Nitrile-aminothiol bioorthogonal near-infrared fluorogenic probes for ultrasensitive in vivo imaging

Abstract

Bioorthogonal chemistry-mediated self-assembly holds great promise for dynamic molecular imaging in living organisms. However, existing approaches are limited to nanoaggregates with ‘always-on’ signals, suffering from high signal-to-background ratio (SBR) and compromised detection sensitivity. Herein we report a nitrile-aminothiol (NAT) bioorthogonal fluorogenic probe (CyNA_P-SS-FK) for ultrasensitive diagnosis of orthotopic hepatocellular carcinoma. This probe comprises a nitrile-substituted hemicyanine scaffold with a cysteine tail dually locked with biomarker-responsive moieties. Upon dual cleavage by tumor-specific cathepsin B and biothiols, the 1,2-aminothiol residue is exposed and spontaneously reacts with nitrile group for in situ intramolecular macrocyclization, enabling near-infrared fluorescence (NIRF) turn-on as well as self-assembly. In living male mice, such ‘cleavage-click-assembly’ regimen allows for real-time and ultrasensitive detection of small cancerous lesions (~2 mm in diameter) with improved SBR (~5) and extended detection window (~36 h), outperforming conventional clinical assays. This study not only presents NAT click reaction-based fluorogenic probes but also highlights a generic dual-locked design of these probes.

Introduction

Optical imaging enables noninvasive, real-time visualization and quantification of biomarkers at molecular levels in living subjects^1,2,3,4,5. By integrating molecular probes with imaging instruments, fluorescence imaging has become an indispensable tool in preclinical research and clinical practice^6,7,8,9. To enhance diagnostic sensitivity, activatable probes of which the fluorescence is switched on by specific stimuli have been developed to translate diseases-associated biomarkers into optical output^10,11,12,13. To date, numerous activatable probes have been constructed for the detection of enzymes^14,15,16,17, chemical mediators^18,19,20,21, and physiological indexes in living animals^{22,23,24,25,26}. Nevertheless, traditional activatable molecular probes encounter challenges associated with rapid diffusion and clearance from diseases sites, which lead to shortened imaging window, reduced signal-to-background ratios (SBRs) and eventually compromised diagnostic accuracy.

In vivo self-assembly involves the spontaneous organization of externally administered small molecules into nanostructures in response to pathological abnormalities, allowing them to accumulate in target tissues^27,28,29. This approach simultaneously capitalizes on the pharmacokinetic advantages of small molecular probes for deep-tissue penetration and broad biodistribution, as well as the benefits of nanoparticles for extended local retention at disease sites^30,31. In normal tissues, the inactive probes remain unassembled and are efficiently cleared out from living bodies^30,32. Thereby, self-assembled probes lead to improved therapeutic outcomes and imaging performance with reduced systematic toxicity. However, due to the sophisticated and dynamic physiological environment in vivo, achieving highly controllable and biocompatible self-assembly of molecular probes in target tissues remains a significant challenge.

Bioorthogonal reactions including Staudinger ligation^33,34, trans-cyclooctene/tetrazine cycloaddition^35,36 and 2-cyano-benzothiazole (CBT)/cysteine (Cys) condensation^37,38,39,40 (Fig. 1a) have been utilized to construct in vivo self-assembled systems for drug delivery, polymer synthesis, and bioimaging. Among them, CBT-Cys click reaction and its biomarker-instructed self-assembly design have given rise to a novel family of nanoprobes for imaging of pH levels, reactive oxygen species and nitrogen species, as well as enzymatic activities^{41,42,43,44,45,46}. However, existing CBT-Cys probes are intrinsically fluorescent prior to forming nanoaggregates for signal enhancement (Fig. 1c), posing a high risk of “false-positive” output and suboptimal SBRs. In contrast, bioorthogonal fluorogenic probes which are natively non-fluorescent but turn on fluorescence exclusively upon specific click reactions provides a powerful solution to address these concerns (Fig. 1b). However, a suitable fluorogenic scaffold and the biomarker-instructed probe design are currently lacking in existing studies.

Fig. 1: Design of NAT bioorthogonal NIR fluorogenic probes.

a, b Schematic of CBT-Cys click reaction and NAT click reaction based fluorogenic hemicyanine in this study. c, d Schematic illustration of the intramolecular cyclization and self-assembly mediated by CBT-Cys click reaction c and NAT click reaction d, the comparison between two strategies in terms of imaging contrast and SBR. e Molecular structures and sensing mechanism of the dual-locked activatable probe (CyNA_P-SS-FK) with GSH/Cat B-triggered intramolecular cyclization and self-assembly for non-invasive NIRF in vivo imaging of orthotopic HCC.

Nitrile-aminothiol (NAT) click reaction is a well-known condensation reaction between nitriles and 1,2-aminothiols to form thiazolines, which occurs readily under physiological conditions with exceptional specificity, fast rates and high yields^47,48,49. For example, the heteroaromatic nitriles, such as 2-cyanoisonicotinamide (CINA), 6-cyanonicotinic acid (CNA) as well as 2-cyano-6-hydroxyquinoline (CHQ) are capable of condensation with N-terminal cysteines with high reactivity and stability at biological environments, thus widely used for protein and peptide synthesis^50,51. However, its utilization as fluorogenic self-assembled probes hasn’t been explored.

Herein we report a dual-locked NAT bioorthogonal fluorogenic probe (CyNA_P-SS-FK) with biomarker-triggered near-infrared fluorescence (NIRF) turn-on and self-assembly for non-invasive diagnosis of hepatocellular carcinoma (HCC) (Fig. 1d, e). CyNA_P-SS-FK comprises a NAT bioorthogonal fluorogenic scaffold (CyNA_P) linked with an aminothiol tail that is jointly caged by glutathione (GSH)-responsive disulfide bond and a cathepsin B (Cat B)-cleavable peptide sequence (Ac-FK) (Fig. 1e). GSH and Cat B are biomarkers co-overexpressed by HCC cells and are highly associated with tumor progression, angiogenesis, and metastasis⁵². Prior to probe construction, the CyNA_X scaffolds (nitrile-substituted hemicyanines) along with the length and rigidity of the spacer chain are carefully screened to optimize click reaction and intramolecular macrocyclization kinetics. CyNA_P-SS-FK is intrinsically non-fluorescent due to the inhibited intramolecular charge transfer (ICT) by nitrile group. Upon removal of both caging groups by Cat B and GSH, the aminothiol head in CyNA_P-SS-FK is exposed to undergo intramolecular cyclization with nitrile through NAT click reaction, which unlocks ICT for the recovery of NIRF emission, and simultaneously enables self-assembly to nanoaggregates. This particular ‘cleavage-click-assembly’ sensing mechanism by CyNA_P-SS-FK allows for real-time ultrasensitive detection of small HCC (~2 mm in diameter) with improved SBR (~5) and prolonged detection window (~36 h), which outperforms the control ‘always on’ probe and clinical assays.

Results

Rational design of NAT bioorthogonal fluorogenic luminophores

Asymmetric hemicyanines with a donor-π-acceptor (D − π − A) motif are frequently used to construct activatable probes in response to various biomarkers^{17,53,54,55,56,57,58}. Its sensing mechanism relies on adjusting the ICT between electron donating phenol group and the electron-withdrawing indole group on the hemicyanine scaffold. When the phenol group is caged with a biomarker-responsive moiety, the electron donating ability is significantly reduced, causing ICT blocking and fluorescence quenching. Upon the removal of caged moieties by specific targets, the probes are activated to ‘turn on’ fluorescence^59,60,61. Given the importance of electron ‘push-pull’ effects in hemicyanine scaffold, we proposed that replacing the electron-donating hydroxyl group with a strong electron-withdrawing nitrile group could block ICT and ‘turn-off’ NIRF signal (Fig. 2a), which could then be restored when nitrile undergoes NAT click reaction to afford thiazoline, as the electron density is dramatically changed.

Fig. 2: Synthesis and screening of NAT bioorthogonal fluorogenic luminophores.

a Chemical structures and sensing mechanisms of nitrile-substituted hemicyanines with free cysteine via NAT click reaction. b Synthetic routes for CyN-1−5. Reagents and conditions: (i) 2-bromocyclohex-1-ene-1-carbaldehyde, Cs₂CO₃, DMF, rt, 12 h. (ii) 1,2,3,3-tetramethyl-3H-indolium iodide, AcONa, Ac₂O, 60 °C, 2−4 h. (iii) vilsmeier haack reagent, EtOH, reflux, 4 h. (iv) 5-formyl-2,4-dihydroxybenzonitrile, K₂CO₃, DMF, 50 °C, 12 h. (v) Cs₂CO₃, CH₃I, ACN, rt, overnight. c The second order rate constant (M⁻¹ s⁻¹) and fluorescence enhancement of CyN-1−5 in the presence of free cysteine. d Fluorescence spectra of CyN-1−5 (20 μM) in the absence or presence of Cysteine (100 μM) in PBS (containing 20% MeOH, pH = 7.4) at room temperature. The fluorescence excitation was set at 580 nm. The experiments in d were repeated independently three times with similar results. Source data are provided as a Source Data file.

To prove the proposed concept, we strategically designed and synthesized a series of nitrile-substituted hemicyanines (termed CyN-X in Fig. 2b and Supplementary Figs. 1, 2). Since the reactivity of the aromatic nitrile is affected by π electron density and steric properties, we synthesized several compounds with different nitrile groups at C-6 (CyN-1–2) and C-7 (CyN-3–5) positions to identify the optimal site for click reaction. As shown in Fig. 2b, the synthesis of CyN-1–3 involved two simple steps: firstly, each benzonitrile derivative was condensed with 2-bromocyclohex-1-ene-1-carbaldehyde in presence of Cs₂CO₃ as base to give compound 2; secondly, the resulted aldehyde was directly condensed with indole quaternary ammonium salt to yield the corresponding fluorophores (CyN-1–3). Differently, CyN-4–5 was obtained by treating a chloro-substituted cyanine with resorcin derivates in the presence of a base. Briefly, Vilsmeier Haack reagent was reacted with an indole quaternary ammonium salt in refluxing ethanol to obtain the chloro-substituted cyanine 4. Resorcin derivative as a nucleophile was substituted on the meso position, resulting in the formation of product CyN-4 via Retro-Knoevenagel reaction. Finally, methylation of CyN-4 by iodomethane yielded CyN-5.

To evaluate the reactivity of nitrile-substituted hemicyanines with 1,2-aminothiols, their optical profiles and the second-order rate constants were determined after the addition of free cysteine in phosphate-buffered saline (PBS) solution. For CyN-1 and CyN-3 with different nitrile group positions, the fluorescence signals were enhanced by 4.1- and 2.7-fold, respectively (Fig. 2c, d and Supplementary Fig. 3). This observation was consistent with previous studies suggesting that the C-6 position is preferable for the electron-donating group of hemicyanine to achieve optimal fluorescence activity⁶². Moreover, CyN-1 has a higher second-order rate constant of 0.309 M⁻¹ s ⁻¹ compared to that of 0.221 M⁻¹ s ⁻¹ for CyN-3 (Supplementary Fig. 4). The proposed mechanism of intermolecular cyclization involves an initial nucleophilic attack by the sulfhydryl group on the nitrile group to produce a thioimidate intermediate (Fig. 2a), followed by condensation reaction with amine to yield the 2-aminothiazolidine intermediate. Finally, ammonia is irreversibly eliminated and results in the formation of the thiazoline product. The detection of ESI-MS spectrum (497.34, 554.35 and 497.36 for CyNC-1−3, respectively) in Supplementary Fig. 5 confirmed the intermolecular cyclization.

While a reduction in electronegativity on the aromatic nitrile would theoretically accelerate the nitrile-aminothiol reaction, a comparison between CyN-1 and CyN-2 reveals that adding one halogen atom on the ortho position significantly decreases the reactivity of aromatic nitrile (0.151 M⁻¹ s⁻¹, Supplementary Fig. 4) and reduces fluorescence enhancement (~1.7-fold). This is likely due to the steric hindrance caused by chlorine atom, which retards the intramolecular cyclization. The reactivity of a hydroxyl-substituted (CyN-4) or an alkoxy-substituted aromatic nitrile (CyN-5) with cysteine was also tested. Since hydroxyl- and alkoxy-substituents offer electron-donating effects, their reactivities are extremely low, as evidenced by negligible fluorescence enhancement (Fig. 2c). Based on these structure-activity studies, CyN-1 stands out with the fastest reacting kinetics and the highest fluorescence enhancement. We herein selected CyN-1 as the optimal bioorthogonal fluorogenic luminophores for the subsequent studies.

Screening of intramolecular NAT bioorthogonal fluorogenic luminophores

To construct and screen intramolecular NAT reaction mediated fluorogenic luminophores, we synthesized luminophores CyNA_X including CyNA₄, CyNA₇ and CyNA_P using CyN-1 with a N-terminal cysteine residue attached to either an alkyl chain or a short PEG chain linked to the indole ring (Fig. 3a). To investigate how the length and rigidity of the chain affect intramolecular macrocyclization kinetics, we varied the lengths of the alkyl or PEG chains (Fig. 3a). The synthetic routes of CyNA_X were shown in Fig. 3b and Supplementary Fig. 6. Briefly, the commercially available raw material potassium phthalimide as amine source was used to convert a serious of dibromo derivatives into the corresponding phthalimide alkylation species 6, which then reacted with the indole to obtain compound 7. The resulted indole derivate 7 was condensed with aldehyde 2a to give hemicyanine derivative 8, followed by deprotection in the presence of N₂H₄•H₂O to yield the primary amine 9. This primary amine was then conjugated with a protected cysteine and then deprotected trityl group to afford the corresponding probe CyNA_X. The chemical structures of the intermediates and final products were characterized by NMR spectroscopy and ESI-MS (see the Supplementary Information).

Fig. 3: Synthesis and screening of CyNA_X for intramolecular cyclization.

a Chemical structures and intramolecular condensation via NAT click reaction. b Synthetic routes for CyNA_X. Reagents and conditions: (i) dibromo derivative, KI, Acetone, reflux, overnight. (ii) 2,3,3-trimethyl-3H-indole, ACN, reflux, 24 h. (iii) compound 2a, AcONa, Ac₂O, 60 °C, 0.5 h. (iv) N₂H₄•H₂O, DCM/MeOH, rt, 2 h. (v) Boc-Cys(Trt)-OH, HBTU, DIPEA, DCM, rt, 2 h. (vi) Et₃SiH, TFA (20% in DCM), rt, overnight. c Molecular orbitals of CyNA_X and their cyclized products CyNA_X-C. The HOMO and LUMO energy levels are indicated. d Hole (red color) and electron (blue color) of optimized excited-state geometries of CyNA_P and CyNA_P-C calculated by TD-DFT method at B3LYP/6-311 G (d, p) level with Grimme’s DFT-D3 correction. e Representative confocal microscopy images of LM3 cells incubated with CyNA_X (30 μM) for 0, 3, 6, 12 and 24 h, respectively. Blue and red fluorescence indicate a cell nucleus stained with DAPI and the signals from CyNA_X, respectively. (Scale bar, 10 μm). f Fluorescence enhancement and condensation reaction rate constants of CyNA_X (start materials, SM) in the presence of 1,2-aminothiols. Log D values of CyNA_X and their cyclized products CyNA_X-C (terminal materials, TM). g Quantification of fluorescence intensity in the panel e. (n =  3, mean ± s.d.). Two-tailed Student’s t test; Control versus probe-treated groups. NS, no statistically significant differences. The experiments in g were repeated independently three times with similar results. h Calculated energy levels of single and triplet states for distinct luminescence behaviors of the CyNA_P and CyNA_P-C molecules. Green arrows represent absorption, red arrows indicate emission, and grey wavy arrows depict nonradiative quenching processes. i Fluorescence spectra of CyNA_X in PBS (containing 20% MeOH, pH = 7.4) at room temperature. The fluorescence excitation was set at 580 nm. The experiments in i was repeated independently three times with similar results. Source data are provided as a Source Data file.

To screen the optimal CyNA_X for intramolecular condensation, their optical profiles and the first-order rate constant were determined in PBS and MeOH solution (Fig. 3f, i and Supplementary Fig. 7). At pH 7.4, both CyNA₄ and CyNA₇, which have alkyl chain substitutions, produced cyclized products, as confirmed by ~4-fold fluorescence enhancement at ~675 nm and the detection of a MALDI-TOF MS peak (536.24 and 578.29) (Supplementary Fig. 8), which corresponded to the mass of cyclized products (CyNA₄-C and CyNA₇-C). The first-order reaction rate constants were determined to be 2.91 × 10⁻⁵ s⁻¹ for CyNA₄ and 0.89 × 10⁻⁵ s⁻¹ for CyNA₇ by high-performance liquid chromatography (HPLC) (Supplementary Figs. 9, 10). In contrast, CyNA_P exhibited a higher fluorescence enhancement (~5.1-fold) and faster kinetics (7.97 × 10⁻⁵ s⁻¹), implying that the flexible PEG linker can promote intramolecular macrocyclization. To provide a theoretical basis for understanding the effects of intramolecular cyclization on their fluorescence enhancement, we calculated the lowest unoccupied molecular orbital (LUMO) and highest occupied molecular orbital (HOMO) energies as well as their spatial distributions in CyNA_X and CyNA_X-C using density functional theory (DFT) (Fig. 3c). For CyNA_X, the LUMOs were distributed on hemicyanine scaffold while the HOMOs were found on the cysteine-containing aminothiol group. In contrast, for the cyclized products CyNA_X-C, both LUMOs and HOMOs were uniformly distributed on hemicyanine-thiazoline backbone, validating the effectiveness of the intramolecular cyclization design in achieving fluorescence ‘Turn-On’. To elucidate the underlying causes of the marked differences in fluorescence between CyNA_P and CyNA_P-C, we embarked on a computational analysis of their emission mechanisms via time-dependent density functional theory (TD-DFT) at the B3LYP/6-311 G (d, p) level with Grimme’s DFT-D3 correction. For CyNA_P, as shown in Fig. 3d the separation between the hole and electron distributions is large, resulting in low electron transport efficiency. Additionally, the predicted optical gap for CyNA_P is 2.20 eV (f = 0.09), which can be ascribed to the S₀ − S₂ transition, rather than the S₀ − S₁ transition with a negligible oscillator strength (f = 0.00) (Fig. 3h). In contrast, in CyNA_P-C, the hole and electron distributions are closer together, facilitating better electron transport. Furthermore, CyNA_P-C exhibits a clear S₀ − S₁ absorption at 2.20 eV (f = 0.03), making it more likely to emit fluorescence from the S₁ state.

To evaluate the optical performance of probes CyNA_X in cells, confocal microscopy was performed on CyNA_X-treated human HCC-LM3 cells. When LM3 cells were incubated with CyNA_X, gradually enhanced fluorescence was observed (Fig. 3e). CyNA₄ and CyNA₇ showed slight fluorescence enhancement over time. In contrast, CyNA_P turned on its fluorescence more rapidly. Fluorescence signal of CyNA_P in LM3 cells was 3.4-fold and 3.6-fold higher than that of CyNA₄ and CyNA₇, respectively (Fig. 3g), at 3 h post-incubation. This could be attributable to CyNA_P’s faster cyclization rate.

‘Click-Assembly’ and specificity of intramolecular condensation

Given that the endogenous free cysteine in cells and plasma ranges from 20 to 100 μM, the specificity of intramolecular condensation was studied in the presence of high concentration of free cysteine. This was done because locally free cysteine may also undergo intermolecular condensation with aromatic nitrile (Fig. 4a and Supplementary Fig. 11). To investigate this, CyNA_P was co-incubated with varying mole ratios of free cysteine, and the condensation products were traced by HPLC. As shown in Fig. 4b, incubating CyNA_P and cystine at a 1:1 molar ratio in PBS resulted in the rapid conversion of ~100% CyNA_P to CyNA_P-C, as confirmed by the detection of an HPLC peak with a retention time of about 20.6 min, corresponding to the intramolecular cyclized products (CyNA_P-C). This high specificity was also observed at a 1:100 and even 1:1000 molar ratios of CyNA_P to cystine, demonstrating that intramolecular condensation occurs preferentially over intermolecular condensation. This was ascribed to the first-order reaction rate constants (7.97 × 10⁻⁵ s⁻¹) of CyNA_P enables promoting intramolecular cyclization preferentially, consistent with previous studies⁴². To evaluate whether the intramolecular condensation can induce the formation of nanoaggregates, we used transmission electron microscopy (TEM) and dynamic light scattering (DLS) assays to analyze NAT click reaction mediated self-assembly of CyNA_P. As shown in Fig. 4c, CyNA_P was converted to the cyclized product CyNA_P-C in PBS solution (pH 7.4), which spontaneously assembled into nanoparticles with a spherical morphology. Furthermore, the average hydrodynamic diameter for CyNA_P-C was measured to be ~100 nm (Fig. 4d). The self-assembly was attributed to the increased rigidity and high hydrophobicity of CyNA_P-C, reflected by a significantly higher Log D value (3.33) relative to CyNA_P (Log D = 2.88; Fig. 3f).

Fig. 4: ‘Click-Assembly’ and specificity of intramolecular condensation.

a Competition studies between intramolecular and intermolecular condensation. b A competitive HPLC assay of CyNA_P incubated with different molar ratio of free cysteine. c Representative TEM images of CyNA_P-SS before and after incubation with GSH in buffer solution (scale bar, 200 nm). d Average diameters of CyNA_P-SS before and after incubation with GSH in buffer solution by DLS. e Chemical structures and condensation mechanisms of CyNA_P-SS and CyNA_P-S in the response of GSH. f Representative confocal microscopy images of LM3 cells incubated with CyNA_P-SS (20 μM) and CyNA_P-S (20 μM) at 37 °C for 8 h in the presence or absence of the GSH synthesis inhibitor BSO. NIRF images acquired at excitation at 594 nm. g Quantification of fluorescence intensity in the panel f. Levels of fluorescence intensity at different groups (n =  5, mean ± s.d.). (Scale bar, 5 μm). Two-tailed Student’s t test; Control versus probe-treated groups. NS, no statistically significant differences. The experiments in b–d and f were repeated independently three times with similar results. Source data are provided as a Source Data file.

To further investigate the specificity of CyNA_P self-cyclization in cells, we synthesized a GSH-responsive probe CyNA_P-SS containing disulfide bond and a control probe CyNA_P-S containing a single sulfide bond insensitive to GSH (Fig. 4e and Supplementary Figs. 12, 13). We hypothesized that the disulfide bond of CyNA_P-SS would be reduced by cellular GSH to expose 1,2-aminothiols residues, which would then react with nitrile groups to form fluorescent cyclized products and self-assemble into nanoparticles. In contrast, CyNA_P-S would not undergo self-cyclization as it is unable to be reduced by GSH (Fig. 4e). To test this, CyNA_P-SS or CyNA_P-S was respectively incubated with LM3 cells for 8 h before confocal fluorescence imaging. As shown in Fig. 4f, NIRF signal of CyNA_P-SS was clearly observed in the cytoplasm of LM3 cells, indicating the formation of fluorescent cyclized products. However, in LM3 cells pre-treated with a GSH synthesis inhibitor buthionine sulfoximine (BSO)⁶³, NIRF signals from CyNA_P-SS decreased to basal levels. Notably, almost no fluorescence was observed in cells incubated with the control probe CyNA_P-S, which was 2.5-fold lower than that of CyNA_P-SS (Fig. 4g and Supplementary Fig. 14). These results confirmed that CyNA_P had a high specificity for intramolecular condensation, even in the presence of endogenous free cysteine in tumor cells.

Design and characterization of dual-locked fluorogenic probes

Encouraged by the robustness of intramolecular cyclization and self-assembly of CyNA_P, we developed a dual-locked activatable probe (CyNA_P-SS-FK) for NIRF imaging and companion diagnosis of orthotopic HCC in living mice. CyNA_P-SS-FK comprises three key units: a NAT bioorthogonal fluorogenic chromophore (CyNA_P), a GSH-responsive disulfide bond and a cathepsin B (Cat B)-reactive peptide moiety (Ac-FK) (Fig. 5b). GSH and Cat B are biomarkers co-overexpressed by HCC cells, promoting tumor growth, angiogenesis, and metastasis^56,64. CyNA_P-SS-FK is intrinsically non-fluorescent and only activate its NIRF signals in the presence of both GSH and Cat B. This ‘dual-locked’ design ensures that the NIRF signal activation is specific to tumor cells but not normal tissues. To rule out ‘cleavage-click-assembly’ induced long-term retention effect in tumor imaging, we further synthesized CyNA_P-T as an ‘always on’ probe for comparison (Supplementary Fig. 15). CyNA_P-T has a similar chemical structure to the cyclized product CyNA_P-C, but differs in that the macrocyclic ring is open. The synthetic route of CyNA_P-SS-FK is shown in Fig. 5a and Supplementary Fig. 16. First, nitrile-substituted hemicyanine 9c was synthesized and then coupled with the protected cysteine derivate to give CyNA_P-SS, which was further conjugated with the protected Cat B substrate (Acetyl-Phe–Lys-O^tBu) to obtain CyNA_P-SS-FK. All the intermediates and probes were characterized by ¹H NMR spectra, LC-MS spectra or High-resolution mass spectrum.

Fig. 5: Design and characterization of dual-locked fluorogenic probes.

a The synthesis route of CyNA_P-SS-FK. b Activation mechanism of the dual-locked molecular probe CyNA_P-SS-FK for LM3 cell detection. c Fluorescence spectra of CyNA_P-SS-FK (20 µM) after incubation with/without GSH, Cat B, or GSH and Cat B at 37 °C in PBS buffer (pH 7.4). Fluorescence excitation at 580 nm. d Fluorescence intensity of CyNA_P-SS-FK (40 µM) after incubation with the indicated enzymes, metal ions (60 µM), ROS (60 µM) in PBS (10 mM, pH 7.4) at 37 °C. Mg²⁺, magnesium chloride; Ca²⁺, calcium chloride; ClO^-, sodium hypochlorite; H₂O₂, hydrogen peroxide, SO₄^2-, sodium sulfate; GSH, glutathione; TCEP, Tris(2-carboxyethyl) phosphine hydrochloride; Cas-3, caspase 3; Galactosidase; Cat B, Cathepsin B. (n =  3, mean ± s.d.). Two-tailed Student’s t test; GSH and Cat B treated group versus other groups. NS, no statistically significant differences. e Mean NIRF intensities of HCC-LM3 after incubation with CyNA_P-SS-FK (with or without BSO) in the panel h (n =  4, mean ± s.d.). n refers to cell imaging experiment repeat times. (Scale bar, 5 μm). Two-tailed Student’s t test; Control versus different incubation time. NS, no statistically significant differences. f Representative TEM image and DLS result of CyNA_P-SS-FK before and after incubation with GSH and Cat B in buffer solution (scale bar, 150 nm). g Schematic illustration of the sensing mechanism of CyNA_P-SS-FK in living cells. h Representative confocal microscopy images of LM3 cells incubated with CyNA_P-SS-FK (25 μM) for 0, 4, 8 and 12 h (with or without BSO), respectively. Blue and red fluorescence indicate a cell nucleus stained with DAPI and the signals from CyNA_P-SS-FK, respectively. NIRF images acquired at excitation at 594 nm. The experiments in c, d, f and h were repeated independently three times with similar results. Source data are provided as a Source Data file.

To investigate the responses of CyNA_P-SS-FK towards GSH and Cat B, the optical properties were characterized in the absence or presence of these biomarkers. CyNA_P-SS-FK initially showed one absorption peak at 550 nm (Supplementary Fig. 17 and was nonfluorescent with a low fluorescent quantum yield of about 0.8%. After incubation with GSH or Cat B alone, CyNA_P-SS-FK remained nonfluorescent as a single cleavage by GSH or Cat B was insufficient to induce NAT reaction-mediated intramolecular cyclization (Fig. 5c). However, when incubated with both GSH and Cat B, the NIR fluorescence of CyNA_P-SS-FK was turned on, showing a ~5-fold increase at 680 nm with an increased fluorescence quantum yield of ~4% (Fig. 5c). DLS and TEM analyses of CyNA_P-SS-FK following incubation with GSH and Cat B further confirmed the intramolecular cyclization and in-situ formation of nanoaggregates with an average size of around 100 nm (Fig. 5f). Moreover, no significant NIRF change was observed in the presence of other analytes including reactive oxygen species, reductant, and enzymes, confirming its high stability and specificity (Fig. 5d). Given its minimal cytotoxicity (Supplementary Fig. 18), the capability of CyNA_P-SS-FK to detect HCC LM3 was investigated. When LM3 cells were incubated with CyNA_P-SS-FK, red fluorescence gradually increased with the incubation time (Fig. 5h). In contrast, minimal NIRF signals were observed in LM3 cells pre-treated with the GSH synthesis inhibitor BSO. Quantification results showed that NIRF signal in LM3 cells was 3.9-fold higher than in cells pretreated with BSO. These results confirmed that the NIRF signals of CyNA_P-SS-FK can be specifically activated by endogenous GSH and Cat B in LM3 cells. The stability of CyNA_P-SS-FK in circulation after injection was also confirmed (Supplementary Fig. 19).

Real-time and in vivo NIRF imaging of HCC

Early diagnosis of cancer is crucial for the treatment and accurate prognosis assessment. However, conventional diagnostic methods often struggle with low specificities and sensitivities, particularly for tumors smaller than 2 mm. The ability of CyNA_P-SS-FK for real-time NIRF imaging of HCC was evaluated in a murine orthotopic HCC model, established according to previous studies⁶⁰. At different time points post-tumor implantation (0, 5, 10, and 15 days), CyNA_P-SS-FK or the control probe CyNA_P-T was intravenously injected into living mice, followed by longitudinal NIRF imaging (Fig. 6a, b). NIRF imaging failed to detect the liver in control mice (0 day) throughout the entire imaging course (Fig. 6c). However, at 5 days post tumor implantation when tumor diameter was ~2 mm (Supplementary Fig. 20), the liver was clearly delineated, suggesting the activation of CyNA_P-SS-FK by upregulated Cat B and GSH during tumor growth. The NIRF signals in the tumor regions increased gradually, peaking at 12 h post-injection of CyNA_P-SS-FK and showing a 1.9-fold signal increase compared to those from healthy control mice. Similar signal patterns were observed at 10- and 15-day post tumor implantation, with maximum tumor signals 2.7 and 3.6 times higher than those from control mice (Fig. 6d), indicating further elevated Cat B and GSH levels.

Fig. 6: Real-time and in vivo NIRF imaging of orthotopic HCC.

a Chemical structure of CyNA_P-SS-FK and the control probe CyNA_P-T. b Schematic illustration of the development of orthotopic liver tumor model and NIRF imaging at different post tumor implantation time points. c Representative NIRF images of living mice after injection of CyNA_P-SS-FK or CyNA_P-T at different post tumor implantation timepoints. The white circles indicate the liver (Li) or bladder (Bl). NIRF images acquired at 710 nm upon excitation at 660 nm. d The dynamic NIRF intensities of liver in living mice as a function of time after injection of CyNA_P-SS-FK or CyNA_P-T for different groups (n =  3, mean ± s.d.). Two-tailed Student’s t test; Day-0 group versus different tumor incubation time groups. NS, no statistically significant differences. e, f Representative confocal fluorescence microscopy images of regional liver slices and photomicrographs of haematoxylin and eosin staining in paraffin-embedded liver sections from different mice groups. The blue, green and red signals originate from 4,6-diamidino-2-phenylindole (DAPI), Cat B antibody staining and CyNA_P-SS-FK, respectively (scale bar, 200 µm). NIRF images acquired at excitation at 594 nm. Each experiment was repeated at least for 3 times. g The IVIS imaging images of the organs of mice after injection of CyNA_P-SS-FK about 12 h. Data are the mean ± SD. n = 3 independent experiments. h Quantified mean fluorescence intensity (MFI) in Fig. 6c (n =  3, mean ± s.d.). Two-tailed Student’s t test; Day-0 group versus different tumor incubation time groups. *p < 0.05, **p < 0.01, ***p < 0.001, NS, no statistically significant differences. i-j Measurements of liver function in the mouse model of orthotopic liver cancer. Levels of ALT and AST at different groups (n =  5, mean ± s.d.). Two-tailed Student’s t test; Day-0 group versus different tumor incubation time groups. NS, no statistically significant differences. k Quantified SBRs of CyNA_P-SS-FK or CyNA_P-T at 12 h post-injection time for different groups. Levels of SBRs at different groups (n =  3, mean ± s.d.). l Schematic illustration of detection timeline of HCC comparing CyNA_P-SS-FK to the clinical and preclinical assays. m Comparison of the retention effect and SBRs of CyNA_P-SS-FK or CyNA_P-T in entire tumor imaging window period. Source data are provided as a Source Data file.

Tumor imaging ability of control probe CyNA_P-T was also studied. After 12 h post-injection, NIR signals of CyNA_P-T in the liver were 2.6-fold lower than those of CyNA_P-SS-FK. Similarly, ex vivo imaging showed that with prolonged tumor burden, the NIFR signal increased accordingly (Fig. 6g). The immunofluorescence studies also showed the NIRF signal of CyNA_P-SS-FK in the liver was ∼2-fold higher than that of CyNA_P-T (Fig. 6e, h and Supplementary Fig. 21). The SBR of CyNA_P-SS-FK was ~5-fold higher than that of the ‘always-on’ probe CyNA_P-T, owing to its active sensing mechanisms (Fig. 6k). Furthermore, at 15-day post tumor implantation, the NIRF signal of mice injected with CyNA_P-SS-FK remained high for over 36 h, showing prolonged tumor retention and an extended imaging window (Fig. 6d, m). In contrast, the NIRF signal from CyNA_P-T rapidly declined and became very weak at 6 h post-injection (Fig. 6d, m), attributed to its rapid diffusion and clearance in vivo. These findings suggested that the intramolecular cyclization and self-assembly of CyNA_P-SS-FK through NAT bioorthogonal reaction enable an enhanced permeability and retention (EPR) effect, thereby allowing prolonged and high-contrast tumor imaging in vivo.

The measurement of serum ALT and AST is routinely used for of liver function assessment in clinics. We thus compared the detection ability of CyNA_P-SS-FK with serum ALT/AST assays in our orthotopic HCC model. A statistically significant increase in ALT and AST was observed at 15-day post tumor implantation when tumor diameter reaches up to 6.5 mm (Fig. 6i, j), which was 3.3- and 3.9-fold higher than control groups, respectively. We also conducted measurements of kidney function in the mouse model to verify CyNA_P-SS-FK could not impair the kidney function (Supplementary Fig. 22). Histological studies showed a small tumor area and a noticeable tumor area were observed at 5- and 10-day post tumor implantation, respectively (Fig. 6f), and no obvious changes of other main organs were found. (Supplementary Fig. 23). The first statistically significant change in CyNA_P-SS-FK-NIRF signals was detected at least 10 days earlier than with ALT/AST assays and 5 days earlier than histological assays (Fig. 6l). These findings demonstrate that CyNA_P-SS-FK not only enables early and sensitive detection of HCC, but also provides improved imaging contrast and a prolonged detection window.

Discussion

A substantial challenge facing diagnosis of diseases via optical imaging is a lack of fluorogenic probes that possess fast intramolecular condensation in vivo to prolong the imaging window periods and enhance SBRs. In situ self-assembly strategy via biorthogonal chemistry allow probes to have longer residence time in the disease site but faster excretion of unwanted small molecules from other organs. Unlike other bioorthogonal reactions such as copper-catalyzed azide-alkyne cycloaddition and palladium-mediated Alloc/Poc cleavage that required metal catalysts, NAT click reaction is readily proceed under physiological conditions without the need of any added catalyst and has been shown to be exquisitely specificity, fast reaction rates and high-yielding. Although CBT-Cys click reaction (analogical to NAT) mediated self-assembly has widely reported, those CBT-Cys probes and their resulting nanoaggregates have ‘always-on’ imaging signals and lack fluorogenicity, giving a relatively low SBR.

To tackle those challenges, we developed NAT bioorthogonal fluorogenic probes for real-time and noninvasive in vivo NIRF imaging. To identify the suitable fluorogenic luminophores, we synthesized a series of nitrile-substituted hemicyanines and screened for the optimal structure, CyN-1, which demonstrated the fastest click reaction rate and best fluorogenicity. We then systematically analyzed the structure–activity relationship of aromatic nitrile in CyN-1 and the aminothiol tail linked through varying spacers, identifying CyNA_P with the highest intramolecular cyclization kinetics and the topmost fluorescence recovery. As the reaction rate constant of CyNA_P’s intramolecular cyclization is orders of magnitude higher than that of its intermolecular reaction with free cysteine, intramolecular condensation preferentially occurs and results in spontaneous self-assembly into nanoaggregates, rather than undergo intermolecular condensation. We further developed a dual-locked activatable probe (CyNA_P-SS-FK) by caging the aminothiol tail of CyNA_P with both GSH-responsive moiety and Cat B-cleavage peptide. In HCC tumor cells with aberrantly high levels of GSH and Cat B, CyNA_P-SS-FK was specifically activated to initiate intramolecular cyclization, igniting NIR fluorescence and simultaneously self-assembling into nanoparticles (~100 nm) for enhanced SBR and prolonged probe retention. Such ‘cleavage-click-assembly’ sensing mechanisms allowed CyNA_P-SS-FK to real-timely and noninvasively imaging orthotopic HCC in living mice with improved SBR (~5, Supplementary Table 1) and prolonged imaging window period, which was unattainable for the control ‘always-on’ and non-cyclization probe CyNA_P-T or the clinical approach the serum ALT/AST assay.

In summary, this study presents a bioorthogonal NIR fluorogenic probe utilizing NAT click reaction for in vivo imaging of HCC. Compared to broadly studied CBT-Cys probes with ‘always-on’ signals, NAT fluorogenic probes with ‘cleavage-click-assembly’ sensing mechanism offers enhanced imaging contrast and prolonged detection window. By simply varying biomarker-responsive moieties, this molecular design can be employed to create a versatile toolkit for ultrasensitive early diagnosis of various diseases beyond HCC.

Methods

General information

¹H NMR and ¹³C NMR spectra were recorded on a Bruker AVANCE III 400 MHz (or 500 MHz) spectrometer (400 MHz or 500 MHz for ¹H NMR, 100 MHz or 125 MHz for ¹³C NMR). Data were reported as follows: chemical shift, integration, multiplicity (s = singlet, d = doublet, t = triplet, q = quartet, m = multiplet, br = broad), coupling constants (Hz), and assignment. Low-resolution mass spectrometry (LRMS) data were recorded on Waters SQD2 with Waters H-Class UPLC. High-resolution mass spectrometry (HRMS) data were recorded on Waters with UPLC quattro premier XE. MALDI-MS spectrometric analyses were recorded at the Bruker ultraflex TOF/TOF instrument. Preparative HPLC was performed on a Shimadzu LC-16P system equipped with an Ascentis® C18 HPLC Column (25 cm × 21.2 mm, 5 μm) using gradient elution. The analytical HPLC was performed on a dual pump Shimadzu LC-16 system equipped with an Ascentis® C18 HPLC column (4.6 × 250 mm, 5 μm) using methanol/water (containing 0.1% Formic acid) as the mobile phase at a flow rate of 1 mL/min. DLS (Dynamic light scattering) experiment was operated on the Dynapro Plate Reader II.

Reactions were monitored by thin-layer chromatography (TLC, silica gel HSGF 254, 0.25 ± 0.02 mm). A silicone oil bath was used for experiments requiring heating. Reaction products were purified by column chromatography on silica gel (Qingdao Haiyang Chemical, zcx-II, 200−300 or 300−400 mesh). Chemical reagents and solvents were purchased from Aladdin Chemicals, Innochem, Energy Chemical and Bidepharm and used as received.

General procedure for the synthesis of 2a–c

A solution of 1 (0.1 mmol, 1.0 equiv.), 2-bromocyclohex-1-enecarbaldehyde (1.2 equiv.) and Cs₂CO₃ (3.0 equiv.) in DMF (2 mL) was stirred at room temperature overnight⁶⁵. The reaction mixture was quenched with H₂O and extracted with EA for three times. The combined organic layer was dried over Na₂SO₄ and concentrated under reduced pressure to afford the 2, which was used without further purification for the next step.

The procedure for the synthesis of 1,2,3,3-tetramethyl-3H-indol-1-ium iodide

A solution of 2,3,3-Trimethylindolenine (10.0 mmol, 1.59 g) and CH₃I (50.0 mmol, 7.1 g) was stirred at 50 °C for 4 h⁵³. After being cooled to room temperature, the reaction mixture was washed with Et₂O for three times and the resulting residue was dried under reduced pressure to afford 1,2,3,3-tetramethyl-3H-indol-1-ium iodide (2,71 g, 9.0 mmol, 90% yield) as pink solid. ¹H NMR (400 MHz, DMSO-d₆) δ 7.92–7.90 (m, 1H), 7.84–7.82 (m, 1H), 7.70–7.54 (m, 2H), 3.98 (s, 3H), 2.77 (s, 3H), 1.53 (s, 6H).

General procedure for the synthesis of CyN-(1-3)

A solution of 2 (0.2 mmol, 1.0 equiv.), 1,2,3,3-tetramethyl-3H-indol-1-ium iodide (90.5 mg, 1.5 equiv.), AcONa (65.6 mg, 4.0 equiv.) in Ac₂O (4 mL) was stirred at 60 °C for 2 h. After being cooled to room temperature, the reaction mixture was quenched with H₂O, and extracted with DCM for three times. The combined organic layer was dried over Na₂SO₄, the filtrated was concentrated under reduced pressure. The purification using column chromatography on silica gel (eluted with DCM/MeOH = 20/1) was performed to afford the CyN-X.

CyN-1

Purplish red solid, 51.9 mg, 66% yield. ¹H NMR (400 MHz, Chloroform-d): δ 8.60 (d, J = 15.2 Hz, 1H), 7.55–7.50 (m, 4H), 7.45–7.43 (m, 2H), 7.39–7.37 (m, 1H), 7.04 (d, J = 15.2 Hz, 1H), 6.93 (s, 1H), 4.27 (s, 3H), 2.94 (t, J = 6.0 Hz, 2H), 2.73 (t, J = 6.4 Hz, 2H), 2.00–1.94 (m, 2H), 1.84 (s, 6H). ¹³C NMR (125 MHz, Chloroform-d): δ 180.0, 156.5, 152.5, 146.2, 142.5, 142.0, 134.7, 129.7, 128.9, 128.3, 127.9, 127.7, 126.1, 125.8, 122.5, 119.0, 117.8, 117.1, 114.0, 113.3, 51.5, 35.9, 29.9, 29.8, 28.0, 25.0, 20.1. LRMS (m/z): [M]⁺ calcd. for C₂₇H₂₅N₂O⁺ 393.20, found 393.37.

CyN-2

Purplish red solid, 48.7 mg, 57% yield. ¹H NMR (500 MHz, Chloroform-d) δ 8.55 (d, J = 15.5 Hz, 1H), 7.56–7.51 (m, 4H), 7.42 (s, 1H), 7.37 (s, 1H), 7.09 (d, J = 15.5 Hz, 1H), 6.81 (s, 1H), 4.30 (s, 3H), 2.95 (t, J = 6.0 Hz, 2H), 2.72 (t, J = 6.0 Hz, 2H), 1.97–1.93 (m, 3H), 1.84 (s, 6H). LRMS (m/z): [M]⁺ calcd. for C₂₇H₂₄ClN₂O⁺ 427.16, found 427.44.

CyN-3

Purplish red solid, 55.0 mg, 70% yield. ¹H NMR (500 MHz, Chloroform-d): δ 8.60 (d, J = 15.5 Hz, 1H), 7.65 (dd, J = 8.5, 2.0 Hz, 1H), 7.60 (d, J = 2.0 Hz, 1H), 7.54–7.46 (m, 4H), 7.36 (d, J = 8.5 Hz, 1H), 6.97 (s, 1H), 6.88 (d, J = 15.5 Hz, 1H), 4.22 (s, 3H), 2.86 (t, J = 6.0 Hz, 2H), 2.71 (t, J = 6.0 Hz, 2H), 1.95–1.90 (m, 2H), 1.81 (s, 6H). ¹³C NMR (125 MHz, Chloroform-d): δ 179.9, 157.7, 154.9, 146.3, 142.3, 141.9, 134.2, 132.9, 131.1, 129.6, 1288, 127.8, 122.7, 122.6, 117.9, 117.1, 116.6, 114.0, 109.2, 108.6, 51.5, 30.2, 29.8, 29.6, 27.9, 24.8, 20.0. LRMS (m/z): [M]⁺ calcd. for C₂₇H₂₅N₂O⁺ 393.20, found 393.37.

4

The 4 was preformed according to the reported literature⁶⁶. A solution of 1,2,3,3-tetramethyl-3H-indol-1-ium iodide (602.2 mg, 2.0 mmol), N-(((E)-2-chloro-3-((phenylamino)methylene) cyclohex-1-enyl)methylene)benzenemin-ium (359.3 mg, 0.5 equiv.), NaOAc (166.0 mg, 1.0 equiv.) and EtOH (30 mL) was refluxed for 4 h. After being cooled to room temperature, the reaction mixture was concentrated under reduced pressure to remove EtOH, the resultant reside was quenched with H₂O, and extracted with DCM for three times. The combined organic layer was dried over Na₂SO₄, the filtrates was concentrated under reduced pressure. The purification using column chromatography on silica gel (eluted with DCM/MeOH = 10/1) was performed to afford the 4 (299.6 mg, 62% yield) as green solid. ¹H NMR (400 MHz, Chloroform-d): δ 8.36 (d, J = 14.0 Hz, 2H), 7.43–7.38 (m, 4H), 7.29–7.20 (m, 4H), 6.22 (d, J = 14.0 Hz, 2H), 3.76 (s, 6H), 2.75 (t, J = 6.4 Hz, 4H), 2.01–1.96 (m, 2H), 1.73 (s, 12H). ¹³C NMR (125 MHz, Chloroform-d): δ 172.9, 150.7, 144.4, 142.8, 141.0, 128.9, 127.8, 125.4, 122.2, 110.9, 101.7, 49.3, 32.7, 28.1, 26.8, 20.7.

CyN-4

The CyN-4 was preformed according to the reported literature⁶⁷. A solution of 4 (30.0 mg, 0.05 mmol), 5-formyl-2,4-dihydroxybenzonitrile (9.8 mg, 1.2 equiv.), K₂CO₃ (13.8 mg, 2.0 equiv.) and DMF (4 mL) was stirred at 50 °C for 12 h. After being cooled to room temperature, the mixture reaction was quenched with H₂O and extracted with DCM for three times. The combined organic layer was dried over Na₂SO₄, the filtrates was concentrated under reduced pressure. The purification using column chromatography on silica gel (eluted with DCM/MeOH = 10/1) was performed to afford the CyN-4 (9.6 mg, 47% yield) as blue solid. ¹H NMR (400 MHz, Chloroform-d): δ 8.40 (d, J = 14.4 Hz, 1H), 7.46 (s, 1H), 7.42–7.35 (m, 2H), 7.15–7.11 (m, 2H), 6.78 (s, 1H), 6.03 (d, J = 14.4 Hz, 1H), 3.64 (s, 3H), 2.64 (t, J = 5.2 Hz, 2H), 2.54 (t, J = 5.2 Hz, 2H), 1.83–1.80 (m, 2H), 1.69 (s, 6H). LRMS (m/z): [M]⁺ calcd. for C₂₇H₂₅N₂O₂⁺ 409.19, found 409.60.

CyN-5

To a solution of CyN-4 (30.0 mg, 0.06 mmol), K₂CO₃ (16.5 mg, 2.0 equiv.) in MeCN (2 mL) was added dropwise CH₃I (25.5 mg, 3.0 equiv.) at room temperature, and then the mixture reaction was stirred overnight. The volatiles were removed under reduce pressure, the resultant residue was quenched with H₂O, and extracted with DCM for three times. The combined organic layer was dried over Na₂SO₄, filtered and concentrated under reduced pressure. The purification using column chromatography on silica gel (eluted with DCM/MeOH = 10/1) was performed to afford the CyN-5 (20.8 mg, 82% yield) as purplish red solid. ¹H NMR (400 MHz, Chloroform-d) δ 8.74 (d, J = 15.6 Hz, 1H), 7.72–7.49 (m, 1H), 7.60 (d, J = 8.8 Hz, 1H), 7.55–7.51 (m, 3H), 7.18 (s, 1H), 7.11 (J = 15.6 Hz, 1H), 6.74 (d, J = 8.8 Hz, 1H), 4.28 (s, 3H), 4.00 (s, 3H), 2.93 (t, J = 6.0 Hz, 2H), 2.71 (t, J = 6.0 Hz, 2H), 1.97–1.94 (s, 2H), 1.83 (s, 6H). LRMS (m/z): [M]⁺ calcd. for C₂₈H₂₇N₂O₂⁺ 423.21, found 423.55.

General procedure for the synthesis of 6a–c

A solution of potassium phthalimide (5, 1.85 g,10 mmol), dibromide derivative (10.0 equiv.), KI (83.0 mg, 5 mol%) and Acetone (30 mL) was refluxed overnight⁶⁸. After being cooled to room temperature, the reaction mixture was filtrated, washed with acetone for three time and the filtrate was concentrated under reduced pressure to afford the 6.

6a (Linker = (CH₂)₄)

White solid, 2.37 g, 84% yield. ¹H NMR (400 MHz, CD₃OD): δ 7.91−7.88 (m, 2H), 7.86−7.83 (m, 2H), 3.75 (t, J = 6.8 Hz, 2H), 3.54 (t, J = 6.8 Hz, 2H), 1.97−1.84 (m, 4H). ¹³C NMR (125 MHz, Chloroform-d): δ 168.5, 134.1, 132.1, 123.4, 37.0, 32.9, 29.9, 27.3;

6b (Linker = (CH₂)₇)

White solid, 2.37 g, 73% yield. ¹H NMR (400 MHz, Chloroform-d): δ 7.82−7.78 (m, 2H), 7.70−7.66 (m, 2H), 3.64 (t, J = 7.2 Hz, 2H), 3.35 (t, J = 6.8 Hz, 2H), 1.84−1.77 (m, 2H), 1.69−1.62 (m, 2H), 1.42−1.31 (m, 6H);

6c (Linker = CH₂CH₂(OCH₂CH₂)₂)

White solid, 1.98 mg, 58% yield. ¹H NMR (400 MHz, Chloroform-d) δ 7.86−7.84 (m, 2H), 7.73−7.71 (m, 2H), 3.91 (t, J = 5.8 Hz, 2H), 3.75 (t, J = 6.0 Hz, 2H), 3.74 (t, J = 6.4 Hz, 2H), 3.66−3.61 (m, 4H), 3.38 (t, J = 6.2 Hz, 2H). ¹³C NMR (125 MHz, Chloroform-d) δ 168.4, 134.5, 134.1, 132.3, 123.8, 123.4, 71.3, 70.6, 70.3, 68.12, 37.4, 30.4.

General procedure for the synthesis of 7a-c

A solution of 6 (1.0 equiv., 2.0 mmol) and 2,3,3-trimethyl-3H-indole (636.9 mg, 2 equiv.) in MeCN (10 mL) was refluxed for three days. After being cooled to room temperature, the reaction mixture was concentrated under reduced pressure. The resulting reside was diluted with MeOH, and then the mixture was added to Et₂O (20 mL) slowly, the precipitates was observed, filtrated and dried under vacuums to afford the 7, which was used without further purification for the next step.

7a (Linker = (CH₂)₄)

White solid, 643.4 mg, 89% yield. ¹H NMR (400 MHz, Chloroform-d) δ 7.85–7.80 (m, 2H), 7.73–7.69 (m, 2H), 7.64–7.54 (m, 4H), 4.69 (t, J = 7.6 Hz, 2H), 3.66 (t, J = 6.8 Hz, 2H), 3.12 (s, 3H), 1.97–1.89 (m, 2H), 1.70–1.63 (m, 2H), 1.66 (s, 6H), 1.60 (s, 3H), 1.49–1.45 (m, 4H), 1.36–1.34 (m, 2H). ¹³C NMR (125 MHz, Chloroform-d): δ 196.5, 168.5, 141.7, 141.0, 134.3, 131.9, 130.3, 129.7, 123.5, 123.4, 115.6, 54.9, 49.7, 37.0, 26.0, 25.2, 23.4, 17.3. LRMS (m/z): [M]⁺ calcd. for C₂₃H₂₅N₂O₂⁺ 361.19, found 361.30.

7b (Linker = (CH₂)₇)

White solid, 645.7 mg, 80% yield. ¹H NMR (400 MHz, Chloroform-d): δ 7.85–7.80 (m, 2H), 7.73–7.69 (m, 2H), 7.64–7.54 (m, 4H), 4.69 (t, J = 7.6 Hz, 2H), 3.66 (t, J = 6.8 Hz, 2H), 3.12 (s, 3H), 1.97–1.89 (m, 2H), 1.70–1.63 (m, 2H), 1.66 (s, 6H), 1.60 (s, 3H), 1.49–1.45 (m, 4H), 1.36–1.34 (m, 2H). ¹³C NMR (125 MHz, Chloroform-d): δ 195.8, 168.6, 141.8, 141.1, 134.1, 132.2, 130.3, 129.7, 123.5, 123.3, 115.5, 54.8, 50.3, 37.8, 28.7, 28.4, 28.0, 26.7, 26.5, 23.3, 17.2. LRMS (m/z): [M]⁺ calcd for C₂₆H₃₁N₂O₂⁺ 403.24, found 403.37.

7c (Linker = CH₂CH₂(OCH₂CH₂)₂)

White solid, 354.0 mg, 42% yield. ¹H NMR (400 MHz, Chloroform-d): δ 7.87–7.85 (m, 2H), 7.77–7.75 (m, 2H), 7.59–7.52 (m, 3H), 4.95 (t, J = 4.8 Hz, 2H), 4.00–3.98 (t, J = 4.8 Hz, 2H), 3.78–3.75 (t, J = 5.6 Hz, 2H), 3.58 (t, J = 5.6 Hz, 2H), 3.52–3.48 (m, 4H), 3.01 (s, 3H), 1.62 (s, 6H). ¹³C NMR (125 MHz, Chloroform-d) δ 198.0, 168.3, 141.5, 134.4, 132.1, 130.1, 129.7, 123.5, 123.2, 115.9, 89.9, 70.6, 70.1, 68.1, 67.2, 54.8, 50.6, 37.3, 23.2, 16.8, 15.4. LRMS (m/z): [M]⁺ calcd. for C₂₅H₂₉N₂O₄⁺ 421.21, found 421.29.

General procedure for the synthesis of 8a–c

A solution of 7 (1 mmol, 1.0 equiv.), 2a (261.0 mg, 1.1 equiv.) and NaOAc (328.1 mg, 4.0 equiv.) in Ac₂O (10 mL) was stirred at 60 °C for 2 h. After being cooled to room temperature, the reaction mixture was quenched with H₂O (50 mL) and extracted with DCM for three times. The combined organic layer was dried over Na₂SO₄, filtrated and the filtrates was concentrated under reduced pressure. The purification using column chromatography on silica gel (eluted with DCM/MeOH = 10/1) was performed to afforded the 8.

8a (Linker = (CH₂)₄)

Violet solid, 412.0 mg, 71% yield. ¹H NMR (400 MHz, Chloroform-d): δ 8.60 (d, J = 15.2 Hz, 1H), 7.82–7.80 (m, 2H), 7.72–7.69 (m, 2H), 7.54–7.40 (m, 7H), 7.06 (d, J = 15.2 Hz, 1H), 6.99 (s, 1H), 4.91 (t, J = 5.6 Hz, 2H), 3.78 (t, J = 5.6 Hz, 2H), 2.98 (t, J = 6.0 Hz, 2H), 2.75 (t, J = 6.0 Hz, 2H), 1.98 (m, 6H), 1.83 (s, 6H). LRMS (m/z): [M]⁺ calcd. for C₃₈H₃₄N₃O₃⁺ 580.26, found 580.46.

8b (Linker = (CH₂)₇)

Violet solid, 385.8 mg, 62% yield. ¹H NMR (400 MHz, Chloroform-d) δ 8.59 (d, J = 15.2 Hz, 1H), 7.82–7.80 (m, 2H), 7.71–7.69 (m, 2H), 7.55–7.38 (m, 7H), 7.05 (d, J = 15.2 Hz, 1H), 6.95 (s, 1H), 4.82 (t, J = 7.2 Hz, 2H), 3.64 (t, J = 7.2 Hz, 2H), 2.97 (d, J = 6.8 Hz, 2H), 2.74 (t, J = 6.4 Hz, 2H), 1.98–1.92 (m, 4H), 1.84 (s, 6H), 1.68–1.42 (m, 12H). LRMS (m/z): [M]⁺ calcd. for C₄₁H₄₀N₃O₃⁺ 622.31, found 622.59.

8c (Linker = CH₂CH₂(OCH₂CH₂)₂)

Violet solid, 473.8 mg, 74% yield. ¹H NMR (400 MHz, Chloroform-d) δ 8.57 (d, J = 15.6 Hz, 1H), 7.80–7.76 (m, 2H), 7.72–7.69 (m, 2H), 7.64 (d, J = 8.0 Hz, 1H), 7.53–7.35 (m, 6H), 7.10 (d, J = 15.6 Hz, f1H), 6.91 (s, 1H), 5.04 (t, J = 4.8 Hz, 2H), 4.03 (t, J = 4.8 Hz, 2H), 3.70 (t, J = 6.0 Hz, 2H), 3.56–3.47 (m, 6H), 2.91–2.88 (m, 2H), 2.2 (t, J = 6.0 Hz, 2H), 1.95–1.92 (m, 2H), 1.83 (s, 6H). LRMS (m/z): [M]⁺ calcd. for C₄₀H₃₈N₃O₅⁺ 640.28, found 640.51.

General procedure for the synthesis of 9a-c

A solution of 8 (1 mmol, 1.0 equiv.), N₂H₄•H₂O (80% conc. in H₂O, 20 equiv.) in DCM/MeOH (v/v = 1:4, 20 mL) was stirred at room temperature for 2 h. The reaction mixture was quenched with H₂O, and extracted with DCM for three times. The combined organic layer was dried over Na₂SO₄, filtrated and the filtrates was concentrated under reduced pressure to afford the 9, which was used without further purification for the next step.

9a (Linker = (CH₂)₄)

¹H NMR (500 MHz, Methanol-d₄) δ 8.70 (d, J = 15.5 Hz, 1H), 7.70–7.69 (m, 1H), 7.65–7.61 (m, 2H), 7.52–7.44 (m, 4H), 7.14 (s, 1H), 6.66 (d, J = 15.5 Hz, 1H), 4.45 (t, J = 7.5 Hz, 2H), 2.91 (t, J = 7.5 Hz, 2H), 2.70 (t, J = 6.0 Hz, 2H), 2.66 (t, J = 6.0 Hz, 2H), 1.95–1.89 (m, 2H), 1.87–1.84 (m, 2H), 1.77 (s, 6H), 1.77–1.73 (m, 2H). LRMS (m/z): [M]⁺ calcd. for C₃₀H₃₂N₃O⁺ 450.25, found 450.41.

9b (Linker = (CH₂)₇)

¹H NMR (500 MHz, Methanol-d₄) δ 8.75 (d, J = 15.5 Hz, 1H), 7.76− (s, 1H), 7.72−7.67 (m, 2H), 7.59−7.52 (s, 4H), 7.20 (s, 1H), 6.73 (d, J = 15.5 Hz, 1H), 4.48 (t, J = 7.5 Hz, 2H), 2.89 (t, J = 7.5 Hz, 2H), 2.78 (t, J = 6.0 Hz, 2H), 2.72 (t, J = 6.0 Hz, 2H), 1.96−1.90 (s, 4H), 1.83 (s, 6H), 1.67−1.61 (m, 2H), 1.50−1.37 (s, 6H). LRMS (m/z): [M]⁺ calcd. for C₃₃H₃₈N₃O⁺ 492.30, found 492.47.

9c (Linker = CH₂CH₂(OCH₂CH₂)₂)

LRMS (m/z): [M]⁺ calcd for C₃₂H₃₆N₃O₃⁺ 510.28, found 510.53.

General procedure for the synthesis of CyNA₄, CyNA₇, CyNA_P

A solution of 9, Cys-NH(Boc)-S(Tri) (556.3 mg, 1.2 equiv.), HBTU (455.1 mg, 1.4 equiv.), DIPEA (181.0 mg, 1.4 equiv.) in DCM (20 mL) was stirred at room temperature for 2 h. The reaction mixture was quenched with H₂O, and extracted with DCM for three times. The combined organic layer was dried over Na₂SO₄, filtrated and concentrated under reduced pressure to afforded the intermedia 10, which was used without further purification for the next reaction.

A solution of 10, Et₃SiH (232.6 mg, 2.0 equiv.) in CF₃COOH (20% in DCM, 10 mL) was stirred at room temperature for 6 h. The reaction mixture was quenched by H₂O, and extracted with DCM for three times. The combined organic layer was dried over Na₂SO₄, filtrated and concentrated under reduced pressure. The purification using column chromatography on silica gel (eluted with DCM/MeOH = 5/1) was performed to afforded the CyNA_X.

CyNA₄

Purplish red solid, 22.1 mg, 4% yield in three steps. ¹H NMR (400 MHz, Methanol-d₄) δ 8.78 (d, J = 15.2 Hz, 1H), 7.77 (s, 1H), 7.71 (td, J = 7.2, 1.6 Hz, 2H), 7.61–7.52 (m, 3H), 7.29–7.16 (m, 2H), 6.75 (d, J = 15.2 Hz, 1H), 4.52 (t, J = 7.6 Hz, 2H), 4.21 (dt, J = 20.8, 5.2 Hz, 1H), 3.97 (t, J = 6.8 Hz, 1H), 3.44 (dt, J = 13.6, 6.8 Hz, 1H), 3.22 (dt, J = 13.6, 6.8 Hz, 1H), 2.98–2.90 (m, 3H), 2.74 (t, J = 6.0 Hz, 2H), 2.03–1.91 (m, 4H), 1.85 (s, 6H), 1.73–1.69 (m, 2H). HRMS (m/z): [M]⁺ calcd. for C₃₃H₃₇N₄O₂S⁺ 553.2632, found 553.2596.

CyNA₇

Purplish red solid, 35.7 mg, 6% yield in three steps. ¹H NMR (400 MHz, Methanol-d₄) δ 8.77 (d, J = 15.2 Hz, 1H), 7.76 (s, 1H), 7.72 (d, J = 7.2 Hz, 1H), 7.68 (d, J = 7.6 Hz, 1H), 7.61–7.54 (m, 3H), 7.20–7.16 (m, 2H), 6.73 (d, J = 15.2 Hz, 1H), 4.48 (t, J = 7.6 Hz, 2H), 4.20 (dt, J = 16.4, 5.2 Hz, 1H), 3.97 (t, J = 6.0 Hz, 1H), 3.21–2.93 (m, 7H), 2.72 (t, J = 6.4 Hz, 2H), 1.96–1.92 (m, 2H), 1.84 (s, 6H), 1.57–139 (m, 8H). HRMS (m/z): [M]⁺ calcd. for C₃₆H₄₃N₄O₂S⁺ 595.3101, found 595.3096.

CyNA_P

Purplish red solid, 18.4 mg, 3% yield in three steps. ¹H NMR (500 MHz, Methanol-d₄): δ 8.77 (d, J = 15.5 Hz, 1H), 7.78 (s, 1H), 7.74–7.69 (m, 2H), 7.60–7.54 (m, 4H), 7.20 (s, 1H), 6.90 (d, J = 15.5 Hz, 1H), 4.71 (t, J = 5.0 Hz, 2H), 3.99 (t, J = 5.0 Hz, 2H), 3.95 (t, J = 6.0 Hz, 1H), 3.60–3,58 (m, 2H), 3.51–3.49 (m, 2H), 3.42–3.36 (m, 3H), 3.24–3.16 (m, 2H), 2.94 (dd, J = 14.5, 5.0 Hz, 1H), 2.89 (dd, J = 14.5, 6.0 Hz, 1H), 2.80 (t, J = 6.0 Hz, 2H), 2.73 (t, J = 6.0 Hz, 2H), 1.97–1.93 (m, 2H), 1.86 (s, 6H). HRMS (m/z): [M]⁺ calcd. for C₃₅H₄₁N₄O₄S⁺ 613.2843, found 613.2808.

CyNA_P-SS

A solution of 9c (51.0 mg, 0.1 mmol, 1.0 equiv.), Cys-NH(Fmoc)-S(S-^tBu) (51.8 mg, 1.2 equiv.), HBTU (53.1 mg, 1.4 equiv.), DIPEA (18.1 mg, 1.4 equiv.) in DCM (5 mL) was stirred at room temperature for 2 h. The reaction mixture was quenched with H₂O, and extracted with DCM for three times. The combined organic layer was dried over Na₂SO₄, filtrated and concentrated under reduced pressure to afforded the intermedia 11, which was used without further purification for the next reaction.

A solution of 11 in piperidine (20% in DCM) was stirred at room temperature for 2 h. The reaction mixture was quenched by H₂O, and extracted with DCM for three times. The combined organic layer was dried over Na₂SO₄, filtrated and concentrated under reduced pressure. The purification using column chromatography on silica gel (eluted with DCM/MeOH = 5/1) was performed to afforded the CyNA_P-SS (9.1 mg, 13% in two-step) as purplish red solid. ¹H NMR (400 MHz, Methanol-d₄) δ 8.82 (d, J = 15.6 Hz, 1H), 7.82 (s, 1H), 7.78–7.74 (m, 2H), 7.65–7.58 (m, 4H), 7.24 (s, 1H), 6.95 (d, J = 15.6 Hz, 1H), 4.76 (t, J = 5.2 Hz, 2H), 4.04 (t, J = 5.2 Hz, 2H), 3.64–3.63 (m, 2H), 3.55–3.52 (m, 2H), 3.45 (t, J = 5.6 Hz, 2H), 2.84 (t, J = 6.2 Hz, 2H), 2.77 (t, J = 6.2 Hz, 2H), 2.01–1.98 (m, 2H), 1.91 (s, 6H), 1.36 (s, 9H). HRMS (m/z): [M]⁺ calcd. for C₃₉H₄₉N₄O₄S₂⁺ 701.3190, found 701.3188.

CyNA_P-S

A solution of 9c (51.0 mg, 0.1 mmol, 1.0 equiv.), Cys-NH(Fmoc)-S(^tBu) (55.63 mg, 1.2 equiv.), HBTU (53.1 mg, 1.4 equiv.), DIPEA (18.1 mg, 1.4 equiv.) in DCM (5 mL) was stirred at room temperature for 2 h. The reaction mixture was quenched with H₂O, and extracted with DCM for three times. The combined organic layer was dried over Na₂SO₄, filtrated and concentrated under reduced pressure to afforded the intermedia 12, which was used without further purification for the next reaction.

A solution of 12 in piperidine (20% in DCM) was stirred at room temperature for 2 h. The reaction mixture was quenched by H₂O, and extracted with DCM for three times. The combined organic layer was dried over Na₂SO₄, filtrated and concentrated under reduced pressure. The purification using column chromatography on silica gel (eluted with DCM/MeOH = 5/1) was performed to afforded the compound CyNA_P-S (10.0 mg, 15% in two-step) as purplish red solid. ¹H NMR (400 MHz, Methanol-d₄) δ 8.82 (d, J = 15.2 Hz, 1H), 7.82 (s, 1H), 7.80–7.73 (m, 2H), 7.66–7.58 (m, 4H), 7.25 (s, 1H), 6.95 (d, J = 15.2 Hz, 1H), 4.77 (t, J = 4.8 Hz, 2H), 4.05 (t, J = 4.8 Hz, 2H), 3.94 (s, 1H), 3.65 (dd, J = 6.0, 3.2 Hz, 2H), 3.56 (dd, J = 6.0, 3.2 Hz, 2H), 3.45 (d, J = 5.3 Hz, 2H), 3.30 (s, 1H), 3.03 (s, 1H), 2.95 (s, 1H), 2.85 (t, J = 6.2 Hz, 2H), 2.78 (t, J = 6.2 Hz, 2H), 2.04–1.99 (m, 2H), 1.92 (s, 6H), 1.37 (s, 9H). HRMS (m/z): [M]⁺ calcd. for C₃₉H₄₉N₄O₄S⁺ 669.3469, found 669.3427.

CyNA_P-T

A solution of compound 9c (20.0 mg, 0.04 mmol), D-Cys (48.4 mg, 10 equiv.) and MeOH/PBS (2.0 mL, V/V = 3:7, pH = 7.4) was stirred at room temperature overnight. The mixture reaction was concentrated under reduced prepress, and the resulted residue was purified with flash column chromatography on silica gel (eluted with DCM/MeOH = 10/1) to afforded the compound CyNA_P-T as blue solid in 6% yield. LRMS (m/z): [M]⁺ calcd. for C₃₅H₄₀N₃O₅S⁺ 614.27, found 614.38.

CyNA_P-SS-FK

A solution of compound CyNA_P-SS (35 mg, 0.05 mmol), Ac-FK (43.5 mg, 2.0 equiv.), HBTU (26.5 mg, 1.4 equiv.), DIPEA (9.0 mg, 1.4 equiv.) in THF (5 mL) was stirred at room temperature for 12 h. The was removed under reduced pressure, and the resultant residue was extracted with DCM for three times. The combined organic layer was dried over Na₂SO₄, filtrated and concentrated under reduced pressure. The purification using fresh column chromatography on silica gel (eluted with DCM/MeOH = 5/1) was performed to afforded the compound CyNA_P-SS-FK (1.1 mg) as purplish red solid in 4% yield. ¹H NMR (500 MHz, Chloroform-d) δ 8.61 (d, J = 15.5 Hz, 1H), 7.70–7.64 (m, 2H), 7.58–7.50 (m, 3H), 7.46 (t, J = 8.0 Hz, 2H), 7.38–7.31 (m, 3H), 7.20 (q, J = 8.0, 7.0 Hz, 2H), 6.99 (d, J = 15.5 Hz, 1H), 6.91 (s, 1H), 5.06 (s, 4H), 4.71 (dd, J = 44.0, 16.5 Hz, 4H), 4.43 (s, 1H), 4.15 (s, 1H), 4.05 (d, J = 10.5 Hz, 2H), 3.57–3.49 (m, 3H), 3.34 (dd, J = 13.5, 3.5 Hz, 1H), 3.22–3.04 (m, 6H), 2.73 (s, 3H), 2.03 (s, 2H), 1.84 (s, 6H), 1.68 (s, 15H), 1.44 (d, J = 5.5 Hz, 2H), 1.31 (s, 9H), 1.27 (d, J = 12.0 Hz, 6H). HRMS (m/z): [M]⁺ calcd. for C₆₁H₈₀N₇O₉S₂⁺ 1118.5453, found 1118.5460.

Kinetics study of intermolecular condensation reaction of CyN-(1−3) with Cysteine

The condensation reaction kinetics were performed according to the report literature⁴⁶. A solution of CyN-X (X = 1, 2, 3) (50 μM), Cysteine (1.0 equiv., 50 μM), TCEP (100 equiv.) and PBS (containing 20% MeOH, adjust the pH of the reaction solution to 7.4 with NaHCO₃ aq.) was attired at room temperature (30 °C). The progress at different time points was monitored by HPLC at 600 nm. The liner analysis of 1/[CyN-X] vs. time gives the 2^nd order rate constant. CyN-1: 0.30956 M⁻¹ s⁻¹; CyN-2: 0.15179 M⁻¹ s⁻¹; CyN-3: 0.22190 M⁻¹ s⁻¹;

Kinetics study of intramolecular condensation reaction of CyNA_X (X = 4, 7, P)

The condensation reaction kinetics were performed according to the report literature⁴⁶. A solution of CyNA_X (X = 4, 7, P) (50 μM), TCEP (100 equiv.) and PBS (containing 20% MeOH, adjust the pH of the reaction solution to 7.4 with NaHCO₃ aq.) was attired at room temperature (30 °C). The progress at different time points was monitored by HPLC at 600 nm with the addition of methylene blue (MB, λ_ex = 700 nm) as internal standard. The liner analysis of In[CyNA_X] vs. time gives the 1^st order rate constant. CyNA₄: −2.91 × 10⁻⁵ s⁻¹; CyNA₇: −0.89 × 10⁻⁵ s⁻¹; CyNA_P: −7.97 × 10⁻⁵ s⁻¹; CyNA₄-C: MALDI-TOF MS: m/z: [M]⁺ Calcd for C₃₃H₃₄N₃O₂S⁺ 536.237; Found 536.241; CyNA₇-C: MALDI-TOF MS: m/z: [M]⁺ Calcd for C₃₆H₄₀N₃O₂S⁺ 578.284; Found 578.291; CyNA_p-C: ¹H NMR (400 MHz, Methanol-d₄) δ 8.82 (d, J = 15.2 Hz, 1H), 7.82 (s, 1H), 7.80 (m, 2H), 7.65–7.59 (m, 4H), 7.25 (s, 1H), 6.94 (d, J = 15.2 Hz, 1H), 4.77 (t, J = 5.2 Hz, 2H), 4.05–4.03 (m, 3H), 3.98 (t, J = 6.0 Hz, 1H), 3.65–3.63 (m, 3H), 3.56–3.53 (m, 3H), 3.45 (m, 2H), 2.98–2.95 (m, 1H), 2.85 (t, J = 6.0 Hz, 2H), 2.78 (t, J = 6.0 Hz, 2H), 2.00 (t, J = 6.0 Hz, 2H), 1.91 (s, 6H). MALDI-TOF MS (m/z): [M]⁺ calcd. for C₃₅H₃₈N₃O₄S⁺ 596.258, found 596.262;

Competitive reaction between intramolecular cyclization and intermolecular condensation of CyNA_P in the presence of varying concentration of free Cys

The condensation reaction kinetics were performed according to the report literature⁴⁶. A solution of CyNA_P (20 μM), Cysteine (20 μM, 2 mM, and 20 mM), TCEP (40 μM, 4 mM, and 40 mM) and PBS (containing 20% MeOH, adjust the pH of the reaction solution to 7.4 with NaHCO₃ aq.) was attired at room temperature (30 °C) for 24 h. The progress was monitored by HPLC at 600 nm.

Computational details

The geometric optimization and frequency analysis of all molecules were conducted using the density functional theory method at the B3LYP/6-311 G(d,p) level with Grimme’s DFT-d3 correction in Gaussian 16 program package. The visualization of frontier molecular orbitals was achieved using the IQmol molecular viewer package.

Measure of fluorescence quantum yields

Methylene Blue was used as a standard with a known fluorescence quantum yield (Φ) value of 0.49 in PBS⁶⁹. Fluorescence quantum yields were calculated using the following equation:

$${\phi }_{s}/{\phi }_{f}=\left({A}_{s}/{A}_{f}\right)\times \left({{Abs}}_{s}/{{Abs}}_{f}\right)\times \left({{\eta }_{s}}^{2}/{{\eta }_{f}}^{2}\right)$$

where Φ_s and Φ_f are the fluorescence quantum yields of the standard and the samples, respectively; A_s and A_f are the emission areas of the standard and the samples, respectively; Abs_s and Abs_f are the absorbance of the standard and the samples at the Wavelength of excitation; η_s and η_f are the refractive indices of the standard and the samples, respectively.

TEM sample preparation

1 μM CyNA_P-SS-FK was incubated with TCEP and Cat B for 12 h in phosphate buffered solution (1×PBS, pH = 7.4), the carbon-coated side of the grid was gently immersed in the dispersion, the photo was taken by Hitachi TEM system. Magnification = 40,000, Accelerating Voltage = 100,000.

Enzyme-responsive experiment

The CyNA_P-SS-FK at a concentration of 20 μM was separately incubated with GSH, Cat B enzyme, and a mixture of both at 37 °C for 6 h. After incubation, fluorescence spectra were collected to calculate the fold increase in fluorescence signal before and after the response.

Specificity experiment

The CyNA_P-SS-FK (40 μM) was incubated with the specified enzymes, metal ions (60 μM), and ROS (60 μM) in PBS (10 mM, pH 7.4) at 37 °C. Mg²⁺ refers to magnesium chloride; Ca²⁺ to calcium chloride; ClO^- to sodium hypochlorite; H₂O₂ to hydrogen peroxide; SO₄^2- to sodium sulfate; GSH to glutathione; TCEP to Tris(2-carboxyethyl) phosphine hydrochloride; Cas-3 to caspase 3; β-Galactosidase (n = 3, mean ± standard deviation).

Cytotoxicity assay

The cytotoxicity of CyNA_P was detected by Cell Counting Kit-8. Human hepatocellular carcinoma cells (HCC-LM3 cells) were obtained from American Type Culture Collection (ATCC). HCC-LM3 cells were grown in DMEM containing 10% fetal bovine serum (FBS), and 1% penicillin & streptomycin, cultured at 37 °C with 5% CO₂. HCC-LM3 cells were implanted in 96-well plates (4000 cells per well in 100 μL complete culture medium) and cultured overnight, incubated with different concentrations of CyNA_P (0, 20, 40, 60, 80, 100 μM) for 12 h. Then, 10 μL CCK-8 solution was added into each wall, followed by further incubation for 1 h. The relative cell viability was performed based on the absorbance of the formazan at 450 nm as OD value. Cells viability = [A (sample) − A (blank)]/[A (control) − A (blank)] × 100%, where A is corresponding to OD value.

Cell imaging study

Human hepatocellular carcinoma cells (HCC-LM3 cells) were obtained from the American Type Culture Collection (ATCC, CBP60654). HCC-LM3 cells were grown in DMEM containing 10% fetal bovine serum (FBS) and 1% penicillin and streptomycin and cultured at 37 °C with 5% CO₂. Various concentrations of CyNA₄, CyNA₇, CyNA_P, CyNA_P-S, CyNA_P-SS, CyNA_P-SS-FK were incubated with HCC-LM3 cells (4000 cells per wall) in a 96-wall confocal plate. The control group was treated with or without inhibitor BSO (CAS S9728), and all groups were cultured overnight. After staining with DAPI, imaging was conducted by a confocal laser scanning microscope FV3000 (Olympus). The excitation and emission wavelengths were 590 ± 10 and 619 nm. Cellular fluorescence intensities were quantified using ImageJ software.

Biodistribution and biocompatibility of CyNA_P in living mice

The mice were all male Balb/c-nu, 4–6 weeks old. All mice experimental procedures were approved by the Institutional Animal Care and Use Committee (IACUC), Sun Yat-sen University (SYSU) with the approval number SYSU-IACUC-2022000220. According to ethical standards, the tumor volume of the experimental mice should not exceed 2000 mm³. All our tumor-bearing experiments adhere to ethical requirements. To determine the clearance pathway and biocompatibility in living mice, 4–6 weeks old mice (Balb/c-nu, 15 mg/kg body weight) was injected with CyNA_P intravenously. All mice were fed with non-fluorescent chow (Xietong Pharmaceutical Bioengineering CO., Ltd, XT19008) and kept in SPF environment. The experimental group and the control group of mice were both randomly assigned to their respective groups. The dynamic NIRF imaging and intensities of abdominal cavity and resected major organs (heart, liver, spleen, lung, kidneys, bladder, intestine, muscle, and skin) was performed and analyzed by IVIS Lumina XR Series III (Perkin Elmer Inc, USA) from mice post-injection at various time. The major organs were collected and placed into 4% paraformaldehyde (PFA) for histological examination.

Stability of the probe CyNA_P-SS-FK

In response to the stability of the probe, we conducted further research. Initially, we injected CyNA_P-SS-FK into the healthy mice via the tail vein. After 12 h post-injection, blood was collected from the heart, centrifuged after standing, and the supernatant was taken. A fluorescent probe stock solution was prepared and quantified by ultraviolet absorbance intensity to match the concentration of the fluorescent probe in the serum. Subsequently, fluorescence spectra were collected to evaluate the stability of the probe’s fluorescence performance by comparing the changes in fluorescence intensity before and after injection. The experimental results, as shown in the figure, indicate that there was hardly any significant change in the fluorescence spectra of the fluorescent probe before and after injection, suggesting that the fluorescent probe possesses a certain level of optical stability.

Establishment of orthotopic HCC animal model in living mice

Balb/c-nu mice (4–6 weeks, male) were randomly selected and utilized for human HCC cells implantation. HCC-LM3 cells were cultured and mixed 1:1 with Matrigel and PBS, temporarily stored on ice. Mice were completely anesthetized and the abdominal cavity was opened to expose their liver. A mixture of cultured HCC-LM3 cells and stromal gel was injected into the liver, followed by closing the abdominal cavity and returning the mice to the SPF environment for further rearing. All animals were fed with non-fluorescent chow.

Real-time NIRF imaging of HCC

Following implantation of tumors in mice with various time points (0, 5, 10, and 15 days), CyNA_P were administered intravenously in the tail vein (15 mg/kg body weight), and NIRF imaging of the major organs was acquired at different time points post injection. Fluorescence images were conducted by an IVIS Lumina XR series III system with an acquisition time of 4.0 s (excitation at 620 ± 10 nm and emission at 710 ± 10 nm). Blood samples from each group were collected and placed at −20 °C for biochemical analysis. Tumors and major organs from each group were collected and placed into 4% paraformaldehyde (4% PFA) at 4 °C for histological examination and immunofluorescence assays.

Histopathologic studies and immunofluorescence assay

All collected tissue were fixed with 4% PFA. After routine dehydration, paraffin-embedded tissues were sectioned (10 μm of thickness) for H&E staining (hematoxylin and eosin staining) according to the instructions. The stained sections were observed under the EVOS M7000 microscope. After fixing the mouse tumor tissue in PFA, 30% sucrose solution was fully dehydrated and embedded in optimal cutting temperature compound (OCT), which was placed into liquid nitrogen for coagulation and frozen sections were made for subsequent staining. Sections were fixed by 100% acetone at −20 °C for 20 min after rewarming at room temperature. Goat serum was blocked for 90 min at room temperature, then primary antibody Cathepsin B Polyclonal antibody (12216-1-AP, Proteintech Group, 1:500 dilution) was added dropwise and incubated overnight at 4 degrees Celsius in the refrigerator. Dropwise addition of secondary antibody (L3016, SAB, 1:500 dilution) for 1 h at room temperature, followed by DAPI staining for 5–10 min and sealing with sealing agent. Immunofluorescence experiments were washed with PBS 3 times for 5 min each step. Fluorescence signal distribution and intensity were observed using a laser confocal microscope after staining.

Biochemical analysis

Blood samples were immediately collected from both health mice and HCC model mice after injected with CyNA_P (15 mg kg⁻¹ body weight) or PBS solution in 24 h. The collected blood samples were centrifuged at 3000 r.p.m. for 15 min at room temperature. The supernatant was stored for subsequent test. ALT (alanine aminotransferase), AST (aspartate aminotransferase) were measured with biological reagent kif (C009-1-1; C010-1-1) from Nanjing Jiancheng Bioengineering Institute. BUN (blood urea nitrogen) and sCr (serum creatinine) were measured with biological reagent kid (#D799850-0100; #D799853-0096) from Sangon Biotech (Shanghai) Company. Processed samples were tested following the instructions provided with the assay kits. UV absorption values were detected by multimode microplate reader (Synergy H1 of BioTek) and calculated by reference standard.

Statistical analysis

The in vitro, in vivo and ex vivo NIFR signal intensity was quantified with region of interest analysis (ROI) analysis by Living Image 4.4 software. All the data was mean ± standard deviation (SD) unless stated otherwise. Investigators were blinded to group allocation during experiments. Statistical differences between two groups were tested with a two-tailed Student’s t test and more than three groups were determined by one-way analysis of variance followed by Tukey’s post hoc test. For all tests, P values less than 0.05 were considered statistically significant. *p < 0.05, **p < 0.01, ***p < 0.001 and ****p < 0.0001. All statistical calculations were performed using GraphPad Prism 9.5.0 and OriginPro 2024, including assumptions of tests used.

Reporting summary

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