Bioorthogonal aggregation-induced emission nanoprobes for targeted photodynamic eradication of Gram-negative pathogens in periodontitis

Abstract

Periodontitis is a bacteria-driven immune–inflammatory disease associated with progressive periodontal tissue destruction and alveolar bone loss. Conventional treatments rely heavily on broad-spectrum antibiotics, which can promote antimicrobial resistance and disrupt the oral microbiota. Herein, we report a bioorthogonal aggregation-induced emission (AIE)-based nanoprobe, AIE@PEG-DBCO, designed for the selective imaging and photodynamic elimination of Gram-negative periodontal pathogens. By exploiting the metabolic incorporation of 8-azido-3,8-dideoxy-D-manno-octulosonic acid (KDO-N₃) into the lipopolysaccharide of Gram-negative bacteria, AIE@PEG-DBCO enables strain-selective fluorescence turn-on and efficient reactive oxygen species generation upon light irradiation. Using Porphyromonas gingivalis as a representative pathogen, the nanoprobe demonstrated effective antibacterial activity and significantly alleviated experimental periodontitis in a rat model. Compared with conventional antibiotic-based therapies, this strategy reduces the risk of drug resistance and minimizes nonspecific disturbance to the oral microbiota by precisely targeting Gram-negative bacteria. This work highlights a materials-based, non-antibiotic approach for periodontal therapy with improved precision and microecological compatibility.

Graphical abstract

Highlights

• •A pathogen-class–specific targeting strategy tailored to the periodontal environment based on metabolic labeling and bioorthogonal chemistry.
• •Integration of AIE-enabled wash-free imaging with targeted photodynamic antibacterial therapy under physiologically relevant conditions.
• •A periodontal therapeutic approach that achieves antibacterial efficacy while limiting disruption to oral microbial homeostasis.

1. Introduction

Periodontitis is a chronic inflammatory disease that worsens with age and is highly prevalent worldwide [1,2]. It is characterized by progressive destruction of the tooth-supporting tissues and alveolar bone resorption, ultimately leading to tooth loosening or loss [3,4]. The disease is initiated by dysbiosis within the local periodontal niche and progressively develops under sustained amplification of a dysregulated host immune response, resulting in periodontal pocket formation and irreversible tissue damage [5,6]. Although broad-spectrum antibiotics are widely used clinically to reduce pathogen burden, such treatments are often accompanied by disruption of oral microbiota homeostasis and the imposition of continuous selective pressure, thereby accelerating the emergence and spread of antimicrobial resistance and leading to microbial imbalance, disease recurrence, and potential systemic adverse effects [[7], [8], [9]]. Therefore, achieving precise identification and selective elimination of Gram-negative periodontal pathogens while preserving the diversity and stability of the oral microbiota has become a critical challenge in the treatment of periodontal infections [[10], [11], [12]] (see Scheme 1).

Scheme 1.

Schematic illustration of the identification of gram-negative bacteria and efficient photodynamic antimicrobial therapy using AIE@PEG-DBCO.

Photodynamic therapy (PDT) inactivates bacteria through the generation of reactive oxygen species (ROS) by photosensitizers under light irradiation, offering advantages such as minimal invasiveness, spatiotemporal controllability, and a relatively low risk of resistance development [[13], [14], [15], [16]]. In addition, photosensitizers can be used for fluorescence-guided imaging. However, conventional photosensitizers are prone to aggregation-caused quenching (ACQ) under physiological conditions, leading to attenuated fluorescence signals and reduced ROS generation efficiency, which limits their application in complex biological environments [17,18]. Aggregation-induced emission luminogens (AIEgens) exhibit fluorescence activation in the aggregated state, effectively overcoming ACQ and providing wash-free, high-contrast imaging together with efficient ROS generation [19,20]. As such, they are considered promising candidates for constructing image-guided antimicrobial PDT systems.

Nevertheless, it should be noted that most existing AIE-based antibacterial photodynamic systems still rely primarily on cationic or hydrophobic interactions to achieve bacterial accumulation and activity [21]. Such targeting strategies are essentially based on nonspecific physicochemical adsorption and often lack pathogen-associated selectivity [22,23]. In complex microecological environments, these approaches may introduce cytotoxicity and cause collateral damage to commensal microbial communities [24,25]. More importantly, although AIE materials have been extensively investigated in the context of antimicrobial photodynamic therapy, systematic analyses addressing whether therapeutic interventions induce global perturbations of oral microecological structure remain relatively limited [26]. This research gap is particularly pronounced in periodontal infections, which are highly dependent on microecological homeostasis [27,28].

Against this background, there is a pressing need for strategies that move beyond nonspecific physicochemical targeting and instead achieve precise recognition based on intrinsic biological features of pathogens, thereby maintaining antimicrobial efficacy while minimizing potential disruption to microecological stability [29,30]. Metabolic glycoengineering introduces unnatural monosaccharides into living cells and provides bioorthogonal chemical handles with minimal functional perturbation, offering a new approach for precise targeting based on pathogen-specific molecular structures [31,32]. In Gram-negative bacteria, lipopolysaccharide (LPS) is a key structural component of the outer membrane and is closely associated with bacterial pathogenicity [33]. The azide-modified sugar KDO-N₃ can be metabolically incorporated into the inner core of LPS, leading to the presentation of azide groups on the bacterial surface, which can subsequently be selectively labeled via copper-free bioorthogonal click reactions [34,35]. This strategy provides a pathogen-class–oriented targeting route that is orthogonal to host tissues and fundamentally distinct from simple electrostatic adsorption.

Based on these considerations, we constructed a multifunctional nanoprobe, AIE@PEG-DBCO, by encapsulating an AIE photosensitizer within 1,2-distearoyl-sn-glycero-3-phosphoethanolamine-polyethylene glycol 2000 (DSPE-PEG₂₀₀₀) functionalized with surface dibenzocyclooctyne (DBCO) groups. The hydrophilic PEG shell endows the system with favorable aqueous stability and biocompatibility, while the surface DBCO moieties enable rapid, copper-free bioorthogonal click reactions with KDO-N₃–metabolically labeled Gram-negative bacteria, thereby achieving selective binding. On this basis, we propose and test the following hypothesis: by leveraging KDO-N₃/DBCO bioorthogonal chemistry, AIE@PEG-DBCO can achieve targeted recognition and wash-free, high-contrast fluorescence imaging of periodontal Gram-negative pathogens, represented by Porphyromonas gingivalis, and maintain efficient ROS generation in the aggregated state under physiologically relevant conditions to enable photodynamic eradication. Furthermore, by enabling precise pathogen regulation with a relatively low ecological cost, this strategy is expected to minimize nonspecific disturbance to the oral microbiota and provide a microecologically compatible therapeutic paradigm for the management of periodontal infections (Scheme 1).

2. Results and discussion

2.1. KDO-N₃-specific metabolic labeling of P. gingivalis

Given that metabolic engineering studies targeting oral pathogenic bacteria remain relatively limited, this study first systematically evaluated the specificity of KDO-N₃ metabolic labeling across different types of oral-related bacteria. In pursuit of this objective, gram-negative P. gingivalis and E.coli, gram-positive S.mutans, and S.aureus were pretreated with KDO-N₃. Among these bacteria, E. coli and S. aureus are classic model bacteria for bacterial metabolic labeling studies [36]. Subsequently, a copper-free click chemistry approach was employed using Cy5-DBCO as the labeling agent. Results revealed that E. coli and P. gingivalis exhibited robust red fluorescence, while S. aureus and S. mutans displayed negligible red fluorescence signals from Cy5-DBCO (Fig. S1A). These findings strongly indicate the successful integration of KDO-N₃ into the lipopolysaccharide (LPS) of P. gingivalis but not S. mutans. This observation aligns with prior outcomes related to KDO-N₃ metabolic labeling of gram-negative bacteria.

Subsequent flow cytometric analysis revealed a pronounced elevation in the fluorescent signal of Cy5-DBCO exclusively within the P. gingivalis group compared with that from S. mutans (Fig. S1B). Notably, the intensity of the fluorescent signal exhibited a positive correlation with the KDO-N₃ incubation concentration (Fig. S1C).

2.2. Synthesis and characterization of AIE@PEG-DBCO

Building on the efficient and selective KDO-N₃ labeling of P. gingivalis, we further explored the feasibility of substituting Cy5-DBCO with a DBCO-functionalized antibacterial system to achieve targeted imaging and antibacterial activity.

The synthesis of AIE@PEG-DBCO is illustrated in Fig. 1A. Briefly, the AIE photosensitizer and DSPE-PEG₂₀₀₀ self-assemble to form a hydrophobic core, while PEG and PEG-DBCO constitute a hydrophilic shell with DBCO groups exposed to the aqueous phase. The AIEgen employed in this study is DPA-SCP, whose synthetic route has been reported previously [37], its molecular structure is shown in Fig. 1B. The encapsulation efficiency of AIE@PEG-DBCO was calculated as 88.61 ± 2.04% based on the DPA-SCP standard curve (Fig. S2). AIE@PEG-DBCO had a diameter of 80.35 nm and displayed excellent dispersion in an aqueous solution (Fig. 1C). TEM analysis revealed uniformly distributed spherical morphology with a consistent size distribution (Fig. 1D).

Fig. 1.

Synthesis and characterization of AIE@PEG-DBCO. (A)Schematic illustration of the preparation of AIE@PEG-DBCO nanoprobes by encapsulating DPA-SCP with DSPE-PEG₂₀₀₀-DBCO via ultrasonication. (B) The structural formula of DPA-SCP. (C) Particle size distribution of AIE@PEG-DBCO. Inset: photograph of AIE@PEG-DBCO in aqueous solution. (D) TEM image of AIE@PEG-DBCO. (E) Changes in particle size and PDI of AIE@PEG-DBCO were kept in a phosphate-buffered saline (PBS) solution for seven days. (F) UV–vis absorption spectra of AIE@PEG-DBCO and its components. (G) Fluorescence emission spectra of AIE@PEG-DBCO. (H) Time-dependent changes in the UV–vis absorption spectra of ABDA (100 μM) incubated with AIE@PEG-DBCO (20 μg mL⁻¹) under light irradiation at an intensity of 50 mW cm⁻², used to evaluate ROS generation. (I) Detection of ¹O₂ generation by AIE@PEG-DBCO using the SOSG probe under light irradiation (100 mW cm⁻²). (J) Quantitative comparison of fluorescence intensity (λ_em = 525 nm) after light irradiation in the presence of SOSG for AIE@PEG-DBCO, KDO-N₃ + AIE@PEG-DBCO + P. gingivalis, hematoporphyrin, and PBS groups, to evaluate ¹O₂ generation. (K) Changes in photoluminescence (PL) intensity of different groups during continuous light irradiation for 10 min at a power density of 100 mW cm⁻². (L) Fluorescence emission spectra of AIE@PEG-DBCO at different water volume fractions. (M) Fluorescence intensity of AIE@PEG-DBCO before and after binding with P. gingivalis via KDO-N₃ targeting. Data are presented as mean ± SD (n = 3). Statistical significance was determined using one-way ANOVA followed by Tukey's multiple comparisons test. Exact p values are shown for statistically significant comparisons only.

Over 7 consecutive days of storage at 4 °C while being shielded from light, no noteworthy alterations in particle size or polydispersity index (PDI) of AIE@PEG-DBCO were observed (Fig. 1E).

Through a comparative analysis of the UV–vis absorption spectra of AIE@PEG-DBCO, AIE@PEG, DSPE-PEG₂₀₀₀-DBCO, and AIEgen, it was evident that the UV–vis absorption profile of AIE@PEG-DBCO encompassed simultaneously the distinct absorption peak characteristics of both DSPE-PEG-DBCO and AIEgen (Fig. 1F). This observation confirms the successful encapsulation of AIEgen by AIE@PEG-DBCO. The fluorescence emission profiles reveal that AIE@PEG-DBCO nanoparticles exhibit a maximum emission wavelength of 590 nm (Fig. 1G) and a substantial Stokes shift, underscoring their promising potential for bioimaging applications.

The reactive oxygen species (ROS) generation capability of AIE@PEG-DBCO was first evaluated using 9,10-anthracenediyl-bis(methylene)dipropanedioic acid (ABDA) as a chemical probe [38]. In the absence of AIE@PEG-DBCO nanoparticles, the absorbance of ABDA remained essentially unchanged under continuous light irradiation (50 mW cm⁻²) for 360 s. Upon introduction of AIE@PEG-DBCO nanoparticles, however, the absorbance of ABDA decreased progressively with increasing irradiation time, indicating effective light-triggered ROS generation (Fig. 1H and Fig. S3).

Next, we employed singlet oxygen sensor green (SOSG) to detect specifically the production of singlet oxygen (¹O₂) to assess the ROS generation capability of AIE@PEG-DBCO comprehensively (Fig. 1I). Remarkably, the fluorescence emission intensity of the AIE@PEG-DBCO group appreciably surpassed that of hematoporphyrin at equivalent concentrations, underscoring the superior ¹O₂ production capacity of AIE@PEG-DBCO compared with hematoporphyrin. Furthermore, the KDO-N₃+AIE@PEG-DBCO + P.gingivalis group exhibited exceptional ¹O₂ production potential (Fig. 1J and Fig. S4). This observation suggests that AIE@PEG-DBCO demonstrates a noteworthy ROS-generating ability in practical antimicrobial applications, a crucial attribute for subsequent investigations into its antimicrobial photodynamic therapy efficacy.

The photostability of AIE@PEG-DBCO under continuous white light irradiation was subsequently evaluated. After 10 min of irradiation, the fluorescence intensity of AIE@PEG-DBCO decreased by only 13.25% ± 1.19%, whereas hematoporphyrin exhibited a pronounced decrease of 40.03% ± 1.49% under identical conditions (Fig. 1K and Fig. S5), demonstrating the superior photostability of AIE@PEG-DBCO.

DPA-SCP is a hydrophobic AIEgen with typical aggregation-induced emission characteristics (Fig. 1L). Encapsulation with amphiphilic polymers (DSPE-mPEG₂₀₀₀ and DSPE-PEG₂₀₀₀-DBCO) yields the water-dispersible AIE@PEG-DBCO nanoparticles, in which intramolecular rotations and vibrations dominate in the dispersed state, resulting in weak fluorescence emission in aqueous solution. Upon bioorthogonal click conjugation with KDO-N₃–labeled P. gingivalis, AIE@PEG-DBCO becomes anchored to the bacterial LPS, imposing spatial constraints on intramolecular motion of the AIEgen and leading to a pronounced fluorescence enhancement (Fig. 1M). This fluorescence “turn-on” behavior enables subsequent bacterial identification by imaging [39].

The biosafety of AIE@PEG-DBCO was assessed by hemolysis assay, CCK-8 assay, and live/dead staining. DSPE-mPEG₂₀₀₀ and DSPE-PEG₂₀₀₀-DBCO induced negligible hemolysis within the tested concentration range, indicating good hemocompatibility of the PEGylated lipid components (Fig. 2A and B). In contrast, free AIEgen showed obvious concentration-dependent hemolysis (Fig. 2C). After encapsulation, AIE@PEG-DBCO exhibited markedly reduced hemolytic activity, suggesting that the PEGylated lipid shell effectively limited direct interaction between the hydrophobic AIEgen and erythrocyte membranes (Fig. 2D).

Fig. 2.

Biocompatibility evaluation of AIE@PEG-DBCO. (A–D) Hemolysis assays of DSPE-mPEG₂₀₀₀ (A), DSPE-PEG₂₀₀₀-DBCO (B), free AIEgen (C), and AIE@PEG-DBCO (D) at different concentrations. PBS and H₂O were used as negative and positive controls, respectively. (E, F) Cell viability of L929 cells and human gingival fibroblasts (HGFs) after treatment with AIE@PEG-DBCO for 24 h (E) and 48 h (F), as determined by CCK-8 assays. (G) Live/dead staining images of L929 cells and HGFs after treatment with different concentrations of AIE@PEG-DBCO for 48 h. Green fluorescence indicates live cells. Scale bars: 500 μm. Data are presented as mean ± SD (n = 5). Statistical significance was determined using one-way ANOVA followed by Tukey's multiple comparisons test. Exact p values are shown for statistically significant comparisons only. (For interpretation of the references to colour in this figure legend, the reader is referred to the Web version of this article.)

CCK-8 assays further showed that AIE@PEG-DBCO had no obvious cytotoxicity toward L929 cells or human gingival fibroblasts within the tested concentrations (Fig. 2E and F).In addition, considering that KDO-N₃ was used as a metabolic labeling reagent in the targeting strategy, its cytocompatibility was also evaluated by 48 h CCK-8 assays, which showed no obvious cytotoxicity in both L929 cells and human gingival fibroblasts (Fig. S6). Consistently, live/dead staining showed predominantly viable cells after treatment (Fig. 2G). These results indicate that AIE@PEG-DBCO has favorable hemocompatibility and cytocompatibility, supporting its further evaluation as a locally administered photodynamic platform for periodontitis.

2.3. AIE@PEG-DBCO selective targeting for bacterial imaging

DBCO-functionalized nanoprobes can be conjugated with azide groups on the outer membrane of Gram-negative bacteria through bioorthogonal copper-free click chemistry, thereby enabling targeted imaging of Gram-negative bacteria (Fig. 3A). In this study, KDO-N₃-pretreated Gram-negative P. gingivalis and Gram-positive S. mutans were separately incubated with AIE@PEG-DBCO at 37 °C, followed by immediate confocal laser scanning microscopy (CLSM) imaging without washing.

Fig. 3.

Bioorthogonal labeling enables selective imaging of gram-negative bacteria with AIE@PEG-DBCO. (A) Schematic: after metabolic incorporation of KDO-N₃ into the inner core of gram-negative bacterial LPS, azide groups are exposed on the surface; the DBCO on AIE@PEG-DBCO undergoes copper-free click reaction with these azides, enabling selective binding and fluorescence “turn-on” (B) After KDO-N₃ pretreatment and reaction with AIE@PEG-DBCO, P. gingivalis shows strong red fluorescence, whereas S. mutans exhibits almost no signal. (C) Specificity controls for the bioorthogonal click reaction (fluorescence microscopy): C1 AIE@PEG-DBCO only; C2 KDO-N₃ only; C3 AIE@PEG-DBCO + KDO-N₃; C4 KDO-N₃ + excess DSPE-PEG₂₀₀₀-DBCO for pre-blocking, followed by AIE@PEG-DBCO. (D) Corresponding TEM controls: D1 AIE@PEG-DBCO only; D2 KDO-N₃ only; D3 AIE@PEG-DBCO + KDO-N₃; D4 KDO-N₃ + excess DSPE-PEG₂₀₀₀-DBCO for pre-blocking, followed by AIE@PEG-DBCO. (For interpretation of the references to colour in this figure legend, the reader is referred to the Web version of this article.)

As shown in Fig. 3B, after 10 min of incubation, P. gingivalis displayed obvious red fluorescence signals on the bacterial surface, mainly distributed along the outer membrane and clearly outlining the bacterial morphology. In contrast, S. mutans showed very weak fluorescence signals, close to the background level (Fig. S7A). These results indicate that AIE@PEG-DBCO can selectively bind to KDO-N₃-labeled Gram-negative bacteria and trigger aggregation-induced emission after binding, allowing rapid, high-contrast imaging without washing.

To evaluate the applicability of this strategy among additional oral-related bacterial strains, we further examined another Gram-negative bacterium, F. nucleatum, as well as two Gram-positive bacteria, E. faecalis and S. gordonii (Fig. S8). Similar to P. gingivalis, KDO-N₃-labeled F. nucleatum exhibited clear red fluorescence signals distributed along the bacterial contour. In contrast, E. faecalis and S. gordonii showed only weak signals close to the background level. These findings further support the preferential labeling of Gram-negative bacteria by AIE@PEG-DBCO and indicate limited nonspecific binding to Gram-positive bacteria, supporting its potential for selective targeted imaging in the oral microbial context.

To further elucidate the interaction mechanism between AIE@PEG-DBCO and the bacterial surface, we performed a series of specificity control experiments (Fig. 3C and Fig. S7B). In the absence of metabolic labeling (C1, AIE@PEG-DBCO only), almost no fluorescence signal was observed, indicating that the probe exhibits low nonspecific adsorption to the bacterial surface. Likewise, treatment with KDO-N₃ alone (C2) did not generate any fluorescence signal. In contrast, when bacteria were metabolically modified with KDO-N₃ and subsequently incubated with AIE@PEG-DBCO (C3), a pronounced fluorescence outlining the bacterial morphology was observed, confirming that KDO-N₃ provides effective binding sites for the copper-free click reaction. Furthermore, competitive blocking with excess DSPE-PEG-DBCO (C4) markedly reduced the fluorescence intensity, indicating that the probe binding relies on the specific azide–DBCO bioorthogonal reaction rather than nonspecific interactions.

TEM analysis further supported click reaction–mediated probe binding (Fig. 3D). In control groups treated with AIE@PEG-DBCO alone or KDO-N₃ alone (D1, D2), the surface of P. gingivalis remained smooth with no obvious attached structures. In contrast, incubation of AIE@PEG-DBCO with KDO-N₃–treated bacteria (D3) resulted in visible probe-associated attachments on the bacterial cell surface, indicating successful localization of the probe to the outer membrane via the click reaction. This surface attachment was markedly reduced in the DSPE-PEG-DBCO pre-blocking group (D4), further confirming that probe binding depends on the azide–DBCO bioorthogonal reaction.

The above results indicate that the introduction of azide groups onto periodontal Gram-negative pathogens through metabolic glycoengineering enables AIE@PEG-DBCO to achieve selective recognition and imaging of P.gingivalis. Unlike most AIE-based antibacterial probes that rely on electrostatic attraction or hydrophobic interactions for bacterial enrichment, the targeting behavior of the present system primarily originates from the metabolic incorporation of KDO-N₃ into the LPS of P. gingivalis and the subsequent bioorthogonal click reaction with DBCO, thereby providing a chemically defined recognition mode directed toward periodontal pathogens [40].

This reaction-driven binding mode helps reduce the nonspecific membrane interactions commonly associated with cationic AIE molecules, allowing relatively stable probe localization on the outer membrane of P. gingivalis and accompanying aggregation-induced fluorescence enhancement, which supports wash-free imaging. In addition, competitive blocking experiments indicate that the probe binding depends on the azide–DBCO bioorthogonal reaction rather than simple physical adsorption, suggesting that the targeting process is suppressible and chemically well defined. Nevertheless, direct nanoscale co-localization with specific outer-membrane markers was not performed in the present study. Therefore, although the current data support bacteria-associated bioorthogonal binding, they do not fully resolve the nanoscale spatial distribution of AIE@PEG-DBCO or completely exclude possible localized colloidal aggregation or precipitation. Future studies combining membrane-specific fluorescent labeling with high-resolution confocal or super-resolution microscopy will be valuable for further clarifying the spatial distribution of AIE@PEG-DBCO and distinguishing membrane-associated bioorthogonal binding from possible localized nanoprobe aggregation [41,42]. Overall, these results support the feasibility of the KDO-N₃/DBCO bioorthogonal strategy for selective imaging of periodontal pathogens using AIE-based photosensitizers and provide experimental evidence for further investigation of related applications.

2.4. Targeted antimicrobial performance of AIE@PEG-DBCO in vitro

Based on the specific and efficient targeting of KDO-N₃–labeled gram-negative bacteria achieved by AIE@PEG-DBCO, we next examined whether such selective binding could be translated into functional antimicrobial activity (Fig. 4A). To this end, we first evaluated the antibacterial efficacy of AIE@PEG-DBCO using SYTO9/PI bacterial live/dead staining before conducting subsequent assays (Fig. S9).

Fig. 4.

Targeted antimicrobial performance of AIE@PEG-DBCO in vitro. (A) Schematic of KDO-N₃–mediated metabolic labeling of gram-negative bacteria and bioorthogonal DBCO–azide targeting by AIE@PEG-DBCO, followed by 450 nm light–induced ROS generation, NPS denotes AIE@PEG-DBCO nanoparticles throughout this figure. CLSM images of SYTO9/PI of P. gingivalis (B) and S. mutans (C) under different experimental protocols. Green fluorescence indicates live bacteria stained with SYTO9, while red fluorescence indicates dead bacteria stained with PI. (D) Photographs of P. gingivalis and S. mutans colony formation under different experimental protocols. Quantification of live/dead P. gingivalis (E) and S. mutans (F) from SYTO9/PI staining. Quantitative analysis of CFUs of P. gingivalis (G) and S. mutans (H) under different experimental protocols. Data are presented as mean ± SD (n = 3). Statistical significance was determined using one-way ANOVA followed by Tukey's multiple comparisons test. Exact p values are shown for statistically significant comparisons only. (For interpretation of the references to colour in this figure legend, the reader is referred to the Web version of this article.)

In the case of P. gingivalis, a substantial population of bacteria exhibited red staining indicative of PI uptake in the KDO-N₃+NPS + Light group, while green fluorescence predominated in the other groups (Fig. 4B and E). Conversely, for S. mutans, none of the groups displayed a significant presence of bacteria stained red with PI; instead, green fluorescence dominated (Fig. 4C and F). This observation underscores that KDO-N₃+NPS + Light induced substantial bactericidal effects exclusively against P. gingivalis, with no significant impact on S. mutans.

Next, we conducted an in-depth evaluation of the antibacterial potential of AIE@PEG-DBCO against gram-negative bacteria using a colony-forming unit (CFU) counting method (Fig. 4D). Across all subgroups, AIE@PEG-DBCO exhibited minimal toxicity in the dark. After a 10-min exposure to white light, the bacterial survival rate in the KDO-N₃+NPS + Light group was a mere 0.17 ± 0.02% within the P. gingivalis group (Fig. 4G), and the bacterial survival rate of S. mutans under the same conditions was 85.09 ± 9.65% (Fig. 4H).

Lower killing efficiencies of photosensitizers against gram-negative bacteria compared with gram-positive bacteria have been reported. This discrepancy arises from the distinctive outer membrane structure of gram-negative bacteria [43]. However, in the present study, the DBCO groups of AIE@PEG-DBCO exhibited a specific and efficient binding affinity for the azide groups present on the outer membranes of gram-negative bacteria through click chemistry. The distinctive interaction of AIE@PEG-DBCO enables it to attach selectively to the surfaces of gram-negative bacteria, enhancing targeted antimicrobial activity against them compared with gram-positive bacteria.

2.5. Antibacterial mechanism of AIE@PEG-DBCO

TEM was used to observe the morphological changes of bacteria under different treatment conditions. After exposure to light, P. gingivalis in the AIE@PEG-DBCO group showed membrane rupture, characterized by contraction, deformation, and concave shape, while the bacterial envelope remained intact in the control group (Fig. 5A). There were no noticeable changes in the morphology of S. mutans in any of the groups (Fig. 5B).

Fig. 5.

Antibacterial mechanism of AIE@PEG-DBCO. TEM images of P. gingivalis (A) and S. mutans (B); (C) Representative DCFH-DA fluorescence images showing intracellular ROS-related fluorescence in P. gingivalis after different treatments under dark or light conditions. (D) Representative DiBAC₄(3) fluorescence images for evaluating membrane depolarization in P. gingivalis after different treatments under dark or light conditions. Increased green fluorescence indicates enhanced DiBAC₄(3) accumulation associated with bacterial membrane depolarization. (E) Extracellular protein leakage from P. gingivalis determined by the bicinchoninic acid (BCA) assay. (F) Membrane permeability of P. gingivalis evaluated by the o-nitrophenyl-β-D-galactopyranoside (ONPG) hydrolysis assay. (G) Intracellular GSH content of P. gingivalis after different treatments. (H) Schematic illustration of the proposed photodynamic antibacterial mechanism. Data are presented as mean ± SD (n = 5). Statistical significance was determined using one-way ANOVA followed by Tukey's multiple comparisons test. Exact p values are shown for statistically significant comparisons only. (For interpretation of the references to colour in this figure legend, the reader is referred to the Web version of this article.)

Subsequently, the intracellular ROS content of bacteria in each group was quantified using DCFH-DA. The green fluorescence intensity in P. gingivalis coincubated with AIE@PEG-DBCO was notably higher than other groups upon light exposure (Fig. 5C and S10, S11). The successful enrichment of AIE@PEG-DBCO into the P. gingivalis outer membrane through click chemistry resulted in a significant increase in intracellular ROS production, leading to the effective elimination of P. gingivalis. Consistent with the elevated ROS level, changes in the membrane potential of P. gingivalis were further evaluated using DiBAC₄(3). Upon light irradiation, the AIE@PEG-DBCO-treated group exhibited markedly enhanced DiBAC₄(3) green fluorescence compared with the other groups (Fig. 5D and S12, S13). Since increased DiBAC₄(3) fluorescence is associated with bacterial membrane depolarization, these results indicate that AIE@PEG-DBCO-mediated photodynamic treatment induced pronounced disruption of the membrane potential of P. gingivalis. This membrane depolarization was consistent with the subsequent increase in membrane permeability, further supporting ROS-mediated membrane dysfunction as an important component of the antibacterial mechanism.

The tests for protein leakage and o-nitrophenyl-β-d-galactopyranoside (ONPG) revealed bacterial content leakage and altered membrane permeability. Specifically, treatment with AIE@PEG-DBCO + Light induced significant cytoplasmic leakage and membrane permeability changes in P. gingivalis (Fig. 5E, F, S14 and S15). GSH content analysis (Fig. 5G) showed that under light exposure, the GSH levels in P. gingivalis treated with AIE@PEG-DBCO were significantly reduced, indicating that the photodynamic therapy induced oxidative stress in the bacteria. In contrast, S. mutans did not show a significant change in GSH levels under the same treatment conditions (Fig. S16).

In summary, we propose that the enrichment of AIE@PEG-DBCO on the surface of Porphyromonas gingivalis via bioorthogonal click chemistry enables efficient generation of ROS upon light irradiation. The locally generated ROS increases membrane permeability, disrupts membrane integrity, and induces leakage of intracellular components, thereby compromising bacterial viability. Meanwhile, a significant decrease in intracellular glutathione levels was observed, indicating disruption of redox homeostasis and attenuation of the bacterial antioxidant defense. Consequently, the combined effects of membrane damage and oxidative stress are likely to contribute synergistically to bacterial inactivation, suggesting that AIE@PEG-DBCO exerts its antibacterial effect primarily through a ROS-mediated photodynamic mechanism (Fig. 5H).

2.6. Treatment in vivo of experimental periodontitis in rats by AIE@PEG-DBCO

Based on the selective recognition and photodynamic antibacterial activity of AIE@PEG-DBCO against P. gingivalis, we further established an experimental periodontitis model in rats by ligation combined with P. gingivalis infection, and evaluated its therapeutic efficacy in vivo (Fig. 6A and S17).

Fig. 6.

In vivo targeted photodynamic therapy of experimental periodontitis using AIE@PEG-DBCO. (A) Schematic illustration of the rat periodontitis model establishment and treatment schedule. (B) Fluorescence imaging of AIE@PEG-DBCO in frozen gingival tissue sections; blue indicates DAPI-stained nuclei, and red indicates AIE@PEG-DBCO. (C) DHE staining of frozen gingival tissue sections for detection of local ROS levels. (D) Representative CFU plates of P. gingivalis recovered from periodontal lesions after different treatments. (E, F) Representative micro-CT images of the maxillary molar region, including buccal views (E) and sagittal sections (F). (G) Quantitative analysis of the DHE-positive fluorescence area in gingival tissue sections. (H) Quantitative analysis of P. gingivalis viability based on CFU counting. (I, J) Quantification of the distance from the cementoenamel junction to the alveolar bone crest (CEJ–ABC) based on micro-CT images, including buccal views (I) and sagittal sections (J). Data are presented as mean ± SD (n = 6). Statistical significance was determined using one-way ANOVA followed by Tukey's multiple comparisons test. Exact p values are shown for statistically significant comparisons only. (For interpretation of the references to colour in this figure legend, the reader is referred to the Web version of this article.)

First, to examine the distribution of AIE@PEG-DBCO within periodontal lesions, fluorescence imaging was performed on frozen gingival sections (Fig. 6B). A clear red fluorescence signal was observed in the lesion area, whereas only weak signals were detected in the control groups, suggesting local enrichment of AIE@PEG-DBCO at the infected site. Together with the in vitro results, this enrichment may be attributed to the bioorthogonal click reaction between azide groups on the outer membrane of KDO-N₃-labeled Gram-negative bacteria and DBCO groups on the nanoprobe surface. The localized fluorescence signal in the lesion area not only reflects the tissue distribution of AIE@PEG-DBCO, but also supports its feasibility as an in vivo targeted imaging probe.

We next evaluated local ROS generation. DHE staining showed that ROS-related fluorescence signals were markedly enhanced in gingival tissues from the AIE@PEG-DBCO light-treated group, and quantitative analysis further confirmed this increase (Fig. 6C and G). These results indicate that AIE@PEG-DBCO, after local enrichment in periodontal lesions, can induce oxidative stress upon light irradiation, providing mechanistic support for its in vivo photodynamic antibacterial effect.

The in vivo antibacterial effect was assessed by counting P. gingivalis CFUs recovered from periodontal lesions. Compared with the periodontitis group, which showed abundant colonies, the AIE@PEG-DBCO light-treated group exhibited a clear reduction in colony numbers, with quantitative analysis showing a significant decrease in P. gingivalis viability (Fig. 6D and H). This result indicates that AIE@PEG-DBCO can effectively reduce the burden of the target pathogen in periodontal lesions in vivo.

Alveolar bone resorption is a major pathological feature of periodontitis progression. Micro-CT analysis showed evident alveolar bone loss and an increased CEJ–ABC distance in the periodontitis group. In contrast, AIE@PEG-DBCO-mediated photodynamic treatment markedly attenuated alveolar bone destruction, with significantly reduced CEJ–ABC distances in both buccal views and sagittal sections (Fig. 6E, F, I, J). These findings suggest that this system not only reduces pathogen burden, but also alleviates periodontitis-associated alveolar bone loss.

Overall, AIE@PEG-DBCO achieved local enrichment and targeted imaging in periodontal lesions, induced ROS generation after light irradiation, reduced P. gingivalis burden, and attenuated alveolar bone resorption. These results support the potential of AIE@PEG-DBCO as a local targeted photodynamic strategy against Gram-negative periodontal pathogens.

To evaluate the effects of AIE@PEG-DBCO-mediated photodynamic therapy on periodontal inflammation and tissue repair, histological and immunofluorescence analyses were performed on gingival tissues from different treatment groups. H&E staining showed intact epithelial structure and minimal inflammatory infiltration in healthy tissues, whereas the periodontitis group exhibited epithelial thickening, elongated rete ridges, disordered basal cell arrangement, and increased inflammatory cells, indicating pronounced local inflammation. AIE@PEG-DBCO treatment markedly improved tissue morphology, restoring epithelial integrity and reducing inflammatory infiltration, with the lamina propria more organized than in the periodontitis group. The minocycline group showed partial improvement, whereas the free AIEgen group showed limited changes (Fig. 7A and Fig. S18).

Fig. 7.

Histological, inflammatory, and tissue repair evaluation after AIE@PEG-DBCO-mediated photodynamic therapy in experimental periodontitis. (A, B) Representative H&E staining (A) and Masson's trichrome staining (B) images of gingival tissues from different treatment groups. (C–E) Immunofluorescence staining of TNF-α (C), IL-6 (D), and IL-10 (E) in gingival tissues. Blue indicates DAPI-stained nuclei, and red indicates the corresponding cytokine signal. (F) Representative immunofluorescence images of CD31 and α-SMA staining in gingival tissues, including merged images and individual α-SMA and CD31 channels. Blue indicates DAPI-stained nuclei, red indicates α-SMA, and green indicates CD31. (G) Quantitative analysis of collagen volume fraction (CVF) based on Masson's trichrome staining. (H–J) Quantitative analysis of TNF-α-positive area (H), IL-6-positive area (I), and IL-10-positive area (J). (K, L) Quantitative analysis of α-SMA-positive area (K) and CD31-positive area (L). (M) Radar plot summarizing the overall therapeutic performance of different treatments in terms of targeted antibacterial activity, collagen restoration, TNF-α suppression, IL-10 expression, and CD31 expression. Data are presented as mean ± SD (n = 6). Statistical significance was determined using one-way ANOVA followed by Tukey's multiple comparisons test. Exact p values are shown for statistically significant comparisons only. (For interpretation of the references to colour in this figure legend, the reader is referred to the Web version of this article.)

Masson's trichrome staining further showed loose and discontinuous collagen fibers in the periodontitis group, suggesting damage to the periodontal connective tissue matrix. In contrast, AIE@PEG-DBCO treatment enhanced the blue collagen signal, with denser and more continuous collagen fiber arrangement. Quantitative analysis of collagen volume fraction (CVF) showed that the AIE@PEG-DBCO group was significantly higher than the periodontitis group and approached the level of the healthy group (Fig. 7B and G), indicating improved restoration of the periodontal connective tissue matrix.

We next examined the expression of the pro-inflammatory cytokines TNF-α and IL-6, as well as the anti-inflammatory cytokine IL-10. TNF-α and IL-6 signals were markedly increased in the periodontitis group, whereas AIE@PEG-DBCO treatment significantly reduced their positive areas, indicating suppression of the local pro-inflammatory response (Fig. 7C, D, H, I). Meanwhile, IL-10 was clearly upregulated in the AIE@PEG-DBCO group (Fig. 7E and J), suggesting a possible shift of the local inflammatory microenvironment toward an anti-inflammatory and repair-associated state.

For tissue repair assessment, CD31 and α-SMA were used to reflect endothelial formation and vascular maturation-related processes, respectively. Both signals were relatively weak in the periodontitis group, whereas AIE@PEG-DBCO treatment markedly increased their positive signals, with quantitative analysis confirming a significant increase in positive areas (Fig. 7F, K, and L). These results suggest that the treatment may promote local angiogenesis and vascular-associated repair.

Taken together, the AIE@PEG-DBCO group showed clear improvements over the periodontitis group in antibacterial activity, inflammatory regulation, collagen restoration, and vascular repair (Fig. 7M). Overall, AIE@PEG-DBCO-mediated photodynamic therapy alleviated local inflammation after reducing periodontal pathogen burden, while promoting collagen matrix reconstruction and vascular-associated repair, providing histological and immunological evidence for its therapeutic effect in experimental periodontitis.

After treatment, the in vivo biosafety of the different therapeutic regimens was further evaluated. Serum biochemical analysis showed no obvious abnormalities in liver function markers, including ALT, AST, ALP, TP, and ALB, or kidney function markers, including CREA, UA, and UREA, among the different groups (Fig. 8A–H), suggesting that AIE@PEG-DBCO-mediated photodynamic therapy did not cause apparent hepatic or renal dysfunction. In addition, hematological analysis showed no significant treatment-related alterations in major blood parameters, including WBC, RBC, HCT, HGB, MCH, PCT, PDW, and PLT (Fig. 8I–P), further indicating the absence of obvious hematological toxicity following repeated local treatment.

Fig. 8.

In vivo biosafety evaluation after different treatments. (A–H) Serum biochemical parameters, including ALT, AST, ALP, TP, ALB, CREA, UA, and UREA. (I–P) Hematological parameters, including WBC, RBC, HCT, HGB, MCH, PCT, PDW, and PLT. (Q) Representative H&E-stained images of the heart, liver, spleen, lung, and kidney. Data are presented as mean ± SD (n = 6).

H&E staining was also performed on major organs, including the heart, liver, spleen, lung, and kidney. The results showed intact tissue architecture in all treatment groups, with no obvious inflammatory infiltration, tissue necrosis, or structural damage (Fig. 8Q). Together, the serum biochemical, hematological, and histopathological results demonstrated no apparent treatment-related systemic toxicity under the tested conditions. These results indicate that AIE@PEG-DBCO exhibited good in vivo biosafety following repeated local administration, supporting its further application as a local photodynamic therapeutic system for periodontitis.

2.7. Impact of AIE@PEG-DBCO treatment on oral microbiota

To evaluate the impact of AIE@PEG-DBCO treatment on the oral microecological environment, 16S rRNA sequencing analysis was performed based on oral swab samples (Fig. 9A). At the genus level, the overall distribution of dominant taxa was comparable between the treatment group and healthy controls, with no evident replacement of major taxa or large-scale community restructuring, indicating that the treatment did not induce a detectable shift in overall community composition (Fig. 9B and C).

Fig. 9.

Evaluation of oral microbiota composition following treatment with the functionalized nanoprobe. (A) Schematic of the experimental design. SD rats were randomly assigned to a control group (Control) or a treatment group (AIE@PEG-DBCO, NPS). After establishment of the periodontitis model, AIE@PEG-DBCO was administered to the treatment group at weeks 1, 2, and 3. Oral swab samples were collected at week 4 for 16S rRNA sequencing. (B) Stacked bar plots showing the relative abundance of oral microbiota at the genus level in each group. (C) Heatmap showing the distribution of dominant genera at the genus level across samples. (D–F) Alpha-diversity indices (Chao1, Shannon index, and Simpson index) used to assess species richness and evenness of the oral microbiota. (G–H) Beta-diversity analysis based on principal component analysis (PCA) and principal coordinate analysis (PCoA), illustrating the overall community structure and sample clustering patterns between groups. (I) Schematic representation of the treatment strategy and its relationship with the oral microbial community structure, with no apparent alteration in overall microbiota composition. Data are presented as mean ± SD (n = 6). Statistical significance was determined using one-way ANOVA followed by Tukey's multiple comparisons test. ns, not significant.

Alpha-diversity analysis further showed that no significant differences were observed between the treatment and control groups in the Chao1, Shannon, or Simpson indices, suggesting that, under the experimental conditions of this study, AIE@PEG-DBCO treatment did not markedly affect the richness or evenness of the oral microbial community (Fig. 9D–F). Beta-diversity analysis using PCA and PCoA demonstrated substantial overlap between the two groups in multivariate space, indicating that the overall community structure remained stable (Fig. 9G and H). In addition, samples from the treatment group exhibited a relatively lower within-group dispersion, suggesting reduced community variability at the current observation time point (Fig. 9I).

Taken together, the microbiota analyses presented in Fig. 9 indicate that, within the defined experimental conditions and observation window, AIE@PEG-DBCO treatment did not lead to detectable restructuring of the overall oral microbial community or significant changes in microbial diversity. These findings suggest favorable microbiota compatibility and minimal ecological disturbance associated with the treatment. Combined with its targeted antibacterial efficacy and therapeutic performance, this strategy may offer a balanced approach for periodontal intervention that achieves pathogen control while limiting disruption to the overall oral microecological structure.

Considering the anatomical and physiological continuity between the oral cavity and the gastrointestinal tract, and the possibility of material ingestion during oral administration, fecal microbiota analysis was included as a complementary assessment [44]. No apparent systemic perturbation of the gut microbiota was observed following AIE@PEG-DBCO treatment (Fig. S19).

Recent advances in biomedical materials have increasingly focused on the integration of biological recognition, disease-related imaging, and therapeutic intervention within a single platform [[45], [46], [47]]. In particular, AIE-based photosensitizers have shown promise for fluorescence-guided photodynamic therapy, benefiting from their aggregation-enhanced emission and ROS-generating capacity [48]. In parallel, activatable fluorogenic probes, biosensing systems, and multifunctional nanoplatforms have been developed for disease detection, imaging, and treatment monitoring [[49], [50], [51]]. In this context, the AIE@PEG-DBCO platform developed in this study differs from conventional nonspecific antimicrobial photosensitizers by linking a defined bacterial glycometabolic feature with bioorthogonal chemistry. Specifically, KDO-N₃ metabolic labeling introduces azide groups into the LPS-containing outer membrane of Gram-negative bacteria, enabling DBCO-mediated recognition, AIE-assisted imaging, and localized photodynamic killing. This design therefore provides a disease-oriented antibacterial strategy that combines Gram-negative bacterial surface recognition, wash-free fluorescence imaging, and targeted photodynamic intervention in the periodontal microenvironment.

Although AIE@PEG-DBCO demonstrated favorable therapeutic efficacy in the treatment of periodontitis, several limitations of the present study should be acknowledged. First, when AIE@PEG-DBCO was administered into periodontal pockets, the precision and stability of local drug delivery were constrained by the technical challenges associated with small-animal handling. Future studies may address this limitation by incorporating delivery platforms such as hydrogels or microneedle patches to enhance local retention and improve therapeutic efficiency [52]. Second, the current system relies on 450 nm irradiation, which has limited tissue penetration compared with longer-wavelength light, particularly in deep periodontal pockets. For clinical periodontal applications, local light delivery may be achieved using periodontal pocket-compatible optical fibers, miniature LED probes, or dental light-delivery devices [53,54]. Nevertheless, further development of red- or near-infrared-activated AIE photosensitizers would be beneficial for improving applicability in deeper periodontal lesions [55]. Third, KDO-N₃ metabolic pre-labeling introduces an additional clinical step. Although local periodontal administration was used in this study to enrich KDO-N₃ at infected sites, future studies should further evaluate its labeling efficiency in complex multispecies oral biofilms, repeated-exposure safety, retention behavior, and patient compliance under clinically relevant treatment schedules. Moreover, although favorable biosafety was observed under the tested four-week repeated local treatment regimen, extended post-treatment follow-up and quantitative assessment of the in vivo fate of KDO-N₃ and AIE@PEG-DBCO were not performed. Future studies should therefore further investigate their biodistribution, metabolic fate, tissue retention, and systemic clearance to better define the long-term safety and translational potential of this bioorthogonal therapeutic system [56].

Compared with other antimicrobial material systems, such as silver-, gallium-, and silver–gallium-based platforms, AIE@PEG-DBCO has distinct advantages and limitations [57,58]. Metal-based antimicrobial materials often exhibit broad antibacterial activity through metal ion release, metal-associated stress, membrane disruption, or interference with bacterial metabolism, but their broad-spectrum effects may also raise concerns regarding pathogen selectivity, commensal microbial disturbance, and dose-dependent cytotoxicity [[59], [60], [61]]. In contrast, AIE@PEG-DBCO emphasizes chemically defined bacterial recognition through KDO-N₃/DBCO bioorthogonal chemistry and light-activated local photodynamic killing, which may help improve pathogen selectivity and reduce unnecessary disturbance to the oral microbiota. Nevertheless, this strategy is procedurally more complex because it requires metabolic pre-labeling and local light irradiation, and further optimization will be needed for future translation.

In addition, the primary focus of this study was to assess whether the therapeutic intervention induced broad alterations in the overall oral microecological environment; therefore, oral swab samples were used for oral microbiota analysis [62]. Nevertheless, the subgingival microbiota represents a critical ecological niche in the pathogenesis and progression of periodontal disease and remains of substantial research interest [63]. In the rat model, technical limitations related to the feasibility and reproducibility of subgingival plaque collection restricted the systematic evaluation of this niche in the present study. Future investigations will aim to further optimize sampling strategies and experimental conditions to enable a more comprehensive assessment of the effects of AIE@PEG-DBCO on the subgingival microbiota, particularly with respect to its differential modulation of pathogenic and commensal bacterial populations.

Overall, this study provides a new perspective for the development of non-antibiotic periodontal therapeutic strategies that balance antibacterial efficacy with microecological compatibility.

3. Conclusion

In the present study, we employed KDO-N₃ for the metabolic labeling of P. gingivalis, and AIEgen served as the core for the design of a nanoprobe (AIE@PEG-DBCO). This nanoprobe was developed to undergo bioorthogonal reactions with azide groups. The methodology established here allows for targeted imaging and photodynamic eradication of gram-negative bacteria, with P. gingivalis as a representative example, effectively mitigating experimental periodontitis in rats. Compared with conventional broad-spectrum antibiotic regimens, our approach avoids the development of drug resistance while preserving the balance of oral flora by selectively eliminating gram-negative pathogens. This approach shows potential for accurately treating periodontitis.

4. Experimental

Detailed experimental procedures are provided in the Supporting Information.

CRediT authorship contribution statement

Shisheng Cao: Writing – original draft, Methodology, Data curation, Conceptualization. Cailing Zhao: Visualization, Methodology, Investigation. Surong Guo: Visualization, Validation, Investigation. Jiashen Hu: Visualization, Validation. Qing Yang: Investigation. Xin Li: Data curation. Yiping Liu: Visualization. Dan Ding: Writing – review & editing, Supervision, Conceptualization. Huijuan Yin: Project administration, Funding acquisition, Conceptualization. Juan Zhang: Funding acquisition, Data curation, Conceptualization.

Declaration of competing interest

The authors declare that they have no known competing financial interests or personal relationships that could have appeared to influence the work reported in this paper.

Appendix A. Supplementary data

The following is the Supplementary data to this article:

Data availability

Data will be made available on request.