Screening and regulation of nanozyme activity via liquid metals coined electron rearrangement and phase engineering

Abstract

Liquid metals (LMs) have been used for design of advanced materials by leveraging their unique liquid and electronic properties. Here, we design a series of liquid metal-buffered amorphous molybdenum sulfide nanozymes with different gallium indium ratios, in which Ga_75.5In_24.5 MoS_X nanozymes exhibit the enzyme-like activity, superior to the crystalline MoS₂. This has been verified and assigned to the roles of liquid metal: (i) serving as the building template, (ii) phase engineering with catalytic active sites, (iii) electron-rich microenvironment for improving catalytic activity. Besides, the amorphous Ga_75.5In_24.5 MoS_X nanozyme possesses oxidase (OXD)-like and nicotinamide adenine dinucleotide oxidase (NOX)-like activity to achieve multiple-enzyme-like cascade catalytic reactions, disrupting intratumoral redox and metabolism homeostasis. As the example of leveraging liquid metal as a well-designed platform to screen high-performance nanozyme with excellent therapeutic effects, this work highlights the roles of tailored electronic structure and phase engineering for regulation of catalytic activity of nanozymes and may also broaden the application of liquid metal in catalysis and biomedical field.

The potential of liquid metals in tuning nanozyme catalytic performance has been underexplored. Here, the authors use liquid metals to screen high-performance nanozymes with excellent therapeutic effects, and liquid metals serve as the building templates, initial phase engineering to yield abundant catalytic active sites, and feature electron-rich microenvironment to improve catalytic activity.

Introduction

Liquid metals (LMs) are enigmatic materials, with a simultaneous metallic and liquid nature^1,2. The most striking features of liquid metals are the amorphous property and unusual electronic microenvironment, which originate from the electron-rich metallic cores and free delocalized electron clouds^3,4. Such characters make liquid metals show great potential to tune the physical, chemical, and electronic properties of materials for boosting their performance in various engineering applications^5–10. More importantly, liquid metals represent a promising technical route for manufacturing modular advanced materials^3,11–19. The obtained combinatorial materials can readily maintain the typical characteristics of liquid metals and may attain better functionality²⁰. In particular, the intrinsic amorphous nature of liquid metals creates a facile pathway for phase engineering of materials, thus deriving more intriguing properties and innovative applications^21,22. Interestingly, liquid metals have also shown promising prospects in catalysis by capitalizing on their unique electronic properties^23–26. For example, recent studies have reported that liquid metals could create dynamic uniform catalytic sites and favorable electronic structures for catalytic metal species to remarkably enhance their catalytic capability^27–29. Similarly, liquid metals have been proposed as a unique electron mediator for temporary electron storage, and promoted the extraction and transfer of electrons, as well as charge separation, greatly improving the optoelectronic property of perovskite films³⁰. Furthermore, the representative Ga-based liquid metal alloys with excellent biocompatibility and attractive flexibility have recently aroused great interest in the biological field, including cancer therapy, antimicrobial treatment, and bioimaging^31–43. These studies indicate the exceptional advantages of liquid metals to create advanced materials for catalytic and biomedicine applications.

Nanozymes are nanomaterials with inherent enzyme-mimicking catalytic properties, and have ignited increasing attention due to their advantages of high catalytic stability and durability, easy preparation, and low cost^44–47. Although promising, the unsatisfactory catalytic efficiency of nanozymes remains a crucial issue for future practical applications^48–52. This demonstrates the vital necessity of developing an appropriate strategy to optimize the nanozyme catalytic performance. From an in-depth perspective, the catalytic active sites and local electronic characteristics determine their catalytic performance to a large extent^{49,50,53–57}. Notably, phase engineering of nanomaterials has demonstrated remarkable prospects in the catalysis field because this process modulates the atomic arrangements and electronic structure of catalysts, hence providing unusual catalytic activity^58,59. Especially, amorphous materials possess a distinctive disordered atomic arrangement that enables them with abundant active sites and optimized electronic configurations, which are beneficial for enhancing their catalytic properties^60,61. Thus, combining the ability of phase engineering and the unique electronic properties, liquid metals are expected to be a dream matrix to construct high-performance nanozymes. However, the potential of liquid metal in regulating nanozyme catalytic activity has rarely been explored mechanistically. In particular, previous research efforts have focused on the application of liquid metal-based nanozyme while overlooking the catalytic potential of liquid metals for screening and regulation of nanozyme activity.^62,63

Here, we design and screen the liquid metals-buffered nanozymes, in which the liquid metals can not only initiate phase engineering to generate an amorphous nanozyme with abundant catalytic active sites, but also afford an electron-rich microenvironment to optimize the electronic structure of nanozymes for improving the catalytic performance (Fig. 1). We first prepare a series of molybdenum sulfide nanozymes by using five liquid metals with the composition of Ga_1-xIn_x (with x = 0, 0.125, 0.245, 0.375, and 0.50) as templates, and these nanozymes feature uniformly core-shell structure and amorphous property. We observe that these nanozymes exhibit peroxidase (POD)-like activities, in which the LM MoS_X nanozymes (LM = Ga_75.5In_24.5) present the best catalytic activity. Then, the role of liquid metals in regulating the enzyme-mimicking catalytic activity of LM MoS_X nanozymes is explored. We demonstrate that, compared to crystalline MoS₂, the amorphous LM MoS_X nanozymes expose abundant active sites and defects, and show about a 10-fold increase in the POD-like catalytic activity, together with a significantly higher glutathione (GSH) consumption ability. Meanwhile, this strategy can turn on the oxidase (OXD)-like and nicotinamide adenine dinucleotide oxidase (NOX)-like activities of LM MoS_X to achieve cascade catalytic reactions, eventually leading to reactive oxygen species (ROS) burst. Both in vitro and in vivo experiments demonstrate that the LM MoS_X nanozymes with multienzyme-mimicking catalytic activity can trigger a highly efficient cascade reaction to disrupt intratumoral redox and metabolic homeostasis, eventually killing tumor cells. In this work, we introduce liquid metal with advantages in synthesis and catalysis to screen and regulate the catalytic and therapeutic effects of nanozymes for efficient tumor inhibition. Our work not only investigates the catalytic activity of liquid metal-buffered nanozyme with different gallium-indium ratios, but also elucidates the mechanism underlying the enhanced enzyme-like catalytic activity through density functional theory (DFT) calculations. This study offers promising insights into the design and regulation of nanozymes, with potential benefits for future biomedical applications of liquid metals.

Fig. 1. Liquid metal-buffered nanozymes-based catalytic therapy.

a The synthetic route of LM MoS_X nanozymes. b Schematic illustration of liquid metal regulating nanozyme catalytic activity. c Schematic representation of antitumor therapy through the highly efficient multienzyme-mimicking cascade catalysis.

Results

The design and preparation of liquid metal-buffered nanozymes

As shown in Fig. 1a, the nanozyme was synthesized through facile probe sonication. Liquid metals with different ratios of gallium and indium were prepared by dissolving a calculated amount of In pellets in liquid Ga. The resultant liquid metal alloys should retain their liquid states at close to room temperature to ensure the subsequent liquid-liquid interfacial reactions⁶⁴. According to the previous reports, the molar fraction of In should below 50 at% in order to ensure their liquid states at close to room temperature^65,66. Specifically, the composition of Ga_1-xIn_x (with x = 0, 0.125, 0.245, 0.375, and 0.50) were chosen and prepared to design a series of liquid metal-buffered nanozymes. Molybdenum is a low-cost and biocompatible transition metal with multiple oxidation states, which endows Mo-based nanomaterials with tunable enzyme-like activity⁶⁷. Among Mo-based nanomaterials, molybdenum sulfide features unique structural and electronic properties and shows superior performance in various applications, especially for catalysis^68,69. The generated nanozymes were denoted as Ga₅₀In₅₀ MoS_X, Ga_62.5In_37.5 MoS_X, Ga_75.5In_24.5 MoS_X, Ga_87.5In_12.5 MoS_X, Ga₁₀₀ MoS_X. During this procedure, liquid metal served as a template to trigger in situ galvanic replacement reactions between liquid metal droplets and MoS₄^2- ions to produce core–shell nanostructures upon ultrasonication. The thermodynamic driving force underlying the galvanic replacement reaction originates from the difference in redox potentials between the two metal-based species^21,64. The size distribution of nanoparticles became narrower with higher sonication intensity applied (Supplementary Fig. 1). Transmission electron microscopy (TEM) images illustrated the well-constructed core-shell structures of the synthesized five nanozymes (Fig. 2a). Then, the energy dispersive spectroscopic (EDS) mapping and X-ray photoelectron spectroscopy (XPS) results demonstrated the existence of Ga, In, Mo and S elements in these nanozymes (Supplementary Fig. 2-3). Next, high-resolution TEM (HR-TEM) images indicated the amorphous nature of five nanozymes (Fig. 2a). According to previous studies, the intrinsic amorphous nature of liquid metals and short reaction time are considered the main reasons for the formation of the amorphous structure^4,11. Dynamic light scattering (DLS) results of five nanozymes revealed that their hydrodynamic diameters were mainly distributed around 300 nm (Fig. 2b). The surface zeta potential measurement showed that these nanozymes were negatively charged (Fig. 2c). Then, we explored the POD-like activity of five nanozymes by monitoring the oxidation of 2, 2’-azino-bis(3-ethylbenzothiazoline-6-sulfonic acid) (ABTS) at 415 nm. As revealed in Fig. 2d, e, obvious ABTS oxidation could be detected, indicating that all five nanozymes possessed POD-like activity. Among them, Ga_75.5In_24.5 MoS_X exhibited the highest catalytic activity, while Ga₅₀In₅₀ MoS_X showed a relatively low activity. According to previous studies, the Indium content in GaIn alloys significantly affects the disordered structure and the charge density distribution of alloys, thus influencing their amorphous structure and electrical conductivity⁷⁰. Specifically, the alloying of Ga with In causes more disordered structure. This amorphous structure is beneficial for the high catalytic performance of catalysts. However, the electrical conductivity of GaIn alloys decreases with the In alloying of Ga because the charge density near the In atoms is significantly smaller, which is unfavorable to the electron transfer and the subsequent catalytic process of the catalysts. Thus, the catalytic activity of the resulting nanozymes could show a trend of initially increasing and then decreasing with increasing In content in the GaIn alloy. These findings could explain the experimental results that the alloying effect of In with Ga favored the catalytic reaction, while lower or higher In contents compromised the catalytic activity. We further employed XPS spectra to characterize the electronic properties of these nanozymes. As displayed in Fig. 2f, the binding energies of Mo 3 d were positively shifted in Ga_75.5In_24.5 MoS_X and Ga₁₀₀ MoS_X nanozymes, in comparison to other nanozymes, revealing the occurrence of electron transfer and rearrangement. Moreover, we have analyzed the surface valence states of Ga_75.5In_24.5 MoS_X and Ga₁₀₀ MoS_X, revealing their different Mo oxidation states (Supplementary Fig. 4). This meant that liquid metals with varying compositions had different effects on the electronic properties and valence states of the resulting nanozyme.

Fig. 2. Characterization of Liquid metal-buffered nanozymes.

a TEM images and HR-TEM images of five liquid metal-buffered MoS_X nanozymes, representative images from n  =  3 independent experiments. b Size distribution of five liquid metal-buffered MoS_X nanozymes. c Zeta potential of five liquid metal-buffered MoS_X nanozymes. d, e The POD-like activity of five liquid metal-buffered MoS_X nanozymes. The control group refers to ABTS + H₂O₂. f Mo 3 d XPS spectra of five liquid metal-buffered MoS_X nanozyme (arb. units, arbitrary units). c, e Data are presented as mean ± standard deviation (SD) (n = 3 independent experiments).

Next, we took Ga_75.5In_24.5 MoS_X (it was denoted as LM MoS_X in the following content unless indicated otherwise) with the best catalytic activity as an example to investigate the role of liquid metal for nanozyme catalysis. First, the structure of LM MoS_X was thoroughly studied using various characterization techniques. TEM image revealed that the obtained LM MoS_X nanozymes exhibited well-constructed core-shell spherical structures (Fig. 3a), with a shell thickness of approximately 40 nm. The elemental mapping images demonstrated the existence of Ga, In, Mo and S elements in LM MoS_X nanozymes (Fig. 3b). Then, the size distribution result revealed that the Z-Average size and polydispersity index (PDI) of LM MoS_X were 309.3 nm and 0.188, respectively (Fig. 3c), indicating that the LM MoS_X had a uniform size and good dispersibility. Moreover, the DLS results, SEM and TEM images, and XPS spectra suggested that the LM MoS_X nanozymes remained monodisperse and stable over 7 days (Supplementary Fig. 5). The elements content in LM MoS_X nanozymes was further analyzed by inductively coupled plasma optical emission spectrometry (ICP-OES, Supplementary Table 1). Subsequently, the amorphous nature of LM MoS_X nanozymes was explored. The selective area electron diffraction (SAED) pattern and the High-resolution (HR)-TEM image revealed that no crystal form was observable in LM MoS_X nanozymes (Fig. 3d, e), and the corresponding 3D pseudo-color surface plot demonstrated the highly disordered arrangement of atoms (Fig. 3f). These results were consistent with the X-ray diffraction (XRD) analysis (Fig. 3g), which showed a single broad peak, further confirming the amorphous structure of LM MoS_X nanozymes. To further characterize the detailed electronic and structural properties of amorphous LM MoS_X nanozymes, crystalline MoS₂ nanospheres were prepared (Supplementary Figs. 6–9). The SAED pattern (Supplementary Fig. 7c), lattice fringes in HR-TEM image (Supplementary Fig. 7) and XRD analysis (Supplementary Fig. 8) proved the crystalline form of MoS₂ nanospheres. Afterwards, XPS was employed to analyze the surface chemical state of the LM MoS_X nanozymes (Fig. 3h–k). According to these XPS results, the amorphous LM MoS_X nanozymes and crystalline MoS₂ had very different valence states and electronic properties. The LM MoS_X nanozymes possessed richer Mo valence states. Figure 3i, k revealed the presence of Mo⁶⁺, Mo⁵⁺, and Mo⁴⁺ oxidation states in LM MoS_X nanozymes, while only Mo⁴⁺ existed in MoS₂ (Supplementary Fig. 9). As shown in Fig. 3j, LM MoS_X nanozymes had many disulfide ligands in the S 2p peak as compared with MoS₂ (Supplementary Fig. 9), which were consistent with the previous reported amorphous MoS_X⁶⁹. These data demonstrated that the LM MoS_X nanozymes exhibited abundant active sites, which were ascribed to their amorphous structure. Meanwhile, we found that the binding energy of Mo 3 d in the LM MoS_X nanozymes was positively shifted when compared to MoS₂ (Fig. 3k), which indicated the electronic perturbations in the Mo 3 d orbitals. Thereafter, electron spin resonance (ESR) measurement was also carried out to probe the electronic information of LM MoS_X nanozymes. The ESR signals in Fig. 3l manifested the existence of S vacancies (g = 2.008), Mo⁵⁺ species (g = 2.01, g = 1.97), and sulfur chain radicals (g = 2.034, g = 2.004), which further confirmed the abundant active sites of LM MoS_X nanozymes⁶⁹. These data elucidated the successful preparation and the delicate structure of LM MoS_X nanozymes.

Fig. 3. Characterization of LM MoS_X nanozymes.

a TEM image of LM MoS_X nanozymes. b HAADF image and elemental mapping of LM MoS_X nanozymes. c Size distribution result of LM MoS_X. d SAED pattern of LM MoS_X nanozymes. e HR-TEM image of LM MoS_X nanozymes and (f) the corresponding 3D pseudo-color surface plot of the selected domain. g XRD spectra of LM MoS_X and MoS₂. h XPS spectra of LM MoS_X. i The Mo 3 d and (j) S 2p XPS patterns of LM MoS_X. k The Mo 3 d XPS analysis of LM MoS_X and MoS₂. l The ESR spectra of LM MoS_X and MoS₂. a, b, d, e Representative of n = 3 independent experiments.

Catalytic performance of the LM MoSX nanozymes

The delicate structure of LM MoS_X nanozymes spurred us to investigate their catalytic performance. As shown in Fig. 4a, a significantly increased absorbance of oxidized ABTS was observed in the LM MoS_X group, while the MoS₂ treated group exhibited weak catalytic activity, indicating that the LM MoS_X nanozymes possessed high POD-like activity. We further used a steady-state kinetic assay to evaluate the catalytic performance of LM MoS_X and MoS₂. The steady-state kinetic parameters, including the Michaelis-Menten constant (K_m) and the maximum initial velocity (V_max) were obtained by Michaelis–Menten saturation curves (Supplementary Figs. 10–12) and were shown in Supplementary Table 2. The apparent K_m value of LM MoS_X with H₂O₂ as the substrate was about 4.33 times lower than MoS₂, and the V_max value of LM MoS_X was about 3.83 times higher than MoS₂. The reduced K_m value indicated that the LM MoS_X had a higher affinity and good substrate selectivity to H₂O₂, and the increased V_max value reflected its superior catalytic activity. Moreover, the apparent K_m value and V_max value of LM MoS_X with H₂O₂ as the substrate were both better than those of MoS₂ and many other reported molybdenum-based nanozymes in Supplementary Table 3. Next, a fluorescence assay was applied to assess the catalytic capability. Terephthalic acid (TA) was used as a probe for capturing •OH (Supplementary Fig. 13). According to the fluorescence spectra in Fig. 4b and Supplementary Fig. 14, the ability to generate •OH radicals of LM MoS_X was 10 times higher than that of MoS₂. As shown in Supplementary Fig. 15, the specific activity (SA) value of LM MoS_X (4.12 U/mg) was 7.5 times higher than that of crystalline MoS₂ (0.55 U/mg), showing that the designed liquid metal-buffered nanozyme had higher catalytic performance. These results were consistent with the ABTS oxidation experiment, both demonstrating the excellent POD-like activity of LM MoS_X nanozymes. Moreover, the nanozyme exhibited consistent POD-like catalytic performance over 7 days (Supplementary Fig. 16), confirming its good stability. To further investigate the enzymatic characteristics of LM MoS_X nanozymes, we utilized 1,3-diphenylisobenzofuran (DPBF) assay to monitor the oxidase-like activity. DPBF is a classical probe for detecting the generation of O₂^•−. Upon O₂^•− oxidation, the intensity of the characteristic absorption of DPBF at around 426 nm would decrease (Supplementary Fig. 17a). As shown in Fig. 4c, a dramatic decrease in DPBF absorbance was observed after LM MoS_X nanozymes treatment, demonstrating the O₂^•− generated by the OXD-mimicking process. By contrast, negligible DPBF degradation was observed in MoS₂ treated group under our experimental conditions (Supplementary Fig. 17b, c). Such OXD-like activity was probably ascribed to the intervalence charge-transfer between Mo⁵⁺/Mo⁶⁺ in the LM MoS_X, which was absent in MoS₂⁷¹. Furthermore, various substrates were selected to explore the substrate selectivity of LM MoS_X. For the POD-like reaction, the substrate selectivity of LM MoS_X nanozymes was examined using some classical POD-like substrates, including 3,3′,5,5′-Tetramethylbenzidine (TMB), ABTS, and 1,2-diaminobenzene (OPD). As shown in Supplementary Fig. 18, the LM MoS_X nanozyme could catalyze these substrates, resulting in characteristic oxidation colors and absorption peaks, which demonstrated its good POD-like activity. For an OXD-like reaction, the substrate selectivity of LM MoS_X nanozymes was carried out in O₂ and N₂-saturated solutions, respectively. As shown in Supplementary Fig. 19, the characteristic absorption of DPBF at around 426 nm substantially decreased in the presence of LM MoS_X nanozymes under a normal or saturated O₂ atmosphere. However, the catalytic ability of LM MoS_X nanozymes was weakened in hypoxia condition, indicating that O₂ was the typical substrate for LM MoS_X nanozymes. To verify the production of ROS, a free radical quenching experiment was performed. Specifically, isopropanol (IPA) and p-benzoquinone (PBQ) served as •OH quencher and O₂^•− quencher, respectively. According to Supplementary Fig. 20, a significant enhancement in fluorescence at 435 nm could be observed for LM MoS_X nanozymes in the presence of H₂O₂ in comparison to other control groups. However, in the presence of IPA, the fluorescence intensity at 435 nm decreased substantially, indicating that LM MoS_X nanozymes could effectively produce •OH. Then, we observed that the absorbance of DPBF at 426 nm in the LM MoS_X + PBQ treated group was comparable to that of the control group, proving the formation of O₂^•− during the catalytic process of LM MoS_X nanozymes (Supplementary Fig. 21). Afterward, the Electron paramagnetic resonance (EPR) experiment was conducted to evaluate the generation of radicals. As depicted in Supplementary Fig. 22, the EPR spectra using 5,5-dimethyl-1-pyrroline N-oxide (DMPO) as probe, exhibiting the distinctive quartet signal (1:2:2:1) of •OH in the LM MoS_X + H₂O₂ group. In addition, an obvious characteristic six-line pattern of O₂^•− was observed after LM MoS_X treatment. These results indicated the brilliant POD-like and OXD-like activities of LM MoS_X.

Fig. 4. Catalytic performance of the LM MoS_X nanozymes.

a The •OH producing capacity of LM MoS_X and MoS₂ measured by ABTS. b The •OH producing capacity of LM MoS_X and MoS₂ measured by terephthalic acid assay. c The time-dependent generation of O₂^•− monitored by DPBF probe. d GSH consumption ability of LM MoS_X nanozymes. e Quantitative analysis of GSH consumption ability of LM MoS_X and MoS₂, Data are presented as mean ± SD (n = 3 independent experiments). f The NOX-like activity of LM MoS_X nanozymes. g Amplex red assay for measuring H₂O₂ generation during the NADH oxidation by LM MoS_X nanozymes. h Diagram showing the multienzyme-like activities of LM MoS_X nanozymes.

In addition to the enzyme-mimicking activities mentioned above, nanozymes have shown great prospects in regulating the biological metabolic processes. Specifically, GSH and nicotinamide adenine dinucleotide (NADH) are important intracellular metabolites upregulated in cancer cells, both serving as the essential reductive factors for maintaining cellular redox homeostasis and modulating cellular metabolism⁷². Moreover, NADH is necessary for preserving cellular bioenergetics, genomic integrity, and other important physiological pathways⁷³. However, these reductive metabolites play a key role in cellular oxidative stress defense mechanisms, which can efficiently scavenge cellular ROS, thereby resulting in limited therapeutic effects. In consideration of the abundant active sites and redox pairs in LM MoS_X nanozymes, we then explored whether the designed nanozymes could work on these critical metabolites. As shown in Fig. 4d, the decrease in GSH was observed after treatment with LM MoS_X nanozymes, revealing that LM MoS_X nanozymes could effectively deplete GSH. The quantitative analysis in Fig. 4e showed that the GSH consumption efficiency of LM MoS_X nanozymes reached 80%, while MoS₂ was about 10%. Next, NADH oxidation in vitro was measured by absorption spectra (Fig. 4f), where the characteristic absorption peak of NADH at 340 nm gradually decreased over time under the treatment of the LM MoS_X nanozymes, and the NAD⁺ peak at 260 nm increased. In contrast, Supplementary Fig. 23 showed that MoS₂ treated group had negligible degradation of NADH, revealing the NOX-mimicking activity of LM MoS_X nanozymes. In the following, the substrate selectivity of LM MoS_X was explored in some biological oxidase substrates. Impressively, NADH could be effectively consumed by LM MoS_X while few (less than 6%) other biomolecules were affected (Supplementary Fig. 24). More significantly, such NOX-mimicking catalytic process was accompanied with H₂O₂ production (Fig. 4g). During this process, NADH served as an electron donor for the synergistic reduction of O₂^•− by LM MoS_X nanozymes to generate H₂O₂, which further contributed to the POD-like activity of nanozyme. In this case, LM MoS_X nanozymes could efficiently induce the cascade reactions with OXD-like, NOX-like, and POD-like catalytic activities, thereby consuming the reductive stress and generating more oxidative stress (Fig. 4h). Moreover, the addition of interfering ion or protein had no obvious influence on POD-,OXD-, and NOX-like activities of LM MoS_X nanozymes, demonstrating the anti-interference ability of the catalytic system (Supplementary Figs. 25–27). These results elucidated that the liquid metal-buffered nanozymes exhibited improved enzyme-like catalytic activity and great multienzyme-mimicking cascade catalytic activity, highlighting the key role of liquid metal in regulating catalytic performance.

DFT Calculations

To gain insights into the reasons for the enhanced enzyme-like catalytic activity of liquid metal-buffered nanozyme, density functional theory (DFT) calculations based on first principles were conducted to investigate their electronic structures. First, three computational surface models representing ideal MoS₂, MoS_X (amorphization) and LM MoS_X were constructed to elucidate their microscopic electron structures. As shown in the charge density maps (Fig. 5a), the amorphous MoS_x exhibited an increased electron density around Mo atoms compared with the MoS₂ model. Notably, the LM MoS_X model achieved the highest electron density, indicating the electron donation from liquid metal to Mo atoms. Moreover, the differential charge density of LM MoS_X in Fig. 5b revealed the electron donation from the liquid metal to the Mo atoms. In addition, the Bader charge of Mo in MoS₂, MoS_X, and LM MoS_X was calculated. The Bader charge analysis indicated a charge gain of 0.52 e^- on the Mo sites in the MoS_X relative to MoS₂. Compared to MoS_X, the electron accumulation on the Mo sites increased by an additional 0.21 e^- in the LM MoS_X, which further confirmed that electron donation from liquid metal to Mo atoms. Then, the density of states (DOS) of MoS₂, MoS_X and LM MoS_X models, as shown in Fig. 5c, revealed that LM MoS_X exhibited a significantly higher DOS near the Fermi level, and several new hybrid electronic states appeared compared to MoS₂ and MoS_X, indicating stronger interaction and charge transfer between LM MoS_X and adsorbates^51,74. Furthermore, the d-band center of the Mo orbital was calculated based on the DOS of MoS₂, MoS_X and LM MoS_X. As shown in Fig. 5d, LM MoS_X nanozymes exhibited an upward shift in d band center position (− 1.686 eV), which was significantly nearer to the Fermi level than those of MoS₂ (− 3.619 eV) and MoS_X (− 2.908 eV). The state near to the Fermi level contributed to the interaction between the adsorbed substrate and LM MoS_X nanozyme, thereby favoring the catalytic reaction⁵⁶. These results were consistent with the kinetic parameters study (Supplementary Table 2), in which the reduced K_m value indicated that the LM MoS_X had a higher affinity to H₂O₂. Based on these results, we confirmed that the pivotal role of liquid metal, i.e., amorphization, and electron transfer, endowing LM MoS_X nanozymes with optimized enzyme-like catalytic activity. In our designed system, on the one hand, liquid metals ingeniously endowed the nanozymes with amorphous property, thus providing abundant active sites for catalysis. On the other hand, the liquid metals matrix and the amorphous character of nanozymes led to a favorable electronic perturbation, which could promote electron transfer and optimize the reaction process, thereby boosting the catalytic behavior. These data indicated the superiority of liquid metal for the construction of high-performance nanozyme.

Fig. 5. DFT Calculations.

a The computational surface models and charge density maps of MoS₂, MoS_X and LM MoS_X. b The differential charge density analysis at the Mo and Ga interfaces of LM MoS_X. c Calculated DOS of the MoS₂, MoS_X and LM MoS_X models. d Calculated Mo 3 d DOS of the MoS₂, MoS_X and LM MoS_X models.

LM MoS_X nanozymes-based catalytic therapy

The excellent catalytic capability of LM MoS_X nanozymes motivated us to explore the intracellular performance. First, cytotoxicity experimental results suggested good biocompatibility of LM MoS_X even at the concentration of 200 μg/mL (Supplementary Fig. 28). Then, LM MoS_X@RB (Rhodamin B) as a model was applied to investigate the cellular uptake behavior of LM MoS_X. As shown in confocal laser scanning microscopy (CLSM) images, the red fluorescence of RB in LM MoS_X@RB treated 4T1 mouse breast cancer cells increased with time and reached the maximum after 9 h (Supplementary Fig. 29), confirming that the LM MoS_X@RB could be effectively internalized by 4T1 cells. Then, we systematically studied the catalytic behavior of LM MoS_X within 4T1 cells. The intracellular NADH and GSH levels were found to gradually decrease with increasing concentrations of LM MoS_X nanozymes (Fig. 6a, b). These results demonstrated that the LM MoS_X nanozymes could efficiently diminish the reductive stress in cancer cells. The intracellular ROS generated by LM MoS_X nanozymes was then evaluated utilizing the 2′,7′-dichlorofluorescein diacetate (DCFH-DA) probe. The 4T1 cells treated with MoS₂ only presented faint fluorescence signals under our experimental conditions. Comparatively, the nonfluorescent DCFH-DA could be efficiently converted into bright green fluorescent DCF inside LM MoS_X nanozymes-treated cells, confirming the burst of intracellular ROS (Fig. 6c, d). Such an elevated level of ROS in cells was partly attributed to the generation of oxidative stress by the POD-mimic and OXD-mimic activities of LM MoS_X nanozymes, and partly caused by the remarkable capability of LM MoS_X nanozymes to eliminate reductive stress, thereby inducing severe redox imbalance within cancer cells. Furthermore, the mitochondrial membrane potential (MMP) of 4T1 cells after different treatments was determined by the JC-1 assay. As seen in Fig. 6e, f, the JC-1 probe displayed an intense red fluorescence signal in PBS, LM NPs (pure Ga_75.5In_24.5, Supplementary Fig. 30) and MoS₂ treated cells. In comparison, the cells exhibited bright green fluorescence in the LM MoS_X nanozymes-treated group, indicating the efficient decrease of MMP and depolarization of mitochondria. These changes were presumably ascribed to the disturbance of the NADH/NAD⁺ cycle, because it was necessary for the mitochondrial electron transport chain (ETC). These experiment results indicated that the LM MoS_X nanozymes could effectively achieve mitochondria damage and ROS burst through cascade catalytic reactions (Fig. 6g).

Fig. 6. LM MoS_X nanozymes-based catalytic therapy in vitro.

a The NADH and (b) GSH content of 4T1 cells after LM MoS_X incubation with various concentrations. c, d Detection of the intracellular ROS in 4T1 cells with different treatments. e, f CLSM imaging of MMP changes with different treatments. g Schematic illustration of the redox imbalance within cancer cells. h Relative cell viability of 293T cells and (i) 4T1 cells with corresponding treatments. j Live and dead co-staining of 4T1 cells with different treatments. c, e, j Representative of n = 3 independent experiments. Two-tailed unpaired Student’s t test (d, i) was used to calculate the statistical significance. a, b, d, h, i Data are presented as mean ± SD (n   =   3 biologically independent samples).

After assessing the catalytic efficacy of LM MoS_X nanozymes in cells, we evaluated the cytotoxicity of the designed nanozymes. First, the in vitro cytotoxicity of nanozymes was measured using the methyl thiazolyl tetrazolium (MTT) method. In general, H₂O₂, GSH, and NADH are elevated in cancer cells, whereas normal cells have comparatively low concentrations. As displayed in Fig. 6h, LM NPs, MoS₂ and LM MoS_X nanozymes exhibited negligible cytotoxicity against 293T human embryonic kidney cells under our experimental conditions. For the 4T1 mouse breast cancer cells (Fig. 6i), the cell viability after LM MoS_X nanozymes treatment substantially diminished, while the MoS₂ treated group showed limited cytotoxicity. As shown in Supplementary Fig. 31, the half maximal inhibitory concentrations (IC50) value of MoS₂ and LM MoS_X nanozymes against 4T1 cells were 105.6 µg/mL and 24.26 µg/mL, respectively. The effective cancer cell killing ability was attributed to the multiple enzyme-mimicking catalytic activity of LM MoS_X nanozymes, which could induce severe oxidative damage and metabolism imbalance in cancer cells. Then, the in vitro therapy efficiency was also evaluated through the LIVE/DEAD cell staining assay, in which apparent cell death (red signals) had occurred in the LM MoS_X treated group (Fig. 6j). Taken together, these results indicated that the LM MoS_X nanozymes showed a significant anticancer efficacy, which was ascribed to the disruption of intracellular redox and metabolism homeostasis caused by the highly efficient enzyme-like catalytic activity of LM MoS_X nanozymes.

The remarkable catalytic effect of LM MoS_X in cancer cells encouraged us to examine the anti-tumor performance in vivo. We prepared hyaluronic acid (HA) coated LM MoS_X nanozymes (denoted as H-LM MoS_X) for anticancer therapy (Supplementary Fig. 32a–e), in which HA modification endowed the nanoplatform with enhanced tumor targeting ability. As illustrated in Supplementary Fig. 32f, H-LM MoS_X exhibited negligible Mo leaching after being soaked in complete cell culture medium for over 7 days. Subsequently, the biotoxicity of the H-LM MoS_X nanozymes was systematically assessed. The hemolytic experiment was first conducted to assure its safety for intravenous injection. As shown in Supplementary Fig. 33, H-LM MoS_X showed negligible hemolysis toxicity. Then, the in vivo biocompatibility of H-LM MoS_X was explored. Mice were injected with H-LM MoS_X via their tail veins to evaluate the toxicology. The blood biochemical test and histological analysis in the main organs of H-LM MoS_X-treated mice were performed, which showed no apparent toxicity and adverse effects compared with the control group (Supplementary Figs. 34, 35). Meanwhile, the mice growth was not impacted within 28 days (Supplementary Fig. 36). These data demonstrated the favorable biocompatibility of H-LM MoS_X nanozymes for bioapplication at an injection dose of 10 mg/kg. The half-life of H-LM MoS_X was determined to be 1.97 h, as shown in Supplementary Fig. 37. Subsequently, 4T1 tumor-bearing mice were established to investigate the in vivo anticancer ability (Fig. 7a). First, the tissue distribution of H-LM MoS_X was evaluated by ex vivo fluorescence. Supplementary Fig. 38 showed that a stronger fluorescence signal was detected in the tumor treated by H-LM MoS_X@Cy3, which revealed its good tumor accumulation capability. The inductively coupled plasma mass spectrometry (ICP-MS) results revealed the enhanced accumulation of H-LM MoS_X in the tumor region (Supplementary Fig. 39). The tumor-targeting capability of H-LM MoS_X was also evaluated through in vivo fluorescence imaging system. As shown in Supplementary Fig. 40, strong red fluorescence signals were observed within the tumor, confirming the effective tumor accumulation of nanoparticles. Then, the overall anti-tumor performance of nanozymes was evaluated. The mice were randomly divided into four groups: (I) Control, (II) H-LM, (III) H-MoS₂, and (IV) H-LM MoS_X. Different formulations were intravenously administered to the mice. As shown in Fig. 7c–e, the tumor growth in the H-LM MoS_X dosed group was significantly inhibited, while H-LM and H-MoS₂ treated groups had little inhibitory effect under our experimental conditions. We further characterized the pathological damage to tumors through the Hematoxylin and eosin (H&E) and terminal deoxynucleotidyl transferase-mediated dUTP nick-end labeling (TUNEL) method (Fig. 7f). These results indicated that H-LM MoS_X treatment could induce apparent tumor histological damage and cancer cell apoptosis. Notably, negligible adverse effects on mice were found during the whole therapy, as verified by the body weight analysis and histological analysis of key organs (Fig. 7b and Supplementary Fig. 41). All the results revealed that the designed H-LM MoS_X nanozymes could attain highly efficient cascade catalysis, thereby presenting remarkable antitumor therapeutic effects with satisfactory biocompatibility.

Fig. 7. LM MoS_X nanozymes-based catalytic therapy in vivo.

a Schematic diagram for the schedule of the in vivo antitumor therapy. b, c Body weight changes and tumor volume curves of the mice during therapy. d The weight of the dissected tumors from different groups. e Photographs of the excised tumors. f H&E and TUNEL analysis of tumor (n = 3 independent samples were collected from different animals). Two-tailed unpaired Student’s t test (c, d) was used to calculate the statistical significance. b–d Data are presented as mean ± SD (n   =   4 mice).

In summary, we leverage liquid metal as a well-designed platform to screen high-performance nanozyme with excellent therapeutic effects. First, we have designed and synthesized a series of liquid metal-buffered nanozymes and screened the most favorable nanozyme to achieve efficient tumor catalytic therapy. In this strategy, the liquid metals possess inherent unique advantages: (i) serving as the building template of nanozymes, (ii) initiating phase engineering to generate an amorphous nanozyme with abundant catalytic active sites, (iii) featuring an electron-rich microenvironment to affect the electronic structure of nanozymes, which are beneficial for regulating nanozyme catalytic activity. The amorphous LM MoS_X nanozymes exhibit superior multienzyme-mimicking cascade catalytic activity, which can induce ROS burst and metabolic imbalance to drive the efficient catalytic therapy of tumors with great biocompatibility. Overall, this work provides insights into understanding the roles of tailored electronic structure and phase engineering for the regulation of catalytic activity of nanozymes. Meanwhile, it may open prospects for future biomedical applications of liquid metals.

Methods

Ethical statement

All studies used healthy female Balb/c mice (Jilin University Laboratory Animal Center, China), which were 7, 8 weeks old and weighed 18–20 g. All animal procedures were approved by the Ethics Committee of Science and Technology of Northeast Normal University (202502088). According to this protocol, the maximal tumor burden is limited to 2000 mm³. In our study, all tumor sizes were kept below 2000 mm³. Sex was not taken into consideration in the study design. Specific pathogen-free conditions were maintained for mouse housing: 26 ± 1 °C, 50 ± 5% humidity, with a 12 h light-dark cycle.

Cell lines

4T1 cell line (Catalog no. CRL-2539, American Type Culture Collection, ATCC), 293T cell line (HEK293T, GDC0067, the China Center for Type Culture Collection, CCTCC).

Preparations of LM MoS_X

Typically, 0.1 g liquid metal alloy was placed in 50 mL (NH₄)₂MoS₄ solution (20 mM). Subsequently, the mixture was sonicated with a tip ultrasonicator at 1000 W and 60% amplitude for 10 min. The ultrasound was applied in 5-second pulses with a 3 s interval. An ice bath was used to keep the mixture temperature constant during ultrasonication. Following the reaction, the largest particles precipitated within seconds. To remove large particles, the remaining nanospheres were subjected to centrifugation (95 × g, 5 min). The product was centrifuged (13708 × g, 8 min) and washed repeatedly with Milli-Q water. To optimize the ultrasonication power, the reaction system was exposed to different power levels.

ICP-OES measurement

The sample (1 mg) was digested with 1 mL of aqua regia (HCl:HNO₃ = 3:1, v/v). After standing at room temperature for 30 min, the mixture was heated (80 °C) until a clear solution was obtained. Elemental analysis was performed using ICP-OES (Agilent 5800).

Peroxidase-like Catalytic Activity Characterization and Kinetic measurements

The experiment was conducted in NaAc-HAc buffer solution (pH =4.0, 25 mM) with LM MoS_X (10 µg/mL) and ABTS (2.5 mM) in the presence of H₂O₂ (0.25, 0.5, 1, 2.5, 5, and 10 mM). The catalytic activity was assessed by monitoring the absorbance of ABTS at 415 nm over time. According to the Michaelis–Menten model, ν = V_max[S]/(K_m + [S]), where ν represents the initial velocity, V_max is the maximal reaction velocity, and [S] is substrate concentration.

EPR Detection of •OH and O₂

For the EPR assay, •OH and O₂^•− were detected using DMPO as the trapping probe. In a typical process, samples were dispersed in HAc-NaAc buffer (pH = 4.0, 25 mM) (solvent for •OH) and methanol (solvent for O₂^•−), respectively. (10 mM H₂O₂, 100 µg/mL LM MoS_X, 100 mM DMPO)

Trapping experiments

To determine the active species generated by LM MoS_X nanozymes, IPA (isopropanol alcohol, 10 μL) and PBQ (1,4-Benzoquinone, saturated PBQ solution) were used as scavengers (10 μL) of •OH and O₂^•−, respectively.

Anti-interference study

NaCl, KCl, CaCl₂, Na₂SO₄, NaNO₃ and NaBr were employed to study the interference of anions, such as Na⁺, K⁺, Ca²⁺ Cl^-, SO₄^2-, NO₃^- and Br^-, and the concentration of each anion was set at 10 mM. For the protein interference study, 10 μg/mL BSA (bovine serum albumin) was employed.

Assessment of Catalytic Stability of LM MoS_X Nanozymes

The nanozyme was dispersed in PBS and stored at room temperature for a certain period of time, and its POD-like activity was measured at a certain time. All catalytic stability experiments were evaluated in the NaAc-HAc buffer solution (pH 4.0, 25 mM), which contained LM MoS_X (10 µg/mL), ABTS (2.5 mM), and H₂O₂ (10 mM).

O₂^•- generation

DPBF (1 mg/mL, 100 µL) solution was added to LM MoS_X solution (50 µg/mL, 600 µL). Absorbance changes were monitored with a UV–Vis spectrophotometer.

Detection of hydroxyl radical (•OH)

•OH generation was probed by the fluorescence of TAOH (E_m= 435 nm). Experiments were conducted in PBS (25 mM, pH 7.4) with the following mixtures: TA (0.5 mM); TA and H₂O₂ (1 mM); TA and nanozymes (50 μg/mL); only nanozymes; TA, nanozymes and H₂O₂. After 12 h, the samples were centrifuged and measured by fluorescence spectra.

In vitro GSH consumption

DTNB (2.0 mg/mL, 100 µL) and GSH (1 × 10⁻³ M, 200 µL) solutions were added into LM MoS_X solutions (1.7 mL) with varying concentrations (0, 20, 40, 60, 80, and 100 μg/mL). After 6 h magnetic stirring (37 °C), the residual GSH was determined by absorbance spectroscopy.

Catalytic reactions of LM MoS_X with NADH

LM MoS_X solution (50 µg/mL, 2.5 mL) was incubated with NADH solution (50 µg/mL) under magnetic stirring at 25 °C. The amount of NADH was determined by absorbance spectroscopy.

Detection of H₂O₂ formation during NADH oxidation

NADH (800 μL, 5 mM) was incubated with LM MoS_X solution (1 mL, 100 µg/mL) and reacted for 5 min. Then, the mixture was transferred to a solution containing Amplex Red (40 μM) and HRP (10 U/mL). Absorbance changes were monitored with a UV–Vis spectrophotometer.

Evaluation of substrate selectivity of LM MoS_X

To study the selectivity of LM MoS_X towards NADH, uric acid (UA, 100 μM), xanthine (XA, 100 μM), L-Tyrosine (100 μM) and NADPH (100 μM) were mixed with LM MoS_X (final concentration: 50 μg/mL), respectively. The uric acid content was detected (λ_max 290 nm) by UV-vis spectroscopy. The xanthine content was monitored (λ_max 270 nm) by UV-vis spectroscopy. To confirm the dityrosine formation (for L-Tyrosine), fluorescence spectra were recorded (E_X = 300 nm). NADPH content was measured at wavelength 340 nm by UV-vis spectroscopy.

Computational method

All DFT calculations employed the Vienna Ab initio Simulation Package (VASP 4.6) software. Amorphous models were obtained by annealing through the ab initio molecular dynamics (AIMD) method. The simulations employed the canonical (NVT) ensemble with a Nosé-Hoover thermostat. The system was initially heated to and maintained at 2000 K for 10 ps using a 1 fs time step, and was finally quenched to 298.15 K. The generalized gradient approximation of Perdew-Burke-Ernzerhof (GGA-PBE) was used to describe the exchange-correlation potential. The ion cores–valence electrons were described using the projector augmented-wave (PAW) method. A plane-wave cutoff energy of 450 eV was used. The structural relaxation was performed with convergence thresholds set to − 0.02 eV/Å for the Hellmann–Feynman forces and 10⁻⁵eV for the energy change. Charge density analysis was obtained from CHGCAR, and all atomic/electronic structures were illustrated with VESTA software.

Definition of d-band center

The d-band center is defined as the energy center of the d-orbital electron density of states (DOS) for transition metal atoms, which is the weighted average energy value of all electronic states in the d-orbital according to their DOS. The algorithm is as follows:

$${\epsilon }_d=\frac{\int \epsilon {\cdot n}_d\left(\epsilon \right)d\epsilon }{\int n_d\left(\epsilon \right)d\epsilon }$$

n_d (ε) is the d-orbital electron density of states, ε is the energy of electronic states, and the integration range covers all electronic states of the d orbital (usually from the binding state below the Fermi level to the antibonding state above).

Cytotoxicity assay

4T1 cells or 293T cells were plated in a 96-well plate (1 × 10⁴ cells/well) and cultured for 24 h, followed by a 24 h treatment with varying concentrations of LM NPs, MoS₂ and LM MoS_X (0, 25, 50, 75, and 100 µg/mL). The MTT solution (0.5 mg/mL) was incubated with cells for 4 h at 37 °C and then replaced with 100 µL of DMSO to dissolve the formazan product. Subsequently, cell viability was analyzed using a microplate reader.

Cellular uptake

4T1 cells were treated with LM MoS_X@RB (40 μg/mL). Afterward, cells were incubated with Hoechst (5 µg/mL) for 15 min and visualized using CLSM.

Determination of ROS generation in vitro

4T1 cells were treated with LM NPs, MoS₂, LM MoS_X (50 μg/mL) for 24 h. Afterward, the cells were incubated with DCFH-DA (10 μM, 30 min), followed by Hoechst staining. The cells were imaged using CLSM.

Detection of intracellular GSH/ NADH level

4T1 cells were seeded in 6-well plates at 5 × 10⁵ cells per well and cultured for 24 h, followed by a 12 h treatment with LM MoS_X at concentrations of 0, 25, 50, 75, and 100 µg/mL. The corresponding GSH contents were quantified using Ellman’s reagent, and the corresponding NADH contents were quantified using the NAD⁺/NADH assay kit.

Live/dead assay

4T1 cells were planted in 6-well plate (5 × 10⁵ cells/well) and incubated for 24 h. 4T1 cells were incubated with various nanoparticles (50 μg/mL) for 24 h. The cells were co-stained with calcein AM (5 µg/mL) and PI (10 µg/mL) for 30 min prior to visualization by CLSM.

Synthesis of H-LM MoS_X

The as-prepare LM MoS_X suspension (2 mg/mL) was mixed with hyaluronic acid (0.1 mg/mL) and stirred at room temperature for 4 h in the dark, followed by harvest and washing with Milli-Q water. H-LM and H-MoS₂ were obtained using the same method.

The biocompatibility of the H-LM MoS_X in mice

Mice were randomly assigned to two groups, Control (PBS), H-LM MoS_X (10 mg/kg, intravenous injection). Body weight was monitored every 4 days for 28 days following injection. After 28 days, mice were euthanized for blood collection (for complete blood panel analysis and serum biochemistry) and organ harvesting. The harvested organs were subjected to H&E staining for microscopic examination.

In vivo pharmacokinetic of H-LM MoS_X

H-LM MoS_X@Cy3 (10 mg/kg) was intravenously administered to healthy mice. At various time points post-injection (0.25, 0.5, 1, 2, 4, 8, 12, and 24 h), mice were anesthetized for blood collection. The collected blood was centrifuged (95 × g, 10 min), and then detected by a fluorescence spectrophotometer.

In vivo anti-tumor study

When tumor volume reached approximately 50 mm³ following subcutaneous inoculation of 4T1 cells, mice were randomly allocated into four experimental groups (n = 4) for intravenous administration of the respective treatments. PBS (I), H-LM (II), H-MoS₂ (III), H-LM MoS_X (10 mg/kg) (IV). Body weight and tumor volume (width²× length/2) were monitored. At the study endpoint, mice were euthanized. Following H&E or TUNEL staining, the harvested tumors and major organs were examined under an optical microscope.

In vivo biodistribution of H-LM MoS_X and LM MoS_X after injection

4T1 tumor-bearing mice received intravenous injections of the various formulations (10 mg/kg, with equal Mo). At the 12 h time point, organs and tumors were harvested and weighed. The samples were digested with 1 mL of aqua regia (HCl:HNO₃ = 3:1, v/v) under heat treatment (80 °C) for 3 h, then cooled and left overnight at room temperature. Subsequently, the digested mixtures were centrifuged (595 × g, 10 min) and filtered. The Mo content was quantified by ICP-MS.

Statistical analysis

All data were expressed in this article as mean result ± standard deviation (SD). Statistical analysis was conducted using OriginPro 2020, GraphPad Prism 8, and Microsoft Excel software. Statistical evaluation was performed using the Student’s t test analysis.

Reporting summary

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

Author contributions

X.Q. and J.R. designed research; W.Z. and J.Z. performed research and analyzed data; W.Z. and X.Q. wrote the paper.

Peer review

Peer review information

Nature Communications 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 presented in the article and Supplementary Information. Source data are provided with this paper. All data underlying this study are available from the corresponding author upon request. Source data are provided in this paper.

Competing interests

The authors declare no competing interests.

Supplementary information

The online version contains supplementary material available at 10.1038/s41467-026-70795-4.