De Novo Lipid Switch Design for Reactive Oxygen Species-Triggered Liposomal Release

Abstract

Liposomal nanocarriers are highly effective for drug delivery since they shield encapsulated therapeutic cargo and enhance cellular infiltration. However, a drawback of liposomal carriers involves limited release of contents after delivery. As such, stimuli-responsive liposomes harnessing synthetic lipid switches that trigger cargo release upon encountering disease-associated conditions exhibit potential for addressing this issue. Reactive oxygen species (ROS) have emerged as prominent target stimuli due to their upregulation in diseased cells. Lipid switches designed for liposome release are often derived from phospholipid precursors, an approach that suffers from high expense and hinders scalability due to the limited availability of phospholipids. To address this issue, we report the design and synthesis of an ROS-responsive lipid switch from first principles that does not utilize phospholipid precursors. This culminated in a scalable non-phospholipid switch that forms stable liposomes yet undergoes clean, H₂O₂-triggered degradation under aqueous oxidative conditions. Using Nile Red release, ANTS-DPX dequenching, and dynamic light scattering (DLS) experiments, we show that liposomes containing lipid switch 4 exhibit minimal basal leakage and maintain structural integrity, while H₂O₂ induces dose-dependent membrane permeabilization with characteristic changes in vesicle size. We also show that liposomes containing 4 demonstrate superior and more consistent results compared to that containing 6, which has longer acyl chains. PC-only negative controls display no comparable response, confirming that release arises from selective boronate oxidation rather than nonspecific degradation. These results highlight the potential of a synthetically accessible, modular non-phospholipid architecture for stimuli-responsive triggered release as a practical platform for redox-responsive therapeutic delivery.

Graphical Abstract

ROS-triggered release of liposomal cargo has been achieved using a conveniently accessible non-phospholipid switch. Created in BioRender. Best, M. (2026).

Introduction

Lipid-based nanocarriers constitute one of the most versatile platforms for drug delivery due to their biocompatibility, membrane-mimetic structure, and clinical utility. Early mechanistic studies of phospholipid bilayers established the principles of membrane permeability and phase behavior, laying the foundation for modern liposome technology.^1-3 These insights ultimately enabled the development of clinically validated formulations such as Doxil^® and multiple lipid nanoparticle (LNP) therapeutics.^2,3 Liposomes remain highly tunable systems in which bilayer composition, size, and surface chemistry can be engineered to control circulation, stability, and cargo release. However, a significant issue with liposomes involves their limited release of therapeutic cargo, which diminishes activity. Indeed, it has been reported that even liposomes and related lipid nanoparticles (LNPs) containing ionizable/cationic lipids designed for endosomal escape deliver less than 5% of their medicinal cargo.^4,5 Stimuli-responsive liposomes have emerged as a promising solution to this longstanding challenge as a means of triggering encapsulated contents. Furthermore, incorporating stimuli-responsive elements into liposomes enhances their precision by enabling drug release only when a specific biochemical signal is present. Temperature-sensitive liposomes offered early proof-of-concept that modest external stimuli can initiate drug release in vivo.⁶ Among endogenous triggers, reactive oxygen species (ROS) are especially compelling. Elevated ROS levels, particularly hydrogen peroxide (H₂O₂), are characteristic of tumor microenvironments, sites of inflammation, and regions of ischemic stress.^7,8 H₂O₂ functions as both a signaling molecule and a driver of oxidative injury, producing concentration gradients that provide an intrinsic mechanism for site-selective nanocarrier activation.⁷

ROS-responsive nanomaterials have thus become a rapidly advancing design area. Pioneering polymeric vesicles from Hubbell demonstrated that oxidative degradation could be harnessed for controlled release,⁹ while Murthy and co-workers achieved in vivo activation of thioketal nanoparticles by inflammatory ROS.¹⁰ Almutairi and co-workers further validated peroxide-triggered disassembly in polymer nanoparticles under physiologically relevant H₂O₂ levels.¹¹ Collectively, these studies highlight the robustness and biological relevance of oxidative activation mechanisms. Within lipid systems specifically, most ROS-responsive designs rely on grafting cleavable motifs, such as boronate esters, thioethers, or selenium/tellurium linkers, onto natural phospholipid scaffolds, commonly through derivatives of dioleoylphosphoethanolamine (DOPE) or phosphatidylcholine (PC).^12-15 These constructs achieve selective H₂O₂-triggered headgroup or linker cleavage, inducing membrane packing disruptions that promote release of hydrophobic and hydrophilic cargos. For example, a DOPE-derived architecture developed by our lab exploited a boronate-protected headgroup that regenerates DOPE and releases both hydrophobic and hydrophilic cargos via H₂O₂-triggered cleavage and curvature changes.^12,15

Despite these successes, phospholipid-dependent strategies suffer from three structural limitations: (i) high costs associated with natural lipid precursors due to their limited availability, (ii) multistep derivatization involving hydrolytically sensitive intermediates, and (iii) limited architectural flexibility due to the fixed phospholipid backbone.^2,3,15 These constraints impede scalability and slow translation, even though the trigger chemistry itself—aryl boronate oxidation—is intrinsically selective, rapid, and tunable.^16-21 This mismatch between versatile molecular trigger chemistry and inflexible materials architecture highlights an underexplored design opportunity: constructing responsive amphiphiles entirely de novo, without inheriting the liabilities of natural phospholipids.²² Aromatic boronate motifs are particularly well-suited for this approach because their oxidative deborylation produces predictable polarity changes that can be directly translated into membrane-level disruption.^16-21 Yet, systematic exploration of non-phospholipid amphiphiles bearing boronate triggers remains limited, and direct comparisons with phospholipid-derived systems under tumor-relevant H₂O₂ conditions are rare.

Here we address this gap by developing a modular, fully non-phospholipid aryl boronate lipid synthesized from inexpensive aromatic and aliphatic precursors. The amphiphile forms stable vesicles when blended with PC and undergoes controlled disruption upon exposure to H₂O₂, resulting in graded release of both hydrophobic and hydrophilic cargos. Combined with a scalable four-step synthesis and robust oxidative response, this architecture establishes a practical foundation for next-generation H₂O₂-responsive liposomal systems independent of natural lipid scaffolds.

Results and Discussion

Design and scalable synthesis of de novo aryl boronate lipid switches 4 and 6

ROS-responsive lipid switches 4 and 6 were designed to balance three key criteria: (i) incorporation of a boronate trigger capable of selective H₂O₂-mediated oxidation, (ii) a straightforward, modular synthetic route from inexpensive starting materials using proven chemistry, and (iii) an overall scaffold that can be prepared on gram scale using standard laboratory equipment. As first-generation non-phospholipid amphiphiles, 4 and 6 were designed to place benzylic ester arms ortho and para to the boronate trigger, an arrangement intended to enable oxidation-initiated self-immolative breakdown while furnishing a two-tail architecture compatible with bilayer formation. The octanoyl/lauroyl groups were selected to vary the chain length of the lipid hydrophobic domain to determine effects on release properties. As summarized in Scheme 1, lipid switches 4 or 6 were accessed in four steps from commercially available 4-bromoisophthalic acid without protecting groups and with minimal chromatographic purification.

Scheme 1:

Scalable four-step synthesis of ROS-responsive lipid switches 4 and 6 from 4-bromoisophthalic acid.

Initial conversion of 4-bromoisophthalic acid to the corresponding acid chloride 1 was achieved by refluxing with thionyl chloride, followed by removal of excess reagent by short-path distillation. The crude acid chloride 1 was used directly in the subsequent reduction, avoiding any chromatographic workup and simplifying scale-up. Reduction under biphasic NaBH₄/K₂CO₃/18-crown-6 conditions, adapted from Kanjilal et al.,²³ provided benzylic diol 2 in 86% combined yield on multigram scale. This operationally simple two-step sequence from 4-bromoisophthalic acid delivers pure 2 using only extractions and drying. It should be noted that a direct reduction of 4-bromoisophthalic acid was initially tried using LiAlH₄, which did not turn out to be successful. This two-step method simplified the process with high efficiency. Esterification of 2 with octanoyl chloride, inspired by a protocol from Lee and co-workers,²⁴ afforded diester 3 in 70% yield after a single silica column. Finally, installation of the boronate trigger was accomplished by a Suzuki–Miyaura borylation using bis(pinacolato)diboron, KOAc, and Pd(dppf)Cl₂ · CH₂Cl₂ in refluxing dioxane to provide lipid switch 4 in 67% yield after column chromatography. Through this approach, we were able to efficiently prepare compound 4 in gram scale. The formation of 5 and 6 followed the same protocols as 3 and 4, respectively. However, both of these steps proceeded with lower yields when we increased the tail length, and particularly the final borylation step, which may result from the greater propensity of long chain lipids to self-assemble. Nevertheless, lipid switches 4 and 6 were both obtained for comparison studies.

Overall, the sequence relies on inexpensive reagents and requires chromatography only in the final two steps. These features collectively underscore the scalability of the synthetic route and distinguish lipid switch 4 and 6 from many phospholipid-based ROS-responsive architectures that depend on costly, pre-functionalized lipid precursors. This combination of high-yielding steps, minimal protecting-group manipulations, and clean analytical signatures supports the practicality of the route for preparative-scale access to lipid 4 and 6.

Immolation of Lipid Switches Upon H₂O₂ Addition

The expected degradation pathway for lipid switches 4 and 6 is shown in Figure 1B. Here, boronate oxidation is designed to generate a phenoxide intermediate that is primed to undergo linker self-immolation to generate triol and octanoate/laurate products. To evaluate this hypothesis, we subjected lipid switch 4 to H₂O₂ to perform time-dependent ¹H and ¹¹B NMR, TLC, and DART-MS (positive and negative mode) experiments. Time-dependent ¹H NMR experiments showed the formation of multiple new peaks between 3.5–5.0 ppm (for benzylic and alcohol groups) as well as in the aromatic region (Figure S21) upon H₂O₂, while ¹¹B NMR exhibited a weakening of the characteristic boron peak for Bpin and formation of the new peaks (Figure S22). TLC showed disappearance of lipid switch 4 and appearance of two new polar spots, with the lower spot co-eluting with octanoate (octanoic acid + base) (Figure S23). Furthermore, DART-MS confirmed formation of (4-hydroxy-1,3-phenylene)dimethanol in positive mode upon the addition of H₂O₂ (Figures S25) and octanoate ion in negative mode (Figure S26). These experiments demonstrate that lipid switch 4 is responsive to H₂O₂ and is cleaved into small molecules products, which we expected would result in the decomposition of liposomes after adding H₂O₂ and the subsequent release of encapsulated cargo.

Figure 1:

(A) Conceptual illustration of H₂O₂-triggered membrane disruption and cargo release from liposomes. Created in BioRender. Best, M. (2026). (B) Proposed pathway for degradation of lipid switches 4 or 6 upon H₂O₂ addition. Mechanistic support for panel B is provided by time-dependent ¹H/¹¹B NMR, TLC, and DART-MS experiments showing H₂O₂-dependent disappearance of 4 and formation of triol and octanoate-derived products (Figures S21-S23 and S25-S26).

H₂O₂-Triggered Release of Hydrophobic Cargo

We next assessed whether scalable aryl boronate lipid 4 can endow liposomes with ROS-responsive permeability using encapsulated Nile Red (NR) as hydrophobic cargo. NR dye is commonly used to screen for liposome triggered release since its fluorescence properties are enhanced within the membrane bilayer, which leads to a decrease in signal upon leakage.²⁵ It should be noted that because NR is highly hydrophobic, it does not freely partition into the aqueous milieu upon escaping the hydrophobic bilayer. Rather, the dye tends to self-associate into nonemisive aggregates, which results in the observed quenching of its fluorescence.^26-28 As shown in Figure 2, 4/PC (50/50) liposomes treated with 98 mM H₂O₂ exhibited a pronounced and continuous loss of NR fluorescence over the course of 6 hours, consistent with peroxide-induced membrane destabilization. In contrast, negative control 4/PC liposomes incubated with Milli-Q (MQ) water displayed only modest fluorescence decay, whereas PC-only liposomes remained highly stable under both MQ water and peroxide conditions. These controls confirm that the observed NR release stems from oxidative cleavage of the boronate moiety in 4, rather than nonspecific peroxide–membrane interactions.

Figure 2:

Time-dependent percentage of initial NR fluorescence intensity at 635 nm for 4/PC 50/50 and PC-only liposomes under various H₂O₂ concentrations.

It should be noted that we studied several formulations consisting of different combinations of (i) 4 and PC, (ii) 4, PC, DOPE, and phosphatidic acid (PA), and (iii) 4, dipalmitoylphosphatidylcholine (DPPC), and cholesterol formulations in an effort to optimize conditions. Further details of the experimented formulations are listed in Tables S1-S3. Among all these formulations, 50/50 4/PC formulation appeared to be the most effective in terms of stability and response to H₂O₂.

Quantification at the 5-hour timepoint (Figure S27) revealed a clear oxidant-dependent trend for both 4/PC and 6/PC liposomes. The comparison in response of these two liposomes is shown in Figure S27A, where 4/PC liposome overall demonstrated superior release properties compared to 6/PC, although 6/PC appeared to release more NR at the highest concentration (98 mM) of H₂O₂. As a result, Figure S27B focuses only on 4/PC liposomal release at 5-hour timepoint. Low H₂O₂ concentrations (0.01–0.1 mM) produced moderate release (24–27%), whereas higher concentrations (1–98 mM) generated substantially greater leakage (38–56%). PC controls again showed negligible NR release. Together, these data demonstrate that lipid switch 4, accessible by the four-step route in Scheme 1, enables dose-dependent, ROS-responsive disassembly of the membrane sufficient to release hydrophobic cargo.

Dynamic Light Scattering Reveals ROS-Induced Structural Destabilization

To correlate the hydrophobic release behavior with changes in vesicle morphology, we analyzed liposomal size and dispersity using dynamic light scattering (DLS). Figure 3A compares the Z-ave of both 4/PC and 6/PC liposomes when treated with different concentration of H₂O₂. 6/PC liposomes did not show any significant changes with changes in H₂O₂. However, for 4/PC liposomes, as shown in Figure S28A, PC-only vesicles and MQ water-treated 4/PC (50/50) vesicles maintained Z-ave diameters of approximately 155–160 nm, which indicated excellent structural stability in the absence of an oxidative trigger. Exposure of 4/PC liposomes to low-to-moderate peroxide concentrations (0.01–1 mM) produced only slight increases in size (160–166 nm), reflecting mild perturbation of the bilayer. In contrast, treatment with 98 mM H₂O₂ caused a dramatic increase in Z-ave to 314 nm, which suggested vesicle swelling, partial fusion, and/or structural disruption consistent with oxidative cleavage of the boronate unit in 4. The raw DLS curves shown in Figure S29 also demonstrate these changes.

Figure 3:

Comparison in (A) Z-ave diameter and (B) corresponding PdI values between 4/PC (50/50) and 6/PC (50/50) liposomes when treated with MQ water (control) and different H₂O₂ concentration.

Polydispersity (PdI) measurements (Figure 3B and S28B) also supported this interpretation: PdI values remained low (0.15–0.20) across all non-disruptive conditions but increased sharply to 0.38 upon addition of 98 mM H₂O₂ to 4/PC liposomes. This transition from monodisperse to heterogeneous size distributions aligns directly with the oxidant-dependent NR release behavior.

H₂O₂-Triggered Release of Hydrophilic Cargo

To validate the generality of the ROS-responsive mechanism, we next evaluated release of hydrophilic cargo. Polar cargo delivery is of critical importance due to the emergence of genetic therapies employing RNA. As a model of hydrophilic cargo release, we evaluated triggered release using the fluorescence resonance energy transfer (FRET) dye pair 8-aminonaphthalene-1,3,6-trisulfonic acid (ANTS) and p-xylenebis-(pyridinium bromide) (DPX). In this assay, ANTS is initially quenched by DPX due to co-encapsulation within the aqueous cores of liposomes, whereas fluorescence is restored when the two dyes are separated by release into solution.²⁹ In this case, polar cargos are non-specifically entrapped during liposome formation and unencapsulated dye is removed using size exclusion chromatography. As shown in Figure 4A, 50/50 4/PC liposomes exposed to 98 mM H₂O₂ displayed strong, time-dependent increases in fluorescence over 12 hours, reflecting dequenching due to probe leakage. In contrast, MQ water-treated 4/PC and 6/PC liposomes exhibited minimal release, and PC-only liposomes remained highly stable even under high peroxide conditions. Figure 4B compares the response of 4/PC and 6/PC lipsomes. While 4/PC showed ~41.5% dequenching at 12 hours, 6/PC demonstrated only ~16.5%, which again indicates that 4/PC liposomes yield superior response. The close similarity between ANTS-DPX and NR release profiles supports a shared oxidative destabilization process initiated at the boronate headgroup of 4.

Figure 4:

(A) Time-dependent ANTS–DPX release over 12 hours. (B) Quantified release comparison between 4/PC and 6/PC liposomes at the 12-hour timepoint

In Figure S30, we showed that peroxide-treated 4/PC liposomes released ~18.8% of encapsulated dye within 5 hours, significantly higher than both the MQ control (~1.3%) and the PC control (~1.5%). The hydrophilic cargo thus mirrors the behavior observed in hydrophobic release experiments, which underscored that the membrane defects produced by oxidative cleavage are not cargo-selective. We also acknowledge that the hydrophilic release profile with this platform is slower and less complete than that reported for some faster ROS-responsive liposomal systems, including the rapid lipid-peroxidation-triggered system described by Luo et al.³⁰ One plausible explanation is that maximization of leakage requires sufficient local conversion of 4 within a given membrane region before destabilization reaches a threshold to support rapid release. Figure S24 supports this explanation in the form of TLC experiments performed at different time points after 4/PC liposomes were treated with H₂O₂. While the octanoate ion appeared to be formed as early as 1 hour after treatment, it took ~11-12 hours for the starting material (lipid switch 4) to be immolated completely. Accordingly, the present formulation should be viewed as an initial non-phospholipid proof-of-concept whose kinetics remain open to further improvement.

Experimental Section

Materials and Methods

The key starting material 4-bromoisophthalic acid was purchased from Oakwood Chemical. Most other reagents and catalysts were also obtained from Oakwood Chemical or Ambeed. PC (L-α-phosphatidyl choline, mixed isomers from chicken eggs), PA (L-α-phosphatidic acid sodium salt from chicken eggs), DOPE (1,2-dioleoyl-sn-glycero-3-phosphoethanolamine), and DPPC (1,2-dipalmitoyl-sn-glycero-3-phosphocholine) were purchased from Avanti Polar Lipids, Inc. (Alabaster, AL). Cholesterol was purchased from Sigma-Aldrich. Dry solvents were obtained from a Pure Solv MD-7 solvent purification system (Innovative Technology, Inc., Newburyport, MA). Column chromatography was performed using 230–400 mesh silica gel purchased from Sorbent Technologies. NMR spectra were obtained using Bruker Avance NEO 500 MHz or Varian VNMRS 500 MHz spectrometers. IR spectra were obtained using a Bruker ALPHA II Compact FT-IR spectrometer. Mass spectra were acquired on a JEOL DART-AccuTOF mass spectrometer. Liposome extruders and polycarbonate membranes were obtained from Avestin (Ottawa, Canada). Ultrapure water was purified using a Millipore system (≥18 MΩ·cm). Osmotic pressure was measured with a Wescor Vapor Pressure Osmometer 5520. Small quantities (< 5 mg) were weighed on a Mettler Toledo XS105 dual-range analytical balance. Fluorescence studies were performed using a Cary Eclipse fluorescence spectrophotometer (Agilent Technologies). Dynamic light scattering (DLS) measurements were carried out on a Malvern Zetasizer Nano ZS instrument equipped with a 4.0 mW laser operating at λ = 633 nm. Data were processed and cleaned in Microsoft Excel 365, and statistical analyses were performed in JMP Pro 17. Plots were generated using Origin Pro 2026 Learning Edition. All error bars denote standard errors calculated from at least three experimental replicates.

4-bromoisophthaloyl dichloride (1)

Commercially purchased 4-bromoisophthalic acid (4.2 g, 17.14 mmol) was added to a 100 mL RBF under argon. Thionyl chloride (8 mL) was added dropwise and the neat reaction was heated to reflux. After 4 hours, excess thionyl chloride was removed using short-path distillation. A small amount of this intermediate 1 was saved for characterization, and the whole remaining part was used right away in the next reduction step without further purification. ¹H NMR (500 MHz, CDCl₃) δ 8.71 (d, J = 2.2 Hz, 1H), 8.12 (dd, J = 8.5, 2.2 Hz, 1H), 7.90 (d, J = 8.5 Hz, 1H). ¹³C NMR (126 MHz, CDCl₃) δ 166.68, 165.19, 136.39, 136.01, 135.43, 134.98, 132.96, 129.38. FT-IR (neat, cm⁻¹): 3094 (Ar =C–H), 1749 (C=O, acyl chloride), 1584, 1556, 1459 (Ar C=C).

(4-bromo-1,3-phenylene)dimethanol (2)

This step was performed following the protocol reported by Kanjilal et al.²³ A mixture of potassium carbonate (7.106 g, 51.423 mmol), 18-crown-6 (20 mol %, 0.091 g, 0.3428 mmol) and a portion of sodium borohydride (0.3 eq., 0.195 g, 5.1423 mmol) in dichloromethane (20 mL) and water (20 mL) was stirred and cooled to 0 °C. A solution of intermediate 1 in dry dichloromethane was added dropwise with additional sodium borohydride (3 eq., 1.95 g, 51.423 mmol) portion-wise simultaneously over 10 minutes. The stirring was continued in an ice bath for an additional 10 minutes. After that, the reaction mixture was diluted with 75 mL of water and the organic layer was separated. The aqueous layer was extracted with ethyl acetate (3×60 mL) and the combined organic layer was washed with brine (1×60 mL). The mixture was then dried over sodium sulfate, filtered and concentrated to obtain 2 as a white solid (3.16 g, 86% yield). ¹H NMR (500 MHz, CD₃OD) δ 7.53 (d, J = 2.3 Hz, 1H), 7.50 (d, J = 8.1 Hz, 1H), 7.16 (dd, J = 8.1, 2.2 Hz, 1H), 4.65 (s, 2H), 4.58 (s, 2H). ¹³C NMR (126 MHz, CD₃OD) δ 142.57, 141.58, 133.34, 128.23, 127.92, 121.40, 64.64, 64.55. FT-IR (neat, cm⁻¹): 3186 (O–H), 2886 (C–H), 1466 (Ar C=C), 1062 (C–O). DART-MS (positive) [M+H–H₂O]⁺ calcd for C₈H₈BrO: 198.9759, found: 198.9756.

(4-bromo-1,3-phenylene)bis(methylene) dioctanoate (3)

Inspired by the protocol reported by Lee et al.²⁴, compound 2 (1.5 g, 6.91 mmol) was dissolved in dry dichloromethane (20 mL) under argon. After cooling down in an ice bath, triethylamine (2.5 eq., 2.4 mL, 17.275 mmol) and 4-dimethylaminopyridine (DMAP) (0.1 eq., 0.084 g, 0.619 mmol) were added and stirred for 10 minutes. Subsequently, octanoyl chloride (2.2 eq., 2.6 mL, 15.202 mmol) was carefully added to the mixture dropwise and the reaction was monitored by TLC and continued in an ice bath for 3 hours. When the starting material was fully consumed, DI water (25 mL) was added to the reaction mixture and the organic fraction was extracted (3×30 mL) by dichloromethane. The combined organic fraction was then dried over anhydrous sodium sulfate, filtered, and evaporated under vacuum. The crude product was subjected to silica column chromatography with EtOAc/hexanes (1:49) to yield compound 3 as a cloudy liquid (2.28 g, 70% yield). TLC (silica, EtOAc/hexanes 1:9, v/v): R_f = 0.4. ¹H NMR (500 MHz, CDCl₃) δ 7.55 (d, J = 8.2 Hz, 1H), 7.37 (d, J = 2.2 Hz, 1H), 7.17 (dd, J = 8.2, 2.2 Hz, 1H), 5.17 (s, 2H), 5.06 (s, 2H), 2.39 (t, J = 7.6 Hz, 2H), 2.34 (t, J = 7.5 Hz, 2H), 1.70–1.60 (m, 4H), 1.34–1.23 (m, 16H), 0.87 (td, J = 7.0, 1.7 Hz, 6H). ¹³C NMR (126 MHz, CDCl₃) δ 173.60, 173.46, 135.96, 135.84, 133.13, 129.55, 129.40, 123.12, 65.55, 65.21, 34.37, 34.34, 31.77, 31.76, 29.22, 29.19, 29.03, 29.01, 25.08, 25.04, 22.70, 22.69, 14.17. FT-IR (neat, cm⁻¹): 2925, 2855 (C–H, alkyl and aryl), 1737 (C=O, ester), 1463 (Ar C=C), 1155 (C–O). DART-MS (positive) [M+H]⁺ calcd for C₂₄H₃₈BrO₄: 469.1953, found: 469.1975.

(4-(4,4,5,5-tetramethyl-1,3,2-dioxaborolan-2-yl)-1,3-phenylene)bis(methylene) dioctanoate (4)

Following the typical protocol for Suzuki–Miyaura cross-coupling,³¹ compound 3 (1.88 g, 4.0 mmol) was dissolved in anhydrous 1,4-dioxane (50 mL). Bis(pinacolato)diboron (1.2 eq., 1.22 g, 4.8 mmol), potassium acetate (3 eq., 1.18 g, 12.0 mmol), and [1,1’-bis(diphenylphosphine) ferrocene] palladium dichloride–dichloromethane complex (0.1 eq., 0.033 g, 0.4 mmol) were added and the reaction was heated to reflux under argon overnight. After cooling down to r.t., the mixture was filtered through a celite column to remove insoluble salts, and the solvent was evaporated using a rotary evaporator. The crude mixture was then redissolved in dichloromethane (40 mL), washed with water (3×30 mL), dried over anhydrous sodium sulfate, filtered, and concentrated. The crude product was then purified by silica column chromatography twice with EtOAc/hexanes (1:49) to obtain lipid 4 as a clear liquid (1.172 g, 67% yield). TLC (silica, EtOAc/hexanes 1:9, v/v): R_f = 0.3. ¹H NMR (500 MHz, CDCl₃) δ 7.83 (d, J = 7.4 Hz, 1H), 7.34 (s, 1H), 7.28 (d, J = 7.4 Hz, 1H), 5.38 (s, 2H), 5.11 (s, 2H), 2.34 (td, J = 7.6, 2.9 Hz, 4H), 1.64 (q, J = 6.9 Hz, 4H), 1.32 (s, 12H), 1.30–1.22 (m, 16H), 0.86 (t, J = 7.0 Hz, 6H). ¹³C NMR (126 MHz, CDCl₃) δ 173.72, 173.67, 142.70, 139.26, 136.62, 127.87, 126.83, 83.97, 65.96, 65.82, 34.54, 34.45, 31.81, 31.80, 31.78, 29.84, 29.31, 29.23, 29.09, 29.05, 25.14, 25.09, 24.96, 22.73, 22.72, 14.19. ¹¹B NMR (160 MHz, CDCl₃) δ 31.09. DART-MS (positive) [M+H]⁺ calcd for C₃₀H₅₀BO₆: 517.3695, found: 517.3728.

(4-bromo-1,3-phenylene)bis(methylene) didodecanoate (5)

The same procedure as 3 was followed to produce 5 starting with 1.0 g of 2 (4.61 mmol) and replacing octanoyl chloride with lauroyl chloride. The product appeared as white solid (1.36 g, 51% yield). TLC (silica, EtOAc/hexanes 1:9, v/v): R_f = 0.45. ¹H NMR (500 MHz, CDCl₃) δ 7.56 (d, J = 8.2 Hz, 1H), 7.38 (d, J = 2.1 Hz, 1H), 7.18 (dd, J = 8.2, 2.1 Hz, 1H), 5.18 (s, 2H), 5.06 (s, 2H), 2.39 (t, J = 7.6 Hz, 2H), 2.35 (t, J = 7.6 Hz, 2H), 1.70–1.60 (m, 4H), 1.29–1.25 (m, 32H), 0.88 (t, J = 6.9 Hz, 6H). ¹³C NMR (126 MHz, CDCl₃) δ 173.63, 173.50, 135.98, 135.86, 133.16, 129.57, 129.41, 123.14, 65.58, 65.23, 34.40, 34.38, 32.05, 29.74, 29.73, 29.61, 29.60, 29.47, 29.41, 29.39, 29.31, 29.28, 25.11, 25.07, 22.82, 14.25.

(4-(4,4,5,5-tetramethyl-1,3,2-dioxaborolan-2-yl)-1,3-phenylene)bis(methylene) didodecanoate (6)

The same procedure as 4 was followed using 0.2 g of 5 to yield 6 as white solid (0.022 g, 10% yield). TLC (silica, EtOAc/hexanes 1:9, v/v): R_f = 0.4. ¹H NMR (500 MHz, CDCl₃) δ 7.83 (d, J = 7.6 Hz, 1H), 7.34 (s, 1H), 7.29 (d, J = 7.6 Hz, 1H), 5.38 (s, 2H), 5.12 (s, 2H), 2.35 (td, J = 7.6, 5.0 Hz, 4H), 1.65 (p, J = 7.6 Hz, 4H), 1.33 (s, 12H), 1.27 (m, 32H), 0.88 (t, J = 6.8 Hz, 6H). ¹³C NMR (126 MHz, CDCl₃) δ 173.73, 173.68, 142.71, 139.27, 136.62, 127.84, 126.82, 83.97, 65.96, 65.83, 34.55, 34.46, 32.06, 29.75, 29.62, 29.60, 29.48, 29.45, 29.41, 29.38, 29.30, 25.15, 25.10, 24.97, 22.83, 14.26. ¹¹B NMR (160 MHz, CDCl₃) δ 30.61.

Preparation of Liposomes for Nile Red (NR) Release Assays

Stock solutions of lipid 4 (10 mM), PC, and NR were prepared in chloroform. Appropriate volumes were combined to obtain total lipid concentrations of 0.7 mM with the desired molar ratios of lipid 4/PC. NR was included at 5 mol% relative to total lipid.

Organic solvents were removed under a gentle stream of nitrogen, and the lipid film was dried under high vacuum for at least 2 h. Hydration was performed at 60°C using 1× TBS (25 mM Tris/Tris HCl, 0.13 M NaCl, 2.7 mM KCl, pH 7.4) for four sets of 15 min with vortexing. Samples underwent ten freeze–thaw cycles (dry ice/acetone bath and 60°C water bath) followed by extrusion through a 200 nm polycarbonate membrane for 21 passes. Liposomes were stored at 4°C and used within 24 h.

Nile Red Release Assays with Hydrogen Peroxide

For release studies, 100 μL of NR-loaded liposomes was transferred into a submicro quartz cuvette. Release was initiated by adding 1 μL of H₂O₂ stock solution (final concentration 0–98 mM). Fluorescence emission at 635 nm was monitored over time (excitation 552 nm; slits 5/5 nm). Data were normalized to the initial fluorescence intensity and reported as a percentage of initial signal.

Controls included (i) lipid 4-containing liposomes treated with Milli-Q water and (ii) PC-only liposomes treated with H₂O₂.

Dynamic Light Scattering (DLS) Analysis

DLS was used to measure liposome diameters before and after H₂O₂ treatment. Samples were diluted 10-fold in the TBS buffer and analyzed at 30°C using backscatter detection at 173°. Measurements were performed on three independently prepared liposome samples for each condition, and results are reported as mean ± standard error.

DLS analysis was performed for NR-loaded lipid 4/PC liposomes as well as the corresponding PC-only control liposomes.

Preparation of Liposomes for ANTS/DPX Hydrophilic Dye Release Assays

ANTS (12.5 mM) and DPX (45 mM) were co-dissolved in 20 mM HEPES buffer (pH 7.4), and osmolality was adjusted to ~ 305 mOsm/kg with NaCl. Lipid mixtures were prepared to yield a final composition of 50% lipid 4 and 50% PC at a total lipid concentration of 1 mM.

Following solvent evaporation and vacuum drying for at least 2 h, the lipid film was hydrated with ANTS/DPX solution at 60°C for four sets of 15 min with intermittent vortexing. Ten freeze–thaw cycles were performed, and samples were extruded through 200 nm membranes for 21 passes.

Unencapsulated dye was removed using a Sephadex G-50 size-exclusion column pre-equilibrated with isotonic HEPES buffer. Fractions (~ 1 mL) were collected, and the second turbid fraction was used for release assays.

ANTS/DPX Release Assays with Hydrogen Peroxide

Purified dye-loaded liposomes (100 μL) were added to a submicro quartz cuvette, followed by addition of 1 μL of 30% H₂O₂ to make a final concentration of 98 mM. Fluorescence emission at 510 nm (excitation 350 nm; slits 10/10 nm) was monitored over time. After each experiment, 1 μL of 20% Triton X-100 was added to lyse vesicles and determine maximum fluorescence. Release was calculated using:

$$\text{Percent Release}=\frac{F_t{-F}_{\text{0}}}{F_{\text{Triton}}{-F}_{\text{0}}}\times 100.$$

Controls included replacing H₂O₂ with Milli-Q water or substituting lipid 4 with PC in the membrane formulation.

Conclusions

We designed a synthetically accessible non-phospholipid aryl boronate lipid that forms stable liposomes and undergoes selective oxidative activation by H₂O₂. The combination of a concise synthetic route that minimizes chromatography and robust ROS-responsive liposomal behavior maximizes the potential of this non-phospholipid aryl boronate scaffold as a scalable platform for oxidant-triggered delivery systems. This architecture provides for reliable vesicle formation, minimal basal leakage, and clear dose-dependent release, as corroborated through Nile Red, ANTS–DPX, and DLS analyses. Taken together, the synthetic route in Scheme 1 and the functional data in Figures 2-4 provide a coherent mechanistic and practical picture. Lipid 4, prepared in four high-yielding steps from an inexpensive aromatic acid and isolated in gram quantities, integrates a boronate trigger that remains silent under non-oxidative conditions but undergoes H₂O₂-mediated oxidative deborylation to the corresponding phenol. This transformation perturbs headgroup polarity and packing, leading to dose-dependent increases in vesicle size and polydispersity and to the release of both hydrophobic (NR) and hydrophilic (ANTS–DPX) cargos. Negative controls remain unresponsive, validating the idea that membrane permeabilization originates from boronate oxidation rather than nonspecific degradation. While these studies establish a strong proof-of-concept, some limitations remain. In particular, efficient leakage currently requires relatively high oxidant concentrations and proceeds more slowly than in some previously reported ROS-responsive liposomal systems. The current study investigates model cargos and simplified oxidative conditions, while the precise pathway for oxidation–permeabilization deserves deeper mechanistic resolution. Further and larger assessments that include biologically relevant ROS conditions and therapeutically meaningful payloads will further define the translational potential of this platform. This non-phospholipid boronate scaffold represents a practical, modular pathway to redox-responsive release. This general approach could also be extended to develop switches that respond to a range of other disease-associated stimuli by swapping out the reactive trigger moiety. Due to its synthetic flexibility and scalability, the platform opens clear opportunities toward sensitivity tuning, incorporation of targeting motifs, and adaptation toward therapeutically oriented formulations.

Supplementary Material

The Supporting Information is available free of charge and includes characterization data (NMR, IR, and mass spectra) for the synthetic compounds.