diameter of ∼500 nm, consistent with LLPS‐driven CVs. In contrast, TEM imaging of the control peptide Biotin‐SK showed micelles with an average diameter of ∼500 nm (Figure 3h), likely reflecting weaker electrostatic complementarity, as lysine carries a localized positive charge, whereas arginine's delocalized guanidinium charge promotes assembly more effectively [12]. In contrast, Biotin‐SE with negatively charged termini, formed ill‐defined fibrous aggregates rather than vesicles (Figure 3i), underscoring the necessity of electrostatic interaction for coacervate assembly. To assess pH sensitivity, Biotin‐SR was examined via cryo‐TEM under acidic (pH 3–5) and basic (pH 10.5–14) conditions. No vesicular structures were observed outside physiological pH (7.4), highlighting the importance of appropriate protonation and deprotonation states of arginine and phosphate groups, respectively, for phase separation (Figure S19).

To elucidate the role of phosphate groups in coacervate formation, Raman spectra were recorded for Biotin‐SR and Biotin‐SE in solution (Figure 3j). Phosphate groups exhibited characteristic Raman bands, particularly in the regions of 950–1000 cm^{− 1} (ν₁, symmetric P–O stretching) and 1010–1150 cm^{− 1} (ν₃, asymmetric stretching) [35]. In Biotin‐SE, the observed peaks at 1001 and 1035 cm^{− 1} correspond to unperturbed phosphate vibrations, indicating negligible intermolecular interaction. On the contrary, Biotin‐SR displayed significant downshifts to 682 and 722 cm^{− 1}, suggesting altered vibrational modes due to strong electrostatic interactions and molecular clustering [12, 36, 37]. These spectral changes confirm the active role of phosphate‐mediated interactions in driving phase separation and stabilizing CV formation in Biotin‐SR.

To evaluate the structural stability of the CVs, time‐dependent imaging studies were performed. Coacervates typically encapsulate diverse molecular species, and this phenomenon was exploited to encapsulate different fluorescence‐active dye molecules, e.g., fluorescein and rhodamine. This was confirmed by the intense fluorescence from respective dye‐loaded droplets, as evidenced by the confocal laser scanning microscopic (CLSM) images (Figure 4a), and encapsulation studies using different fluorescent dyes also revealed enhanced peripheral fluorescence intensity, further supporting the hollow vesicular architecture of the coacervate assemblies (Figure S23). Notably, Biotin‐SR coacervate vesicles remained structurally stable after 48 h, as evidenced by bright‐field microscopy (Figure S20). FRAP analysis revealed a vesicular assembly exhibiting ∼56% fluorescence recovery for a freshly obtained CV after photo bleaching, indicating substantial molecular mobility (Figure S22). Notably, fluorescence recovery was much lower (∼18%) for the sample that was photo‐bleached following an additional 1 h incubation (Figure 4b,c), consistent with progressive interfacial rigidification and restricted molecular diffusion. In both instances of the FRAP analysis, data were collected till the plateau was achieved. These observations demonstrate the emergence of mechanically stabilized, membrane‐like CVs with tenable dynamics, underscoring their potential for controlled molecular encapsulation, synthetic protocell construction, and stimuli‐responsive delivery applications. Figure 4d demonstrates the hollow nature of the vesicles, as evidenced by the radial fluorescence intensity profile with the dye molecule preferentially localized at the peripheral region of the condensate [11]. Atomistic simulations of a 10 nm diameter vesicle in aqueous solution over 120 ns demonstrated structural stability (Figure 4e–h and Video S1).

FIGURE 4.

Time‐dependent stability, dynamic behavior, and molecular architecture of Biotin–SR coacervate vesicles. (a) Fluorescence and bright‐field microscopy images illustrating the temporal stability of Biotin–SR coacervate vesicles, monitored over time in the absence and presence of encapsulated fluorescein and rhodamine dyes. (b) Fluorescence recovery after photobleaching (FRAP) analysis of dye‐loaded vesicles. (c) Quantitative FRAP recovery profile showing limited fluorescence recovery after photobleaching, indicating the restricted dynamic nature of the vesicular membrane. (d) Radial fluorescence intensity analysis across individual vesicles confirming their hollow interior (aqueous lumen). (e,f) Molecular dynamics simulations of phosphorylated peptide‐based coacervate vesicles in water, revealing structural integrity and stability over a 100 ns simulation period. (g) Cross‐sectional snapshot at 100 ns showing a well‐defined aqueous lumen; arrowheads indicate phosphate groups preferentially oriented toward the inner vesicular surface. (h) Volume‐rendered cross‐section depicting the atomic‐scale architecture of the vesicle (carbon atoms shown in green; standard element coloring otherwise; nonpolar hydrogens and solvent molecules omitted for clarity).

¹H NMR spectra of Biotin‐SR and Biotin‐SE were recorded in DMSO‐d₆ and in a DMSO‐d₆ :D₂O (94:6 v/v) mixture. A comparison of the spectra reveals pronounced up field shifts across both aromatic and aliphatic resonances (Figure 5b,c and Figure S12) in the spectrum recorded for Biotin‐SR in DMSO‐d₆ :D₂O. This can be rationalized by the aggregation‐induced shielding effects arising from its self‐assembly. In contrast, Biotin‐SE exhibited minimal spectral changes under identical conditions, consistent with its inability to form CVs, which could be attributed to unfavorable electrostatic interactions [12, 38].

FIGURE 5.

(a) The chemical structure of the FRET pair FITC‐SR (donor, the excitation and emission wavelengths are 491/516 nm) and RhB‐SR (acceptor, the excitation and emission wavelengths are 546/567 nm), and the schematic illustration of their multivalent interaction to make the vesicle. Partial ¹H NMR spectra of Biotin‐SR (DMSO‐d₆ and DMSO‐d₆ : D₂O (94:6)) and Biotin‐SE (DMSO‐d₆ and DMSO‐d₆ : D₂O (94:6)) in the (b) aromatic and the (c) aliphatic region. The whole spectrum is shown in the Supporting Information. Distinct upfield shifts for the aromatic and aliphatic protons of Biotin‐SR in DMSO‐d₆ : D₂O (94:6) compared to only DMSO‐d₆ indicate the coacervate formation. In contrast, Biotin–SE exhibited only a marginal upfield shift, indicating the absence of coacervate formation. (d) FRET experiment from the 1:1 mixture of the FITC‐SR and RhB‐SR (100 µm concentration each in water) shows significant FRET occurs between FITC‐SR and RhB‐SR, and the FRET efficiency increases with time. (e) Fluorescence microscopic images of the coacervate vesicles. The green and red colors show the presence of FITC‐SR and RhB‐SR respectively in the coacervate vesicles, and their merged images showing the significant co‐localization. This supports the coacervate vesicle formation by the FITC‐SR and RhB‐SR through intermolecular multivalent interaction. Mechanism of self‐assembly. (f) The phosphorylated peptide in solution showing intramolecular interactions (pointed with arrows) and the electrostatic potential (ESP) map. Standard color representation used for elements, with carbon in grey. (g) Free energy profile as a function of intermolecular distance, r, showing a favorable intermolecular interaction of the phosphorylated peptide. (h) A lowest energy conformation showing antiparallel bimolecular assembly of the phosphorylated peptide, driven by favorable interactions between Arg and the phosphate moiety. Additional intermolecular hydrogen bonding with backbone atoms could also be observed. Ball‐and‐stick models with standard element colors are shown; nonpolar hydrogens are omitted for clarity. (i,j) Phase separation of the phosphorylated peptide into droplets in an aqueous environment. Solvent is omitted for clarity.

The upfield chemical shift changes (Δδ) for the key proton signals were quantified, and the values are summarized in Table S1 (Biotin‐SR) and Table S2 (Biotin‐SE). Notably, the aromatic proton resonances and guanidinium‐associated proton signals (c, f, g) exhibited significant upfield shifts, indicating strong involvement of π–π stacking and electrostatic interactions during the self‐assembly process. In addition, several aliphatic proton signals (i, k, l, m), corresponding to arginine, serine, and biotin moieties, also showed appreciable chemical shift changes, suggesting their participation in the supramolecular aggregation process.

To validate the proposed intermolecular interactions and the hypothesized multivalent interaction, including electrostatic, cation‐π self‐assembly of Biotin‐SR, further reinforced by π–π stacking between the diphenylalanine moieties—an equimolar mixture of FITC‐SR and RhB‐SR (100 µm each, in water) exhibited a time‐dependent increase in FRET efficiency. This progressive enhancement indicates decreasing donor‐acceptor separation, consistent with close spatial organization and the gradual formation of coacervate vesicles (Figure 5d). Furthermore, confocal fluorescence microscopy confirmed CV formation by the equimolar FITC–SR/RhB–SR mixture, where the green and red fluorescence signals correspond to FITC–SR and RhB–SR, respectively, and the merged image reveals significant co‐localization within the coacervate vesicles, supporting the intermolecular multivalent interactions driving self‐assembly (Figure 5e). Atomic simulation of the phosphorylated peptide exhibits strong intra‐ and inter‐molecular interactions; however, intermolecular interactions are favored over intramolecular interactions by ∼ −6 kcal/mol free energy, leading to spontaneous self‐assembly, driven by the electrostatic attractions between the Arg and the phosphate groups, backbone hydrogen bonding and the pi‐pi stacking interactions between the Phenyl group of the side chains. These collective interactions facilitate phase separation into membranous or vesicular structures. This orientation facilitates lateral growth and ultimately drives phase separation into membranous or vesicular structures, rather than micellar droplets with a hydrophobic core (Figure 5f–j, Video S2). A few earlier reports also support such a head‐to‐tail orientation [39, 40, 41, 42].

As discussed earlier, the phosphoester linkages within Biotin‐SR CVs are susceptible to enzymatic dephosphorylation by ALP [43], a process expected to disrupt the electrostatic interactions that stabilize the CVs. Incubation of Biotin‐SR CVs with ALP (0.5 U mL^{− 1}, 37°C, PBS, pH 7.4) induced a time‐dependent morphological transition, as visualized by cryogenic transmission electron microscopy (cryo‐TEM) (Figure 6a). Spherical droplets gradually elongated and transformed into fibrous nanostructures, reaching complete conversion after 16 h, consistent with ALP‐triggered dephosphorylation [44]. The reaction kinetics, using a solution of Bioton‐SR (1 mg/mL) incubated with ALP (0.5 U mL^{− 1}) in mili Q water, monitored by reverse phase HPLC in C‐18 column, revealed a progressive decline of the Biotin‐SR peak and the concomitant increase of a new peak at a higher retention time, corresponding to the dephosphorylated species (Figure 6d) [23, 45]. As anticipated, zwitterionic Biotin‐SR CVs exhibited a near‐neutral surface charge, which shifted to +15.8 mV after ALP treatment, reflecting the increased net positive charge due to the dephosphorylation process (Figure 6e). These results demonstrate an ALP‐triggered phase transition from coacervate droplets to fibrous assemblies through phosphoester bond cleavage.

FIGURE 6.

ALP‐triggered dephosphorylation and disassembly of Biotin‐SR coacervate vesicles. (a) Cryogenic transmission electron microscopy (cryo‐TEM) images showing time‐dependent morphological evolution of Biotin‐SR coacervates upon ALP treatment (0.5 U mL^{− 1}), revealing transformation from spherical vesicles to fibrous structures. (b,c) Kelvin probe force microscopy (KPFM) images of Biotin‐SR and Biotin‐SR + ALP, highlighting surface potential changes accompanying enzymatic dephosphorylation. (d) HPLC chromatograms of Biotin–SR ( 1mg/ml, mili Q water) at successive time points after ALP exposure, showing decay of the parent peak and emergence of a higher‐retention dephosphorylated product. (e) Zeta‐potential measurements of Biotin‐SR with and without ALP (0.5 U mL^{− 1}), confirming charge neutralization and vesicle disassembly. (f) Fluorescence spectra of GOx–Biotin‐SR mixtures in water monitored over 0 – 12 h, showing a gradual decrease in GOx emission (λ_em = 343 nm) consistent with encapsulation within Biotin‐SR CVs. (g) Release kinetics of GOx from GOx@Biotin‐SR in the presence of ALP, where the progressive recovery of GOx fluorescence reflects ALP‐induced cargo release. (h) ESP map of the vesicle surface. (i) ESP map of the vesicle inner surface, shown in cross‐section. (j) Interaction of the phosphorylated peptide (ball‐and‐stick) with the glucose oxidase (GOx) dimer (ESP map). (k) Free energy profile of the bimolecular interaction of the dephosphorylated peptide. (l) A lowest energy conformation of the dephosphorylated peptide, in contrast to the phosphorylated form. Backbone zipping and cross‐β formation are observed (Figure S19). The arrowhead indicates an i–i′ cross‐backbone hydrogen bond. A counterion stabilizes Arg in the parallel configuration. (m) AlphaFold3‐predicted model of cross‐β fibers formed by the dephosphorylated peptide.

To confirm the dephosphorylation of Biotin‐SR, Kelvin Probe Force Microscopy (KPFM) was performed on drop‐cast films prepared on FTO‐coated substrates. Untreated Biotin‐SR assemblies exhibited a surface potential in the range of ∼104–107 mV, consistent with a zwitterionic structure. Upon ALP treatment, the surface potential increased significantly to ∼840–842 mV, corroborating further the dephosphorylation process induced by ALP (Figure 6b,c).

GOx has an intrinsic fluorescence originating from its tightly bound flavin adenine dinucleotide (FAD) cofactor [46]. To investigate the encapsulation efficiency of GOx (1 mg/mL), it was co‐incubated with Biotin‐SR (2 mg/mL) in an aq. medium. Time‐dependent steady state fluorescence spectroscopy (λ_ex/λ_em: 280/343 nm) revealed a gradual decrease in emission intensity over 12 h, indicative of GOx encapsulation within the coacervate vesicles (Figure 6f).

To exclude the possibility of intrinsic self‐quenching of GOx during prolonged fluorescence monitoring, a control experiment was performed using free GOx under identical experimental conditions (Figure S25). Only a minor decrease in fluorescence intensity (∼5.17%) was observed over 12 h, indicating negligible self‐quenching of the fluorophore itself. In contrast, GOx@Biotin‐SR displayed a significantly greater reduction in fluorescence intensity, suggesting that the observed fluorescence decrease primarily originates from encapsulation within the coacervate phase rather than intrinsic fluorophore quenching. These results confirm that the fluorescence changes reported in Figure 6f are predominantly due to GOx encapsulation. After accounting for the background quenching effect originates from free GOx, the overall encapsulation efficiency of GOx in GOx@Biotin‐SR CVs was calculated to be approximately 68%. Finally, ALP‐triggered release of encapsulated GOx was examined. Upon treatment with ALP (0.5 U mL^{− 1}) with encapsulated GOx within Biotin‐SR CVs (GOx@Biotin‐SR), a time‐dependent increase in fluorescence intensity was recorded (Figure 6g), confirming the controlled enzymatic release of GOx from the deassembled CV matrix. The cumulative release efficiency was determined to be ∼59% of the initially encapsulated enzyme after 12 h, demonstrating the potential of this system for stimulus‐responsive drug delivery applications [47]. To verify encapsulation at the molecular level, Diffusion‐Ordered Spectroscopy (DOSY) NMR was performed in D₂O [48]. Free GOx and Biotin‐SR exhibited individual diffusion coefficients of 7.37 × 10^{− 1 1} m²/s and 10.52 × 10^{− 1 1} m²/s, respectively. In contrast, the mixture incubated for 12 h exhibited a single diffusion coefficient of 1.32 × 10^{− 10} m²/s, suggesting formation of a supramolecular adduct, GOx@Biotin‐SR (Figures S13–S15).

Electrostatic potential (ESP) mapping of the vesicle surfaces, obtained from computational studies, revealed a balanced distribution of negative and positive charges arising from the orientation of the (RO)PO₃ ^{2 −} groups and the cationic arginine residues toward the water interface (Figure 6h,i). The simulations further demonstrated that the phosphorylated peptide interacts with GOx through preferential binding of the phosphate groups to positively charged surface patches on the enzyme (Figure 6j and Video S3). Importantly, these interactions do not involve the catalytic site of GOx, indicating that the enzyme can be suitably incorporated into the coacervate vesicles without compromising its activity, thereby rendering the system suitable for cargo encapsulation and delivery applications. Interestingly, dephosphorylation of Biotin‐SR by ALP shifts the assembly preference from an antiparallel arrangement to a parallel alignment. The latter configuration enables i→i’ cross‐backbone hydrogen bonding, which subsequently propagates through a zipping mechanism to form ordered cross‐β structures that extend into fibrillar aggregates (Figure 6k–m and Figure S27). Counter ions further stabilize the charged side chains within these parallel assemblies.

To probe enzyme encapsulation and ALP‐triggered release of GOx from GOx@Biotin‐SR CVs, circular dichroism (CD) spectroscopy was performed [49]. CD spectra of Biotin‐SR (1 mm) and GOx (1 mg mL^{− 1}) in aqueous 1X PBS (pH 7.4) displayed the anticipated random‐coil and α‐helical conformations, respectively (Figure 7a). Quantitative BeStSel analysis indicated ∼ 51% random‐coil content for Biotin‐SR and 60.1% α‐helicity for GOx (Figure S24). In GOx@Biotin‐SR CVs, the CD profile was dominated by random‐coil features (∼ 45%), reflecting suppression of the enzyme's α‐helical signature due to confinement within the coacervate matrix. Upon ALP treatment (0.5 UmL^{− 1}, 12 h), dephosphorylation triggered CV disassembly and the subsequent release of GOx into solution, evidenced by the re‐emergence of a distinct α‐helical band (∼47% helicity). This spectrum closely resembled that of the native enzyme, confirming structural recovery and successful enzymatic release following ALP‐mediated CV disassembly.

FIGURE 7.

Enzymatic stability, NO generation, and cellular response of GOx@Biotin‐SR coacervate vesicles. (a) Circular dichroism (CD) spectra of GOx, Biotin‐SR, GOx@Biotin‐SR, and GOx@Biotin‐SR + ALP in PBS (pH 7.4), showing recovery of the α‐helical signature of GOx after ALP treatment. (b) Enzymatic activity comparison of free GOx and GOx@Biotin‐SR using the RhBNs fluorescent probe. (c) In‐vitro NO release from GOx@Biotin‐SR in the presence of glucose and ALP, monitored by the RhBNs fluorescence titration assay. (d) Griess assay for NO quantification, showing a time‐dependent (0.5–8 h) increase in absorbance at 540 nm upon incubation of GOx@Biotin‐SR with glucose + ALP; the signal further intensified in a H₂O₂‐concentration‐dependent manner (1 and 2 mm). (e) EPR spectra of GOx@Biotin‐SR + glucose + ALP using PTIO as the NO spin trap, confirming characteristic seven‐line splitting. (f) Time dependent DLS study (g) Serum stability of Biotin–SR determined by HPLC after incubation in human serum (37°C), showing ≈ 61% peptide retention after 24 h. (h) Flow‐cytometric analysis of intracellular NO generation in GOx@Biotin‐SK and GOx@Biotin‐SR‐treated MCF‐7 cells. (i) Cell viability of MCF‐7 cells treated with GOx@Biotin‐SR, GOx@Biotin‐SK and GOx@Biotin‐SE for 24 h (MTT assay), showing ∼9.5%, 42.23% and 78.35% viability at 100 µm, respectively. (j) Fluorescence microscopy images from the DCFDA assay showing pronounced green fluorescence upon GOx@Biotin‐SR treatment (8.5 µm), indicative of elevated NO‐derived ROS. (k) Quantitative ROS analysis using DCFDA, revealing 5.55‐ and 2.77‐fold higher green emission for GOx@Biotin‐SR and GOx@Biotin‐SK, respectively, relative to untreated cells. Significances are marked as **p < 0.01, vs. the control. Scale Bars correspond to 50 µm.

Encapsulation attenuates direct solvent exposure of GOx, thereby preserving its catalytic activity compared with the free enzyme [50]. To examine the enzymatic activities of free GOx and GOx@Biotin‐SR CVs were compared by monitoring glucose‐driven nitric oxide (NO) generation in situ using a rhodamine‐based probe (RhBNs) [51], arising from arginine oxidation to citrulline (Figure 7c) [26, 27, 52]. Details of RhBNs synthesis, characterization, and assay protocols are provided in Figures S10 and S11. The synthesis and characterization of RhBNs, as well as the experimental procedure used to monitor the in situ NO generation by GOx and GOx@Biotin‐SR, are provided in the Supporting Information. Fluorescence measurements of GOx@Biotin‐SR/RhBNs/ALP (0.5 U mL^{− 1}) mixtures showed negligible emission at 595 nm in the absence of glucose, whereas the stepwise addition of glucose (25–200 µm) produced a progressive fluorescence enhancement. Generation of H₂O₂ via glucose oxidation to gluconic acid is well documented [53, 54, 55], and the progressive fluorescence enhancement as a function of glucose was attributed to H₂O₂‐mediated NO generation via oxidation of arginine to citrulline within the reaction milieu (Figure 7c). As discussed earlier, Biotin‐SR CVs undergo a dephosphorylation process when treated with ALP, and this triggers the release of encapsulated GOx with concomitant generation of H₂O₂. Importantly, glucose‐dependent fluorescence enhancement was not observed for control experiments performed in the presence of sodium pyruvate (200 µm), a potent H₂O₂ scavenger (Figure S28) [56]. These in vitro assays confirmed the essential role of H₂O₂ in the catalytic cascade, as well as the superior catalytic activity of GOx@Biotin‐SR in generating H₂O₂ via oxidation of glucose to gluconic acid, compared to free GOx under identical conditions (Figure 7b; Supporting Information for details). These findings indicate that CV‐encapsulation offers a protective microenvironment that preserves enzymatic stability and the functional advantages of GOx@Biotin‐SR under physiologically relevant conditions.

Nitric oxide (NO) generation was further verified by Griess assay (Figure 7d and Figure S29) [57] and electron paramagnetic resonance (EPR) spectroscopy using a PTIO‐based spin‐trapping assay (PTIO = 2‐phenyl‐4,4,5,5‐tetramethyl‐1,3‐imidazoline‐1‐oxyl‐3‐oxide) [58]. The characteristic five‐line EPR spectrum of PTIO (3290–3340 G) evolved into a seven‐line pattern upon incubation with GOx@Biotin‐SR, glucose and ALP, confirming NO formation (Figure 7e). To confirm the roles of glucose and ALP in NO generation, two control experiments were performed: GOx@Biotin–SR + glucose (without ALP) and GOx@Biotin–SR + ALP (without glucose). For the experiment performed in the absence of ALP, only a minimal NO generation was observed. Fluorescence assay for NO detection showed a fluorescence intensity remained ∼2.5‐fold lower after 4 h compared to the ALP‐treated system (Figure S26b,c), highlighting the essential role of ALP in promoting NO production. In contrast, negligible fluorescence changes were observed in the absence of glucose even at higher ALP concentrations (up to 25 µL), due to the lack of GOx‐mediated H₂O₂ generation required for NO release (Figure S26a). These results confirm that efficient NO generation depends on both ALP activation and glucose availability, rather than spontaneous leakage.

The stability of Biotin‐SR CVs was evaluated using dynamic light scattering (DLS). The average hydrodynamic diameter remained within the range of ∼400–500 nm over 48 h (2, 4, 6, 12, 24, and 48 h), indicating good colloidal stability of the CVs (Figure 7f). Time‐dependent morphological stability was further examined by bright‐field microscopy at different intervals up to 48 h (Figure S20). Additionally, quantitative particle size analysis using ImageJ showed minimal variation in vesicle size throughout the incubation period, further corroborating the DLS results (Figure S21). Serum stability is a critical determinant of the therapeutic potential of peptide‐based assemblies. Upon incubation of Biotin‐SR in human serum at 37°C, HPLC analysis showed that approximately 61% of the peptide remained intact after 24 h, indicating a substantial circulation half‐life of at least 31.93 h (Figure 7g and Figure S45). A significant stability of Biotin‐SR toward proteolytic degradation further underscores its suitability for therapeutic applications [44].

To evaluate the therapeutic potential, the cytotoxic effect of GOx@Biotin‐SR was systematically investigated against live MCF‐7 cells. MCF‐7 breast cancer cells are known to exhibit elevated alkaline phosphatase (ALP) activity compared to non‐malignant breast epithelial counterparts, rendering them a suitable in vitro model for probing ALP‐dependent biochemical processes. Elevated ALP levels are often associated with dysregulated phosphatase signalling during tumor progression and metabolic reprogramming. Mechanistically, ALP catalyzes the hydrolytic cleavage of phosphoester linkages, which disrupts complementary electrostatic interactions responsible for stabilizing vesicular architectures formed through phosphate‐mediated assembly. This enzymatic property provides a rational framework to evaluate ALP‐triggered destabilization and activation mechanisms in enzyme‐responsive anticancer delivery platforms such as GOx@Biotin‐SR. To further validate the ALP‐responsive behavior, ALP activity in MCF‐7 cells was quantitatively evaluated using an Alkaline Phosphatase (ALP) Activity Kit according to the manufacturer's protocol. Following treatment, cells were carefully harvested and lysed to obtain intracellular protein extracts, which were subsequently incubated with a chromogenic substrate specific for ALP. Enzymatic conversion of the substrate generated a quantifiable colorimetric signal that was monitored spectrophotometrically. A standard calibration curve prepared using defined standards enabled precise determination of ALP activity in the experimental samples. Notably, treated MCF‐7 cells exhibited a substantial increase in ALP activity compared to untreated controls, as evidenced by increased absorbance values and a corresponding upward shift relative to the calibration standards. This enhancement confirms the successful induction of ALP activity under the applied experimental conditions (Figure S41). Importantly, the elevated ALP activity strongly supports the proposed enzyme‐triggered activation mechanism, wherein increased phosphatase activity promotes cleavage of the phosphoester functionality, resulting in destabilization of the coacervate assembly and subsequent activation of the therapeutic system. Overall, these findings validate the functional responsiveness of the cellular model and establish the critical role of ALP in mediating activation and therapeutic efficacy of the designed enzyme‐responsive coacervate platform, thereby reinforcing its potential for targeted anticancer applications.

MTT assays were performed following incubation with varying concentrations of GOx@Biotin‐SR (3.125–100 µm) for 24 h with live MCF‐7 cells in DMEM media. A significant dose‐dependent reduction in cell viability was observed with an IC50 of 8.5 µm (Figure S30). Further, to validate the role of biotin in facilitating cellular internalization of the biotinylated coacervates, a competitive inhibition assay was systematically performed using avidin as a high‐affinity biotin‐sequestering agent. MCF‐7 cells (1 × 10⁵) were pre‐incubated with avidin (3 µm) to occupy and block surface‐accessible biotin‐binding sites prior to treatment with GOx@Biotin‐SR (10 µm). This pre‐saturation strategy was designed to inhibit ligand–receptor interactions at the cell membrane, thereby blocking the receptor‐mediated entry pathway. Experimental results revealed that, in the absence of avidin pre‐treatment, GOx@Biotin‐SR exhibited pronounced intracellular uptake (Figure S39), indicating efficient interaction of the biotin moiety with the biotin receptor, namely the Sodium‐Dependent Multivitamin Transporter (SDMT), followed by subsequent cellular internalization. In stark contrast, live MCF‐7 cells pre‐treated with avidin, showed a significant reduction in intracellular fluorescence intensity (Figure S40), demonstrating minimal coacervate uptake and confirming effective receptor blockade. To further substantiate these findings and directly confirm intracellular delivery, cell lysates from untreated and avidin‐treated MCF‐7 cells were analyzed using HPLC and ESI‐MS. A characteristic peak corresponding to GOx@Biotin‐SR was clearly detected in samples without avidin treatment, whereas no detectable signal was observed in avidin‐blocked samples, corroborating the inhibition of cellular uptake. The marked difference observed for the untreated and pretreated MCF‐7 cells, supported by both imaging and analytical data, strongly indicates a receptor‐mediated endocytic mechanism rather than nonspecific internalization. Thus, these findings establish that the cellular uptake of GOx@Biotin‐SR is predominantly governed by specific recognition of the biotin receptor (SDMT), thereby validating both the cellular entry pathway and the targeting efficiency of the designed coacervate system.

To further validate the role of biotin in receptor‐mediated tumor targeting in vivo, tris(2,2′‐bipyridine)ruthenium(II) ([Ru(2,2′‐bipy)₃]^{2 +}) was encapsulated within the CVs and employed as an ICP‐MS‐traceable probe [59]. Remarkably, [Ru(2,2′‐bipy)₃]^{2 +}‐loaded Biotin‐SR CVs exhibited nearly 10‐fold higher Ru accumulation in tumor tissues compared with non‐biotinylated [Ru(2,2′‐bipy)₃]^{2 +}@FITC–SR CVs, strongly supporting the crucial role of biotin in enhancing receptor‐mediated tumor accumulation in vivo (Figure S42).

To delineate the role of NO in modulating cell viability, a control study using GOx@Biotin‐SK was performed. Treatment with 100 µm GOx@Biotin‐SK, GOx@Biotin‐SE resulted in 78% and 42% cell viability respectively, compared to 9.5% viability for GOx@Biotin‐SR (Figure 7i). This marked difference underscores the synergistic contribution of glucose‐depletion‐induced starvation and NO‐mediated gas therapy, and highlights the critical role of the arginine residue in inducing NO generation [60]. Intracellular NO generation by CVs of GOx@Biotin‐SR was evaluated in live MCF‐7 cells using the NO‐responsive RhBNs probe and quantified by flow cytometry. Cells treated with GOx@Biotin‐SR exhibited a pronounced increase in fluorescence intensity, indicating efficient intracellular NO production. In contrast, treatment with GOx@Biotin‐SK resulted in only negligible fluorescence, confirming its minimal NO‐generating capability (Figure 7h). To assess intracellular oxidative stress, a DCFDA (2′,7′‐Dichlorodihydrofluorescein Diacetate) assay was performed after incubating live MCF7 cells with GOx@Biotin‐SR (8.5 µm) for 12 h. Strong intracellular green fluorescence corresponding to oxidized DCF was observed in GOx@Biotin‐SR treated cells, indicative of substantial RNS generation (Figures 7j and Figure S32). In contrast, cells treated with GOx@Biotin‐SK exhibited markedly lower fluorescence under analogous conditions, consistent with the absence of the arginine residue and its role in promoting NO formation. The RNS/ROS generated by GOx@Biotin‐SR caused pronounced disruption of mitochondrial membrane potential and structural integrity. JC‐1 (5,5′,6,6′‐tetrachloro‐1,1′,3,3′‐tetraethylbenzimidazolyl carbo cyanine iodide) staining revealed a clear fluorescence shift from red to green in MCF‐7 cells treated with GOx@Biotin‐SR for 12 h, indicative of the depolarization of mitochondrial membrane potential (ΔΨm) [23, 61, 62]. In contrast, GOx@Biotin‐SK treated cells exhibited a comparatively lower red‐to‐green fluorescence shift, consistent with partial preservation of ΔΨm (Figure 8a). These results confirm that GOx@Biotin‐SR induces mitochondrial dysfunction through coupled ROS/RNS generation and collapse of intracellular energy generation [63]. This behavior is attributed to the presence of cationic guanidinium groups of arginine at physiological pH (7.4), which promote preferential mitochondrial targeting. Consistent with this, colocalization studies using RhB‐SR and MitoTracker Green yielded a high Pearson's correlation coefficient (∼0.92), confirming mitochondrial localization (Figure S33).

FIGURE 8.

(a) Assessment of mitochondrial membrane potential (MMP) depolarization in live MCF‐7 cells using JC‐1 after 12 h incubation with GOx@Biotin‐SK or GOx@Biotin‐SR. Intense green fluorescence at 488 nm in GOx@Biotin‐SR–treated cells indicates pronounced MMP depolarization compared to untreated and GOx@Biotin‐SK–treated controls. (b) Mitochondrial ROS levels probed by MitoSOX in live MCF‐7 cells following identical treatments. Strong red fluorescence in GOx@Biotin‐SR–treated cells confirms elevated mitochondrial ROS generation, whereas GOx@Biotin‐SK shows only weak signals. (c) Mitochondrial morphology analysis of live MCF‐7 cells after treatment with GOx@Biotin‐SK and GOx@Biotin‐SR. GOx@Biotin‐SR induces mitochondrial swelling and reduced fusion, reflecting disrupted membrane potential and impaired mitochondrial distribution within the cytosol. (d,f) Effect of Biotin‐SR, Biotin‐SK and Biotin‐SE in normal HEK‐293 and MCF‐7 cell lines respectively. (e) Effect of GOx@Biotin‐SR and GOx@Biotin‐SK in HEK‐293 cell. Scale Bars correspond to 50 µm.

Consistently, live‐cell imaging with the mitochondrial superoxide probe MitoSOX Red revealed intense ROS accumulation in GOx@Biotin‐SR treated MCF‐7 cells (Figure 8b, Figures S34 and S35), whereas untreated cells and those treated with GOx@Biotin‐SK exhibited minimal fluorescence. The low ROS levels observed for GOx@Biotin‐SK are primarily attributed to oxidative stress arising from glucose depletion alone [64]. Collectively, these findings demonstrate that GOx@Biotin‐SR induces an enhanced mitochondrial oxidative stress via a synergistic interplay between glucose deprivation and concurrent ROS/RNS generation. In contrast, GOx@Biotin‐SK caused minimal mitochondrial alterations, underscoring the essential role of NO generation in the therapeutic mechanism [65]. Notably, only Biotin‐SR, Biotin‐SK and Biotin‐SE alone were non‐toxic to both normal HEK‐293 and MCF‐7 cells, confirming the biocompatibility of the carrier system (Figure 8d,f). In contrast, GOx@Biotin‐SR and GOx@Biotin‐SK did not exhibit significant cytotoxicity in normal HEK‐293 cells, non‐tumorigenic MCF‐10A human mammary epithelial cells, and NIH‐3T3 mouse fibroblasts, indicating their cancer cell‐selective activity (Figure 8e and Figure S31).

Beyond mitochondrial perturbation, GOx@Biotin‐SR treatment also resulted in a marked loss of microtubule integrity, indicative of downstream cytoskeletal destabilization driven by intracellular oxidative and nitrosative stress. Live MCF‐7 cells were incubated with 10 µm GOx@Biotin‐SR for 12 h and subjected to immunofluorescence imaging. Images revealed disruption of kinetochore‐microtubule attachments and complete microtubule disruption by 24 h (Figure 9a and Figure S36) [55]. Immunocytochemical analysis further showed upregulation of mitotic checkpoint proteins MAD2 and BUBR1 (Figure 9b,c, Figures S37 and S38), suggesting that GOx@Biotin‐SR induced mitotic arrest through spindle assembly checkpoint activation. This stabilization of aberrant kinetochore–microtubule interactions prevent timely microtubule detachment and error correction during mitosis. Suppression of normal attachment turnover results in persistent, improperly tensioned spindle connections, ultimately leading to defective chromosome segregation [66, 67, 68]. The failure to resolve faulty interactions activated the spindle assembly checkpoint, as evidenced by increased BUBR1 and MAD2 expression, resulting in mitotic arrest and anticancer effects. These findings highlighted the role of oxidative stress in GOx@Biotin‐SR‐mediated antiproliferative effects. To further elucidate the molecular response to GOx@Biotin‐SR treatment, we examined the expression of tumor suppressor protein p53 and its downstream effector p21, key mediators of cell cycle arrest under oxidative stress [69]. Immunoblot analysis revealed a 55% and 84% upregulation of p53 (45 kDa) and p21 (21 kDa) compared to untreated MCF‐7 cells, respectively, in GOx@Biotin‐SR‐treated MCF‐7 cells compared to untreated controls (Figure 9d). These findings, together with cytoskeletal protein disruption and mitotic checkpoint activation, highlight a coordinated stress response leading to cell cycle arrest and apoptosis, underpinning the potent anticancer activity of GOx@Biotin‐SR.

FIGURE 9.

Mitotic spindle disruption and cell‐cycle arrest induced by GOx@Biotin–SR. (a) Confocal fluorescence images of live MCF‐7 cells treated with GOx@Biotin‐SR (10 µm, 12 and 24 h), showing pronounced disruption of the microtubule network. Microtubules were immunostained using anti‐tubulin (EP1332Y, 1:300 v/v) primary and corresponding secondary antibodies. (b,c) Immunofluorescence imaging of spindle checkpoint proteins Mad2 and BubR1 in MCF‐7 cells after GOx@Biotin‐SR exposure, showing strong fluorescence signals consistent with checkpoint activation. (d) Immunoblot analysis of p53 and p21 expression in GOx@Biotin‐SR treated cells relative to untreated controls, demonstrating upregulated apoptotic and cell‐cycle arrest markers, and quantitative densitometric analysis of p53 and p21 confirming significant expression enhancement. (f) Flow‐cytometric histograms showing DNA content distribution in MCF‐7 cells treated with GOx@Biotin‐SR, GOx@Biotin‐SK, and control samples, revealing distinct cell‐cycle arrest. (g) Quantitative analysis of cell‐cycle phase distribution derived from flow‐cytometric data. The results are represented as the mean ± SD, n = 3 per group from three independent experiments. Significances are marked as ^** p < 0.01, vs. the control. Scale Bars correspond to 20 µm.

Cancer cell migration is a critical contributor to metastasis and therapeutic failure [70]. To evaluate the anti‐migratory efficacy of GOx@Biotin‐SR, acell migration assay was conducted with a monolayer of live MCF‐7 cell treated with 2 µm GOx@Biotin‐SR. Notably, treated cells exhibited negligible cellular migration after 12 h, indicating the efficacy of GOx@Biotin‐SR in inhibiting migration (Figure 10a). In contrast, untreated and GOx@Biotin‐SK treated cells showed complete wound closure, demonstrating intact migratory potential. Quantitative analysis of the wound healing assay revealed a marked difference in migratory behavior across treatment conditions in MCF‐7 cells. The untreated control group exhibited efficient wound closure of approximately 89%, indicating robust cell migration and proliferation. In contrast, cells treated with GOx@Biotin‐SK showed a moderate reduction in wound closure to 68%, suggesting partial inhibition of migratory capacity. Notably, treatment with GOx@Biotin‐SR resulted in a drastic suppression of wound closure, with only ∼9% closure observed, highlighting a pronounced inhibitory effect on cell migration. This significant reduction in wound healing ability underscores the enhanced therapeutic efficacy of GOx@Biotin‐SR in impairing cellular motility, likely contributing to its anticancer potential by limiting processes associated with tumor progression and metastasis. These results further confirmed the therapeutic potential of GOx@Biotin‐SR for suppressing metastatic progression.

FIGURE 10.

(a) Biotin receptor mediated cellular internalization of Biotin‐SR in MCF7 cell. (b) in vitro cell migration assay of confluent MCF‐7 monolayers treated with GOx@Biotin‐SR (2 µm) or GOx@Biotin‐SK (2 µm) to evaluate anti‐migratory activity. GOx@Biotin‐SR–treated cells showed negligible migration and failed to close the wound gap within 12 h, whereas untreated and GOx@Biotin‐SK–treated cells completely healed the scratch. Scale Bars correspond to 400 µm. (c) 3D spheroid growth inhibition of MCF‐7 cells after treatment with GOx@Biotin‐SR and GOx@Biotin‐SK (8.5 µm). Bright‐field images recorded on days 3, 5, and 7 reveal pronounced suppression of spheroid growth and progressive size reduction upon GOx@Biotin‐SR treatment, while GOx@Biotin‐SK caused only moderate inhibition relative to the untreated control. Scale Bars correspond to 50 µm. (d) Flow‐cytometric apoptosis assay of dissociated tumor tissue reveals increased apoptotic populations following GOx@Biotin‐SR and GOx@Biotin‐SK treatment.

To better mimic the in vivo tumor microenvironment, we evaluated the therapeutic efficacy of GOx@Biotin‐SR in a three‐dimensional (3D) multicellular tumor spheroid (MCTS) model using MCF‐7 cells. MCF‐7 cells were cultured on agarose‐coated, hydrophobic polystyrene surfaces to promote uniform spheroid formation [71]. After 3 days of growth, the spheroids were treated with 8.5 µm of GOx@Biotin‐SR or GOx@Biotin‐SK, while an untreated group served as the control. To obtain a comprehensive assessment of spheroid morphology and treatment response, Z‐stack images were acquired and computationally reconstructed to generate side‐view profiles, enabling robust estimation of the total tumor spheroid volume. Over a 7‐day treatment period, GOx@Biotin‐SR‐treated spheroids exhibited a pronounced reduction in size and volume, indicating potent therapeutic activity in a 3D context [72]. In contrast, untreated spheroids remained largely unchanged in both structure and volume, whereas GOx@Biotin‐SK‐treated spheroids showed only moderate growth inhibition, likely attributable to glucose deprivation effects alone. These observations were further supported by volumetric reconstruction from Z‐axis imaging, which confirmed substantial structural collapse and shrinkage exclusively in the GOx@Biotin‐SR‐treated group (Figure S43). Additionally, flow cytometric analysis of dissociated spheroids revealed a significant increase in apoptotic cell populations following GOx@Biotin‐SR treatment, establishing apoptosis as the primary mechanism of tumor cell death. Together, these observations indicate the superior anticancer efficacy of GOx@Biotin‐SR in a physiologically relevant 3D tumor model and highlight its potential for effective tumor penetration and therapeutic activation.

To evaluate the in vivo therapeutic efficacy of GOx@Biotin‐SR, a xenograft breast cancer model was established in female Wistar rats (2–3 months old, 250–300 g, n = 10) using MDA‐MB‐231 cells [73]. Prior to this, these animals were immunosuppressed by intramuscular administration of cyclophosphamide (30 mg kg^{− 1} day^{− 1}) for seven consecutive days. Tumor progression was monitored over 20–25 days by serial measurement of tumor volume. After Palpable tumor nodules were formed then systemic administration of GOx@Biotin‐SR (10 mg kg^{− 1}) was initiated.

Following intravenous administration, HPLC analysis of plasma samples collected at different time intervals revealed that Biotin‐SR exhibited an in vivo half‐life of approximately 9.25 h (Figures S46, S48–S51). In addition, ∼15% of Biotin–SR remained in systemic circulation, while urinary excretion studies showed that only ∼25% of the administered Biotin‐SR underwent direct clearance within 24 h (Figure S47), indicating prolonged systemic retention. Based on these observations, a 48 h dosing interval was selected to ensure sustained therapeutic efficacy while minimising potential accumulation and reducing dosing frequency. The prolonged in vivo persistence of GOx@Biotin–SR was further corroborated by HPLC‐assisted urinary excretion analysis (Figures S52–S54).

Treatment with GOx@Biotin‐SR led to a marked suppression of tumor growth and a significant reduction in final tumor mass relative to untreated controls (Figure 11a), confirming its strong antitumor efficacy. Kaplan–Meier survival analysis revealed prolonged survival, with 85% of Biotin‐SR and 40% with Biotin‐SK treated animals remaining alive throughout the observation period respectively, compared to rapid mortality in the untreated and Biotin‐SE treated group (Figure 11f) [74, 75].

FIGURE 11.

In vivo tumor growth inhibition studies. (a) GOx@Biotin‐SR treatment markedly suppressed tumor progression in the Wistar rat tumor model over 10 and 21 days compared with GOx@Biotin‐SK and GOx@Biotin‐SE treatments. (b) Expression levels of TNF‐α following different treatments. (c) Variation in tumor size at different time intervals. (d) Changes in animal body weight during treatment up to 21 days. (e) Final tumor weight demonstrating significant tumor growth inhibition by GOx@Biotin‐SR compared to the control groups; error bars represent SD. (f) Kaplan–Meier survival analysis showing prolonged survival, with a 40% survival rate in GOx@Biotin‐SR‐treated rats compared to complete mortality in the untreated group. (g) H&E‐stained histological sections of major organs (Brain, lungs, kidney, spleen and liver) from treated rats exhibiting normal tissue morphology, confirming negligible systemic toxicity of GOx@Biotin‐SR. Data are presented as mean ± SD (n = 3 per group) from three independent experiments. Statistical significance is indicated as ^* p < 0.05 vs. the control group. Scale bars correspond to 50 µm.

Elevated TNF‐α expression in tumor tissue following Biotin‐SR treatment indicated enhanced ROS generation and effective anticancer activity, whereas untreated and Biotin‐SE groups showed no significant TNF‐α expression. Biotin‐SK induced comparatively lower TNF‐α levels (Figure 11b) [76, 77, 78]. To evaluate systemic safety, GOx@Biotin‐SR was administered via tail vein injection in a separate cohort of mice, followed by histological examination of major organs (brain, liver, kidney, spleen, and lungs) using hematoxylin and eosin (H&E) staining. No pathological abnormalities or signs of organ toxicity were observed (Figure 11f), indicating a favorable biocompatibility profile and minimal systemic toxicity.

3. Conclusions

In summary, we demonstrate that a sequence‐defined short peptide can encode enzyme‐responsive phase behavior to generate functional coacervate vesicles capable of programmable chemical output. The GOx@Biotin‐SR system translates ALP–mediated dephosphorylation into supramolecular reconfiguration, triggering spatially confined vesicle disassembly and controlled enzyme release within tumor environments. and thereby establishing a coordinated oxidative–nitrosative reaction network. Mechanistically, the coordinated depletion of glucose and accumulation of ROS/RNS impose severe metabolic and redox stress, leading to mitochondrial dysfunction, collapse of membrane potential, and disruption of oxidative phosphorylation. Concomitantly, redox imbalance induces microtubule disruption, stabilizing aberrant spindle structures and activating the spindle assembly checkpoint, as evidenced by upregulation of MAD2 and BUBR1. The resulting prolonged mitotic arrest, together with p53–p21 signalling, drives apoptotic cell death and suppresses cancer cell proliferation and migration. These effects are consistently observed across two‐dimensional cultures, three‐dimensional tumor spheroids, and in vivo xenograft models, where the platform induces marked tumor regression and survival benefits without detectable off‐target toxicity. Beyond the specific therapeutic context, this study establishes a general chemical framework in which enzymatic cues are transduced into structural and functional outputs through sequence‐programmed coacervation. By coupling phase behavior to reaction networks, enzyme‐activated peptide coacervates emerge as adaptive supramolecular systems capable of integrating recognition, compartmentalization and chemical transformation. More broadly, these findings suggest that sequence‐programmed coacervation can serve as a versatile logic gate, where enzymatic signals are decoded into complex, multi‐stage chemical responses within living systems.

Conflicts of Interest

The authors declare no conflicts of interest.

Supporting information

Supporting File 1: adma73804‐sup‐0001‐SuppMat.docx

Supporting File 2: adma73804‐sup‐0002‐VideoS1.mp4.

Supporting File 3: adma73804‐sup‐0003‐VideoS2.mp4.

Supporting File 4: adma73804‐sup‐0004‐VideoS3.mp4.

Data Availability Statement

The data that support the findings of this study are available from the corresponding author upon reasonable request.