Encoding Biological Selectivity Through Positional Isomerism of Phosphindole Oxide AIEgens for Targeted Photodynamic Therapy

ABSTRACT

Discriminating biological targets via their intrinsic biophysical signatures offers a promising alternative to classical lock‐and‐key recognition, yet remains constrained by simplistic design. Herein, we present a programmable molecular strategy that leverages configurational isomerism on a phosphindole oxide (PIO) scaffold as a decisive variable for encoding biological selectivity. Strategic substitution at geometrically defined positions generates isomeric pairs with distinct conformational preferences, aqueous self‐assembly behaviors, and surface electrostatic landscapes. These isomer‐dependent divergences translate into sharply differentiated biological staining patterns. One isomer enables selective labeling of cancer over normal mammalian cells and preferential engagement with bacterial versus mammalian cells, while its counterpart functions as a broad‐spectrum staining agent. By integrating with the inherent photodynamic activity of the PIO framework, we achieve effective tumor suppression and accelerated healing of infected wounds in vivo with favorable biosafety. This work establishes positional isomerism as a generalizable design dimension for encoding biophysical selectivity, providing molecular‐level guidance for next‐generation theranostic and precision medicine platforms.

Positional isomerism on a phosphindole oxide scaffold programs conformational preferences, aqueous assembly architectures, and surface electrostatics, directing differential biological engagement and underpinning target photodynamic therapy against tumors and infected wounds with favorable biosafety.

1. Introduction

The precise discrimination of biological targets is a cornerstone of advanced therapeutics [1, 2]. While canonical lock‐and‐key recognition paradigms, such as those based on antibody–antigen and nucleic acid hybridization interactions, offer high specificity, their practical application is often constrained by cost, complexity, immunogenic risks, and, critically, an inherent susceptibility to heterogeneity in target expression [3, 4, 5]. An emerging complementary paradigm seeks to exploit innate biophysical disparities (e.g., fluidity, tension, permeability, surface charge) between cell types (e.g., healthy vs. cancerous, mammalian vs. bacterial) as a robust basis for differentiation [6, 7, 8, 9, 10]. This provides a cost‐effective and broadly applicable route for early‐stage screening and targeted intervention. Current mainstream strategies within this paradigm often rely on simplistic adjustments, like varying chain length, charge, or rigidity [11, 12, 13, 14, 15, 16, 17], which limit chemical diversity, functional scope, and mechanistic depth. Although recent advancements have begun exploiting more subtle structural variations, such as regioisomeric, constitutional, configurational, and chiral isomerism [13, 18, 19, 20, 21, 22], there is still room to go beyond empirical trial‐and‐error modification by explicitly linking these subtle structural variations to cell‐type‐specific biophysical targeting.

A greater challenge lies in the integration of such biophysics‐guided selectivity and potent therapeutic function within a single molecular framework. This integration is especially critical in the context of photodynamic therapy (PDT) [23, 24, 25, 26], where therapeutic safety and efficacy are intrinsically tied to the precise target and spatially controlled activation of photosensitizers (PSs) [27, 28, 29]. To address this fundamental design challenge, we introduce a molecular design strategy that employs positional isomerism as a decisive and programmable variable to direct interfacial interactions and supramolecular assembly in biological environments. This strategy is implemented on a phosphindole oxide (PIO) scaffold [30, 31] as a platform for its well‐established role as a high‐performance PS core [32, 33] with favorable photophysical and photochemical properties [34, 35] and its geometrically and electronically distinct substitution sites providing precise handles for isomeric design [36, 37, 38].

As illustrated in Figure 1, the α‐site on PIO enforces a planar, rigid geometry, whereas the β‐site permits a twisted, flexible conformation. By strategically allocating a nonpolar, electron‑donating triphenylamine derivative group (TPA/MeTPA) and a polar, electron‑accepting pyridinium unit (PyMe/PyBu) to these sites, we construct a modular PIO‐linked donor–acceptor (D–A) system whose isomers share identical chemical modules and differ solely in their α/β site connectivity, thereby yielding four positional isomer pairs (Scheme S1) as a molecular library. As exemplified by the MeTPAPyBu isomers, positional isomerism is expected to dictate their aqueous self‐assembly outcome [39, 40]. In α‑MeTPA‑PyBu, the rigid MeTPA‑PIO backbone drives a compact hydrophobic interior, leaving the flexible PyBu units adaptively exposed at the periphery, conferring good structural stability but limited plasticity and restricted cationic exposure. This architecture may favor intact translocation through biological membranes rather than adaptive interfacial integration. This yields good structural stability but limited plasticity and cationic accessibility, thus favoring intact transmembrane translocation over adaptive interfacial integration. Conversely, β‑MeTPA‑PyBu embodies an inverted architecture with rigid PIO‑PyBu periphery encasing a highly adaptive MeTPA interior. This architecture confers high structural adaptability and cationic exposure, allowing the assembly to undergo dynamic reconfiguration that facilitates efficient integration with and penetration through biological membranes. Consequently, this deliberate isomeric programming is expected to underpin differential interaction profiles with cell‐type‐specific biophysical signatures (e.g., fluidity, permeability, and surface charge), thereby allowing for tailored selectivity across cell types.

FIGURE 1.

Schematic illustration of design strategy, encompassing positional isomerism of the PIO‐linked D–A system, the resulting assembly architecture characteristics (illustrated with MeTPA‐–PyBu isomers, PF₆ ^– counterions omitted for clarity), and the encoded biological selectivity enabling targeted PDT applications.

2. Results and Discussions

2.1. Design and Synthesis

To validate the proposed design principles, the MeTPAPyMe/PyBu series was selected for experimental investigation. This decision was driven by the limitations of the previously reported analogues (α‐TPAPyMe/PyBu), whose excessive hydrophilicity undermined the formation of stable nanoaggregates in aqueous media [12]. To address this, our design strategically incorporated extended alkyl chains (TPA → MeTPA and PyMe → PyBu). This modification enhanced the overall molecular hydrophobicity, thereby restoring an optimal amphiphilic balance to enable controlled self‐assembly in water.

The synthesis of the target PIO derivatives followed a modular route, as shown in Scheme S2. The key step involved a silver(I) oxide‑mediated coupling between a pre‑synthesized MeTPA–ethynyl–pyridine intermediate and ethyl phenylphosphinate, affording a separable mixture of α‑ and β‑regioisomers. Following chromatographic purification, the cationic pyridinium moiety was introduced via N‑alkylation of the pyridine unit with either iodomethane or 1‑iodobutane, producing the corresponding PyMe or PyBu iodide salts, respectively. A final anion exchange with potassium hexafluorophosphate (KPF₆) yielded the desired PF₆ ^– salts, which offer improved counterion stability. Detailed synthetic procedures and full characterization data are provided in the Electronic Supporting Information (ESI) [41].

2.2. From Molecular Structure to Macroscopic Self‐Assembly Behavior

Static water contact angle measurements were conducted to comparatively assess the surface hydrophilicity/hydrophobicity of the PIO derivatives (Figure S1). The results demonstrate that elongation of the alkyl chains (TPA → MeTPA and PyMe → PyBu) systematically enhances hydrophobicity, a trend consistent with the corresponding increase in ClogP values. For the PyMe derivatives, β‑TPA‑PyMe and β‑MeTPA‑PyMe exhibit greater hydrophilicity compared to their α‑counterparts, likely attributable to the more rigid presentation of the PIO‐PyMe moiety, which promotes the formation of a well‐ordered, polar interface. In contrast, the distinction between the MeTPA‑PyBu isomers is less pronounced at the macroscopic level of contact angle measurement, where the overarching hydrophobicity of the butyl chains predominates.

The aqueous self‐assembly behaviors of the PIO derivatives were subsequently investigated by transmission electron microscopy (TEM, Figure S2) and dynamic light scattering (DLS, Figure 2A,B). All four PIO compounds form well‐defined nanoaggregates in water, confirming the successful attainment of amphiphilic balance through molecular design. The assembly outcomes were found to be exquisitely governed by two key structural variables: alkyl chain length and positional isomerism. The MeTPA‑PyMe isomers assemble into smaller nanoparticles and exhibit near‐zero/modest zeta potentials. In contrast, the MeTPA‑PyBu isomers formed larger nanoparticles and significantly higher surface charges. This pronounced difference is attributed to the enhanced hydrophobic driving force imparted by the extended butyl chains, which promotes more extensive aggregation, yielding nanoparticles with a denser hydrophobic core and a more pronounced cationic shell. Furthermore, within each isomeric pair, the β‐MeTPA derivatives consistently exhibited higher zeta potentials than their αcounterparts. Remarkably, β‐MeTPAPyBu also displayed a broader size distribution compared to α‐MeTPAPyBu, suggesting its participation in a more dynamic assembly equilibrium with a more accessible, charged interfacial environment.

FIGURE 2.

(A) Particle size distribution and (B) zeta potential of the MeTPA‐PyMe and MeTPA‐PyBu isomers (10 µM, in water with 1 vol % DMSO). (C) Potential energy profiles as a function of dihedral angle for the MeTPA and PyBu segments in the MeTPA‐PyBu isomer. (D) Representative snapshots of the MeTPA‐PyBu isomers (counterions included) at 0, 100, 200, 300, and 400 ns. Enlarged: Representative equilibrium‐state structures of the MeTPA‐PyBu isomer aggregates (pyridinium groups highlighted in ball‐and‐stick) with size parameters. (E) Time‐dependent evolution of molecular count in the largest cluster of the MeTPA‐PyBu isomer aggregates, with the average values over the equilibrium‑state portion indicated (counterions included). (F) Time‐dependent evolution of dRMSD and R _g of the MeTPA‐PyBu isomer aggregates, with the average values over the equilibrium‑state portion indicated (counterions excluded). (G) ESP maps and (H) comparison of key structural and energetic parameters, including N (H‐bonding), SASA, and Coulomb potential (V (Coulomb)) of the equilibrium‐state MeTPA‐PyBu isomer aggregates (counterions excluded).

To elucidate the molecular origin of the isomer‐dependent assembly behaviors, we performed theoretical calculations on the representative MeTPAPyBu isomers. α‐MeTPA‐PyBu adopts dihedral angles of 36.8° and 46.2° at the α‐ and β‐sites, respectively, with bounding box dimensions of approximately 18.2 Å × 16.1 Å × 11.8 Å (Figure S3). In contrast, β‐MeTPA‐PyBu adopts dihedral angles of 28.5° and 49.4° at the α‐ and β‐sites, revealing a more conformationally asymmetric molecular geometry with dimensions of approximately 18.7 Å × 17.2 Å × 9.6 Å. For both isomers, the α‐site dihedral is smaller than the β‐site one, confirming stronger conjugation via α‐site coupling. Although their overall molecular dimensions are comparable, α‐MeTPA‐PyBu exhibits a relatively compact and volumetrically balanced molecular dimension, while β‐MeTPA‐PyBu displays pronounced anisotropy with a more elongated and flattened molecular profile, which is expected to introduce steric mismatch during molecular packing while favoring the formation of directional and loosely packed supramolecular assemblies.

Potential energy surface analysis further verifies the asymmetry between rigidity at the α‐site and flexibility at the β‐site. As shown in Figure 2C, the rotation about the α‐site dihedral is associated with a higher energy barrier, reflecting a relatively rigid backbone enforced by α‐site conjugation. In contrast, the β‐site dihedral accesses a broader low‐energy conformational region, conferring enhanced conformational flexibility and adaptability.

Molecular dynamics (MD) simulations were performed to track the aggregation process of the MeTPA‑PyBu isomers in an aqueous environment. As shown in Figure 2D, both isomers initially undergo a similar nucleation phase, subsequently forming an elongated cluster from randomly dispersed monomers. Despite this similar onset, their subsequent assembly trajectories diverge markedly (Figures 2E,F and S4). β‑MeTPA‑PyBu demonstrates faster aggregation kinetics, reaching an equilibrium state in 100 ns while maintaining an elongate supramolecular morphology throughout the simulation, with its longest axis extending to 8.87 nm. In contrast, α‐MeTPAPyBu undergoes a pronounced structural collapse, requiring over 300 ns to reorganize into a quasi‐spherical core with dimensions confined to 4.29–6.20 nm. The supramolecular packing divergence was further quantified by radius of gyration (R _g), a measure of structural compactness, and distance root‐mean‐square deviation (dRMSD), which reflects internal atom distance fluctuations. αMeTPAPyBu exhibits a smaller R _g (∼1.92 nm) but a higher dRMSD (∼3.23 nm), confirming a compact interior packing accompanied by noticeable peripheral fluctuation. In contrast, β‐MeTPA‐PyBu exhibits a larger R _g (∼2.26 nm) together with a small dRMSD (∼3.04 nm), consistent with loosely packed architecture but relatively rigid outer periphery. This divergence fundamentally shapes the interfacial exposure and electrostatic presentation of the resulting assemblies. Quantitative analysis reveals that the β‐MeTPA‐PyBu aggregate displays a larger solvent‐accessible surface area (SASA, 148.5 nm²) and forms more hydrogen bonds (N (H‐bonding), 82.8) than α‐MeTPA‐PyBu (SASA, 144.5 nm²; N (H‐bonding), 64.0) (Figures 2H and S4), confirming its enhanced solvent exposure and interfacial accessibility. More importantly, electrostatic analysis shows that the β‐MeTPA‐PyBu aggregate exhibits a highly heterogeneous electrostatic potential (ESP) distribution with intensified positive charge regions and a substantially higher Coulomb potential (+40 485 kJ mol⁻¹), compared to the more homogeneous ESP profile and lower Coulomb potential (+29 302 kJ mol⁻¹) of α‐MeTPAPyBu (Figure 2G). The patchy, anisotropic electrostatic landscape of the β‐MeTPA‐PyBu aggregate favors strong, directional interactions with negatively charged biological membranes, whereas α‐MeTPA‐PyBu exhibits weaker and less spatially focused electrostatic interactions.

Collectively, the intrinsic conformational preferences dictated by positional isomerism (α vs. β) govern the supramolecular architecture (compact vs. loose) during assembly, which in turn defines the surface electrostatic landscape (uniform/weak vs. heterogeneous/strong). This provides a robust foundation for understanding their subsequent divergent biological interactions, moving the analysis from correlation to mechanistic causation.

2.3. Photophysical Behaviors and Reactive Oxygen Species Generation

The photophysical properties of the PIO derivatives were characterized to understand how configurational isomerism modulates optical signatures. Absorption spectra reveal a pronounced divergence between the isomers (Figure 3A). The α‑MeTPA derivatives exhibit a stronger absorption maximum at 445–450 nm, indicative of a more effective π‑conjugation between the MeTPA donor and PIO acceptor. In contrast, the β‑MeTPA derivatives display a weaker and notably red‑shifted absorption band at 480–485 nm (Figure 2B), which is characteristic of a charge‐transfer (CT) state, arising from the twisted conformation between MeTPA and PIO.

FIGURE 3.

(A) Absorption (10 µM, in DMSO) and (B) emission (100 µM, in water with 5 vol % DMSO) spectra of the MeTPA‐PyMe and MeTPA‐PyBu isomers. (C) Plots of I/I ₀ versus concentration of the MeTPA‐PyMe and MeTPA‐PyBu isomers in DMSO and in water (with 5 vol % DMSO), where I ₀ is the emission intensity at 10 µM. (D) Plots of I/I ₀ vs f _Hex of the MeTPA‐PyMe and MeTPA‐PyBu isomers in acetone/Hex mixtures with different f _Hexs, where I ₀ is the emission intensity at f _Hex = 0. Inset: PL spectra of α‐MeTPAPyBu in acetone/Hex mixtures with different f _Hexs (10 µM). (E) Plots of relative PL intensity of DCF for ROS detection in the presence of the MeTPA‐PyMe and MeTPA‐PyBu isomers versus irradiation time (10 µM, in water with 1 vol % DMSO). White light: 20 mW cm⁻². (F) The highest occupied molecular orbital (HOMO) and the lowest unoccupied molecular orbital (LUMO) of the MeTPA‐PyBu isomers. (G) Calculated excited‐state energy levels of the MeTPA‐PyBu isomers with spin–orbit coupling (SOC) constants.

Regarding photoluminescence, the α‑MeTPA derivatives show weak but detectable emission in dilute dimethyl sulfoxide (DMSO) solution. In contrast, their β‑MeTPA counterparts are essentially non‑emissive across a broad concentration range from 10 µM to 100 µM (Figure S5), consistent with their twisted, CT‐prone conformation, which strongly suppresses radiative recombination. Remarkably, upon aqueous aggregation, the MeTPA‑PyBu isomers display a pronounced, concentration‐dependent enhancement of deep‑red emission (λ _max = 675–692 nm, Figure 3B), indicative of aggregation‑induced emission (AIE) [42, 43] features (Figure 3C), with β‐MeTPA‐PyBu exhibiting a significantly stronger enhancement than its α‐counterpart. In comparison, β‐MeTPA‐PyMe exhibits a modest increase with increasing concentration, while α‐MeTPA‐PyMe remains largely weakly emissive across the entire concentration range. The superior AIE response of the β‑MeTPA derivatives in aqueous media reflects the effective restriction of intramolecular motion (RIM) [44, 45, 46] of their flexible MeTPA units within the hydrophobic aggregate core. Conversely, the weaker aqueous AIE response of the α‑MeTPA counterparts signifies the insufficient constraint of their flexible pyridinium groups positioned at the hydrated aggregate periphery. To further verify the AIE effect in a nonpolar environment, we examined the emission behavior in acetone/hexane (Hex) mixtures (Figures 3D and S6). Upon aggregation in Hex (Figure S7), the MeTPA‑PyBu isomers exhibit a dramatic emission enhancement at high Hex fraction (f _Hex), while the MeTPA‑PyMe isomers display relatively minimal variation. These observations corroborate the intrinsic AIE nature of the MeTPA‑PyBu isomers, reflecting a more effective RIM effect from the PyBu unit, relative to that in the PyMe counterpart.

The photodynamic potential of the PIO derivatives was evaluated by their ability to generate reactive oxygen species (ROS) under white‐light irradiation (Figures 3E and S8). A clear correlation is observed, with ROS generation efficiency closely following their AIE activity [46]. The MeTPA‑PyBu isomers exhibit higher ROS generation ability than the MeTPA‑PyMe isomers, and the α‑MeTPA derivatives outperform their α‑TPA counterparts, which is rationalized by alkylation‐enhanced aggregation, indicating a pronounced aggregation‐induced ROS generation (AIG‐ROS) behavior. In addition, the β‐MeTPA derivatives show higher ROS yields than their α‐MeTPA counterparts. This is mechanistically explained by their favorable excited‐state dynamics: the twisted CT state (Figure 3F) in the β‐MeTPA derivatives reduces the singlet‐triplet energy gap (ΔE _ST), and enhances spin–orbit coupling (SOC, Figure 3G), thereby facilitating intersystem crossing (ISC) [47, 48]. Furthermore, ROS production also increases with alkyl‐chain extension (PyMe < PyBu, TPA < MeTPA), directly linking enhanced molecular hydrophobicity to more robust aggregation and consequently amplified photodynamic output. Overall, the MeTPA‑PyBu isomers exhibit decent AIE activity and exceptional ROS production, positioning them as promising candidates for further photodynamic therapy applications.

2.4. Cell‑Selective Staining and Biocompatibility Profiles

The biological interaction profiles of the PIO derivatives were first examined by confocal laser scanning microscopy (CLSM), which revealed that fluorescence signals are predominantly localized in the cytoplasm, along with a general trend of enhanced fluorescence in cancer mammalian cells (HeLa and MCF‐7) relative to normal mammalian cells (L929 and NIH/3T3) (Figure S9). To quantitatively validate this observation, we performed flow cytometry and calculated the relative mean fluorescence intensity (MFI) values normalized to untreated controls (Figure 4A). The results confirm consistently higher fluorescence in cancer cells than in normal cells, likely associated with the distinct biophysical characteristics of cancer cells, particularly their elevated negative surface charge density and increased membrane fluidity [6, 7, 8, 9, 10], which favor interactions with cationic and structurally adaptable nanoaggregates. Within this general trend, the relative MFI values vary markedly, being strongly dependent on positional isomerism. Specifically, the α‐MeTPA derivatives yield exceptionally low fluorescence signals in normal cells (relative MFI ∼1.3–2.7), approaching background levels, yet exhibit a pronounced and selective enhancement in cancer cells. For α‑MeTPA‑PyMe, the relative MFIs reach ∼5.0–12.4 in cancer cells, yielding a cancer‑to‑normal contrast of ∼3.6–9.5 fold. More impressively, α‑MeTPA‑PyBu shows a further increase to ∼12.4–15.7, affording a robust contrast of ∼5.8–7.9 fold. In contrast, β‐MeTPAPyMe produces more comparable signals across cell types, giving a substantially lower cancer‑to‑normal contrast of only ∼1.3–3.2 fold. Although β‐MeTPAPyBu exhibits a moderately higher contrast of ∼2.4–5.6 fold, it suffers from substantial staining in normal cells (relative MFI ∼9.5–18.3), which compromises their specificity. This discrepancy can be attributed to the distinct supramolecular architectures of the two isomer families. The less exposed cationic periphery and restricted conformational adaptation of the α‐MeTPA derivatives raise a higher kinetic barrier for membrane interaction. Consequently, effective engagement requires a cellular environment with higher charge density and increased membrane fluidity, as typically observed in cancer cells, to overcome this barrier. Conversely, the fully accessible cationic interface and adaptable hydrophobic interior of β‑MeTPA derivatives facilitate efficient but less selective membrane association, enabling broad membrane interaction across different cell types. Collectively, these results identify α‑MeTPA‑PyBu as a highly selective probe for cancer cells over normal mammalian cells, highlighting the critical role of positional isomerism in encoding cell‐type specificity.

FIGURE 4.

(A) Flow cytometry histograms of various mammalian cells after treatment with the MeTPA‐PyMe and MeTPA‐PyBu isomers (5 µM, 12 h), with relative MFI values normalized to untreated controls indicated. (B) Cell viability of L929 cells after treatment with the MeTPA‐PyMe and MeTPA‐PyBu isomers at varied concentrations (24 h). Flow cytometry histograms of (C) MRSA and (D) C. albicans after treatment with the MeTPA‐PyMe and MeTPA‐PyBu isomers (5 µM, 30 min), with relative MFI values normalized to untreated controls indicated.

These divergent engagement patterns displayed by the PIO derivatives observed in normal mammalian cells prompted an evaluation of biosafety. Cytotoxicity assays reveal that the MeTPAPyMe isomers are well tolerated at concentrations up to 15 µM, and α‐MeTPA‐PyBu maintains cell viability above 80% with only minor fluctuation (Figures 4B and S10), underscoring their suitability as a biocompatible probe. Differently, β‐MeTPAPyBu induces a substantial loss of viability, with reductions exceeding 50% at 15 µM, which notably aligns with its potent, non‐selective accumulation in normal cells. Complementary hemolysis assays (Figure S11) indicate that the MeTPAPyMe isomers and α‐MeTPA‐PyBu exhibited negligible hemolytic activity with lysis <5% even at 20 µM. Conversely, β‑MeTPA‑PyBu triggers pronounced, dose‐dependent erythrocyte lysis, exceeding the conventional safety threshold (5%) at concentrations as low as 10 µM. This confirms its strong intrinsic membrane‐disruptive potential, a direct consequence of its exposed cationic surfaces and dynamic interfacial packing.

A striking reversal of this eukaryotic selectivity pattern was observed in the prokaryotic model of methicillin‐resistant Staphylococcus aureus (MRSA), with the α‐MeTPA derivatives (relative MFI ∼4.4–7.9) outperforming their β‐counterparts (relative MFI ∼1.8–2.7) (Figures 4C and S12). In addition, alkyl chain elongation (PyMe → PyBu) further enhances the fluorescence intensity across both isomeric series, making α‐MeTPAPyBu the brightest probe in MRSA. Colocalization imaging with the nucleic acid stain Hoechst 33342 indicates that the MeTPA‐PyBu isomers localize predominantly at the bacterial periphery (Figure S13), rather than accumulating within the cytoplasm, likely through both electrostatic and hydrophobic interactions with the Gram‐positive bacterial cell envelope. The dense, highly crosslinked peptidoglycan layer acts as a selective structural barrier [49, 50], which favors the penetration and retention of the uniform‐sized, compact, and less hydrophilic nanoaggregates formed by the α‐MeTPA derivatives, while impeding the passage of the heterogeneous, looser, and more hydrophilic assemblies characteristic of the β‐MeTPA derivatives.

In contrast, only limited fluorescence enhancement was observed in the Gram‐negative bacterium Escherichia coli (E. coli) (Figures S14 and S15), with α‐MeTPA‐PyBu being the only one exhibiting a discernible yet heterogeneous signal increase. This weak staining efficiency is consistent with previous reports on cationic PIO derivatives [12, 31, 32, 35], and can be attributed to the unique cell envelope structure of Gram‐negative bacteria, where an additional lipopolysaccharide‐rich outer membrane acts as an effective permeability barrier that restricts the access of cationic nanoaggregates [49, 50]. This partial enhancement observed for α‐MeTPA‐PyBu suggests that, overall suboptimal staining, rational structural optimization is worth pursuing in future work to enable effective E. coli labeling in future work.

To extend the evaluation to fungal systems, CLSM and flow cytometry were performed on the yeast‑form Candida albicans (C. albicans) stained with the PIO derivatives (Figures 4C and S16). The PyMe derivatives exhibit weak fluorescence (relative MFI ∼1.7–1.8), whereas the PyBu derivatives show pronounced enhancement, with β‑MeTPA‑PyBu displaying a higher relative MFI (∼11.2) than its α‑counterpart (∼6.0). This behavior likely arises from the negatively charged and relatively permeable (in contrast to the rigid peptidoglycan network of Gram‐positive bacteria) fungal cell envelope [51], which facilitates electrostatic capture and efficient uptake of cationic, structurally nanoaggregates, in a manner reminiscent of mammalian cells. This preliminary result suggests β‐MeTPAPyBu as a promising scaffold for fungal imaging and antifungal PDT, warranting further investigation in future work.

Together, these results establish that positional isomerism is the primary structural driver of biological staining efficiency, which in turn dictates cell‐type selectivity and biocompatibility. This effectively positions α‐MeTPA‐PyBu as targeted and biosafe probes that exhibit selective staining toward cancer cells over normal mammalian cells and toward Gram‑positive bacteria over mammalian cells. In contrast, β‑MeTPA‑PyBu functions as a broad‐spectrum staining agent, albeit with its compromised biocompatibility.

2.5. From In Vitro to In Vivo Anticancer Therapy

The photodynamic efficacy of the PIO derivatives against cancer cells was systematically evaluated. In cellular assays, all compounds exhibit dose‐dependent phototoxicity in Hela and MCF‐7 cells. Interestingly, while the four compounds show comparable phototoxic effects toward HeLa cells (Figure S17), their activities diverge in MCF‐7 cells (Figures 5A and S18). The α‐MeTPA derivatives display significantly stronger phototoxicity than their β‐MeTPA counterparts. Notably, α‐MeTPAPyBu emerges as the most potent candidate, demonstrating the strongest photo‐induced cytotoxicity alongside minimal dark toxicity, a combination that underscores its favorable therapeutic window for on‐demand PDT applications.

FIGURE 5.

(A) Cell viability of MCF‐7 cells after treatment with the MeTPA‐PyBu isomers at varied concentrations (24 h) without or with light irradiation (applied after 12‐h treatment, 40 mW cm⁻², 20 min). Data are presented as mean ± s.d. (n = 6). *p < 0.05, **p < 0.01, and ***p < 0.001. (B) In vivo fluorescence imaging of MCF‐7 tumor‑bearing mice after intratumor injection of the MeTPA‐PyBu isomers (10 µL, 20 µM, 2 h). (C) Tumor volume growth curves during treatment employing the MeTPA‐PyBu isomers (10 µL, 20 µM) without or without light irradiation (applied after 2‐h treatment, 200 mW cm⁻², 15 min). (D) Tumor weights and representative photographs of excised tumors after treatment. Data are presented as mean ± s.d. (n = 5). ns = no significant, *p < 0.05, **p < 0.01, and ^# p < 0.05 vs. PBS (light +) group. (E) H&E (for morphology), Ki‑67 (for proliferation), and TUNEL (for apoptosis) staining of tumor tissues after tumor treatment.

The in vitro performance, particularly of α‐MeTPAPyBu, prompted further evaluation in an MCF‐7 tumor‐bearing mouse model. Following intratumoral injection, fluorescence imaging reveals pronounced and sustained retention of the MeTPAPyBu isomers within tumors (Figure 5B), delineating a clear therapeutic window for localized light activation. Strikingly, α‐MeTPAPyBu exhibits significantly stronger intratumoral fluorescence than β‑MeTPA‑PyBu, directly visualizing its enhanced cancer cell selectivity. Tumor growth curves demonstrate the light‐dependent suppression of tumor progression by both MeTPAPyBu isomers (Figure 5C). The therapeutic effect was further supported by significant reductions in both volume and weight of the excised tumors from the MeTPAPyBu + light treatment groups compared to the controls (Figure 5D). Histopathological examination of tumor tissues provides additional mechanistic insights into the treatment response, revealing extensive necrosis (via H&E staining), significantly suppressed proliferation (via Ki‐67 staining), and enhanced apoptosis (via TUNEL assay) in the treated tumor tissues (Figure 5E). Notably, α‐MeTPAPyBu elicits a more pronounced antitumor response than β‐MeTPAPyBu, an outcome directly consistent with its superior intratumoral accumulation and cancer cell selectivity. These results collectively substantiate the potential of α‐MeTPAPyBu as an effective candidate for targeted anticancer photodynamic therapy.

To evaluate systemic safety, histopathological analyses were performed on major organs (heart, liver, spleen, lung, and kidney) following the treatment. The examinations reveal no significant morphological abnormalities, inflammatory infiltration, or tissue damage across all organs assessed (Figure S19). These findings support a favorable biosafety profile for the MeTPAPyBu isomers and underscore their potential for further translational development in photodynamic oncology.

2.6. From In Vitro to In Vivo Antibacterial Therapy

The photodynamic antibacterial efficacy of the PIO derivatives was systematically evaluated against methicillin‐resistant S. aureus (MRSA). Against planktonic bacteria, the MeTPAPyMe isomers exhibit negligible bactericidal activity in the dark, while showing significantly enhanced killing effect upon light irradiation (Figure S20). Differently, the more hydrophobic MeTPAPyBu isomers display potent bactericidal activity even without light, which is further intensified upon irradiation, achieving complete sterilization at 5 µM (Figure 6A). Even stronger photodynamic bactericidal activity is observed against the standard S. aureus (Figure S21). In parallel, against established MRSA biofilms, the MeTPAPyMe isomers show limited influence (Figure S22), whereas the MeTPAPyBu isomers display considerable biofilm‐disrupting capability in the dark, with a marked increase observed following light exposure (Figure 6B). These results demonstrate that the MeTPAPyBu isomers function as effective dual‐mode antibacterial agents, combining inherent bactericidal activity in the dark with substantially augmented PDT efficacy.

FIGURE 6.

(A) Bacterial colony photographs of floating MRSA after treatment with the MeTPA‐PyBu isomers at varied concentrations (30 min) without or with light irradiation (40 mW cm⁻², 30 min). Bacterial survival rates were indicated. (B) Biomass of MRSA biofilm after treatment with the MeTPA‐PyBu isomers at varied concentrations for 30 min without or with light irradiation (40 mW cm⁻², 30 min). Data are presented as mean ± s.d. (n = 3). *p < 0.05, ***p < 0.001. (C) Representative photographs of MRSA‐infected wounds and (D) wound area analysis during treatment employing Van or α‑MeTPA‑PyBu (100 µL, 20 µM) without or without light irradiation (applied after 30‐min treatment, 100 mW cm⁻², 30 min). Data are presented as mean ± s.d. (n = 3). *p < 0.05, **p < 0.01, ^### p < 0.001 vs. PBS (light +) group. (E) H&E (for morphology) and Masson's trichrome (for collagen and connective tissue) staining of wound tissues after wound treatment.

Given its superior biocompatibility, α‐MeTPAPyBu was selected as the lead candidate for in vivo antibacterial evaluation. In a MRSA‑infected wound model, the therapeutic effects of α‐MeTPAPyBu were evaluated both in the absence and presence of localized light irradiation. Compared to the PBS control, wounds treated with α‐MeTPAPyBu alone show significant improvement in healing, comparable to the antibiotic Vancomycin (Van) (Figure 6C,D). This demonstrates that α‐MeTPAPyBu exerts intrinsic bactericidal activity independent of photoactivation. Notably, wounds treated with α‑MeTPA‑PyBu followed by light irradiation exhibit markedly accelerated contraction and tissue repair, surpassing the outcomes achieved with Van and α‐MeTPA‐PyBu alone, and approaching near‑complete healing. This regimen also leads to visibly reduced redness and swelling, faster scab detachment, and better restoration of skin texture relative to all other groups. Histological analysis of wound tissue sections further corroborates the superior regenerative outcome in the α‑MeTPA‑PyBu (light+) group, which shows markedly reduced inflammatory cell infiltration, well‑restored epidermal architecture (H&E staining), and more organized collagen deposition (Masson's trichrome staining) (Figure 6E). Together, these findings demonstrate that αMeTPAPyBu functions dually by combining inherent bactericidal activity with a potent PDT effect against MRSA wound infection, actively promoting tissue repair. This showcases its promise as an innovative non‐antibiotic treatment modality.

Histopathological assessment of major organs (heart, liver, spleen, lung, and kidney) posttreatment reveals normal tissue architecture without notable lesions or inflammation across all groups (Figure S23). These findings support the favorable systemic biosafety profile for α‐MeTPAPyBu and its potential for further therapeutic translation.

3. Conclusion

In summary, this work establishes positional isomerism as a decisive molecular variable for programming biological selectivity, thereby enabling precision photodynamic therapy. Strategic assembly of hydrophobic (MeTPA) and hydrophilic (PyMe/Bu) units at the geometrically constrained, electronically distinct α/β‐sites on a PIO scaffold generates isomeric pairs with divergent conformational preferences, supramolecular architectures, and surface electrostatic landscapes. Specifically, the α‑MeTPA derivatives possess a rigid PIO‐MeTPA backbone with a flexible PyMe/Bu unit, driving the formation of compact nanoaggregates with a limited, homogeneous cationic exposure. Conversely, the β‑MeTPA derivatives adopt a rigid PIO‑PyMe/Bu periphery encasing a highly adaptive MeTPA interior, yielding a loose nanoaggregate with structural adaptability and an anisotropic, highly accessible cationic surface. This precise structural control translates directly into divergent biological engagement: β‑MeTPA‑PyBu functions as a broad‑spectrum agent, whereas α‑MeTPA‑PyBu achieves cancer‑over‑normal and bacteria‑over‑mammalian selectivity. Accordingly, α‑MeTPA‑PyBu displays minimal cytotoxicity and hemolytic activity, thus enabling precise photodynamic therapy against tumors and infected wounds. Beyond the specific compounds reported here, this work establishes a generalizable design paradigm for encoding biological selectivity that can be extended to other molecular scaffolds featuring multiple distinct substitution sites. Furthermore, integrating other isomeric strategies offers a route to engineer orthogonal selectivity across multiple biological targets, while also enabling the combination of complementary therapeutic modalities to synergistically enhance treatment efficacy and safety. Collectively, this work provides a blueprint for expanding the functional scope of next‐generation theranostic systems and precision medicine.

Funding

The National Natural Science Foundation of China (U23A20594 and 22375066) and the GuangDong Basic and Applied Basic Research Foundation (2023B1515040003).

Conflicts of Interest

The authors declare no conflicts of interest.

Supporting information

Supporting File: adma74291‐sup‐0001‐SuppMat.pdf.

Data Availability Statement

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