Biomimetic Phosphohydrolase Nanozyme Based on Defect-Engineered Metal–Organic Framework

Open AccessCCS ChemistryRESEARCH ARTICLES10 Jan 2024Biomimetic Phosphohydrolase Nanozyme Based on Defect-Engineered Metal–Organic Framework Xiaoxue Kou†, Yuhong Lin†, Yong Shen, Linjing Tong, Rui Gao, Suya Liu, Siming Huang, Fang Zhu, Guosheng Chen and Gangfeng Ouyang Xiaoxue Kou† MOE Key Laboratory of Bioinorganic and Synthetic Chemistry, School of Chemistry, Sun Yat-sen University, Guangzhou 510006 , Yuhong Lin† MOE Key Laboratory of Bioinorganic and Synthetic Chemistry, School of Chemistry, Sun Yat-sen University, Guangzhou 510006 , Yong Shen MOE Key Laboratory of Bioinorganic and Synthetic Chemistry, School of Chemistry, Sun Yat-sen University, Guangzhou 510006 , Linjing Tong MOE Key Laboratory of Bioinorganic and Synthetic Chemistry, School of Chemistry, Sun Yat-sen University, Guangzhou 510006 , Rui Gao MOE Key Laboratory of Bioinorganic and Synthetic Chemistry, School of Chemistry, Sun Yat-sen University, Guangzhou 510006 , Suya Liu Shanghai NanoPort, Thermo Fisher Scientific, Pudong District, Shanghai 200120 , Siming Huang Guangzhou Municipal and Guangdong Provincial Key Laboratory of Molecular Target and Clinical Pharmacology, The NMPA and State Key Laboratory of Respiratory Disease, School of Pharmaceutical Sciences and The Fifth Affiliated Hospital, Guangzhou Medical University, Guangzhou 511436 , Fang Zhu MOE Key Laboratory of Bioinorganic and Synthetic Chemistry, School of Chemistry, Sun Yat-sen University, Guangzhou 510006 , Guosheng Chen *Corresponding author: E-mail Address: [email protected] MOE Key Laboratory of Bioinorganic and Synthetic Chemistry, School of Chemistry, Sun Yat-sen University, Guangzhou 510006 and Gangfeng Ouyang MOE Key Laboratory of Bioinorganic and Synthetic Chemistry, School of Chemistry, Sun Yat-sen University, Guangzhou 510006 https://doi.org/10.31635/ccschem.023.202303541 SectionsSupplemental MaterialAboutAbstractPDF ToolsAdd to favoritesDownload CitationsTrack Citations ShareFacebookTwitterLinked InEmail Phosphodiester bonds are the backbone linkages of all the nucleic acids which store and transfer biological information. The hydrolysis of phosphodiester bonds is of great importance to the replication, recombination, and damage repair of DNA as well as the maturation and processing of RNA. However, the spontaneous scission of individual phosphodiester linkages is kinetically challenging under physiological conditions, with an estimated half-life of 30 million years. Here, we discover a defect-engineered Zr-metal–organic framework (Zr-MOF) nanozyme containing cluster-missing reo defects, named UiO-66 REO, that possesses intrinsic phosphodiesterase-like activity that allows for the cleavage of the phosphodiester bond in physiological condition (pH = 7.4 and 37 °C), outperforming the activity-recording Zr-MOFs. The atomic-level structure of the UiO-66 REO nanozyme has been directly identified by the synchrotron radiation absorption spectrum and integrated differential phase contrast-scanning transmission electron microscopy. We find that both the defective boundary regions and unusual Zr6-clusters formed in the reo phase of the UiO-66 REO nanozyme contribute to the improvement in efficiency of phosphodiesterase-like hydrolysis, which realizes the cleavage of DNA in mild conditions. This work offers a new insight into biomimetic enzymology using crystal and defect engineering and opens up new possibilities for the development of low-cost, structurally stable and sustainable phosphodiesterase nanozymes for different biological applications. Download figure Download PowerPoint Introduction Nanomaterials with enzymatic activities are an emerging class of biomimetic catalysts considered to be excellent substitutes for their natural counterparts, also called nanozymes.1,2 Nanozymes have attracted tremendous attention due to their sustainable synthesis and long-term storage ability, high catalytic efficiency, high environmental resistance, and cost-effectiveness.3–6 According to the type of catalytic reaction they involve, nanozymes can be classified as an oxidoreductase mimic, such as oxidase, peroxidase, catalase, superoxide dismutase, as well as a hydrolase mimic, and so on.7–9 In particular, as a family of over 200 enzymes, hydrolase is responsible for an enormous amount of substrates, allowing them to play a crucial part in the catalytic breakdown of lipids, carbohydrates, and proteins.10,11 For instance, phosphatase is a particularly important enzyme for catalyzing the cleavage of phosphate groups from molecules and has a role in energy transfer and blood sugar regulation, and so on in biological systems.12–14 Especially, phosphodiester bonds are the backbone linkages of all the nucleic acids which store and transfer biological information. Without any catalyst, the scission of individual phosphodiester linkages in DNA is kinetically challenging under physiological conditions, with an estimated half-life of 30 million years.15 Despite the remarkable resistance to hydrolysis, it is self-evident that the cleavage of phosphodiester linkages is central to a series of biological processes, including replication, recombination, and damage repair of DNA, as well as the maturation and processing of RNA.16 Nevertheless, after a decade of development of nanozymes, the engineering of hydrolase-mimic nanozymes is still rare (a small fraction of 7.1%), especially compared with the impressive success of oxidoreductase mimics (with the majority of 92.9%).5,17,18 Metal–organic frameworks (MOFs), an emerging class of porous crystalline materials, possess highly active metal nodes and well-defined pore structures.19,20 In addition, the unambiguous topologies of MOF serve as an ideal platform to uncover the catalytic mechanism in depth. Therefore, MOFs are structurally favorable for engineering nanozymes, which integrate the biomimetic active center and spatial microenvironment.21 In fact, inspired by the catalytic center of O-bridged bimetal structure of phosphatase, scientists have considered the Zr-MOF with Zr-OH-Zr active site as a phosphohydrolase nanozyme for degrading organophosphorus pesticides/chemical warfare agents by cleaving the phosphate ester bond.22–24 Further enlarging pore apertures (for better diffusion) and lowering node connectivity (to increase accessibility to active sites), scientists have reported series Zr-MOFs like NU-1000 and MOF-808 for efficient phosphorolysis of chemical warfare agents,25–27 which lay the firm foundation to access the holy grail of biomimetic phosphorolysis in a sustainable manner.28–31 These hydrolytic reactions, however, usually perform in strong basic environments (pH ≥ 10),32 which are barely comparable to the natural hydrolases of which the hydrolysis proceeds under physiological conditions. As a result, this significantly limits the diversity of the Zr-MOFs nanozymes that can be explored in biological scenarios, and consequently, the efficient cleavage of the phosphodiester bond in physiological conditions remains challenging. Herein, we demonstrate that a low-cost MOF nanozyme containing cluster-missing reo defects (UiO-66 REO) has real phosphohydrolase-like function (Scheme 1) and showcase the first MOF nanozyme enabling the cleavage of phosphodiester bond under physiological conditions (pH = 7.4 and 37 °C). The phosphodiesterase-like activity of the UiO-66 REO nanozyme significantly outperforms the activity-recording Zr-MOFs including MOF-808, NU-1000, MOF-818. and UiO-66-NH2 as well as UiO-66 (Ce). The UiO-66 REO nanozyme successfully demonstrates the cleavage of stable phosphodiester linkages in DNA, of which the phosphodiester linkages are extremely resistant to hydrolysis under neutral conditions. Scheme 1 | Schematic illustration of defect-engineered UiO-66 nanozyme with reo topology domain for cleaving phosphodiester bond. Download figure Download PowerPoint Experimental Methods Synthesis of the UiO-66 ideal, UiO-66 REO, and UiO-66 MW The synthesis of the UiO-66 ideal followed a previous report.33 ZrCl4 (189 mg, 0.81 mmol), 35% HCl (0.143 mL, 1.62 mmol), and terephthalic acid (H2BDC, 270 mg, 1.62 mmol) were dissolved in 4.87 mL N,N-dimethyl formamide (DMF). The mixture was sonicated for 30 min until completely transparent. Then the mixture was transferred to a Teflon-lined autoclave and heated at 220 °C for 20 h. The resulting white powder was collected by centrifugation, washed extensively with DMF and methanol, and vacuum dried at 60 °C. The UiO-66 REO was synthesized by adjusting the reaction temperature, resulting in cluster-missing reo defects. ZrCl4 (189 mg, 0.81 mmol), 35% HCl (0.143 mL, 1.62 mmol), and H2BDC (270 mg, 1.62 mmol) were dissolved in 4.87 mL DMF. The mixture was sonicated for 30 min until completely transparent. Then the solution was transferred to a Teflon-lined autoclave and heated at 100 °C for 20 h. The resulting white powder was collected by centrifugation, washed extensively with DMF and methanol, and vacuum dried at 60 °C. The UiO-66 MW was synthesized based on microwave-assisted strategy.34 ZrCl4 (291 mg, 1.25 mmol), H2BDC (208 mg, 1.25 mmol), acetic acid (2.1 mL, 37.5 mmol), and water (0.135 mL, 7.5 mmol) were dissolved in DMF (10 mL) and transferred to a 30 mL glass vial. The mixture was capped with a septum and sonicated for 15 min before being introduced into a microwave reactor heating lid. The temperature was set at 120 °C with a 15 min hold time. The resulting white powder was collected by centrifugation, washed extensively with DMF and methanol, and vacuum dried at 60 °C. Integrated differential phase contrast-scanning transmission electron microscope characterization The integrated differential phase contrast-scanning transmission electron microscope (iDPC-STEM) image acquisition was carried out on a double Cs-corrected Themis Z scanning transmission electron microscope (ThermoFisher Scientific, Shanghai, China). A spherical aberration corrector (SCORR) was utilized for the electro probe, which was aligned using a standard sample before imaging. The imaging process was operated at 300 kV with a convergence semiangle set at 10.8 mrad. For image export, the image integration (of four images) was implemented by a high-pass filter dark-field (DF4) detector, which allowed for the reduction of low-frequency information. The electron-beam current was lower than 1 pA. The collection angles were set at 5∼19 mrad for the DF detector and 20∼120 mrad for the high-angle annular dark-field (HAADF) detector. The dwell time was 5 μs while the pixel size was 0.26 Å × 0.26 Å, and the electron dose was approximately 60 e−/Å2 in the iDPC-STEM imaging process. Diffuse reflectance infrared Fourier transform spectroscopy measurements After the specific reaction time, the MOF catalysts were immediately centrifuged at 12,000 rpm followed by drying and then removed into an in-situ heating sample cell at a specific temperature. The sample-loaded cell was purged with N2 gas for 1 h prior to the diffuse reflectance infrared Fourier transform spectroscopy (DRIFTS) measurement. 31P-nuclear magnetic resonance spectra of catalytic processes Hydrolysis profiles were recorded via 31P-NMR measurement. The MOF catalyst (50 μg) was loaded into a 2-[4-(2-hydroxyethyl)piperazin-1-yl]ethanesulfonic acid (HEPES) solution (pH = 7.4, 10 mM, and 0.45 mL deionized water/0.05 mL D2O) and stirred for homogeneous dispersion. The bis(4-nitrophenyl) phosphate (BNPP) (1.5 mM) was added to the above buffer solution and stirred at 37 °C for 24 h. Then the reaction mixture was transferred to an NMR tube for spectrum measurement. General catalytic procedures In a general catalytic system, 500 μL MOFs suspension was added into BNPP or other substrates (4-nitrophenyl butyrate (p-NPB), 4-nitrophenyl sulfate (p-NPS), and phosphomonoester bond-contained 4-nitrophenyl phosphate (p-NPP) or phosphotriester bond-contained tris(4-nitrophenyl) phosphate (TNPP)) solution (300 μM). The concentration of UiO-66 REO and other MOFs was maintained at 50 μg mL−1. The reaction was conducted in HEPES (10 mM and pH = 7.4) at 25 °C. UV–vis spectroscopy was conducted to monitor the absorbance for the formation of p-nitrophenol (p-NP) at 400 nm. The data was collected using the kinetic mode. For the catalytic kinetic parameters' determination, a 400 μL catalyst suspension (50 μg) was directly added to the buffer solution in a 1 mL cuvette, and various concentrations of BNPP (0.01–0.5 mM) were then introduced. After thorough mixing, the catalytic activity was recorded as quickly as possible by measuring the absorbance at 400 nm with a spectrophotometer. The data was collected using the kinetic mode. General procedures at different reaction pH The effects of reaction pH on the hydrolysis of BNPP was investigated over a pH range of 2–10. The reaction pH was tuned by adding the NaOH or HCl at different concentrations. At low pH values, it is difficult to detect the product p-NP with the UV–vis spectrophotometer, so high performance liquid chromatograph (HPLC, Supporting Information Table S4) was employed to monitor the catalytic kinetics on a C18 column with the UV detector at 400 nm.35 Dye permeation experiment A dye permeation experiment was performed to evaluate the accessibility of the UiO-66 ideal and UiO-66 REO interior. The two MOF samples (1 mg each) were soaked in 5 mL rhodamine b (Rh B) solution (10 μg mL−1) at room temperature. After being incubated for 4 h, the precipitate was collected by centrifugation and then resuspended in 5 mL ethanol for rapid sample preparation. Confocal laser scanning microscopy (CLSM) was conducted to observe the dye distribution within the nanoparticles. Results and Discussion Synthesis and characterization of UiO-66 nanozyme with reo defects UiO-66 is a Zr-MOF known for its high water stability, cost-effectiveness, and ease of scale-up production.36,37 We chose UiO-66 as an MOF platform to design phosphodiesterase nanozyme by means of defect engineering. The UiO-66 REO with reo defects was prepared through a cluster-missing strategy. As a comparison, a defective-free fcu UiO-66 framework (denoted as UiO-66 ideal) was synthesized by a standard solvothermal method (Figure 1a).33 The crystallinity and structure of these samples were measured by powder X-ray diffraction patterns (PXRD), as shown in Figure 1a,b. And the diffractions at ca. 4° and 6° (2θ) were identified in UiO-66 REO, which were allocated to the (100) and (110) lattice planes, respectively. The two broad peaks were assigned to the formation of UiO-66 with defect nanoregions containing some cluster-missing reo phases38,39 In the UiO-66 ideal sample, there were no obvious diffuse scattering features in the low-angle region. As shown in Figure 1c, the typical thermogravimetric analysis (TGA) curves of the two samples displayed three main steps.40 First, the desolvation process consisting of the removal of the solvent is usually completed under 150 °C. Second, the dihydroxylation process consists of the removal of two structural water molecules and compensating ligand from the hexanuclear zirconium (Zr6) nodes in the range of ca. 150–400 °C. Finally, the organic parts are decomposed, and the framework eventually yields ZrO2. Based on these pyrolysis processes, we deduced that the UiO-66 ideal has a BDC linker to the Zr6 node ratio of 6. While the UiO-66 REO has just four linkers per node, it is in good agreement with the eight-connected reo topology (the detailed calculation is described in Supporting Information). In addition, the insight into the changes of the N2 adsorption amount and the pore size distribution between UiO-66 REO and UiO-66 ideal also clarified the defective structure ( Supporting Information Figure S1 and Table S1). Further, the Fourier transform infrared (FT-IR) spectra of the UiO-66 samples showed strong bands at 1578 and 1390 cm−1, assigned to the asymmetric and symmetric stretch of carboxylate groups from the MOF ligand ( Supporting Information Figure S2), respectively, which consisted of the syn–syn bridging mode of carboxylate groups.41,42 Furthermore, the distinct shoulder band at 1550 cm−1 in the FT-IR spectra was recorded in the UiO-66 REO but not in the UiO-66 ideal. This represents an asymmetric stretch for monodentate coordinated carboxylates, indicating that the unsaturated Zr6-cluster of UiO-66 REO formed. Figure 1 | UiO-66 REO synthesis and characterization. (a) PXRD pattern of the UiO-66 ideal and UiO-66 REO. (b) Low-angle region PXRD patterns from 2° to 10°. (c) TGA curves of UiO-66 ideal and UiO-66 REO. (d) EXAFS spectra of the Zr K-edge for UiO-66 idea, UiO-66 REO samples, and reference oxides. The WT spectra of the Zr K-edge for UiO-66 ideal (e), UiO-66 REO (f). WT means wavelet transform. Download figure Download PowerPoint Furthermore, X-ray absorption fine structure (XAFS) spectroscopy was used to ascertain the coordination environment of Zr. The Zr K-edge X-ray absorption near-edge structure spectra after normalization are shown in Supporting Information Figure S3. The Fourier-transformed X-ray absorption fine structure (FT-EXAFS) spectra reveal the local coordination environment around the Zr atoms, which is dominated by two peaks at ∼1.5 and ∼3.1 Å, attributed to the Zr-O and Zr–Zr shell, respectively (Figure 1d and Supporting Information Figure S4).43,44 The slight decrease of the peak at 1.8 and 3.1 Å indicated the loss of Zr–O bond and Zr…Zr scatters, respectively. Compared to the UiO-66 ideal, the fitting R-space EXAFS data of UiO-66 REO showed that the coordination numbers of Zr–O and Zr…Zr scattered monotonically decreases ( Supporting Information Table S2), in line with the cluster-missing reo topology. In addition, the wavelet transform (WT) spectra of the Zr K-edge also support the cluster-missing defects phase in UiO-66 REO. As shown in Figure 1e,f, the WT contour plot of UiO-66 REO exhibits one maximum peak at around 3.1 Å (assigned to the Zr…Zr scatters) slightly lower than that of UiO-66 ideal in intensity. This difference revealed that the Zr species also had different local environments supporting the cluster-missing defects. Phosphodiesterase-like hydrolytic activity Among the slowest hydrolytic reactions, the cleavage of phosphodiester is top-ranked on the list.45 In water, a diphosphate anion undergoes the hydrolytic process (P–O bond cleavage) with a t1/2 value of 30 million years at room temperature.14,15,46 This phosphodiesterase-like activity has been evaluated by monitoring the hydrolysis of BNPP (a widely used substrate for phosphodiesterase).4731P-NMR spectra were employed to monitor the BNPP hydrolytic process (Figure 2h). Interestingly, when the catalysis was under physiological conditions (pH 7.4 and 37 °C), we found that the UiO-66 REO displayed a high phosphodiesterase-like catalytic activity, in which the characteristic peak of BNPP gets reduced and converts into p-NPP in 24 h (Figure 2h and Supporting Information Figure S5). Meanwhile the UiO-66 ideal displayed no phosphodiesterase-like catalytic activity under similar conditions. To clarify the advantages of the phosphodiesterase-like hydrolysis of UiO-66 REO under physiological conditions, we also carefully synthesized other well-known phosphatase-mimicking MOFs like low-connected node and large-pored Zr-MOFs (MOF-808, NU-1000, etc.), as well as rare-earth metal cluster Ce-MOFs (scanning electron microscopy (SEM) and PXRD patterns shown in Figure 2a–g). In similar physiological conditions, very limited hydrolytic products were found when using the same dosage of these reported MOFs as catalyst (50 μg mL−1). Additionally, the BNPP hydrolytic kinetics by different MOFs were also conducted using UV absorption spectroscopy. As expected, utilizing an identical amount of catalysts, the other MOFs yielded little or no phosphodiesterase-like activity under physiological settings, which were the opposite of UiO-66 REO ( Supporting Information Figure S6). This was in line with the reported result, that is, strong basic media is required to activate the hydrolytic activity of these MOFs.32 In addition, we noticed that the UiO-66(Ce) exhibited moderate phosphodiesterase-like activity. This can be interpreted from the extremely strong Lewis acidity of cerium(IV) cation, which confers the Ce nanomaterials or complex enzyme-like activity of hydrolytic reactions.1,48–50 Figure 2 | Phosphodiesterase-like hydrolytic bioactivity. Crystal structure, SEM images, and X-ray diffraction patterns of Zr-MOFs and Ce-MOFs, including UiO-66 REO (a), UiO-66 ideal (b), UiO-66-NH2 (c), NU-1000 (d), MOF-808 (e), MOF-818 (f), and UiO-66 (Ce) (g). (h) 31P-NMR spectra of the catalytic hydrolytic process for BNPP by different MOFs in physiological conditions (pH = 7.4, 37 °C, and 24 h). (i) Model substrates used in the hydrolytic experiment. (j) The catalysis profiles of p-NPP, BNPP, TNPP, p-NPB, and p-NPS by UiO-66 REO. (Substrate concentration: 300 μM; nanozyme dosage: 50 μg mL−1; HEPES 10 mM, pH = 7.4, and 25 °C.) Download figure Download PowerPoint The phosphodiesterase-like catalytic kinetics of UiO-66 REO were further investigated. In this experiment, UV–vis spectroscopy was used to monitor the production of p-NP from the broken P–O bond ( Supporting Information Figure S7). Supporting Information Figure S8a describes the initial rates of BNPP cleavage by UiO-66 REO as a function of the concentration of BNPP, and the Lineweaver-Burk plot depicts the characteristic linear fit ( Supporting Information Figure S8b). Fitting the Supporting Information Figure S8a data with the Michaelis–Menten equation,51 the kinetic parameters of Vmax (maximum reaction rate) and Km (Michaelis constant Km, reflecting the affinity of nanozyme toward a substrate) were calculated to be 0.44 × 10−3 mM min−1 and 0.7 mM, respectively. We next chose other model substrates including phosphomonoester and phosphotriester to explore the catalytic selectivity of UiO-66 REO (Figure 2i,j). All the reaction products for these substrates were p-NP and also can be detected by UV–vis spectra. We surveyed the catalytic efficiency of UiO-66 REO catalysts in terms of the generating rate of p-NP during the first 100 seconds. Under the same conditions, the catalytic conversion of phosphodiester-contained BNPP to p-NP was efficient with an excellent rate of 11.7 μM min−1, while very limited catalytic product of p-NP was recorded in the case of phosphomonoester of p-NPP or phosphotriester of TNPP (Figure 2j). And the reaction rate was 7.7 and 5.9 times lower than that of BNPP for p-NPP and TNPP, respectively ( Supporting Information Table S3). When the hydrolytic reaction was maintained for one hour, we also noted that UiO-66 REO had a more efficient catalytic conversion rate for BNPP than both p-NPP and TNPP ( Supporting Information Figure S9). This suggests that the hydrolytic selectivity for phosphodiester bonds using this UiO-66 REO catalyst. In vivo, C–O and S–O bonds are also two types of important chemical bond. UiO-66 REO indicated almost no catalytic activity for p-NPS with the type of S–O bond, and for the 4-nitrophenyl butyrate (p-NPB) with the type of C–O bond (Figure 2j and Supporting Information Table S3). These results evidence that this UiO-66 REO nanozyme displayed desirable hydrolytic selectivity for phosphodiester, as with the function of natural phosphodiesterase. Microstructure profile by high-resolution electron microscopy All the characterizations of PXRD, TGA, and XAFS provided overall structural information of the UiO-66 REO. However, the local structure, for example, the boundary of a crystal that influences the entrance of catalytic substrates, remained unknown. Seeing is believing. We visualized the microstructure of UiO-66 REO utilizing the iDPC-STEM technique, an advanced low-dose-electron imaging technique for beam-sensitive materials.52,53 We utilized iDPC-STEM to directly identify the high-resolution structures of the UiO-66 ideal (Figure 3a) and UiO-66 REO (Figure 3e), wherein the dark spots and bright spots in the lattice correspond to empty pore and Zr6 clusters, respectively. The ordered linkage of the UiO-66 ideal (fcu topology phase) was observed throughout the framework (Figure 3b,c and Supporting Information Figure S10), while the cluster-missing reo phase was identified in UiO-66 REO (Figure 3f,g and Supporting Information Figure S11), which was in line with the PXRD, TGA, and XAFS data. Importantly, the enlarged images of boundary regions found that the UiO-66 ideal had a highly crystalline boundary (Figure 3d and Supporting Information Figure S10). However, the boundary (Figure 3h and Supporting Information Figure S11) and interior of UiO-66 REO crystals ( Supporting Information Figure S12) were partially amorphous. The disappeared lattice and weak diffracted intensity in the fast Fourier-transform (FFT) pattern also confirm it (the inset in Figure 3h and Supporting Information Figures S11 and S12). The amorphous surface termination resulted in abundant inhomogeneous cavity with large size, which broadened the pore window and accelerated the entrance of catalytic substrates. Figure 3 | Microstructure imaging. Theoretical structural models along the [001] (a) and [110] (e) zone axes of fcu UiO-66. The iDPC-STEM image of an UiO-66 ideal (b) and an UiO-66 REO (f) particle oriented along the [001] and [110] zone axis, respectively. The structural analysis of the intact fcu phase in UiO-66 ideal (c) and the reo defect in UiO-66 REO sample (g). The enlarged images (d, h) of the domains highlighted in yellow in panels (b) and (f). Inset is the FFT patterns of the corresponding domains. Plots of fluorescence intensity along UiO-66 ideal particles (i) and UiO-66 REO particles (m) after dye permeation experiment, respectively: The selected particles were highlighted in green in panels (j) and (n). CLSM images of the UiO-66 ideal particles (j) and the UiO-66 REO particles (n) after dye permeation experiment and the corresponding bright field (k, o) and merging (l, p) images. The CLSM imaging was implemented under 516 nm emission. Download figure Download PowerPoint We further designed a dye (Rh B, ca. 18.5 × 8.7 × 13.4 Å dimension, Supporting Information Figure S13) permeation experiment to evaluate the accessibility to the interior of two MOF samples. The molecular dimension of Rh B is comparable with that of BNPP (11.1 × 6.5 × 9.1 Å dimension, Supporting Information Figure S13). In a defect-free UiO-66 ideal particle, the crystallographic window is roughly 8 Å ( Supporting Information Figure S14). Thus, the UiO-66 ideal excluded the entrance of Rh B because of the size exclusion, as evidenced by the limited fluorescence on the surface of UiO-66 ideal particles (Figure 3i–l). As a comparison, UiO-66 REO, with amorphous surface termination unambiguously identified by iDPC-STEM, allowed the free entrance of Rh B. This was confirmed by the CLSM results, in which the bright red fluorescence was recorded throughout the UiO-66 REO particles (Figure 3m–p). Based on these structural analyses, we can infer that the amorphous surface termination in UiO-66 REO facilitated the diffusion of catalytic substrates, while the unique and unsaturated Zr6-clusters, which resulted from the reo topology phase, offering favorable active sites for phosphodiesterase-like catalytic transformations (vide infra). The proposed mechanism for phosphodiesterase-like activity The redox Zr6-clusters in Zr-MOFs are highly reactive in phosphate ester hydrolysis.31,32 We noted that the individual Zr6 clusters of UiO-66 (without topological linkage) presented limited phosphodiesterase-like catalytic activity under physiological conditions ( Supporting Information Figure S15). We then speculated that the unique Zr6-clusters in the reo defect phase are the catalytic center contributing to the extraordinary phosphodiesterase-like hydrolysis. By using a heat treatment method, we gradually "collapsed" the cluster-missing defects of the reo topology phase.33 Specifically, we heated UiO-66 REO to various increased temperatures, as shown in Figure 4a. The heat treatment resulted in the collapse of the reo topology phase by degrees, along with maintaining the corresponding d