Steric confinement-induced emission probe for monitoring protein conformations in live cells

Abstract

Protein conformational changes drive signal transduction to regulate cellular activities, yet monitoring of these changes in living cells remains challenging. Here, we introduce BIOSCE (BIOprobe based on Steric Confinement-induced Emission), a technique that enables tracking of individual protein conformations in living cells across millisecond-to-minute timescales. BIOSCE reports protein conformational changes via steric confinement–induced luminescence switching from non-luminescent to luminescent states. We demonstrate that BIOSCE rapidly senses calmodulin conformational changes triggered by intracellular calcium fluctuations. The BIOSCE platform achieved millisecond-resolution monitoring of single-protein conformations within cellular signaling pathways, as evidenced by its sensitive detection of rapamycin-dependent FKBP (FK506-binding protein)-FRB (FKBP-rapamycin binding) interactions regardless of the labeled partner. Furthermore, we applied BIOSCE to track the spatial distribution of SNAP25 (25 kDa synaptosomal nerve-associated protein) during botulinum neurotoxin A (BoNT/A) intoxication, revealing differential catalytic processing of its cleavage fragments. This generalizable approach provides a robust platform for investigating single-molecule conformational changes with high spatiotemporal resolution and enables direct evaluation of transient cellular events.

Protein conformational changes are crucial for signal transduction in cells, yet tracking these changes in living cells remains challenging. Here, the authors develop BIOSCE (BIOprobe based on Steric Confinement-induced Emission) by binding a steric sensitive fluorophore (MTPABP-Cl) to a HaloTag, which uses steric confinement-induced luminescence to monitor protein conformations in living cells with millisecond resolution.

Introduction

Cellular activities rely on precisely coordinated protein–protein interactions, with conformational states across spatial and temporal scales driving nearly all biological processes, from initiation to modulation and termination of signaling pathways^1–5. For example, monitoring external stimuli-induced protein-protein interactions and conformational states^6,7, or the dynamic distribution of proteins, enables the tracking of cellular physiological activities. However, labeling and tracking the behavior of dynamic proteins in living cells remains challenging.

At present, the detection of calmodulin conformational changes, which reflect intracellular Ca²⁺ fluctuations, has proven to be a powerful strategy for monitoring calcium-dependent physiological processes. This calmodulin-inspired optogenetic and chemogenetic approach has led to the development of genetically encoded calcium indicators (GECIs), including optimized green calcium-modulated protein (GCaMP) series (GCaMP1 to jGCaMP8)^8,9, which exploit the precise coupling between calmodulin’s structural dynamics and calcium signaling^10,11. GECIs enable visualization of neural activity and other Ca²⁺-mediated cellular events, but their fundamental design principle exhibits limited transferability to other protein systems.

In addition, extensive advances in single-molecule FRET (smFRET) platforms have enabled real-time monitoring of nanometer-scale conformational dynamics, providing quantitative insights into folding intermediates, transition pathways, and dynamic heterogeneity⁷. It is achieved by detecting the exposure of a genetically encoded peptide tag strategically embedded within the target protein, enabling conformation-specific reporting. This method is theoretically adaptable to a wide range of protein structures, but its practical application remains constrained by the persistent challenge of identifying optimal tag insertion sites that ensure accurate reporting without disrupting native function. In addition, extensive advances in single-molecule FRET (smFRET) platforms have enabled real-time monitoring of nanometer-scale conformational dynamics, providing quantitative insights into folding intermediates, transition pathways, and dynamic heterogeneity^12–15. These smFRET-based studies further highlight both the power and the practical constraints of dual-fluorophore systems, motivating the development of single-label strategies such as HaloTag.

Circularly permuted HaloTag (cpHaloTag) scaffold offers a versatile platform for fluorescent biosensors by binding diverse fluorophores and modulating their emission in response to protein-protein interactions¹⁶. Building upon this scaffold, the split-HaloTag system represents a strategic evolution from static interaction detection to dynamic monitoring of biological processes^6,17,18. This platform operates through the interaction between an inactive fragment (cpHalo∆) and a small peptide (Hpep), with activation contingent upon specific biological events⁶. This platform functions as a sensitive molecular recorder, capturing dynamic protein interactions with high specificity¹⁹. However, its utility is limited by insufficient temporal resolution to resolve subsecond protein conformational changes in living cells⁶. Collectively, FRET-based biosensors and the split-HaloTag system are intrinsically limited in capturing the diverse temporal scales and spatial distributions of protein conformational changes during cellular activities^6,14,20. These limitations underscore the pressing need for a generalizable sensing platform that combines high spatiotemporal resolution, minimal invasiveness, and robust signal output.

In this study, we developed BIOSCE (BIOprobe based on Steric Confinement-induced Emission), a fluorogenic strategy based on HaloTag protein tags^{16–18,21–25}. The fundamental design principle of BIOSCE centers on a steric confinement-induced emission mechanism, implemented through a rationally designed small-molecule fluorophore, MTPABP-Cl. This fluorophore features molecular rotor units whose intramolecular rotation and thus fluorescence quantum yield are highly sensitive to changes in local spatial hindrance^26–29. When MTPABP-Cl is covalently bound to a HaloTag (HT) fused to a protein of interest, conformational changes of the target protein directly alter the steric confinement around the fluorophore. This modulates the freedom of intramolecular rotation, thereby transducing structural shifts into quantifiable fluorescence signals. This unique design permits not only the detection of activity-dependent conformational changes but also enables simultaneous tracking of subcellular localization and protein movement with high spatiotemporal resolution. The photostability of MTPABP-Cl supports extended live cell imaging, facilitating the observation of sustained biological events without signal attenuation. Our approach provides a scalable method adaptable for recording diverse biological activities, expanding the toolbox for visualizing conformational changes in cellular events.

Results

Design strategy of BIOSCE

As a starting point for monitoring protein conformational changes in living cells, we used the self-labeling protein HaloTag (HT) as a bridge that covalently binds to specific small-molecule ligands, enabling targeted protein labeling and tracking. To facilitate imaging in living cells, a steric confinement-induced emission (SCE) dye was selected as the fluorescent ligand for the HaloTag (HT) in our molecular design. Specifically, 4,4′-dimethoxytriphenylamine was chosen as the luminescent donor unit, while 2,1,3-benzothiadiazole served as the acceptor. The 2,1,3-benzothiadiazole unit was covalently linked to 1-bromo-6-chlorohexane, forming the HT ligand dye, MTPABP-Cl (Supplementary Scheme 1). This structural arrangement ensures that the molecule remains weakly emissive in its free state. However, upon binding to HaloTag, steric confinement is induced, which restricts intramolecular motions (RIM), leading to an increase in quantum yield. The structure of the MTPABP-Cl was confirmed through nuclear magnetic resonance (NMR) spectra and high-resolution mass spectrometry (Supplementary Fig. 1–6).

We employed MTPABP-Cl as a fluorescent probe for imaging protein behaviors, and named the method “BIOSCE” (Fig. 1a). To assess the reliability of the method, we systematically characterized the photophysical properties of MTPABP-Cl. It exhibited strong absorption at 460 nm, and its emission wavelength peaked at 660 nm (Fig. 1b, Supplementary Data 1). Subsequently, the properties of configuration were investigated in DMSO/H₂O mixtures. As illustrated, the emission intensity of the configuration was enhanced with the increasing fraction of water (Fig. 1c, d, Supplementary Data 1). In addition, the photoluminescence quantum yield (PLQY) of MTPABP-Cl increases upon HT binding (27.5%) and rises further when the ligand is accommodated within HT-fusion proteins (38.6%) that impose a more compact local pocket, thereby directly reporting differences in steric confinement (Supplementary Fig. 7, Supplementary Data 2), demonstrating steric confinement-induced emission characteristics. In contrast, Si-rhodamine, JF₆₃₅, and AIEgen-based probes, QMSO₃Cl, also exhibit fluorescence enhancement upon HaloTag binding^16,30, but their luminescence mechanisms differ markedly from that of MTPABP-Cl. JF₆₃₅ brightens because HaloTag binding shifts the lactone-zwitterion equilibrium toward the emissive zwitterionic state, producing a binary fluorescence turn-on response. QMSO₃Cl becomes fluorescent mainly through aggregation-induced restriction of intramolecular motions, yielding strong signals particularly during protein clustering. MTPABP-Cl operates through a steric-confinement–induced emission mechanism, in which HaloTag binding imposes spatial restriction of intramolecular motions. This produces a continuous, confinement-dependent fluorescence increase rather than a simple on/off response, enabling BIOSCE to sensitively report subtle differences in local protein packing and conformational dynamics.

Fig. 1. Photophysical properties and molecular dynamics analysis.

a Schematic diagram of steric confinement-induced emission in the “BIOSCE” method. b Absorbance and fluorescence spectra of 10 μM MTPABP-Cl. c Photoluminescence (PL) spectra of MTPABP-Cl in water/DMSO mixtures with different water fractions (fw). The concentration of MTPABP-Cl was 10 μM, and the excitation wavelength was 460 nm. d Changes in the relative fluorescence intensity(I/I_0-1) of MTPABP-Cl at 662 nm with an increase in the water fraction of the water/THF mixture. e–h Molecular dynamics analysis of MTPABP-Cl with HaloTag, HaloTag fusion protein (HaloCaMP, -Ca²⁺), and HaloTag fusion protein (HaloCaMP, +Ca²⁺). e The root mean square deviation (RMSD), f root mean square fluctuation (RMSF), g radius of gyration (Rg), h solvent-accessible surface area (SASA) value.

To further explain the observed fluorescence change of MTPABP-Cl, which is indeed due to the steric confinement, we performed molecular docking experiments and molecular dynamics (MD) simulations. Molecular docking results revealed a higher binding affinity for MTPABP-Cl towards the HT fusion protein (HaloCaMP, −Ca²⁺, −8.6 kcal mol⁻¹), HT fusion protein (HaloCaMP, +Ca²⁺, −8.7 kcal mol⁻¹) compared to HT alone (−7.8 kcal mol⁻¹), primarily mediated by multiple intermolecular interactions (Supplementary Fig. 8a–c). Furthermore, the HaloTag fusion protein (HaloCaMP, +Ca²⁺) compared with the HaloTag fusion protein (HaloCaMP, −Ca²⁺) establishes multiple Pi-Cation interactions with MTPABP-Cl, contributing to the stability of the complex.

The MD simulations results showed that the root mean square deviation (RMSD, 0.084 ± 0.01 nm for MTPABP-Cl/HT complex, 0.358 ± 0.1 nm for MTPABP-Cl/HT fusion protein (HaloCaMP, −Ca²⁺), and 0.149 ± 0.05 nm for MTPABP-Cl/HT fusion protein (HaloCaMP, +Ca²⁺) (Fig. 1e, Supplementary Data 1), the root mean square fluctuation (RMSF, 0.082 ± 0.04 nm for MTPABP-Cl/HT complex, 0.265 ± 0.1 nm for MTPABP-Cl/HT fusion protein (HaloCaMP, −Ca²⁺), and 0.259 ± 0.08 nm for MTPABP-Cl/HT fusion protein (HaloCaMP, +Ca²⁺) (Fig. 1f, Supplementary Data 1) and radius of gyration (Rg, 1.787 ± 0.00 nm for MTPABP-Cl/HT complex, 2.791 ± 0.03 nm for MTPABP-Cl/HT fusion protein (HaloCaMP, −Ca²⁺), and 0.804 ± 0.02 nm for MTPABP-Cl/HT fusion protein (HaloCaMP, +Ca²⁺) (Fig. 1g, Supplementary Data 1), indicating that the fusion partner in the presence of Ca²⁺ induces a conformational change around the bound MTPABP-Cl to form a more compact local pocket. This increased local steric confinement, coupled with a smaller solvent-accessible surface area (SASA, 114.686–127.564 nm² for MTPABP-Cl/HT complex, 205.224–227.448 nm² for MTPABP-Cl/HT fusion protein (HaloCaMP, −Ca²⁺), and 20.62–68.99 nm² for MTPABP-Cl/HT fusion protein (HaloCaMP, +Ca²⁺) (Fig. 1h, Supplementary Data 1), allows for more pronounced and specific non-covalent interactions that restrict the intramolecular motions of MTPABP-Cl in a more effective manner. These results indicated that the binding interface in the HaloTag fusion protein (HaloCaMP, +Ca²⁺) complex imposes greater steric hindrance on MTPABP-Cl. This enhanced confinement restricts intramolecular motions that facilitate non-radiative decay, thereby favoring radiative transition and leading to the observed increase in fluorescence quantum yield.

MTPABP-Cl exhibits a low background in the intracellular environment

We systematically evaluated the ability of MTPABP-Cl to covalently bind to HT protein under in vitro conditions. To determine the binding kinetics, the HT protein was successfully purified from Escherichia coli cells (Supplementary Fig. 9a, b and Supplementary Fig. 16a, Supplementary Data 2) and incubated with MTPABP-Cl. The labeling kinetics were monitored by measuring the fluorescence response using a microplate reader. Covalent conjugation of MTPABP-Cl to HT protein was assessed by sodium dodecyl sulfate–polyacrylamide gel electrophoresis (SDS-PAGE) and typhoon fluorescence laser scanning. Furthermore, to better quantify the dose dependence and labeling efficiency of the BIOSCE-based MTPABP-Cl toward intracellular HT fusion proteins, we performed surface plasmon resonance (SPR) experiments to characterize the binding affinity. The biocompatibility of MTPABP-Cl was assessed via a CCK-8 cytotoxicity assay.

As shown, binding between MTPABP-Cl (100, 1000 nM) and HT protein (1.5 μM) reached equilibrium within 20 min (Fig. 2a, Supplementary Data 1), indicating rapid interaction. In addition, after loading the same amounts of HT fusion protein (1.5 μM) onto the SDS-PAGE, the samples containing different doses of the MTPABP-Cl (1, 1.5, 2, 2.5, 3 μM) revealed minimal fluorescence background fluorescence (Fig. 2b and Supplementary Fig. 16b). The protein and fluorescent signal location matched well, indicating a successful coupling of the MTPABP-Cl to HT fusion protein with high conjugation efficiency (Fig. 2b and Supplementary Fig. 16b). SPR analysis revealed that the association rate constant (k_a) of MTPABP-Cl was 3.45 × 10² M⁻¹ s⁻¹ (Fig. 2c, Supplementary Data 1), indicating favorable binding kinetics toward HaloTag protein. Additionally, the cytotoxicity results indicated even when incubated with 15 µM MTPABP-Cl, the viability of HeLa cells remained above 95% (Fig. 2d, Supplementary Data 1), indicating low cytotoxicity and good biocompatibility.

Fig. 2. Labeling characteristics of MTPABP-Cl in vitro.

a Fluorescence intensity of MTPABP-Cl (0.1–1000 nM) in 1.5 μM HT protein substrate (N  =  3 individual replicates). b 1.5 μM HT fusion protein was labeled with 1–3 μM MTPABP-Cl and analyzed by SDS-PAGE. c Affinity and kinetic analysis of MTPABP-Cl (3.13, 12.5, 50 μM) binding to HT protein (50 μg mL⁻¹). K_a, association constant. d Hela cells were treated with 0–15 μM MTPABP-Cl for 24 h and subjected to CCK-8 assays (N  =  4 individual replicates). e Fluorescence intensity of MTPABP-Cl (0.1–1000 nM) bound to 1.5 μM HT protein in the presence of 600 μM sucrose (N  =  3 individual replicates). f Fluorescence intensity of MTPABP-Cl (0.1–1000 nM) bound to 1.5 μM HT protein in the presence of 4 μM dextran-70 (N  =  3 individual replicates). All data were presented as means ± standard error of the mean (SEM).

To mimic intracellular crowding conditions, we introduced sucrose (342.30 Da) and dextran-70 (~70 kDa) into the HT protein system^31–34. Notably, MTPABP-Cl (100, 1000 nM) binding to HT protein (1.5 μM) reached equilibrium within 20 min. Similarly, equilibrium was also reached within 20 min in the mixed system of HT protein with 150, 300, 600 μM sucrose or 2, 4 μM glucan-70, but the signal in sucrose was nearly twice that in dextran (Fig. 2e, f and Supplementary Fig. 10a–i, Supplementary Data 1–2). This difference is likely due to the smaller molecular size of sucrose, which creates a more homogeneous and compact crowding environment around the HT-MTPABP-Cl complex. Such tighter steric confinement enhances restriction of intramolecular motion and thus amplifies SCE-based fluorescence, whereas the larger dextran-70 produces a more heterogeneous microenvironment that imposes weaker confinement, resulting in a comparatively lower fluorescence response. These results demonstrate that MTPABP-Cl binds rapidly and with high affinity to HT protein, produces low fluorescence background upon conjugation, and remains effective under physiologically relevant crowding conditions, supporting its suitability for labeling of protein conformational changes in living cells.

Specificity and stability of MTPABP-Cl for labeling in living cells

BIOSCE approach utilizes HaloTag (HT) protein as a genetic fusion tag to label HT-fused proteins in living cells with the chloroalkane-based probe MTPABP-Cl, allowing monitoring of protein conformational states. We used HT tag and localization sequences as the protein genetic expression pattern for gene coding, enabling the target proteins to achieve spatial localization. To verify the labeling specificity of MTPABP-Cl within the BIOSCE framework, we utilized three different subcellular localization models, the cell membrane, cytoplasm, and nucleus, for overexpression of HT-fusion proteins in living cells. Given the good cell membrane permeability of MTPABP-Cl, live-cell imaging was performed after incubating cells with 5 µM of the probe.

We verified that MTPABP-Cl would rapidly and selectively label HT-fusion proteins without perturbing their native subcellular localization. Subsequently, we transfected the lyn11_EGFP_FKBP_Halo, EGFP_Halo, and H2B_Halo plasmids into HEK 293T cells, achieving successful expression of these proteins. We found that MTPABP-Cl exhibits distinct colocalization characteristics with HT fusion proteins expressed in different cellular compartments, and the fluorescent signals corresponded accurately to the expected subcellular locations through nuclear staining as a control (Fig. 3a–c). To further verify the universality of the phenomenon, we repeated the experiment in HeLa cells and obtained the same results (Fig. 3a–c). These results indicate that MTPABP-Cl labeling does not interfere with the intrinsic localization of the fusion proteins, confirming its suitability for non-disruptive live-cell imaging.

Fig. 3. Specificity and stability of MTPABP-Cl for labeling in living cells.

Subcellular localization of HEK 293 T (top) and HeLa (bottom) cells overexpressing lyn11_EGFP_FKBP_Halo (a), EGFP_Halo (b), H2B_Halo (c), and labeled with MTPABP-Cl (5 μM, 30 min). Scale bars, 10 μm. d Fluorescence imaging of HeLa cells expressing with or without HT fusion protein and labeled with MTPABP-Cl (5 μM, 30 min) at 0 min (left) and 180 min (right). Scale bars, 20 μm. e Fluorescence intensity of MTPABP-Cl in HeLa cells without HT fusion protein in (d) (N  =  3 individual biological replicates, n  =  19 cells). f Fluorescence intensity ratio of MTPABP-Cl to EGFP in HeLa cells in (d). Red line and light gray areas represent the mean value and SEM, respectively. All data were presented as means ± standard error of the mean (SEM).

Protein behaviors in cellular activities have diverse time scales; long-term continuous laser imaging requires stable fluorescence labeling tools. We further assessed the stability of MTPABP-Cl under long-term continuous laser imaging. HeLa cells expressing EGFP_Halo were labeled with MTPABP-Cl, while untransfected cells served as a background control (Fig. 3d). The background fluorescence from MTPABP-Cl in untransfected cells remained low and stable (Fig. 3e, Supplementary Data 1). In contrast, the MTPABP-Cl/EGFP fluorescence intensity ratio showed a linear increase (R² = 0.97) at the same laser intensity irradiated for 3 h in the same cells, and the fluorescence of MTPABP-Cl binds to EGFP_Halo protein remained stable with almost no attenuation (Fig. 3f, Supplementary Data 1). These results ensured minimal background and high signal fidelity, enabling reliable tracking of dynamic protein processes in living cells.

Sensing protein conformational states during cellular activity

Cellular transient events enable rapid responses to external stimuli and play essential roles in processes such as neurotransmitter release and fluctuations in intracellular Ca²⁺ levels. Among these, calmodulin (CaM) acts as a key Ca²⁺ sensor that transduces calcium signals into physiological responses by activating or modulating downstream effectors³⁵. Conventional genetically encoded calcium indicators based on green fluorescent protein (GCaMPs) only report Ca²⁺ dynamics primarily through CaM conformational changes⁸. Given the outstanding photophysical properties of MTPABP-Cl, we explored whether MTPABP-Cl of the BIOSCE, based on steric confinement-induced emission could detect CaM conformational changes upon Ca²⁺ binding, taking GCaMP as a reference.

We engineered SCECaMP, a chemigenetic Ca²⁺ indicator constructed by fusing CaM to the N-terminus of HaloTag (HT) and a CaM-binding peptide to its C-terminus, building upon established design principles^8,36. Similar to “HaloCaMP,” SCECaMP reports intracellular Ca²⁺ levels in the presence of a fluorescent substrate^16,37 (Fig. 4a). It has been reported that HaloCaMP can reflect neuronal activity in primary cultured hippocampal neurons under different action potential stimulations¹⁶. Here, we evaluated the performance of SCECaMP in detecting rapid Ca²⁺ transients in different cellular models.

Fig. 4. MTPABP-Cl of BIOSCE senses conformational changes in SCECaMP.

a Schematic of the SCECaMP Ca²⁺ sensor showed the domain organization (top) and primary structure (bottom). Normalized fluorescence (ΔF/F₀) responses of MTPABP-Cl in HEK 293T cells expressing SCECaMP (b) or GCaMp6s (c) in the absence or presence of 70 mM KCl (N  =  3, 3 individual biological replicates, n  =  9, 9 cells, respectively). Blue and magenta lines represent the mean values, while light blue and light gray shaded areas indicate the SEM, respectively. Normalized fluorescence (ΔF/F₀) responses of MTPABP-Cl in HT22 cells expressing SCECaMP (d) or GCaMp6s (e) in the absence or presence of 70 mM KCl (N  =  3, 3 individual biological replicates, n  =  6, 6 cells, respectively). Green and red lines represent the mean values, while light green and light gray shaded areas indicate the SEM, respectively. Approximately 100 frames (absence of KCl) were averaged to establish the F₀ baseline value. ΔF was calculated as ΔF = F − F₀, where F (presence of KCl, total 600 frames) represents the average fluorescence intensity of the regions of interest (ROIs) in each frame. All data were presented as means ± standard error of the mean (SEM).

We expressed SCECaMP in HEK 293T cells and labeled with MTPABP-Cl to report the process of CaM conformational changes, and cells expressed GCaMP6s were used as a control. To quantify the fluorescence responses, cells were stimulated with 70 mM KCl to evoke intracellular Ca²⁺ elevations³⁸. Intracellular calcium imaging revealed that MTPABP-Cl-based BIOSCE signaling responded rapidly to KCl-induced Ca²⁺ transients, with fluorescence increasing upon Ca²⁺ rise and gradually decaying as Ca²⁺ levels returned to baseline (Fig. 4b, Supplementary Data 1). GCaMP6s also exhibited a rapid fluorescence increase followed by decay during Ca²⁺ clearance (Fig. 4c, Supplementary Data 1). The enlarged results indicated that SCECaMP labeled with MTPABP-Cl displayed Ca²⁺ response kinetics comparable to those of GCaMP6s (Fig. 4b, c, Supplementary Data 1). However, the fluorescence decay of MTPABP-Cl was slower than that of GCaMP6s after Ca²⁺ decrease (Fig. 4b, c, Supplementary Data 1). Similar experiments conducted in mouse hippocampal neuron cells (HT22), and the results were consistent with the findings mentioned above (Fig. 4d, e, Supplementary Data 1), supporting the broad applicability of MTPABP-Cl for protein labeling. Notably, MTPABP-Cl also exhibited Ca²⁺ response speeds similar to those reported for Si-rhodamine-based HaloTag ligands^16,39. Taken together, these results confirmed that MTPABP-Cl of BIOSCE labeled the dynamic proteins proximity via the steric confinement-induced emission principle, proving its feasibility and consistent with our conclusion. It further demonstrated the utility of the BIOSCE approach for studying cellular processes at the molecular level.

Monitoring protein behaviors in the FKBP-rapamycin-FRB model

Studies on the mTOR signaling pathway have revealed that rapamycin (an antifungal antibiotic macrolide) binding to FK506-binding protein (FKBP) facilitates the subsequent interaction between FKBP and FKBP-rapamycin binding (FRB), forming a ternary complex that inhibits mTOR activity^40–42. This discovery underpinned the development of the FKBP-rapamycin-FRB ternary complex protein interaction model^43–45. However, the dynamic characteristics of FKBP–FRB interactions on relevant timescales within living cells remain largely unexplored. Given the high specificity and sensitivity of BIOSCE binding to HT fusion proteins in cells, we further verified whether BIOSCE could provide insights into rapamycin-dependent FKBP-FRB proximity interactions.

To evaluate whether BIOSCE can be applied to the FKBP-rapamycin-FRB model, we examined the labeling of FKBP and FRB by MTPABP-Cl in the presence or absence of rapamycin. We expressed (i) lyn11-FKBP-Halo and FRB (Fig. 5a), (ii) lyn11-FKBP and FRB-Halo (Fig. 5f) proteins in HeLa cells and analyzed MTPABP-Cl labeled results using fluorescence microscopy and flow cytometry. Fluorescence imaging showed that cells labeled with MTPABP-Cl exhibited significant enhancement in the presence of rapamycin (Fig. 5b, g). The quantitative results of flow cytometry further confirmed the fluorescence intensity of MTPABP-Cl was significantly upregulated in the presence of rapamycin (Fig. 5c–e, h–j; Supplementary Fig. 11a–d and Supplementary Fig. 15a–c; Supplementary Data 1).

Fig. 5. Labeling and quantification of protein behaviors in the FKBP-rapamycin-FRB model using MTPABP-Cl of BIOSCE.

a, f Schematic of BIOSCE application in the FKBP-rapamycin-FRB model. Fluorescence micrographs of HeLa cells co-expressing lyn11_FKBP_Halo, FRB proteins in (b), and lyn11_FKBP, FRB_Halo proteins in (g). Labeling with MTPABP-Cl (5 μM, 30 min) was observed only in the absence (left) or presence(middle) of rapamycin (100 nM), and fluorescence response (right) to 100 nM rapamycin. Scale bar, 10 μm. Representative flow cytometry analysis of HeLa cells co-expressing lyn11_FKBP_Halo, FRB proteins in (c) and lyn11_FKBP, FRB_Halo proteins in (h), which were incubated with MTPABP-Cl (5 μM, 30 min) in the absence or presence of rapamycin (100 nM). d, i Cell counts peak map of MTPABP-Cl fluorescence intensity in the absence (left) or presence (right) of rapamycin. Quantitative data of MTPABP-Cl bound to HT fusion protein in HeLa cells in the Q1 region of (c, h) in (e, j), respectively (N  =  3, 3 individual biological replicates, respectively). Data were presented as violin plots showing the quartile and median, the center line represents the median, the top and bottom line represents upper and lower quartiles, and whiskers are the minimum and maximum values. Statistical significance was determined by a two-tailed unpaired Student’s t-test.

We evaluated the FKBP–FRB proximity dynamic interactions in the presence or absence of rapamycin. We constructed recombinant plasmids and expressed them in HeLa cells in the following combination: (i) lyn11-FKBP-Halo, FRB (Fig. 6a); (ii) lyn11-FKBP, FRB-Halo (Fig. 6f). Subsequently, MTPABP-Cl of BIOSCE was incubated with cells expressing these proteins during imaging. In cells harboring the above constructs, the normalized fluorescence intensity of MTPABP-Cl increased within milliseconds in the presence of rapamycin compared to its absence (Fig. 6b, g, Supplementary Data 1). Furthermore, quantitative analysis of the peak fluorescence intensity after MTPABP-Cl labeling of FKBP and FRB showed increases of 132 and 111% relative to the absence of rapamycin, respectively (Fig. 6c, h, Supplementary Data 1).

Fig. 6. MTPABP-Cl of BIOSCE responds to dynamic protein proximity.

a Schematic of BIOSCE response to FKBP_Halo and FRB protein interactions. b Normalized fluorescence responses (ΔF/F₀) of MTPABP-Cl labeled HeLa cells expressing FKBP_Halo and FRB proteins in the absence or presence of rapamycin (N  =  3 individual biological replicates, n = 16 cells). Red line and light gray areas represent the mean value and SEM, respectively. c Statistical comparison of the peak MTPABP-Cl fluorescence in the absence or presence of rapamycin in cells co-expressing FKBP_Halo and FRB in (b) (N  =  3 individual biological replicates, n = 16 cells). Representative FRAP images of FKBP_Halo, FRB (d); FKBP, FRB_Halo (i) proteins and experimental treatments, respectively. Images before (before), immediately following (bleach), and after recovery (post) are shown. Scale bars are indicated. Quantification of FRAP experiments for FKBP_Halo, FRB (e); FKBP, FRB_Halo (j) proteins and experimental conditions, respectively (N  =  3 individual biological replicates, n = 10, 5 cells, respectively). Blue and magenta lines represent the mean values, while light gray shaded areas indicate the SEM, respectively. f Schematic of BIOSCE response to FKBP and FRB_Halo protein interactions. g Normalized fluorescence responses (ΔF/F₀) of MTPABP-Cl labeled HeLa cells expressing FKBP and FRB_Halo proteins in the absence or presence of rapamycin (N  =  3 individual biological replicates, n = 16 cells). Purple line and light gray areas represent the mean value and SEM, respectively. h Statistical comparison of the peak MTPABP-Cl fluorescence in the absence or presence of rapamycin in cells co-expressing FKBP and FRB_Halo in (g). (N  =  3 individual biological replicates, n = 16 cells). Approximately 200 frames (absence of rapamycin) were averaged to establish the F₀ baseline value. ΔF was calculated as ΔF = F − F₀, where F (presence of rapamycin, total 501 frames) represents the average fluorescence intensity of the regions of interest (ROIs) in each frame. All data were presented as means ± standard error of the mean (SEM). Statistical significances (i.e., p-value) were calculated using Welch’s corrected two-tailed t-tests.

We verified that the differences in MTPABP-Cl labeled FKBP and FRB fluorescence intensities arose from protein mobility. To test this, we performed fluorescence recovery after photobleaching (FRAP) experiments. As expected, FRAP revealed significant differences in FKBP and FRB diffusion mobility before and after rapamycin treatment (Fig. 6d, i and Supplementary Fig. 12a–d, Supplementary Data 2). Rapamycin induced higher mobility of FKBP- or FRB-HaloTag fusion proteins compared to untreated cells (Figs. 6e, j, Supplementary Data 1). Given the structure and luminescence mechanism of MTPABP-Cl, we speculated that BIOSCE could be applied to sense conformational changes of single HaloTag fusion proteins through steric confinement. Such conformational changes would generate continuous fluorescence responses in BIOSCE, leading to pronounced fluorescence fluctuations. Taken together, these findings demonstrate that MTPABP-Cl within the BIOSCE platform might be an ideal candidate dye for recording dynamic protein-protein proximity interactions in cellular activities.

Tracking of SNAP25 mobility during botulinum toxin intoxication

Supramolecular protein assemblies, as evolutionary innovations, are distributed throughout biological systems to execute complex functions. One such magnificent protein assembly is the neuronal SNARE (Soluble N-ethylmaleimide-sensitive factor Attachment REceptor) complex (trimeric supramolecular organization of SNAP25, syntaxin, and VAMP), which plays a central and fundamental role in neuronal signaling^46–48. SNAP25 is a helical protein and the prime contributor in the complex assembly^46,48. Botulinum neurotoxins (BoNTs) are neurotoxic proteins produced by Clostridium botulinum and related bacterial species, and serotype A (BoNT/A) is one of the serotypes of BoNTs^49,50. It has been reported that the light chain (LC) of BoNT/A binds and cleaves the SNAP25 between the residues glutamine (Q197) and arginine (R198) in cells, might cause the vesicle fusion process to fail to proceed normally and thereby blocking the release of neurotransmitters^51–53. However, the dynamic distribution of SNAP25 N-terminal and C-terminal fragments during botulinum toxin intoxication in living cells remains largely unexplored.

To enable selective interrogation of the different fragments generated after BoNT/A cleaves SNAP25, we designed two complementary fusion orientations: Halo_SNAP25 and SNAP25_Halo. These constructs ensured that the HaloTag remained attached to only one side of the cleavage site (Q197-R198). HaloTag covalently bound the chlorinated ligand MTPABP-Cl, which might allow fragment-specific tracking following toxin-mediated proteolysis (Fig. 7a). To further verify, we expressed SNAP25 protein in HT22 cells and confirmed their expression by western blotting (Supplementary Fig. 13 and Supplementary Fig. 16c). Cells expressing SNAP25 protein were then incubated with recombinant BoNT/A for 24 h, and BoNT/A activity was assessed by immunoblotting. The results showed that SNAP25 protein was cleaved to varying degrees at different time points after BoNT/A treatment (Fig. 7b; Supplementary Fig. 14a, b and Supplementary Fig. 16d, Supplementary Data 2). This confirmed that SNAP25 enabled expressed in HT22 cells and the recombinant BoNT/A retained biological activity.

Fig. 7. Tracking SNAP25 cleavage and distribution during BoNT/A intoxication using MTPABP-Cl of BIOSCE.

a Schematic of BIOSCE tracking BoNT/A intoxication and the design of Halo_SNAP25 and SNAP25_Halo fusion protein constructs. b HT22 cells were transfected with SNAP25 plasmids for 48 h in the absence of BoNT/A. Then, cells were treated with 25 nM BoNT/A for 1, 6, 12, 24 h, lysed, and analyzed for the indicated proteins by immunoblots. c, d Top row: schematic diagram of the cells expressing SNAP25-HT fusion proteins and labeled with (i) MTPABP-Cl (5 μM, 30 min), BoNT/A (25 nM, 1 h), (ii) BoNT/A (25 nM, 1 h), MTPABP-Cl (5 μM, 30 min). Middle row: representative images of HT22 cells expressing SNAP25_Halo (left) or Halo_SNAP25 (right), and incubated with (i) and (ii). White circles were the HT22 cell membrane, and blue circles were the outline of the nucleus. Scale bar, 5 μm. Bottom row: representative trajectories of SNAP25_Halo or Halo_SNAP25 proteins labeled with MTPABP-Cl in (i) and (ii) cells. Scale bar, 10 μm. Complementary cumulative distribution function (CCDF) graphs of apparent diffusion coefficients, as determined by single-particle tracking, of the indicated proteins in (e, f) (n = 153, 272, 369, and 82 molecules for SNAP25_Halo (left) or Halo_SNAP25 (right) proteins labeled with MTPABP-Cl in (c, d) cells, respectively). Normalized fluorescence responses (ΔF/F₀) of MTPABP-Cl in HT22 cells expressing Halo_SNAP25 (left), SNAP25_Halo (right) proteins in 5 μM MTPABP-Cl, 25 nM BoNT/A and 50 μM Zncl₂ (g), 25 nM BoNT/A, 5 μM MTPABP-Cl and 50 μM Zncl₂ (h). (N  =  3 individual biological replicates, n = 7, 6, 5, 5 cells, respectively). Blue and magenta lines represent the mean values, while light gray shaded areas indicate the SEM, respectively. All data were presented as means ± standard error of the mean (SEM).

BoNT/A binds and cleaves the SNAP25 between the residues glutamine (Q) and arginine (R) in cells, which is critical for interpreting the behavior of Halo_SNAP25 versus SNAP25_Halo constructs. Therefore, we investigated the dynamic distribution of SNAP25-HT fragments labeled with MTPABP-Cl within living cells during BoNT/A treatment. Specifically, we determined the apparent diffusion coefficients of Halo_SNAP25 and SNAP25_Halo by particle tracking, calculating the mean squared displacement of individual protein trajectories under different treatment sequences of BoNT/A and MTPABP-Cl treatment, and plotted their distributions (Fig. 7c, d, Supplementary Data 1). As expected, when cells were first incubated with MTPABP-Cl and then treated with BoNT/A, the N-terminal of Halo_SNAP25 and C-terminal fragments of SNAP25_Halo were sparsely distributed at the cell membrane and in the cytoplasm, respectively (Fig. 7c, Supplementary Data 1). In addition, when cells were first treated with BoNT/A and then incubated with MTPABP-Cl, the N-terminal of Halo_SNAP25 and C-terminal fragments of SNAP25_Halo were densely distributed at the cell membrane and in the cytoplasm, respectively (Fig. 7d, Supplementary Data 1). The differential distribution is a direct consequence of the BoNT/A cleavage mechanism and the subsequent behavior of the protein fragments. The N-terminal Fragment (SNAP25 1-197aa) retains the palmitoylation sites that are naturally present in full-length SNAP25. Palmitoylation is a lipid modification that serves as a membrane anchor Dong. Therefore, upon cleavage, this N-terminal fragment remains tethered to the plasma membrane. The C-terminal Fragment (SNAP25 198-206aa) generated by the cleavage lacks any membrane-anchoring domains. As a result, it is released from the membrane into the cytoplasm and diffuses freely.

Apparent diffusion coefficients ranged from 0.03 to 0.07 μm² s⁻¹for N-terminal of Halo_SNAP25 and the C-terminal fragments of SNAP25_Halo in the presence of BoNT/A (Fig. 7e, f, Supplementary Data 1). The speed of fragment migration affects the difference in fragment distribution and depends on whether BoNT/A takes effect before or after MTPABP-Cl labeling. When MTPABP-Cl is added first, the intact Halo_SNAP25 or SNAP25_Halo is labeled in its relatively slow membrane-associated state; after BoNT/A cleavage, the newly generated fragments disperse more slowly (0.0491 μm² s⁻¹, 0.0384 μm² s⁻¹) from their original positions, resulting in a sparser distribution. In contrast, when BoNT/A precedes MTPABP-Cl labeling, the cleavage fragments now freed from the intact SNARE assembly exhibit faster initial migration (0.0681 μm² s⁻¹, 0.0536 μm² s⁻¹) and wider dispersal, allowing subsequent MTPABP-Cl labeling to capture a larger population of rapidly moving N-terminal and C-terminal fragments and producing a denser spatial distribution. These results demonstrated that MTPABP-Cl of BIOSCE enable effectively track the spatial distribution of proteins.

Zn²⁺ induces conformational changes in the loop region of SNAP25 protein⁵⁴. To investigate the effect of Zn²⁺ on transient SNAP25 conformational changes during BoNT/A cleavage, we expressed Halo_SNAP25 and SNAP25_Halo recombinant proteins in HT22 cells and performed confocal laser scanning microscopy using a live-cell workstation. Quantitative analysis of normalized fluorescence intensity showed that MTPABP-Cl rapidly responded to conformational changes at the N-terminal of SNAP25 during Zn²⁺ treatment (Fig. 7g, h, Supplementary Data 1). We speculate that this rapid response may result from increased exposure of SNAP25 regions following BoNT/A cleavage. Overall, these results provide insights into the dynamic behavior of SNAP25 during BoNT/A cleavage and demonstrate that BIOSCE can effectively track transient protein conformational changes in living cells.

Discussion

Proteins are essential for maintaining cellular function, both by enabling individual processes and as part of carefully regulated complex protein networks that orchestrate the wider interactions throughout the cell^55,56. Many protein functions depend on dynamic biomolecular events such as transient conformational changes and activation⁵⁷. Therefore, the development of biosensor systems is critical for monitoring protein behaviors in vivo and pushing the frontier of biological imaging. In this study, we designed the BIOSCE as a method, transforming this widely used HT labeling tag into an application for recording the process of protein conformational changes in cellular physiological events. Key characteristics of BIOSCE include rapid responsiveness to Ca²⁺-dependent protein conformational changes, the ability to monitor dynamic protein behaviors in the FKBP-rapamycin-FRB model, and the capacity to track the distribution of SNAP25 fragments after BoNT/A treatment.

The dynamic distribution of proteins during cellular activities provides critical insights into the functions and interactions of cellular regions and organelles, as well as the spatial regulation of signaling pathways and the spatiotemporal specificity of cellular responses^58–60. Specially, tracking the mobility of SNAP25 during botulinum toxin intoxication offers valuable information about synaptic vesicle docking and fusion mechanisms^53,61. Effective methods for tracking protein conformational changes require high sensitivity and spatiotemporal resolution^62–64. Proximity ligation assay (PLA) and fluorescence correlation spectroscopy (FCS) provide complementary insights but face fundamental constraints. PLA requires cell permeabilization, precluding live-cell applications^65–69, while FCS yields ensemble-averaged information and cannot resolve conformational states of individual proteins^70–74. Molecular dynamics simulations can estimate biomolecular parameters but cannot fully capture conformational processes in living cells⁷⁵. Probe-based labeling approaches, such as biotinylation, peroxidase tagging, and photoaffinity labeling, provide valuable insights but often suffer from instability and phototoxicity^76–79.

Our strategy employs cell-permeable small-molecule HT ligands to selectively label target proteins and report their conformational states in living cells without perturbing spatiotemporal protein expression. Unlike GFP-like HT ligand dyes, which frequently exhibit concentration-dependent quenching^17,30, MTPABP-Cl avoids this limitation through its steric confinement–induced emission design, demonstrating robust anti-quenching properties. MTPABP-Cl of BIOSCE design strategy is scalable; prototype indicators could be designed by selecting steric confinement-induced emission luminescence structure with better solubility and lower fluorescent background in established sensors with HT protein. The BIOSCE married the genetic specificity of HT-based sensors with the superior photophysical properties of small-molecule fluorophores, provided a reference for the design of chemigenetic biosensors that overcome limitations of current fluorescent protein-based sensors. The steric confinement-induced emission mechanism of MTPABP-Cl might address a gap between existing probe classes. Dyes like JF₆₃₅ function as superb fiducial markers, as binding to the HaloTag protein results in large increases in absorption and fluorescence, but their signal saturates upon tag occupancy¹⁶. AIEgen probes such as QMSO₃Cl effectively map protein clustering or disassembly but are insensitive to protein conformational changes³⁰. We confirmed that MTPABP-Cl and its steric confinement-induced emission are neither a binary switch nor an aggregate sensor, but rather a quantifiable readout of conformational strain, enabling the monitoring of rapid conformational changes in signaling proteins as demonstrated herein.

In summary, BIOSCE provides a dynamic HaloTag-based sensing platform for tracking protein conformational states across spatial and temporal scales, offering broad utility in both basic biology and medical research. Based on the steric confinement–induced emission of MTPABP-Cl, the BIOSCE method can not only report Ca²⁺-related cellular activities but also, in principle, reflect protein conformational changes associated with diverse physiological processes in vivo. At present, the method has been validated through probe-based labeling of proteins in cellular models; we will extend its application to tissues and in vivo systems in our future work.

However, several key challenges must be overcome to realize this translational vision. A primary limitation lies in probe delivery, as biological barriers and heterogeneous tissue environments may impede uniform distribution and labeling efficiency in deeper tissues⁸⁰. Additionally, despite the benefits of far-red emission, photon scattering and limited penetration depth could constrain imaging resolution at the whole-animal level. Furthermore, while MTPABP-Cl exhibits low cytotoxicity in vitro, potential phototoxicity and altered protein dynamics under prolonged in vivo illumination remain to be systematically evaluated⁸¹.

In order to mitigate these challenges, we will focus on: (1) developing amphiphilic or carrier-assisted probe formulations to enhance delivery; (2) optimizing optical properties for deeper tissue penetration; and (3) rigorously assessing biocompatibility in animal models. Subsequently, we will engineer an amphiphilic derivative of MTPABP-Cl and design natural protein probes based on steric confinement-induced emission, aiming to achieve precise sensing of protein proximity in vivo⁸². The realization of these goals will empower us to sample dynamic protein processes along signaling pathways, which will provide critical data for AI-based virtual cell modeling and establish a transformative tool for linking protein distribution with function. We anticipate that this comprehensive strategy will inspire a generation of protein-labeling tools, thereby greatly expanding the bioanalytical toolbox for both basic life science research and clinical applications.

Methods

Synthesis of (4-((6-chlorohexyl)oxy)phenyl)boronic acid (1)

(4-hydroxyphenyl)boronic acid (689.65 mg, 5 mmol) and K₂CO₃ (1.0365 g, 7.5 mmol) were added into a 100 mL of two-necked round-bottom flask. The flask was vacuumed and purged with nitrogen for three times. Acetonitrile (40 mL) and 1-bromo-6-chlorohexane (1.12 mL, 7.5 mmol) were injected under a nitrogen atmosphere separately, and the mixture was heated to reflux and stirred for 12 h. Afterwards, iced water was added, and the mixture was extracted with dichloromethane three times. The organic phase was combined, dried with Na₂SO₄, and the solvent was evaporated under reduced pressure. The crude product was purified by silica gel chromatography (eluent: dichloromethane/hexane 1/2) to afford (4-((6-chlorohexyl)oxy)phenyl)boronic acid as a white solid (85.5%, 1.095 g). ¹H NMR (400 MHz, Chloroform-d) δ 8.15 (d, J = 8.0 Hz, 2H), 7.00 (d, J = 8.0 Hz, 2H), 4.05 (t, J = 6.4 Hz, 2H), 3.57 (t, J = 6.6 Hz, 2H), 1.83 (p, J = 6.5 Hz, 4H), 1.58–1.50 (m, 4H).

Synthesis of 4-(7-bromobenzo[c][1,2,5]thiadiazol-4-yl)-N,N-bis(4-methoxyphenyl)aniline (2)

(bis(4-methoxyphenyl)amino)phenyl)boronic acid (349.2 mg, 1 mmol), 4,7-dibromobenzo[c][1,2,5]thiadiazole (293.97 mg, 1 mmol), Pd(PPh₃)₄ (12 mg, 0.01 mmol) and K₂CO₃ (414.6 mg, 3 mmol) were added into a 100 mL of two-necked round-bottom flask. The flask was vacuumed and purged with nitrogen for three times. Then tetrahydrofuran (20 mL), ethanal (20 mL), and water (5 mL) were added and the mixture was heated to reflux and stirred for 24 h. Afterwards, the mixture was cooled to room temperature, and the mixture was extracted with dichloromethane three times. The organic phase was combined, dried with Na₂SO₄, and the solvent was evaporated under reduced pressure. The crude product was purified by silica gel chromatography (eluent: dichloromethane/hexane 1/4) to afford 4-(7-bromobenzo[c][1,2,5]thiadiazol-4-yl)-N,N-bis(4-methoxyphenyl)aniline as a red solid (88.2%, 457.3 mg). ¹H NMR (400 MHz, DMSO-d₆) δ 7.95 (d, J = 8.7 Hz, 2H), 7.89 – 7.81 (m, 4H), 7.10 (dt, J = 8.6, 4.3 Hz, 6H), 6.98 – 6.92 (m, 4H), 6.88 (d, J = 8.8 Hz, 2H), 4.04 (t, J = 6.5 Hz, 2H), 3.76 (s, 6H), 3.65 (t, J = 6.6 Hz, 2H), 1.76 (s, 4H), 1.51 – 1.42 (m, 4H). ¹³C NMR (101 MHz, DMSO) δ 159.24, 156.50, 153.97, 148.99, 140.22, 131.77, 131.34, 130.72, 130.30, 129.62, 128.57, 127.97, 127.54, 119.01, 115.52, 114.98, 67.90, 55.73, 45.85, 32.47, 29.01, 26.52, 25.29.

Synthesis of 4-(7-(4-((6-chlorohexyl)oxy)phenyl)benzo[c][1,2,5]thiadiazol-4-yl)-N,N-bis(4-methoxyphenyl)aniline (MTPABP-Cl)

4-(7-bromobenzo[c][1,2,5]thiadiazol-4-yl)-N,N-bis(4-methoxyphenyl)aniline (259.21 mg, 0.5 mmol), (4-((6-chlorohexyl)oxy)phenyl)boronic acid (192.4 mg, 0.75 mmol), Pd(PPh₃)₄ (6 mg, 0.005 mmol) and K₂CO₃ (207.3 mg, 1.5 mmol) were added into a 50 mL of two-necked round-bottom flask. The flask was vacuumed and purged with nitrogen for three times. Then tetrahydrofuran (10 mL), ethanal (10 mL), and water (2 mL) were added and the mixture was heated to reflux and stirred for 24 h. Afterwards, the mixture was cooled down to room temperature, and the mixture was extracted with dichloromethane three times. The organic phase was combined, dried with Na₂SO₄, and the solvent was evaporated under reduced pressure. The crude product was purified by silica gel chromatography (eluent: dichloromethane/hexane 1/2) to afford 4-(7-(4-((6-chlorohexyl)oxy)phenyl)benzo[c][1,2,5]thiadiazol-4-yl)-N,N-bis(4-methoxyphenyl)aniline as a red solid (72.13%, 234.5 mg). ¹H NMR (400 MHz, DMSO-d₆) δ 7.95 (d, J = 8.7 Hz, 2H), 7.89–7.81 (m, 4H), 7.10 (dt, J = 8.6, 4.3 Hz, 6H), 6.98–6.92 (m, 4H), 6.88 (d, J = 8.8 Hz, 2H), 4.04 (t, J = 6.5 Hz, 2H), 3.76 (s, 6H), 3.65 (t, J = 6.6 Hz, 2H), 1.76 (s, 4H), 1.51–1.42 (m, 4H). ¹³C NMR (101 MHz, DMSO) δ 159.24, 156.50, 153.97, 148.99, 140.22, 131.77, 131.34, 130.72, 130.30, 129.62, 128.57, 127.97, 127.54, 119.01, 115.52, 114.98, 67.90, 55.73, 45.85, 32.47, 29.01, 26.52, 25.29. m/z: calcd. for C₃₈H₃₆ClN₃O₃S: 649.2166; found: 649.2167.

Measurement of absorption and emission spectra

Precisely weigh MTPABP-Cl and prepare a 10 mM DMSO stock solution, then dilute it with DMSO to a 10 μM working solution for spectral measurements. Absorption and emission spectra were measured using a UV-Vis spectrophotometer and a PL spectrometer, respectively.

Absolute photoluminescence quantum yield (PLQY) measurement

Prepare 0.15 mL of a 10 mM MTPABP-Cl stock solution. In subsequent tests, the stock solution can be diluted to a 10 μM concentration with DMSO in the presence of HaloTag or HaloTag fusion protein. The Hamamatsu Quantum Yield Spectrometer (Quantaurus-QY, C11347) was then used to determine the absolute PLQYs of MTPABP-Cl solutions. The excitation wavelength was set at 461 nm. The integrating sphere can detect photons from 400 to 1000 nm. Experiments were performed by three independent replicates.

Molecular docking

The computational analyses were conducted utilizing AutoDock Tools (1.5.6) (Scripps Research US). The three-dimensional structure of HaloTag or HaloTag fusion protein (HaloCaMP is consistent with the SCECaMP we used in our experiment) was obtained from the RCSB Protein Data Band under the PDB ID code 5UY1 and 6U2M, respectively. The three-dimensional structure of MTPABP-Cl was obtained from Ligprep module (LigPrep, 2018) in Maestro software. These proteins and ligand were imported and preprocessing, which included the removal of charges and the addition of polar hydrogen atoms. The grid parameter file and docking parameter file were subsequently generated. The optimal docking results were evaluated on the basis of the orientation and conformation that resulted in the lowest binding score, indicating the best affinity. Additionally, the conformation with the lowest Vina score was selected for molecular docking visual analysis using AutoDock-Vina (1.1.2) and PyMOL (3.2) software.

Molecular modeling analysis

Molecular dynamics (MD) simulations were conducted utilizing the GROMACS (2023.5) software under conditions of constant temperature and pressure, along with periodic boundary conditions. The system is built using the Amber14SB force field and the TIP3P water model to characterize the MTPABP-Cl and HaloTag or HaloTag fusion protein (HaloCaMP is consistent with the SCECaMP we used in our experiment) in the absence and presence of Ca²⁺. These systems were solvated in a cubic box, ensuring a minimum distance of 12 Å between all atoms and the edges of the water box to prevent interactions between molecules in different boxes due to periodic boundary conditions. Then Na⁺ and Cl^- ions were added to neutralize the charge. Following system construction, the steepest descent method was applied to minimize the energy of the two systems, thereby alleviating any excessive atomic proximity. Electrostatic interactions were computed using the particle-mesh Ewald (PME) method, whereas the cut-off for van der Waals intermolecular interactions was established at 12 Å, with updates occurring every 20 steps. The temperature of the system is held at 300 K using a V-rescale thermostat, while the pressure is regulated at 1 bar using a Parrinello–Rahman barostat. Subsequently, simulations are conducted for 0.1 ns under the NVT and NPT ensembles to equilibrate the system, allowing for the initiation of formal simulations. Ultimately, the composite systems were subjected to MD simulation for 100 ns, and the trajectory data were saved every 10 ps.

Molecular cloning

We synthesized the target genes from a commercial vendor (GenScript) and used a molecular cloning kit to insert the target DNA fragment into pcDNA3.1(-) or pcDNA3.4 vector according to the manufacturer’s protocol. All constructed plasmids were transformed and amplified in E. coli strains (TransGen). Endotoxin-free plasmid DNA was purified using the FastPure EndoFree Plasmid Mini Kit (Vazyme). The concentrations of construct plasmids were determined by UV5 Nano (Mettler Toledo) and verified by Sanger sequencing (GenScript).

Protein expression and purification

The His tag, TEV recognition site, EGFP, and HT cDNA sequence were inserted into the pET-15b plasmid sequentially. Transformed into E. coli strain BL21 (DE3) (Vazyme) competent cell was grown in LB supplemented with 100 μg mL⁻¹ampicillin for 24 h, monoclonal colonies were selected, and shook at 220 r. p. m at 37 °C to an optical density at 600 nm (OD₆₀₀) of 0.4–0.6. Protein expression was induced by the addition of 0.5 mM Isopropyl β-D-thiogalactoside (IPTG), and cells were incubated for an additional 10 h at 16 °C.

Cells were harvested by centrifugation (4500 g, 10 min, 4 °C), lysed by sonication, and the lysate was cleared by centrifugation (75,000 g, 10 min, 4 °C). Purification from lysate was achieved by nickel affinity chromatography, followed by Ni-NTA 6FF agarose column (Sangon). Beads were washed with 20 volumes of binding buffer (20 mM Tris-HCl, 500 mM NaCl, 10 μM ZnCl₂, 10% glycerol), containing 10 mM imidazole, pH 7.2. Protein was eluted with 10 volumes of binding buffer complemented with 40 mM imidazole. Correct size of protein was assessed by SDS PAGE.

Proteins were concentrated using Amicon Ultra-30 centrifugal filter devices (Merck) with a molecular weight cut-off (MWCO) smaller than the protein size. Subsequently, proteins were concentrated using an ultrafiltration centrifuge tube, and the solution was replaced with PBS. Protein concentrations were estimated using Omni-Easy™ ready-to-use BCA protein quantitation kit. Protein was aliquoted and stored with 10% glycerol at −80 °C.

Surface plasmon resonance (SPR)

SPR measurements were performed using a 2-channel Reichert 2SPR instrument fitted with an autosampler and a degassing pump (Reichert Technologies). SPR buffers containing PBST (4% DMSO), 10 mM Sodium acetate, pH = 4.34, 1 M Ethanolamine, pH = 8.5, 10 mM Glycine, pH = 2.0 were prepared, sterile filtered, and degassed in autoclaved glassware prior to each experiment. PBST (4% DMSO) was used to prime and run both the sample and syringe pump reservoirs (25 μL min⁻¹). Gold sensor chip modified with carboxymethyl dextran hydrogel was installed and equilibrated under flow conditions (25 μL min⁻¹). Activate with EDC/NHS at a 4:1 (w/w) ratio for 7 min, immobilize 10 μM HT protein onto the chip for 10 min, and block with 1 M ethanolamine (pH 8.5) for 10 min. Blank PBST tubes (4% DMSO) and the MTPABP-Cl sample tubes with low to high concentrations alternately in the injection area. Set the association time to 90 s, the dissociation time to 150 s, and perform detection with 12 loop repetitions according to the manufacturer’s instructions. Kinetic parameters were determined by curve-fitting using TraceDrawer software, fit with a one-to-one one-state model⁷⁵.

SDS-PAGE analysis of HT protein labeled with MTPABP-Cl

HT fusion proteins were incubated with excess MTPABP-Cl at different concentrations (1.5–3 μM) in activity buffer overnight at room temperature. Reactions were stopped by adding 4 × laemmli buffer (Bio-Rad) and heating denaturated at 95 °C for 5 min. The same concentration samples were loaded on SDS-PAGE gels and run for 50 min at 150 V. Gels were imaged on a Typhoon fluorescence laser-scanner (488 nm excitation laser, 720 nm emission filter, 20 nm bandpass).

In vitro titration of MTPABP-Cl bound to HT protein

The ligand reactions were in a 1.5 μM HT protein, PBS, and water system, and 300 μM, 600 μM sucrose, and 2 μM, 4 μM dextran-70 in 1.5 μM HT, PBS, water, respectively. Add 0.1, 1, 10, 100, 1000 nM MTPABP-Cl of BIOSCE in equal proportions, and the reaction system was 100 μL. Fluorescence was measured at 37 °C for 1 h (interval of 2 min, 30 cycles) by a plate reader with excitation wavelength of 460 nm and emission wavelength of 660 nm.

Cell culture

Human embryonic kidney 293T (HEK 293 T, ATCC, CRL-3216) cells, mouse hippocampal neuron cells (HT22, CL-0697, Pricella), and HeLa (ATCC, CCL-2) cells were previously preserved in the laboratory. HEK 293T cells were cultured and maintained in Dulbecco’s Modified Eagle medium (DMEM) supplemented with 10% heat-inactivated fetal bovine serum (FBS) and 1% Penicillin-Streptomycin (×100) at 37 °C in a humidified atmosphere of 5% CO₂. HeLa cells cultured in low-glucose Dulbecco’s modified Eagle’s medium with L-glutamine and phenol red (low-glucose DMEM) containing 10% (v/v) heat-inactivated FBS and 1% (v/v) Penicillin-Streptomycin at 37 °C in a humidified atmosphere of 5% CO₂. Unless otherwise noted, cells were cultured with 5 mL leach medium in a T-25 flask, 2 mL of each medium in a 35 mm glass-based dish, and a microplate 6-well with lid, and 15 mL leach medium in a T-75 flask, 1 mL each medium in a microplate 24-well with lid.

Transient transfection

To express EGFP_Halo, EGFP_FKBP_Halo, H2B_Halo, FRB/FRB_Halo, SCECaMP, GCamp6s, Halo_SNAP25 and SNAP25_Halo proteins in living cells. Three days before imaging, HEK 293T, mouse hippocampal neuron cell (HT22), and HeLa cells were seeded to more than 70% confluence on 35-mm glass-bottomed dishes (Cellvis) in MEM or DMEM supplemented with 10% FBS. After 12 h, plasmids were transiently transfected into cells utilizing Lipofectamine 3000 (Thermo Fisher Scientific) according to the manufacturer’s protocols, with plasmid DNA and lipo3000 at a 2:7 ratio in Opti-MEM (Gibco).

Cell counting kit-8 (CCK-8) for cell viability

HeLa cells were used to study cell toxicity. Briefly, HeLa cells were seeded in 96-well plates at a density of 5000 cells per well and cultured overnight. Then, the cells were incubated with various concentrations of the MTPABP-Cl in fresh media. After 24 h incubation, the media was removed and washed with PBS for 3 times. Cells were then incubated with fresh serum-free medium containing 10% CCK-8, followed by another incubation for 1 h. Finally, the absorbance at 450 nm was measured by a microplate reader. Experiments were performed by four independent replicates.

Subcellular localization of HT proteins labeled with MTPABP-Cl

For the subcellular localization imaging assay, pCDNA3.4_EGFP_Halo, pCDNA3.1(-)_EGFP_FKBP_halo, pCDNA3.4_H2B_Halo were transfected into HEK 293 T and HeLa cells using lipofectamine^TM 3000 according to the manufacturer’s protocol, respectively. After 8 h incubation, the medium was replaced, and cells were grown for an additional 24 h and washed once with 1 × PBS (pH 7.4). Cells were treated with 5 μM MTPABP-Cl for 30 min at 37 °C, 5% CO₂. Cells were washed 3 times with fresh medium containing 1 µM HT protein to scavenge the substrate. Subsequently, the cells were living cell Hoechst 33342 Staining Solution (100 × ) for 30 min at 37 °C, 5% CO₂, and washed 3 times with 1 × PBS (pH 7.4) before confocal fluorescence microscope analysis.

Confocal microscopy was performed on a Nikon Eclipse Ti2-E inverted microscope equipped with a DUVB spectral detector, and living cell imaging was performed at 37 °C with 5% CO₂ atmosphere in a humidified chamber. A 60 × /1.4 oil immersion objective was used at to image a single plane at 1024 × 1024 pixels resolution (EGFP 488 nm ex., 500–530 nm em., MTPABP-Cl 488 nm ex., 600–720 nm em., Hoechst 488 nm ex., 400–470 nm em., lyso-Track 488 nm ex., 490–530 nm em., Mito-Track 488 nm ex., 490–550 nm em.). Images were used without further processing.

Sensing calmodulin conformational changes induced by Ca²⁺

SCECaMP and GCaMP6s constructs were transiently transfected into HEK 293T and HT22 cells using lipofectamine^TM 3000 according to the manufacturer’s protocol, respectively. After 8 h incubation, the medium was replaced, and cells were grown for an additional 24 h. For imaging of Ca²⁺ concentration changes in HEK 293T cells transiently expressing SCECaMP proteins, cells were incubated with 5 μM MTPABP-Cl for 30 min in 37 °C, 5% CO₂. Cells were washed 3 times with fresh medium containing 1 µM HT protein to scavenge the substrate. For GCaMP6s imaging, HEK 293T and HT22 cells were used without MTPABP-Cl treatment. Then, the fluorescence fluctuations of MTPABP-Cl in BIOSCE during protein conformational changes were recorded following stimulation with 70 mM KCl at a flow rate of 1 mL min⁻¹, simultaneously illuminated with a 488 nm laser and imaged using confocal with a peristaltic pump under no-wash conditions.

Time series images were obtained at a frame rate of 4 frames per second using a Nikon Eclipse Ti2-E inverted microscope equipped with a DUVB spectral detector, and living cell imaging was performed at 37 °C with 5% CO₂ atmosphere in a humidified chamber. A 60 × /1.4 oil immersion objective was used at to image a single plane at 256 × 256 pixels resolution (MTPABP-Cl 488 nm ex., 600–720 nm em., GFP 488 nm ex., 500–530 nm em). Calcium imaging experiments were performed by three independent replicates. Images were used without further processing.

Monitoring protein behaviors in the rapamycin-dependent FKBP–FRB model

FKBP_HT and FRB or FKBP and FRB_HT protein co-expression constructs were transiently transfected into HeLa cells using Lipofectamine 3000 according to the manufacturer’s protocol, respectively. After 8 h of incubation, the medium was replaced, and the cells were cultured for an additional 24 h. Cells were incubated with 5 μM MTPABP-Cl for 30 min in 37 °C, 5% CO₂. Cells were washed 3 times with fresh medium containing 1 μM HT protein to scavenge the substrate. The fluorescence fluctuations of MTPABP-Cl in BIOSCE during protein conformational changes were recorded following stimulation of 100 nM RAPA at a flow rate of 1 mL min⁻¹ using confocal imaging with a peristaltic pump.

Time series images were obtained at a frame rate of 30 frames per second using a Nikon Eclipse Ti2-E inverted microscope equipped with an A1R spectral detector, and living cell imaging was performed at 37 °C with 5% CO₂ atmosphere in a humidified chamber. A 60 × /1.4 oil immersion objective was used at to image a single plane at 256 × 256 pixels resolution (MTPABP-Cl 488 nm ex., 600–720 nm em). FKPB and FRB interactions imaging experiments were performed by three independent replicates. Images were used without further processing.

Tracking SNAP25 mobility during BoNT/A cleavage

Halo_SNAP25 and SNAP25_Halo constructs were transiently transfected into HT22 cells using Lipofectamine 3000 according to the manufacturer’s protocol, respectively. After 8 h incubation, the medium was replaced, and cells were cultured for an additional 24 h. For mapping the BoNT/A cleaved SNAP25, cells were incubated with 5 μM MTPABP-Cl for 30 min in 37 °C, 5% CO₂. Then, cells were washed 3 times with fresh medium containing 1 μM HT protein to scavenge the substrate. Imaging was performed with Nikon Eclipse Ti2-E (60 × objective, 1024 × 1024 pixels) living cells workstation in the presence of 25 nM BoNT/A duration 1 h. To record the BoNT/A cleaved SNAP25 dynamic process under Zncl₂, the same methods were used transfected cells. Additionally, we imaged the 25 nM BoNT/A cleavage SNAP25-HT fusion protein and the process of SNAP25-HT fusion protein conformational changes in the presence or absence of 50 μM Zncl₂.

Time series images were obtained at a frame rate of 30 frames per second using a Nikon Eclipse Ti2-E inverted microscope equipped with a DUVB spectral detector, and living cell imaging was performed at 37 °C with 5% CO₂ atmosphere in a humidified chamber. A 60 × /1.4 oil immersion objective was used at to image a single plane at 512 × 512 pixels resolution (MTPABP-Cl 488 nm ex., 600–720 nm em). Experiments were performed by three independent replicates. Images were used without further processing.

Imaging data analysis

Image analysis was performed with Fiji (Image J)⁸³. Regions of interest (ROIs) were hand-segmented, and mean fluorescence intensities from individual ROIs were derived for multiple fields-of-views. Background intensities were measured as mean intensities from cell-free regions and subtracted for background correction. For recording of rapamycin-dependent FKPB-FRB interactions in HeLa cells, the mean intensity of prior to rapamycin stimulation defined as F₀. For recording intracellular calcium levels in HEK 293T and HT22 cells with SCECaMP, GCaMP6s, the mean intensity of prior to KCl stimulation is defined as F₀. For mapping the BoNT/A cleaved SNAP25_Halo or Halo_SNAP25 fusion protein in HT22 cells, the mean intensity of prior to Zn²⁺ stimulation is defined as F₀. ΔF was calculated as ΔF = F − F₀, where F represents the fluorescence intensity of the ROI in each frame after treatment with rapamycin, KCl, or Zn²⁺, respectively.

Fluorescence recovery after photobleaching (FRAP)

Nikon Eclipse Ti2-E with 60 × objective, oil immersion was used to perform FRAP experiments on FKBP_Halo, FRB, FRB_Halo, and FKBP proteins labeled with 5 μM MTPABP-Cl in living cells. The acquisition mode, laser power, time interval between frames, total number of frames, and other FRAP-specific settings were customized for each protein of interest (POI) such that each experiment would satisfy four criteria: (i) have sufficient signal, (ii) have sufficient duration to capture recovery, (iii) have sufficient temporal resolution, (iv) endure minimal photobleaching throughout the time course⁸⁴. For each single FRAP acquisition course, several frames were first recorded to establish pre-bleach levels of signal, followed by photobleaching with 100% laser power of a square region.

After photobleaching, fluorescence recovery was recorded over time. Raw image series were processed Fiji (Image J), followed by drift correction using the plugin of linear stack alignment with SIPT. Averaged intensity measurements from an unbleached region were further used to correct for the photobleaching occurring during the image acquisition. All FRAP experiments were performed using a total of three biologically independent samples.

Protein isolation and western blot

The HeLa cells were seeded into 12-well culture plates in complete MEM media (0.5 mL) and incubated at 37 °C under 5% CO₂ until 70–80% confluence was reached. Cells were transfected with SNAP25-HT fusion proteins using lipofectamine^TM 3000 reagent according to the manufacturer’s protocol. After 48 h, removed media in wells and washed with PBS for 3 times. Then, the cells were treated with 25 nM BoNT/A for 24 h in the presence and absence of BoNT/A, the medium was discarded, and then washed with PBS for 3 times, and then the cells were lysed with radio-immunoprecipitation assay (RIPA, Beyotime) buffer supplemented with protease and phosphatase inhibitors.

After centrifugation, the concentration of the supernatants was determined by bicinchoninic acid assay (BCA). Samples were boiled in SDS-PAGE sample buffer and analyzed by SDS-PAGE. Proteins were transferred to polyvinylidene fluoride (PVDF) membranes, blocked in blocking Buffer (Beyotime), and immunoblotted with appropriate primary and secondary antibodies. Specifically, SNAP25-HT fusion proteins were first blotted with rabbit polyclonal SNAP25 Cell Signaling Technologies, 5308, 1:1000 dilution) primary antibody at 4 °C overnight, washed with Tris-buffered saline plus 0.1% (v/v) Tween 20 (TBST) three times, then blotted with horseradish peroxidase (HRP)-labeled Goat anti-Rabbit IgG secondary antibody (Absin, abs20002, 1:2000 dilution) at room temperature for 1 h. Immunoblots were washed with TBST for three times and developed using BeyoECL Plus chemiluminescence reagent. GAPDH was used as a loading control and immunoblotted in the same method using anti-GAPDH-HRP (Beyotime, P0971M, 1:500 dilution).

Flow cytometry

Plasmids of EGFP_FKBP_halo, FRB_YFP, and FKBP_halo, FRB, FKBP, FRB_Halo were co-transfected into HeLa cells using lipofectamine^TM 3000 according to the manufacturer’s protocol, respectively. After 8 h incubation, the medium was replaced, and cells were grown for an additional 24 h in the 6-well plates. These cells were incubated with 5 μM MTPABP-Cl for 30 min in 37 °C, 5% CO₂. Then, cells were washed 3 times with fresh medium containing 1 µM HT protein to scavenge the substrate. Then, cells were washed once with 1 × PBS (pH 7.4), digestion was performed with 0.25% trypsin at 37 °C for 1 min, followed by the addition of MEM medium containing 10% FBS (3 volumes of trypsin), and centrifugation at 1000 rpm for 3 min. Cells were re-suspended in a 300-mesh cell sieve in 1 × PBS containing 2% FBS and transferred to a flow tube (5 mL FACS tube, FALCON).

Samples were subjected to the autosampler of a BD FACSAria III flow cytometer. A total of 20,000 cells were collected immediately to detect the cell population. Cell populations were gated for living (SSC-A/FSC-A) and single cells (SSC H/SSC-A). Fluorophores were recorded as follows: FITC (488 nm excitation, 530/30 nm emission), PerCP-Cy5-5 (488 nm excitation, 695/40 nm emission). Photomultiplier tube detectors and fluorescence compensation parameters were adjusted such that signal saturation and signal bleed-through were avoided.

Flow cytometry data analysis

Raw data from quantitative flow cytometry measurements were imported into FlowJo (10.10.0) and processed as follows. Gates for live cells (SSC-A/FSC-A) and single cells (SSC-H/SSC-A) were applied. Dot plots displaying entire populations or ratio density plots were generated using FlowJo’s layout editor.

Single particle tracking

Particle Tracker plugin in Fiji (ImageJ) was used to track individual particles from the videos³⁴. First, the radius, cutoff, and percentile parameters were adjusted for each frame to distinguish particles from the background. The linking range and displacement parameters were then optimized to connect particle positions across frames. All detections and links were visually verified, and trajectories spanning more than 30 frames were selected for further analysis³⁴. Dynamical analysis of the trajectories was performed with custom code in MATLAB. The mean squared displacement of protein trajectories of Halo_SNAP25 and SNAP25_Halo were determined by single particle tracking, and the apparent diffusion coefficients⁸⁴ calculations were done by custom MATLAB code.

Quantification and statistical analysis

Schemes and figures were made with Adobe Illustrator. Biochemical experiments were performed in three technical replicates. Unless stated otherwise, cell experiments were performed in three independent biological replicates, including technical replicates. For mammalian cell experiments, biological replicates are defined as different passage of cell lines. Statistical analyses were performed with GraphPad Prism 9.5.1⁸⁵. All data were presented as means ± standard error of the mean (SEM). Statistical significances (i.e., p-value) were calculated using Welch’s corrected two-tailed t-tests. The different p-values were classified as **** for p-value ≤ 0.0001, *** for p-value ≤ 0.001, ** for p-value ≤ 0.01, * for p-value ≤ 0.05, and non-significant (ns) for p-value > 0.05.

Reporting summary

Further information on research design is available in the Nature Portfolio Reporting Summary linked to this article.

Author contributions

This study was conceived by K.Z., B.H., and C.S. H.J., L.Y., B.H., and K.Z. designed the study. H.J., L.Y, and Y.Y performed experiments. J.L Provided material support. H.J. and K.Z. wrote the original draft. H.J., L.Y., Y.Y., H.L., C.S., Z.D., B.H., and K.Z review and edit the paper with comments from all authors.

Peer review

Peer review information

Communications Chemistry thanks the anonymous reviewers for their contribution to the peer review of this work. A peer review file is available.

Data availability

The data that support the findings of this study are available within the paper and the supplementary information.

Competing interests

K.Z., B.H., H.J., and L.Y. are listed as inventors of patents (A labeling probe for quantifying protein dynamics. Chinese Patent Application No. 202511320520.5) related to labeling technologies filed by Beijing Normal University. The other authors declare no competing interests.

Supplementary information

The online version contains supplementary material available at 10.1038/s42004-026-01914-x.