Synthesis and characterization of Zn-based SAzymesAs shown in Fig. 1a, the Zn-SA/CNCl SAzyme was synthesized by the NaCl-co-pyrolysis strategy, during which NaCl and Zeolite Imidazole Framework-8 (ZIF-8) were pyrolyzed under argon atmosphere at 950  °C. The characterization of ZIF-8 was exhibited in Supplementary Figs. 1–3. As shown in Supplementary Fig. 4, the NaCl in a separate ceramic boat as the chlorine source was placed in the upstream direction of ZIF-8 powder. During the pyrolysis of ZIF-8 at 950  °C, the isolated Zn-N4 sites were atomically dispersed on the N-doped carbon substrates. Above the melting point of NaCl (801  °C), volatile Cl species from the melted NaCl evaporated and was captured by partial Zn-N4 sites, with the formation of Zn-N4Cl1 sites. As shown in Fig. 1b, the high angle annular dark field scanning transmission electron microscopy (HAADF-STEM) image of Zn-SA/CNCl SAzyme excluded the existence of Zn-based nanoparticles after pyrolysis. The corresponding energy dispersive X-ray (EDX) spectroscopy elemental mapping of Zn-SA/CNCl SAzyme in Fig. 1c confirmed the homogeneous elemental distribution of C, N, Zn and Cl elements on the sample. The atomic ratio of Zn to Cl was 1.8:1 by EDX spectroscopy elemental analysis, confirming that only partial Zn ISAS existed as Zn-N4Cl1 sites. The aberration-corrected scanning transmission electron microscopy (AC-STEM) of Zn-SA/CNCl SAzyme in Fig. 1d also confirmed the existence of Zn ISAS. We also synthesized the Zn-SA/CN SAzyme and Zn-SA/CNCl-HCl SAzyme as reference samples. The Zn-SA/CN SAzyme was obtained by pyrolysis of pure ZIF-8 without NaCl at 950  °C. The Zn-SA/CNCl-HCl SAzyme was prepared by etching the Zn-SA/CNCl SAzyme under concentrated hydrochloric acid at 100  °C for 24 h. As shown in Fig. 1e and Fig. 1f, the HAADF-STEM image and corresponding EDX spectroscopy elemental mapping of Zn-SA/CN SAzyme also excluded the existence of Zn-based nanoparticles and confirmed the uniform elemental distribution of C, N and Zn on the Zn-SA/CN SAzyme. The Zn ISASs from Zn-SA/CN SAzyme were directly observed by AC-STEM image in Fig. 1g. As shown in Supplementary Fig. 5, after etching the Zn-SA/CNCl SAzyme in concentrated hydrochloric acid at 100  °C for 24 h, the density of Zn ISASs from Zn-SA/CNCl-HCl SAzyme sharply decreased, confirmed by EDX spectroscopy elemental mapping and AC-STEM image. The Zn contents of Zn-SA/CNCl SAzyme, Zn-SA/CN SAzyme and Zn-SA/CNCl-HCl SAzyme were 1.06 wt%, 4.89 wt% and 0.20 wt%, respectively, determined by inductively coupled plasma optical emission spectrometry (ICP-OES) measurement. The X-ray diffraction (XRD) results of Zn-based SAzymes were exhibited in Supplementary Fig. 6. The two broad peaks around 25o and 44o from Zn-based SAzymes were ascribed to the characteristic carbon (002) and (100)/(101) diffractions, respectively. The elemental compositions of C, N, Cl, and Zn of Zn-based SAzymes were analyzed by X-ray photoelectron spectroscopy (XPS) measurement in Supplementary Figs. 7–9. In Supplementary Fig. 7, the C-Cl bond and Zn-Cl bond co-existed in the Cl 2p spectrum of Zn-SA/CNCl SAzyme. Considering the atomic ratio of Zn to Cl was 1.8:1, therefore partial Zn ISAS existed as Zn-N4Cl1 sites. After etching Zn-SA/CNCl SAzyme to obtain Zn-SA/CNCl-HCl SAzyme by concentrated hydrochloric acid, the peak intensity of Zn 2p spectrum and Zn-Cl bond in the Cl 2p spectrum in Supplementary Fig. 8 sharply decreased, indicating that the Zn-N4Cl1 sites were etched by hydrochloric acid. The nitrogen sorption isotherm experiments were performed to determine the Brunauer-Emmett-Teller (BET) surface areas of Zn-based SAzymes in Supplementary Figs. 10–13. The BET surface areas of Zn-SA/CNCl, Zn-SA/CNCl-HCl, and Zn-SA/CN were 1101 m2 g−1, 1253 m2 g−1 and 1029 m2 g−1, respectively. The Zn-SA/CNCl and Zn-SA/CN had similar BET surface areas while the BET surface area of Zn-SA/CNCl-HCl increased after etching Zn-SA/CNCl by HCl. The Zn-SA/CNCl and Zn-SA/CN also had similar distribution of mesopore while the Zn-SA/CNCl-HCl had more mesopore after HCl etching. The Raman spectra of Zn-based SAzymes were exhibited in Supplementary Fig. 14. The disorder peak (D) at around 1340 cm−1 was attributed to graphite nanocrystalline boundaries and inplane defects. While the graphitic peak (G) at around 1580 cm−1 was relative to the degree of graphitization. The ratios between the intensities of D peak and G peak (ID: IG) of Zn-SA/CNCl, Zn-SA/CN and Zn-SA/CNCl-HCl were 1.00, 1.02 and 1.04, respectively, indicating the similar structure of carbon substrates of Zn-based SAzymes.Fig. 1: The synthetic procedure and characterization of Zn-based SAzymes.a The schematic illustration of NaCl-co-pyrolysis strategy. The gray, blue, yellow and green balls represented C, N, Zn and Cl atoms, respectively. b,c The HAADF-STEM image of Zn-SA/CNCl SAzyme and the corresponding EDX spectroscopy elemental mapping results. d The AC-STEM image of Zn-SA/CNCl SAzyme. e,f The HAADF-STEM image of Zn-SA/CN SAzyme and the corresponding EDX spectroscopy elemental mapping results. g The AC-STEM image of Zn-SA/CN SAzyme.In order to reveal the coordination structures of Zn catalytic sites at atomic level, the X-ray absorption near-edge structure (XANES) and extended X-ray absorption fine structure (EXAFS) measurements at Zn K-edge were performed. The normalized XANES curves at Zn K-edge of Zn-SA/CNCl, Zn-SA/CN, Zn foil and ZnO were compared simultaneously in Fig. 2a. The locations of the near-edge absorption of Zn-SA/CNCl and Zn-SA/CN samples located between those of Zn foil and ZnO, indicating that the Zn sites from Zn-SA/CNCl and Zn-SA/CN samples were in positive valence states. Compared with the Zn-SA/CN, after introducing the Cl element into Zn-SA/CNCl, the valence state of Zn element from Zn-SA/CNCl slightly increased, which was confirmed by the positive shift in the XANES spectra in Supplementary Fig. 15a and the Zn 2p XPS spectra in Supplementary Fig. 15b. To analyze the coordination shell of the Zn samples, the corresponding k3-weighted Fourier transform EXAFS (FT-EXAFS) curves in R space without correction of radical distance phase were exhibited in Fig. 2b. The dominant peak at around 2.2 Å from the FT-EXAFS curve of Zn foil was assigned to Zn-Zn bond. The two prominent peaks at around 1.5 Å and 2.9 Å from ZnO were ascribed to the first coordination shell of Zn-O pathway and the second coordination shell of Zn-O-Zn pathway, respectively. By comparison, only one dominant peak at around 1.5 Å existed in Zn-SA/CNCl and Zn-SA/CN samples, demonstrating the sole existence of Zn ISAS from both Zn-SA/CNCl and Zn-SA/CN, without formation of Zn or ZnO. As shown in Fig. 2c, the peak intensity at around 1.8 Å of the FT-EXAFS curve of Zn-SA/CNCl was stronger than that of Zn-SA/CN, which was attributed to the formation of Zn-Cl bond from partial Zn ISAS of Zn-SA/CNCl.Fig. 2: The characterization of Zn-based SAzymes by XANES and EXAFS measurements.a, b The XANES spectra and corresponding FT-EXAFS results in R space of Zn-SA/CNCl, Zn-SA/CN, Zn foil and ZnO. c The comparison of FT-EXAFS spectra in R space of Zn-SA/CNCl and Zn-SA/CN. d, e The fitting results of Zn-SA/CNCl and Zn-SA/CN in R space. f The fitting result of Zn-SA/CNCl in k space. g–i The WT analysis of Zn-SA/CNCl, Zn-SA/CN, and Zn foil.As analyzed by Supplementary Figs. 16–22, the Zn-N4Cl1/Zn-N4(OH2) and Zn-N4(OH2)/Zn-N4(OH2) models were utilized to simulate the Zn-SA/CNCl and Zn-SA/CN, respectively. As shown in Supplementary Figs. 23–25, by combination of the in-situ EXAFS experiment and the EXAFS experiment at Cl K-edge, we confirmed the adsorption of water molecules on Zn atoms and the existence of Zn-Cl bond in Zn-SA/CNCl SAzyme. The corresponding fitting results of the FT-EXAFS curves of Zn-based samples were shown in Fig. 2d–f, Supplementary Fig. 26 and Supplementary Table 1. The fitting results exhibited the coordination numbers of Zn-N/O bond and Zn-Cl bond were 4.55 and 0.55 in Zn-SA/CNCl, respectively, revealing the Zn-N4Cl1/Zn-N4(OH2) model was a rational model to simulate the Zn-SA/CNCl. By comparison, the coordination numbers of Zn-N bond and Zn-Cl bond were 3.97 and 0.57 in Zn-SA/CNCl-heating during in-situ EXAFS experiment, revealing the Zn-N4Cl1/Zn-N4 model was a rational model to simulate the Zn-SA/CNCl-heating. The wavelet transform (WT) analysis of Zn samples at the Zn K-edge was also performed owing to its unique advantages with powerful resolutions in both k and R spaces. The WT contour plots of Zn-SA/CNCl, Zn-SA/CN and Zn foil were exhibited in Fig. 2g, Fig. 2h and Fig. 2i, respectively. As shown in Fig. 2g, the intensity of the dominant peak at around 1.53 Å in R space and 3.90 Å−1 in k space from the WT contour plot of Zn-SA/CNCl was stronger than that of Zn-SA/CN in Fig. 2h (1.50 Å in R space and 3.90 Å−1 in k space), which was ascribed to the existence of Zn-Cl bond from Zn-SA/CNCl. The prominent peak at around 2.30 Å in R space and 6.80 Å−1 in k space from the WT contour plot of Zn foil in Fig. 2i was attributed to the Zn-Zn bond, which were absent in both Zn-SA/CNCl and Zn-SA/CN. Therefore, we revealed that the Zn-N4(OH2) sites and Zn-N4Cl1 sites co-existed in Zn-SA/CNCl and the Zn-N4(OH2) sites solely existed in Zn-SA/CN by combination of FT-EXAFS fitting results and WT analysis.Peroxidase-like activities of Zn-based SAzymesThen, we tested the peroxidase-like activities of Zn-SA/CNCl SAzyme and Zn-SA/CN SAzyme by the oxidation of TMB with H2O2 as oxidant. The blue-colored oxidized TMB (oxTMB) as the product had a characteristic absorbance peak at 652 nm. Figure 3a showed the reaction-time curves of the TMB oxidation reaction catalyzed by Zn-based SAzymes. The Zn-SA/CNCl SAzyme exhibited much superior peroxidase-like activity than Zn-SA/CN SAzyme, with a sharp improvement of reaction rate by introducing Zn-N4Cl1 sites as the catalytic regulator of Zn-N4 catalytic sites. The Zn-SA/CNCl-HCl SAzyme, which was obtained by etching the Zn ISAS from Zn-SA/CNCl SAzyme with concentrated hydrochloric acid, had an obvious decay in peroxidase-like activity compared to Zn-SA/CNCl SAzyme, indicating that the Zn ISASs served as the catalytic sites. In Fig. 3b and Fig. 3c, we investigated the effects of the pH values and the catalytic temperatures on the peroxidase-like activity of Zn-SA/CNCl SAzyme. As shown in Fig. 3b, the peroxidase-like activity of Zn-SA/CNCl SAzyme exhibited a volcanic trend as the pH values increasing from 2 to 9, with the optimal pH value of 4, indicating that Zn-SA/CNCl SAzyme exhibited high peroxidase-like activity under a wide pH range and possessed a pH-dependent activity similar with natural enzymes. In Fig. 3c, the Zn-SA/CNCl SAzyme also showed excellent peroxidase-like activity under a wide range of temperatures (30-45 °C), with the optimal catalytic temperature of 39 °C, demonstrating the high stability of Zn-SA/CNCl SAzyme in harsh conditions. Except for TMB as peroxidase substrate, we also investigated the oxidation of other peroxidase substrates, such as 2,2’-azino-bis (3-ethylbenzothiazoline-6-sulfonic acid) (ABTS) and o-phenylenediamine (OPD) catalyzed by Zn-based SAzymes. In Fig. 3d and Fig. 3e, Zn-SA/CNCl SAzyme also exhibited much higher peroxidase-like activity than Zn-SA/CN SAzyme without Zn-N4Cl1 catalytic regulators during the oxidation of ABTS and OPD. Similarly, the Zn-SA/CNCl-HCl SAzyme also had inferior peroxidase-like activity than Zn-SA/CNCl SAzyme, demonstrating the catalytic role of Zn-ISAS. As shown in Fig. 3f and Fig. 3g, we investigated the kinetic parameters of peroxidase substrate TMB catalyzed by Zn-SA/CNCl SAzyme by changing the concentration of TMB (0-1.0 mM). The Vmax and the Michaelis constant (Km) of TMB were calculated to be 7.27 × 10−4 M min−1 and 1.01 mM, respectively, according to Michaelis-Menten equation (Fig. 3g). Similarly, as shown in Fig. 3h and Fig. 3i, we also investigated the kinetic parameters of H2O2 by changing the concentrations of H2O2 (0-3.0 M) during catalysis. As shown in Fig. 3i, the Vmax and the Km of H2O2 were determined as 3.46 × 10−4 M min−1 and 147 mM, respectively. For comparison, we also measured the kinetic parameters (Vmax and Km) of substrates (TMB and H2O2) catalyzed by Zn-SA/CN SAzyme and Zn-SA/CNCl-HCl SAzyme in Supplementary Figs. 27–28. The corresponding kinetic parameters, such as Vmax, Km, the catalytic constant (kcat) and the catalytic efficiency (kcat/Km) catalyzed by different Zn-based SAzymes were summarized in Fig. 3j. The Vmax (7.27 × 10−4 M min−1), kcat (897.5 min−1), and kcat/Km (8.89 × 105 M−1 min−1) of TMB catalyzed by Zn-SA/CNCl SAzyme were 346 times, 1496 times, and 133 times those of Zn-SA/CN SAzyme without the Zn-N4Cl1 catalytic regulators (Vmax = 2.10 × 10−6 M min−1, kcat = 0.6 min−1 and kcat/Km = 6.67×103 M−1 min−1), demonstrating that the atom-pair engineering of the Zn-N4 catalytic sites and the Zn-N4Cl1 catalytic regulator effectively boosted the peroxidase-like activities of Zn-SA/CNCl SAzyme. Besides, for the kinetic parameters of H2O2, the Vmax (3.46 × 10−4 M min−1), kcat (427.2 min−1), and kcat/Km (2.91 × 103 M−1 min−1) catalyzed by Zn-SA/CNCl SAzyme were 177 times, 854 times, and 523 times those of Zn-SA/CN SAzyme (Vmax = 1.95 × 10−6 M min−1, kcat = 0.5 min−1 and kcat/Km = 5.56 M−1 min−1). After etching the Zn-SA/CNCl SAzyme in concentrated hydrochloric acid at 100 °C for 24 h, the Zn content of Zn-SA/CNCl-HCl SAzyme was 0.20 wt% Zn, a sharp decrease compared to that of Zn-SA/CNCl SAzyme with 1.06 wt% Zn. The Vmax of TMB and H2O2 catalyzed by Zn-SA/CNCl SAzyme were 4.8 times and 2.4 times those of Zn-SA/CNCl-HCl SAzyme, further indicating that the Zn ISAS served as catalytic sites. We compared the kinetic parameters of TMB and H2O2 catalyzed by Zn-SA/CNCl SAzyme and other reported SAzymes in Supplementary Tables 2-3 and Supplementary Fig. 29. As shown in Supplementary Fig. 29, the Vmax of TMB and H2O2 catalyzed by Zn-SA/CNCl SAzyme were excellent among other reported noble and non-noble metal SAzymes. Especially, compared with the reported Zn-N4 SAzyme (PMCS)48, the Vmax of TMB and H2O2 catalyzed by Zn-SA/CNCl SAzyme were increased by 114 times and 47 times, respectively, demonstrating that the atom-pair engineering of Zn-SA/CNCl SAzyme effectively boosted the peroxidase-like activity of Zn-N4 catalytic sites. As shown in Supplementary Figs. 30–31, we characterized the Zn-SA/CNCl SAzyme after catalysis by HAADF-STEM, AC-STEM and XPS measurements, revealing the atomic dispersion of Zn element from Zn-SA/CNCl SAzyme after catalysis. After catalysis, the Zn content of Zn-SA/CNCl SAzyme was 0.92 wt%, determined by ICP-OES measurement.Fig. 3: The peroxidase-like activity and kinetics of Zn-based SAzymes.a The reaction-time curves of the TMB colorimetric reaction catalyzed by Zn-based SAzymes. The Error bars in Fig. 3 represent s.d. obtained from three independent experiments. b, c The effects of pH values and temperatures during TMB colorimetric reaction catalyzed by Zn-SA/CNCl SAzyme. d, e The reaction-time curves of the ABTS and OPD colorimetric reactions catalyzed by Zn-based SAzymes. f, g The reaction-time curves of the TMB colorimetric reaction with different concentration of TMB catalyzed by Zn-SA/CNCl SAzyme and the corresponding Michaelis-Menten curves fitted by ordinary Least square. h, i The reaction-time curves of the TMB colorimetric reaction with different concentration of H2O2 catalyzed by Zn-SA/CNCl SAzyme and the corresponding Michaelis-Menten curves fitted by ordinary Least square. j The comparison of the kinetics based on Zn active sites of Zn-based SAzymes. [E] is the molar concentration of the Zn active sites of Zn-based SAzymes. Km is the Michaelis constant. Vmax is the maximal reaction velocity. kcat (kcat = Vmax/[E]) is the catalytic constant. The kcat/Km value indicates the catalytic efficiency of the Zn-based SAzymes.As shown in Supplementary Fig. 32, we compared the specific activities of Zn-based SAzymes. The specific activity of Zn-SA/CNCl SAzyme was 66.7 U mg−1, which was around 334 times that of Zn-SA/CN SAzyme (0.2 U mg−1), demonstrating that Zn-N4Cl1 catalytic regulator remarkably enhanced the peroxidase-like activity of Zn-N4 catalytic sites. After HCl etching of Zn ISAS, the specific activity of Zn-SA/CNCl-HCl SAzyme exhibited an obvious decay to 23.2 U mg−1, revealing the catalytic role of Zn ISAS in Zn-SA/CNCl SAzyme during catalysis. As exhibited in Supplementary Table 4, the specific activity of Zn-SA/CNCl (66.7 U mg−1) was excellent among most of the reported SAzymes. In order to analyze the reactive oxygen species during the activation of H2O2, electronic spin resonance (ESR) experiment was performed. As shown in Supplementary Fig. 33, we observed the formation of ·OH during the activation of H2O2 by Zn-SA/CNCl SAzyme with 5,5-dimethyl-1-pyrroline N-oxide (DMPO) as a spin trapping agent. For comparison, the signal intensity of ·OH was negligible during the activation of H2O2 by Zn-SA/CN SAzyme, which well explained the superior peroxidase-like activity of Zn-SA/CNCl SAzyme than Zn-SA/CN SAzyme during catalysis.We further investigated the kinetic parameters of peroxidase substrate ABTS and OPD catalyzed by Zn-based SAzymes. As shown in Supplementary Fig. 34, for the oxidation of ABTS, the Vmax (5.33 × 10−4 M min−1), kcat (658.0 min−1), and kcat/Km (3.12 × 104 M−1 min−1) of ABTS catalyzed by Zn-SA/CNCl SAzyme were 70 times, 329 times, and 40 times those of Zn-SA/CN SAzyme (Vmax = 7.60 × 10−6 M min−1, kcat = 2.0 min−1 and kcat/Km = 7.81 × 102 M−1 min−1). Similarly, for the oxidation of OPD in Supplementary Fig. 35, the Vmax (1.02 × 10−4 M min−1), kcat (125.9 min−1), and kcat/Km (4.37 × 10 4M−1 min−1) of OPD catalyzed by Zn-SA/CNCl SAzyme were 182 times, 839 times, and 585 times those of Zn-SA/CN SAzyme (Vmax = 5.59 × 10−7 M min−1, kcat = 0.15 min−1 and kcat/Km = 74.7 M−1 min−1). Taken together, with atom-pair engineering strategy by simultaneously constructing both Zn-N4 catalytic sites and Zn-N4Cl1 catalytic regulators, we have achieved the transformation of Zn-SA/CN SAzyme with low peroxidase-like activity to high-performance Zn-SA/CNCl SAzyme, demonstrating the tremendous potentials of atom-pair engineering for boosting the activity of SAzyme.Inhibition of tumor cell growth in vitro and in vivoConsidering the excellent peroxidase-like activity of Zn-SA/CNCl SAzyme and the acidic micro-environment of tumor cell, we further explored the therapeutic efficacy of Zn-SA/CNCl SAzyme for tumor cell suppression. Firstly, we compared the killing effects of Zn-SA/CNCl SAzyme and Zn-SA/CN SAzyme on different tumor cells in vitro, such as 4T1, MCF7, A549 and A375 cells. As shown in Fig. 4a, as the concentrations of SAzymes increasing from 0 μg ml-1 to 200 μg ml−1, the 4T1 cell viability gradually decreased. At a concentration of 200 μg ml−1 SAzymes, the 4T1 cell viability was 13.8% killed by Zn-SA/CNCl SAzyme, much lower than that of Zn-SA/CN SAzyme with 66.4% of 4T1 cell viability, indicating that the Zn-SA/CNCl SAzyme effectively suppressed the viability of 4T1 cell in vitro. Similarly, for killing the MCF7, A549 and A375 tumor cells, the Zn-SA/CNCl SAzyme also exhibited superior inhibitory effects than Zn-SA/CN SAzyme for suppressing the viability of tumor cells (Fig. 4b–d). As shown in Supplementary Fig. 36, we also evaluated the killing effects of Zn-based SAzymes on normal mammalian cells, revealing the good compatibility of Zn-SA/CNCl SAzyme. As shown in Fig. 4e, f, we directly observed the Zn-SA/CNCl SAzyme and Zn-SA/CN SAzyme in 4T1 cells by transmission electron microscopy (TEM) measurement, demonstrating the successful phagocytosis of SAzymes by 4T1 cells. The SAzymes mainly localized into the endosome and lysosome in 4T1 cells. To compare the content of cytosolic reactive oxygen species (ROS) in 4T1 cells after phagocytosis of SAzymes, we employed the laser scanning confocal microscopy by utilizing the fluorescent probe H2DCFDA. The fluorescence intensity inside the 4T1 cells with Zn-SA/CNCl SAzyme (Fig. 4g) was much stronger than that of the 4T1 cells with Zn-SA/CN SAzyme (Fig. 4h), revealing the Zn-SA/CNCl SAzyme was more advantageous for the formation of ROS species in 4T1 cells. As shown in Fig. 4i, the corresponding flow cytometry results demonstrated that the signal intensity of intracellular ROS inside 4T1 cells treated by Zn-SA/CNCl SAzyme obviously increased compared to that of 4T1 cells treated by Zn-SA/CN SAzyme and control group without SAzyme, revealing the Zn-SA/CNCl SAzyme effectively facilitated the formation of ROS by the atom-pair engineering of SAzyme.Fig. 4: The inhibition of tumor cell growth by Zn-based SAzymes in vitro and in vivo.a–d The cell viabilities of 4T1, MCF7, A549 and A375 cells after incubation with different concentration of Zn-SA/CNCl and Zn-SA/CN SAzymes. e, f The TEM images for visualization of the internalization of Zn-based SAzymes in 4T1 cells. g, h The intracellular reactive oxygen species (ROS) in 4T1 cells treated with Zn-based SAzymes. The 4T1 cells were stained with H2DCFDA and we observed the change in fluorescence by confocal microscopy. i The intracellular ROS levels of 4T1 cells treated with 100 μg ml−1 Zn-based SAzymes and control group without Zn-based SAzymes by flow cytometry results. j, k The changes of tumor volume and body weight of 4T1-tumor-bearing mice during 15 day treatment. (****P-value < 0.0001, ***P-value: 0.0003; ns not significant.) The Error bars in Fig. 4 represent s.d. obtained from five independent experiments.To investigate the antitumor effect of Zn-based SAzyme in vivo, we carried out the antitumor tests of the 4T1-tumor-bearing mice, which were treated with Zn-SA/CNCl SAzyme and Zn-SA/CN SAzyme by intratumoral injection during 15 day treatment. We subcutaneously implanted 1.0 × 106 4T1 cells into the right flank of 6–8 weeks old Balb/c female mice. The Zn-based SAzyme (12.5 mg kg−1) was injected intratumorally into mice after the volume of 4T1 tumor reached around 50 mm3. As shown in Fig. 4j, after 15 day post-treatment by Zn-SA/CNCl SAzyme, the average tumor volume was 702 mm3, about 43% that of the phosphate-buffered saline (PBS) controls (1641 mm3), indicating that Zn-SA/CNCl SAzyme obviously inhibited the 4T1 tumor growth in vivo. By comparison, the average tumor volume reached 1354 mm3 treated by Zn-SA/CN SAzyme, about 83% that of PBS controls, demonstrating the Zn-N4Cl1 catalytic regulators from Zn-SA/CNCl SAzyme effectively enhanced the inhibitory effect of Zn-N4 catalytic sites by the atom-pair engineering of SAzyme. To evaluate the in vivo toxicity of the Zn-based SAzymes during tumor treatment, we measured the loss of body weights of mice during treatment, as shown in Fig. 4k. Compared with the PBS controls, the body weights of mice treated by Zn-SA/CNCl SAzyme and Zn-SA/CN SAzyme exhibited no obvious changes during the 15 day treatment. We also excluded the damage of Zn-SA/CNCl SAzyme on the normal cells surrounding the tumor tissue during treatment, as shown in Supplementary Fig. 37. To evaluate the biosafety of the Zn-based SAzymes, the pathological analysis of the main organs of tumor-bearing mice was compared in Supplementary Fig. 38, there was no abnormal organ pathological changes after treating the tumor-bearing mice with Zn-SA/CNCl SAzyme and Zn-SA/CN SAzyme, demonstrating the good biosafety of the Zn-based SAzymes.DFT studies on the peroxidase-like activity of SAzymesTo understand the regulation mechanism of Zn-N4Cl1 catalytic regulator on the Zn-N4 catalytic site, the DFT calculation was performed to study the catalytic pathways and the regulation mechanism. As exhibited in Supplementary Fig. 39, we excluded the catalytic role of isolated Zn-N4Cl1 site for adsorption and activation of H2O2 reactant. As shown in Fig. 5, we constructed four possible models of catalytic sites (Zn2-N6Cl1 in Fig. 5a, Zn2-N8Cl1-1 in Fig. 5b, Zn2-N8Cl1-2 in Fig. 5c, and Zn2-N8Cl1-3 in Fig. 5d), containing the Zn-N4 catalytic sites and the adjacent Zn-N4Cl1 catalytic regulator (Zn2-NxCl1 models). We compared the Zn-Zn distance between ZIF-8 and the Zn2-NxCl1 models in Supplementary Figs. 40–41. As shown in Supplementary Figs. 42–43, we measured the Zn-Zn distances of Zn-Zn atom-pairs in the AC-STEM image of Zn-SA/CNCl, revealing the Zn-Zn distances in Zn2-NxCl1 models were rational. The corresponding Zn2-Nx models without Zn-N4Cl1 catalytic regulator were also constructed as references. As shown in Supplementary Fig. 44, compared with the Zn2-Nx models without Zn-N4Cl1 sites, the net Bader charges of both Zn-N4 site and Zn-N4Cl1 site from Zn2-NxCl1 models slightly increased, indicating the slightly higher valence state of the Zn2-NxCl1 models, which was consistent with the experimental results of XANES analysis and XPS measurement.Fig. 5: The DFT calculations on the peroxidase-like activity of Zn-based SAzyme.a–d The catalytic pathways and the corresponding energy changes catalyzed by four possible models of catalytic sites (Zn2-NxCl1 and Zn2-Nx catalytic sites). The brown, light blue, gray and green balls represented C, N, Zn and Cl atoms, respectively.The catalytic pathway for peroxidase-like catalysis had three steps: the adsorption of H2O2, the release of ·OH radicals and re-exposure of Zn-N4 catalytic sites. We defined the total energies of each Zn2-Nx model and H2O2 before catalysis as 0 eV, respectively. As shown in Fig. 5a–d, compared with the Zn2-Nx models without Zn-N4Cl1 catalytic regulator (the blue lines), the introduction of Zn-N4Cl1 catalytic regulator adjacent to the Zn-N4 catalytic sites effectively lowered the energies of the entire catalytic pathway (the red lines), including * + H2O2, *H2O2, *OH + ·OH, and * + 2·OH, indicating that it was advantageous for catalysis on Zn2-NxCl1 model than Zn2-Nx models. In all steps, only the adsorption of H2O2 was the exothermal step while the re-exposure of Zn-N4 catalytic sites in the last step had the highest energy changes as the rate-determining step (RDS). The energy changes for the adsorption of H2O2 on Zn2-NxCl1 models in Fig. 5a–d were -0.92 eV, -0.79 eV, -0.81 eV and -0.76 eV, respectively, lower than those on the corresponding Zn2-Nx models, with -0.77 eV, -0.41 eV, -0.36 eV and -0.34 eV, respectively, demonstrating the easier adsorption of H2O2 on Zn2-NxCl1 models than the corresponding Zn2-Nx models. In the last step as RDS, the energy changes for re-exposure of Zn-N4 catalytic sites on Zn2-NxCl1 models in Fig. 5a–c were 2.67 eV, 3.64 eV, and 3.04 eV, respectively, lower than those on the corresponding Zn2-Nx models, with 3.06 eV, 3.70 eV, and 3.19 eV, respectively. While the energy change in the last step as RDS on Zn2-N8Cl1-3 in Fig. 5d (3.35 eV) was similar to that on Zn2-N8-3 in Fig. 5d (3.31 eV). These results demonstrated the easier adsorption of H2O2 and easier re-exposure of Zn-N4 catalytic sites on Zn2-NxCl1 models than Zn2-Nx models during catalysis. The corresponding optimized structures of *H2O2 and *OH on different Zn2-NxCl1 models and Zn2-Nx models were shown in Supplementary Figs. 45–48.As shown in Supplementary Fig. 49, to illustrate the stronger adsorption of H2O2 on Zn2-NxCl1 models than the corresponding Zn2-Nx models, we measured the Cl-H distance between the Cl atom from Zn-N4Cl1 site and the H atom from the *H2O2 on the Zn-N4 site. The Cl-H distances of *H2O2 on Zn2-N6Cl1, Zn2-N8Cl1-1, Zn2-N8Cl1-2, and Zn2-N8Cl1-3 were 2.14 Å, 2.03 Å, 1.99 Å and 2.14 Å, respectively, shorter than the corresponding Cl-H distance between the Cl atom from Zn-N4Cl1 site and the H atom from the *OH on the Zn-N4 site (2.99 Å, 3.62 Å, 4.92 Å and 5.15 Å), indicating the existence of interaction between the Cl atom from Zn-N4Cl1 site and the H atom from the *H2O2 on the Zn-N4 site. The Cl-H interaction facilitated the adsorption of H2O2 on Zn2-NxCl1 models than Zn2-Nx models and accelerated the activation of H2O2 during catalysis. We further analyzed the Cl-H interaction of adsorbed H2O2 on Zn2-NxCl1 models by charge density differences in Supplementary Fig. 50, also revealing the existence of Cl-H interaction. We compared the d band centers of Zn2-NxCl1 models and Zn2-Nx models in Supplementary Fig. 51. The d band centers of Zn2-NxCl1 models had an obvious upward-shift compared with Zn2-Nx models, which strengthened the adsorption of H2O2 on Zn2-NxCl1 models. As shown in Supplementary Fig. 52, we revealed that the contribution from the adjacent C-Cl bond for improving the catalytic activity of Zn-N4 site was weak. Therefore, the adjacent Zn-N4Cl1 catalytic regulators effectively facilitated the adsorption of H2O2 on Zn-N4 catalytic site and re-exposure of Zn-N4 catalytic sites, accelerated the activation of H2O2 and improved the reaction rate during catalysis.
