English

• Enzymes are indispensable catalysts across a broad spectrum of applications, ranging from pharmaceutical synthesis and clinical diagnostics to food fermentation and environmental remediation, owing to their remarkable catalytic efficiency under mild conditions [1-5]. However, the structural sensitivity of enzymes, stemming from their inherently dynamic three-dimensional architectures governed by weak interactions such as hydrogen bonds, van der Waals forces, and hydrophobic interactions, renders them susceptible to destabilization under in vitro conditions [2,6,7]. One promising approach to enhance the biocatalytic stability of enzymes, while also imparting them recyclability, is the encapsulation of enzymes within porous materials [8-11]. Among these, reticular porous frameworks such as metal-organic frameworks (MOFs) have garnered significant attention due to their modular synthesis, high surface area, customizable pore structures, and adjustable chemical compositions [9,12,13].

  Since the pioneering work by Liang et al. [14], which demonstrated the biomimetic encapsulation of enzymes within a MOF-ZIF-8, numerous studies have extended this concept, leading to the development of enzyme@MOF biocatalysts [9,12], including enzyme@ZIF-90, enzyme@MAF-7, enzyme@MAF-6, enzyme@MIL-100, and enzyme@UiO-66-F₄, using in situ biomimetic encapsulation strategies [15-19]. These strategies leveraged interfacial interactions between metal precursors and enzyme surface amino residues (e.g., acidic residues) through electrostatic attraction, thereby facilitating the in situ growth of the MOF framework around the enzyme template in a manner reminiscent of natural biomineralization. Despite the effectiveness of this biomimetic encapsulation, the strong coordination bonds within the MOF framework often require high temperatures to achieve a thermodynamically stable crystalline phase, particularly for MOFs with larger pore sizes [20,21]. Given the thermal instability of enzymes, only a limited number of small-pore MOFs have been demonstrated to support effective enzyme encapsulation in situ.

  In contrast to MOFs, hydrogen-bonded frameworks (HOFs) represent a novel class of porous materials assembled from discrete organic linkers via intermolecular hydrogen bonding, alongside other weak interactions such as π-π stacking and electrostatic interactions [22-24]. HOFs inherit the structural advantages of MOFs, yet their weak hydrogen-bonding interactions confer significantly greater flexibility in molecular connectivity. This flexibility enables HOFs to crystallize under mild, environmentally benign conditions [22,23], thus avoiding high temperatures that can lead to enzyme deactivation during encapsulation. Furthermore, the metal-free nature of HOFs enhances their biocompatibility [25], making them an attractive alternative for biomimetic enzyme encapsulation. Over the past five years, our group has focused on developing biomimetic encapsulation strategy using HOFs to synthesize biocatalysts [26-31]. We have found that 4-connected carboxylic acid-based linkers with pyrene cores, such as 1,3,6,8-tetrakis(p-benzoic acid)pyrene (H₄TBAPy), can facilitate the interfacial interaction with enzymes through hydrogen bonding and electrostatic interactions. The subsequent self-assembly of H₄TBAPy via hydrogen bonding and π-π stacking results in hydrogen-bonded biohybrid frameworks (HBFs), which exhibit significantly higher enzyme loading efficiency compared to the MOF-based approaches. Moreover, these HBFs possess long-range ordered channels that are larger than those found in biomineralized MOFs, enhancing substrate accessibility to the encapsulated enzymes and thus improving their bioactivity [32-34]. Despite these advancements, a major challenge remains in the encapsulation of enzymes with low isoelectric point (pI), whose surfaces are predominantly dominated by acidic residues. These residues tend to strongly dispel the H₄TBAPy ligands through electrostatic repulsion, thus limiting enzyme loading and reducing the overall efficacy of the resultant HBF catalysts.

  In this study, we present a site-specific surface modification strategy to efficiently encapsulate low-pI enzymes into HBF while preserving their biocatalytic activity. We employed density functional theory (DFT) calculations to identify the preferential binding sites of enzyme surface residues with H₄TBAPy and demonstrated that the NH₃⁺-moiety in the enzyme's amino acid residues was thermodynamically favorable for binding with the partially deprotonated H₄TBAPy ligand in aqueous solutions, while avoiding further deprotonation of the carboxyl groups of H₄TBAPy otherwise will lead to the failure in the hydrogen-bonded assembly of H₄TBAPy (Scheme 1). Guided by these insights, we selectively aminated the acidic residues of low-pI enzymes with ethylenediamine (EDA) via a N-(3-dimethylaminopropyl)-N′-ethylcarbodiimide (EDC)-mediated activation method. The synergetic electrostatic and hydrogen bonding interactions between the -NH₃⁺ groups of the modified enzymes and the partially deprotonated -COOH (-COO⁻) groups of the organic linker facilitate the growth of HOF crystals around the enzyme surface, enabling efficient in situ encapsulation with up to 100% enzyme loading efficiency, surpassing the previously reported methods. Compared to free enzymes, the resultant HBF biocomposites exhibited significantly enhanced stability under challenging conditions, including exposure to high temperature, organic solvents, denaturants, hydrolytic enzymes and different pH conditions. By harnessing both the high biocatalytic performance of the HBF biocomposite and the fluorescence property of quantum dots (QDs), we successfully developed a sensitive fluorescent sensor for glucose detection, with a quantification limit of 5 µmol/L. This work provides a simple yet effective enzyme surface engineering strategy for the one-pot synthesis of HOF biocatalysts with superior catalytic activity and stability, offering great potential for diverse applications that integrate the functionalities of enzymes and nanomaterials.

  Scheme 1

  

  Scheme 1.  Schematic of the unsuccessful encapsulation of low-pI enzyme and the facilitated capsulation of the aminated enzyme. (a) Schematic of unsuccessful encapsulation of low-pI enzyme into the HOF framework. (b) Schematic of the facilitated encapsulation of the enzyme upon amination via EDA molecules.

  DownLoad: Full-Size Img PowerPoint

  The π-conjugated carboxylate linker, H₄TBAPy, has previously been shown to form HBF-1 composites with proteins under mild conditions [27,32]. The robust face-to-face π-π stacking and hydrogen bonding interactions confer exceptional stability to the HBF-1 framework, making it highly promising for various applications. Motivated by these findings, we selected H₄TBAPy as the organic molecular module to construct hybrid frameworks with high enzyme loading. However, the theoretical pK_a value of H₄TBAPy (6.90) [32] indicates that it undergoes partial deprotonation in neutral aqueous solution (Fig. 1a(i)). This partial deprotonation creates an electrostatic repulsion between the linker and enzymes with low pI characteristics, leading to the presence of a mass of unbound linkers in the solution and reducing the encapsulation efficiency. To investigate this effect, we selected glucose oxidase (GOx, from Aspergillus niger), an enzyme with a pI of approximately 4 [35], as a model enzyme.

  Enzyme surfaces are characterized by a diverse array of functional groups, making it crucial to identify the specific groups that interact most favorably with the organic linker. To this end, we conducted DFT calculations to explore the affinity between various functional groups (methyl (-CH₃), phenyl (-pH), hydroxyl (-OH), thiol (-SH), carboxyl (-COOH), and amino (-NH₂)) and the H₄TBAPy linker, both in its neutral and partially deprotonated states. Given that the carboxyl and amino groups of enzymes typically exist as deprotonated (-COO−) and protonated (-NH₃⁺) groups respectively in aqueous solutions, we focused on these species in the calculations (Fig. S1 in Supporting information and Fig. 1b). From the calculations (Fig. S1), we found that the carboxylate (-COO⁻) group of the enzyme exhibited the highest binding energy (−61.62 kJ/mol) with the non-deprotonated H₄TBAPy linker. However, this interaction triggered a proton transfer from the -COOH group of the linker to the -COO⁻ group of the enzyme, leading to the deprotonation of the linker and a corresponding loss of hydrogen bonding sites required for HBF framework assembly through carboxyl dimers (Fig. 1a(ⅱ)) [24].

  Figure 1

  

  Figure 1.  Principles of the failed encapsulation of the pristine low-pI enzyme and the facilitated encapsulation of modified enzyme. (a) Schematic of the partially deprotonated H₄TBAPy linker and low-pI enzyme with negative charge in water solution, existing a charge repulsion (ⅰ) and a proton transfer that facilitated the deprotonation of the linker (ⅱ), both of which hinder the efficient enzyme encapsulation. (b) The binding mode and binding energy between the partially deprotonated linker and the amino acid residue with -NH₃⁺ group.

  DownLoad: Full-Size Img PowerPoint

  Further analysis of the binding between the partially deprotonated carboxyl group of the linker and the -NH₃⁺ group of the enzyme revealed a high binding energy of −51.48 kJ/mol, with a proton transfer from the -NH₃⁺ to the -COO⁻ group (Fig. 1b). This interaction suggests that the recuperation of neutral linker is more favorable for the assembly of the HBF framework [32]. We hypothesize that the combination of the -COO⁻ group of the linker and the -NH₃⁺ group on the enzyme surface represents the optimal binding site. The enhanced hydrogen bonding between -COOH and -NH₂ groups, along with the reduced deprotonation of the linker, facilitates the enzyme-driven assembly of the HBF structure (Fig. 1b). These findings support the idea that the introduction of -NH₃⁺ groups onto the enzyme surface can significantly promote more efficient assembly of HBFs, thereby enhancing the overall enzyme loading and stability within the hybrid framework.

  Upon analysis of the DFT results, we employed an amination approach to enhance the density of -NH₃⁺ groups on the enzyme surface. The aminated glucose oxidase (GOx), referred to as GOx-NH₂, was synthesized through site-specific coupling of -COOH groups in GOx with the -NH₂ groups in EDA molecules based on an EDC activation method [36], as depicted in Fig. 2a, with the detailed chemical reaction mechanism illustrated in Fig. S2 (Supporting information). This modification was expected to increase the electropositivity of GOx by concomitantly depleting the -COOH groups and introducing -NH₂ groups, which would protonate to form -NH₃⁺ in aqueous solution. As anticipated, the zeta potential of purified GOx-NH₂ increased to +13.13 ± 2.58 mV, compared to −19.8 ± 1.34 mV of the unmodified GOx (Fig. 2b), providing the first evidence of the successful enzyme amination.

  Figure 2

  

  Figure 2.  Enzyme surface amination and structure and activity characterization. (a) Schematic of GOx amination by site-specifically reacting to the carboxyl groups in the enzyme via EDC activation, increasing the -NH₃⁺ groups of the enzyme in water solution. (b) Zeta potentials of GOx, GOx-NH₂ and GOx/EDA. (c) IEF electrophoresis photographs of GOx and GOx-NH₂, respectively. (d) Catalytic curves of GOx and GOx-NH₂ based on a typical GOx-HRP cascade reaction. Conditions: enzyme, 10 µg/mL; ABTS, 0.2 mmol/L; glucose, 0.4 mmol/L; HRP, 1 µg/mL. (e) FT-IR spectra of GOx and GOx-NH₂ that the scanning ranges from 1800 cm^-1 to 1200 cm^-1. (f) CD spectra of GOx and GOx-NH₂, in which the bands around 210 nm and 218 nm were assigned to the α-helix and β-sheet structure of the enzyme, respectively.

  DownLoad: Full-Size Img PowerPoint

  Subsequently, we investigated the change in the pI value of the aminated enzyme using an isoelectric focusing (IEF) electrophoresis approach. As shown in Fig. 2c, the pI values of GOx and GOx-NH₂ were found to be in the ranges of 3.8–4.2 and 4.0–4.5, respectively, as well indicating an increase in electropositivity of the enzyme after amination. Such an elevation was possibly due to the protonation of the introduced amino groups on the enzyme surface in aqueous solutions. Further, the content of free amino groups in GOx and GOx-NH₂ was quantified using an o-phthalaldehyde (OPA) method (Fig. S3 in Supporting information) [37], with the results summarized in Table S1 (Supporting information). The free amino content in GOx was determined to be 189.1 ± 6.9 nmol/mg protein, while GOx-NH₂ exhibited a significant increase in free amino content, measuring 406.7 ± 5.7 nmol/mg protein. This corresponded to an increase of approximately 17 amino groups per GOx molecule, consistent with the results from matrix-assisted laser desorption/ionization mass spectrometry (MALDI-MS) analysis (Fig. S4 in Supporting information). To validate the specificity of the amination, we performed a control experiment where GOx was treated with EDA in the absence of EDC activation (termed GOx/EDA). As shown in Fig. 2b, the zeta potential of GOx/EDA remained negative (−9.94 ± 0.32 mV), and the free amino content was measured at 192.0 ± 5.8 nmol/mg protein, which is similar to the unmodified GOx. These findings suggest that EDA molecules preferentially undergo chemical grafting to the enzyme surface when EDC activation was employed, rather than simply adsorbing to the enzyme through non-covalent interactions.

  After ascertaining the successful amination, we next assessed the enzymatic activity and structural integrity of GOx-NH₂. The enzyme activity was evaluated via a typical GOx-horseradish peroxidase (HRP) cascade reaction, using glucose as the substrate and 2,2′-azino-bis(3-ethylbenzothiazoline-6-sulfonic acid) (ABTS) as the chromogenic reagent. As shown in Fig. 2d, GOx-NH₂ exhibited 100% catalytic activity under the same enzyme dosage as the original GOx, indicating that amination did not compromise the enzyme's catalytic function. The enzyme structure change was analyzed by spectroscopic methods including ultraviolet-visible (UV–vis), Fourier transform infrared (FT-IR), circular dichroism (CD) and fluorescence spectroscopies. As depicted in Fig. S5 (Supporting information) and Fig. 2e, FT-IR spectroscopy of GOx-NH₂ showed a slight blue shift of the amide I band from 1634 cm−¹ (for native GOx) to 1644 cm−¹, suggesting a small disturbance in the enzyme's secondary conformation. CD spectroscopy further confirmed this change, showing a slight decrease in the α-helix content of GOx-NH₂ (Fig. 2f and Table S2 in Supporting information). This change in secondary structure might be attributed to alterations in the enzyme's hydrophobic, hydrophilic, and hydrogen bonding interactions subsequent to amination [38]. However, the tertiary structure analysis of GOx-NH₂ using fluorescence and UV–vis spectroscopy indicated negligible changes in the overall 3D architecture of the enzyme (Figs. S6 and S7 in Supporting information). Importantly, the enzyme's catalytic activity was maintained, suggesting that the chemical modification primarily affected the secondary structure without disrupting the essential 3D architecture or the active site of the enzyme. In summary, while the amination of GOx resulted in minor changes to its secondary structure, the enzyme's macroscopic 3D conformation, including its coenzyme, was largely preserved, allowing for the retention of full catalytic activity.

  We next investigated the feasibility of GOx-NH₂ to mediate the in situ growth of HOF and synthesize GOx-NH₂-embedded HBF-1 biocomposites (denoted as GOx-NH₂@HBF-1), with unmodified GOx used as a comparison (denoted as GOx@HBF-1). Upon the addition of GOx-NH₂ solution (2.5 mg dissolved in 4.5 mL pure water with pH ~7) to the H₄TBAPy solution (5 mg dissolved in 0.5 mL DMF), a significant increase in precipitate production was observed (about 5 times higher than that of unmodified GOx), indicating a GOx-NH₂-mediated HBF generation (Fig. 3a). Afterwards, the powder X-ray diffraction (PXRD) patterns (Fig. 3b) and scanning electron microscopy (SEM) images (Fig. S8 in Supporting information) confirmed that both the crystallinity and morphology of GOx@HBF-1 and GOx-NH₂@HBF-1 were consistent with the pure HOF (PFC-1) crystal, suggesting successful formation of the HBF-1 framework.

  Figure 3

  

  Figure 3.  Biomimetic encapsulation of enzyme within porous HOF and structure, activity and stability characterization. (a) Photographs of the generation of GOx@HBF-1 and GOx-NH₂@HBF-1 biocomposites. (b) PXRD patterns of GOx@HBF-1 and GOx-NH₂@HBF-1 as compared to the simulated PFC-1 crystal. (c) CLSM images of the GOx-NH₂@HBF-1, where the GOx-NH₂ was pre-labeled by AF350. (d) N₂ adsorption/desorption isotherms of the PFC-1 and GOx-NH₂@HBF-1 composite. Inset was the curve of pore volume against pore size distribution. (e) Catalytic curves of free GOx, GOx@HBF-1 and GOx-NH₂@HBF-1. Conditions: The enzyme concentration in the GOx or GOx-NH₂@HBF-1 was 50 µg/mL. The GOx@HBF-1 and GOx-NH₂@HBF-1 amounts were kept at 89 µg/mL. Other conditions: 3.0 mmol/L glucose, 0.2 mmol/L ABTS, 8 µg/mL HRP. (f) The kinetic curve of GOx-NH₂@HBF-1 fitted by Michaelis-Menten equation under a series of glucose concentrations. Conditions: 50 µg/mL GOx-NH₂, 0.2 mmol/L ABTS, 8 µg/mL HRP. (g) The retained activity of GOx, GOx-NH₂ and GOx-NH₂@HBF-1 after different treatments. The treatment time was kept at 30 min.

  DownLoad: Full-Size Img PowerPoint

  Quantitative analysis of protein content via Bradford assays (Fig. S9 in Supporting information) revealed that 100% of GOx-NH₂ was incorporated into the HOF, which was 13.8 times higher than the encapsulation efficiency of unmodified GOx (Table S3 in Supporting information). The calculated enzyme loading for GOx-NH₂ within HBF-1 was 56.1% (w/w), significantly surpassing the 20.5% (w/w) achieved from GOx. These results highlighted the enhanced encapsulation efficiency of GOx-NH₂ facilitated by a surface chemistry modulation strategy, allowing for effective incorporation of the enzyme into the HOF framework under mild aqueous conditions. GOx-NH₂ incorporation in HBF-1 was further verified by FT-IR spectroscopy and thermogravimetric analysis (Figs. S10 and S11 in Supporting information). To further confirm the encapsulation, GOx-NH₂ was pre-labeled with a blue AF350 dye to form a fluorescent GOx-NH₂@HBF-1 composite, which was subsequently examined using confocal laser scanning microscopy (CLSM). The CLSM images showed that the fluorescence dye-labeled GOx-NH₂ was uniformly distributed throughout the HBF-1 framework (Fig. 3c). Additionally, the nitrogen (N₂) adsorption/desorption isotherms demonstrated a marked decrease in N₂ uptake and pore volume upon enzyme incorporation, confirming that GOx-NH₂ was successfully embedded within the HOF (Fig. 3d).

  The bioactivity of the GOx-NH₂@HBF-1 composite was then evaluated using the GOx-HRP cascade reaction. Due to the 2D long-range ordered mesoporous channels (ca. 2.0 nm) of the HBF-1 framework [32,33], as well as the substrate enrichment effect in a confined environment [26,39], both the catalytic rate and overall activity of GOx-NH₂@HBF-1 were found to be comparable to free GOx at various glucose concentrations (Fig. 3e and Fig. S12 in Supporting information). In contrast, GOx@HBF-1 exhibited only 0.32% to 3.87% of the catalytic activity compared to GOx-NH₂@HBF-1, which might be due to the lower enzyme content and less compact enzyme distribution within the framework (Fig. 3e and Fig. S12). Enzyme kinetics were further assessed under a series of glucose concentrations, and correspondingly the Michaelis-Menten constants (K_m) and maximum reaction velocity (V_max) were determined. The K_m of GOx-NH₂@HBF-1 was calculated to be 1.94 mmol/L (Fig. 3f and Fig. S13 in Supporting information), which was higher than that of free GOx (K_m = 0.56 mmol/L, Fig. S14 in Supporting information), suggesting a reduced substrate affinity for the encapsulated enzyme. Such a decrease might be owing to the mass transfer hindrance imposed by the HOF "shield" outside the enzyme. However, the V_max of GOx-NH₂@HBF-1 was 4.99 µmol L^-1 s^-1 (Fig. 3f), which was comparable to that of free GOx (V_max = 4.72 µmol L^-1 s^-1, Fig. S14). These findings demonstrated that although the affinity between the enzyme and substrate was weakened after encapsulation, the confined environment within the HBF could enhance the catalytic rate, leading to a comparable activity to free enzyme.

  The inherent instability of free enzymes remains a significant challenge for their practical applications. Thus, we investigated the stability of GOx-NH₂@HBF-1 under harsh conditions, including heating (70 ℃), exposure to organic solvents (methanol and acetone), denaturant (6 mol/L urea), hydrolytic enzyme (4 mg/mL trypsin) and different pH (2–11) conditions. As expected, the encapsulation within the HOF framework could provide substantial protection for GOx-NH₂, which maintained over 70% of its original activity under these challenging conditions (Fig. 3g and Fig. S15 in Supporting information). PXRD and SEM analyses further confirmed the structural integrity of GOx-NH₂@HBF-1 after exposure to these treatments (Figs. S16 and S17 in Supporting information). In contrast, free GOx lost > 90% of its activity after the same treatments, further demonstrating the protective effect of the HOF framework. Additionally, after amination, GOx-NH₂ exhibited improved resistance to heating and trypsin hydrolysis compared to native GOx. It suggested that chemical modification may enhance the enzyme's stability, possibly due to the more robust conformation induced by amination [40,41]. Finally, the recyclability of GOx-NH₂@HBF-1 was assessed by performing repeated catalytic cycles. As shown in Fig. S18 (Supporting information), the biocomposite retained over 80% of its original activity after seven cycles, highlighting its potential for long-term application in catalytic processes.

  We then extended our approach to other low-pI enzymes, which are typically difficult to trigger or accelerate the de novo assembly of HBF-1. Specifically, we explored the amination of xanthine oxidase (XOD from bovine milk, pI ~5.3) [42] and β-galactosidase (β-GAL from Aspergillus oryzae, pI ~4.7) [43] using EDA as well. The successful amination of XOD and β-GAL to obtain XOD-NH₂ and β-GAL-NH_2, respectively, was confirmed by zeta potential, OPA and MALDI-MS tests (Figs. S19-S21 and Table S1 in Supporting information). Additionally, the compatibility of this chemical modification on the enzymes in terms of structural integrity and catalytic activity was also corroborated (detailed results and discussion were seen in Figs. S22-S25 in Supporting information).

  We next investigated the self-assembly of H₄TBAPy linkers induced by XOD-NH₂ and β-GAL-NH₂. Notably, both XOD-NH₂ and β-GAL-NH₂ could facilitate the growth of the HBF-1 biocomposites (denoted as XOD-NH₂@HBF-1 and β-GAL-NH₂@HBF-1, respectively), with enzyme loading as high as 56.0% (w/w) (Fig. S26 and Table S3 in Supporting information), and the standard curves for protein content quantification using Bradford assays were shown in Figs. S27 and S28 (Supporting information). In contrast, the unmodified XOD failed to induce HBF-1 formation, as no product was collected from the reaction. However, XOD-NH₂ successfully generated the HBF-1 biocomposite with an encapsulation efficiency of 90.0% and a loading content of 56.0% (w/w), highlighting the importance of surface chemistry modification for enhancing the enzyme incorporation into the HBF-1 framework (Table S3). Further characterizations using PXRD, SEM, TGA and FT-IR confirmed the successful crystallization of HBF-1 framework, as well as the embedding of XOD-NH₂ and β-GAL-NH₂ within the composites (details seen in Figs. S29-S32 in Supporting Information). Furthermore, the catalytic activities of XOD-NH₂@HBF-1 and β-GAL-NH₂@HBF-1 could be well retained, demonstrating that the encapsulation process did not impair their enzyme functionality (Figs. S33 and S34 in Supporting information).

  Fluorescence sensors have emerged as powerful tools for rapid and sensitive detection of biomarkers due to their fast response time, high sensitivity, resistance to light scattering, and ease of use [44-47]. In this study, we developed a fluorescence sensor based on a combination of hydrogen peroxide (H₂O₂)-sensitive CdS QDs and GOx-NH₂@HBF-1 for glucose analysis. The QDs were synthesized based on previously reported protocols, which have demonstrated their H₂O₂-induced fluorescence quenching feature [48,49]. Dynamic light scattering (DLS) and transmission electron microscopy (TEM) characterization of the QDs revealed an average particle size of approximately 7.5 nm (Figs. S35a and b in Supporting information), and high-resolution TEM images confirmed the characteristic {110} reflection of CdS (Fig. S35c in Supporting information), consistent with previous studies [49,50]. The fluorescence property of the QDs was examined under 365 nm excitation, which exhibited an orange-red emission (Fig. 4a) with a maximum emission peak at 596 nm (Fig. 4b). While the GOx-NH₂@HBF-1 framework emitted green fluorescence (Fig. 4a), with a maximum emission peak at 529 nm (Fig. 4b). Interestingly, when QDs were mixed with GOx-NH₂@HBF-1 (denoted as QDs+GOx-NH₂@HBF-1), a significant blue shift in the fluorescence was observed, with the maximum emission at 441 nm (Figs. 4a and b). This result suggested that the QDs were not merely mixed with the HOF but were incorporated into the HBF-1 framework through an interface interaction.

  Figure 4

  

  Figure 4.  Fabrication of QDs+GOx-NH₂@HBF-1 glucose senser and sensing performance. (a) Photographs of QDs, GOx-NH₂@HBF-1 and the QDs+GOx-NH₂@HBF-1 mixture solution under a UV lamp (365 nm), respectively. (b) The fluorescence spectra of QDs, GOx-NH₂@HBF-1 and the QDs+GOx-NH₂@HBF-1 mixture solution, respectively. Excitation wavelength was set at 365 nm. (c) Schematic of the QDs+GOx-NH₂@HBF-1 fluorescence sensor for glucose detection. (d) The linear range for glucose detection based on the QDs+GOx-NH₂@HBF-1 sensor.

  DownLoad: Full-Size Img PowerPoint

  Structural integrity and morphology of GOx-NH₂@HBF-1 upon QD incorporation were first confirmed, in which the crystallinity and rod-like morphology have maintained (Fig. S36 in Supporting information). Zeta potential measurements (Fig. S37 in Supporting information) and TEM analysis (Fig. S38 in Supporting information) further supported the incorporation of QDs onto the surface of GOx-NH₂@HBF-1. In addition, the UV–vis spectra of QDs+GOx-NH₂@HBF-1 composite revealed a blue shift compared to the pure GOx-NH₂@HBF-1, indicating a metal-to-ligand charge transfer between the QDs and the HOF (Fig. S39 in Supporting information) [51]. This interfacial interaction was responsible for the observed transformation in the fluorescence property of the composite.

  Afterwards, a fluorescence sensor for glucose detection was fabricated based on the QDs+GOx-NH₂@HBF-1. Briefly, the detection principle relied on the fluorescence quenching phenomenon of the QDs induced by H₂O₂, which is generated through the catalytic activity of GOx-NH₂@HBF-1 (Fig. 4c). We first designed a rhodamine B (RhB) adsorption experiment to investigate whether the QDs would hinder mass transfer and consequently result in inefficient sensing. The adsorption kinetics curves revealed that, while the incorporation of QDs resulted in a slight decrease in the adsorption rate, the overall adsorption capacity remained unaffected (Fig. S40 in Supporting information). These results demonstrated that the pore channels of PFC-1 remained accessible to substrate molecules even after QDs modification, suggesting that the encapsulated enzymes could retain their catalytic functionality. The sensor exhibited a wide glucose linearity ranging from 5 µmol/L to 2000 µmol/L (linear equation: y = (4.6626 × 10^–4)x + 0.04474, R² = 0.999, Fig. 4d and Fig. S41 in Supporting information), with a limit of quantification of 5 µmol/L which was lower than other recently reported colorimetric and fluorescent sensors (Table S4 in Supporting information). After glucose sensing, the morphology and crystallinity of GOx-NH₂@HBF-1 remained intact, confirming the stability of the sensor (Fig. S42 in Supporting information). Additionally, the selectivity of this sensor for glucose detection was also validated (Fig. S43 in Supporting information). These results demonstrated the utility of the GOx-NH₂@HBF-1-based fluorescence sensor as a multifunctional biohybrid catalyst in combination with a fluorescence probe for rapid, sensitive, and reliable glucose detection, with potential applications in clinical glucose monitoring.

  In summary, this study reports a site-specific surface animation strategy to modulate surface chemistry of low-pI enzymes, enabling the facile formation of interfacial electrostatic and hydrogen bonding interactions with carboxylic organic linker precursors of HOF. As a consequence, the animated enzymes are able to mediate the in situ growth of HOF using itself as the biotemplates, which were difficult to realize using pristine enzymes. This biomimetic HOF encapsulation strategy allows up to 100% enzyme encapsulation, while well preserving the biocatalytic function. In addition, the HOF frameworks show a remarkable protective effect for the embedded enzymes under challenging conditions, making them good candidates for diverse applications. Furthermore, the biocatalytic activity of the HBF biocomposites in combination with the fluorescence quenching feature of a H₂O₂-sensitive QD was harnessed for the development of a sensitive glucose sensor, with a wide linearity ranging from 5 µmol/L to 2000 µmol/L alongside with a low quantification limit of 5 µmol/L. This work provides new insights to understand the enzyme surface chemistry for biomimetic HOF encapsulation and offers an effective strategy for designing highly stable and efficient HOF biocomposites with broad implications for advancing biocatalysis.

  Declaration of competing interest

  The authors declare that they have no known competing financial interests or personal relationships that could have appeared to influence the work reported in this paper.

  CRediT authorship contribution statement

  Haoquan Huang: Methodology, Investigation, Formal analysis, Data curation. Haiting Chen: Methodology, Investigation, Formal analysis, Data curation. Xinran Dong: Formal analysis, Data curation. Yanbin Xu: Formal analysis, Data curation. Anlian Huang: Formal analysis, Data curation. Qiaoyi Cen: Formal analysis, Data curation. Huairou Zhu: Formal analysis, Data curation. Guosheng Chen: Writing – review & editing, Validation, Investigation, Funding acquisition, Conceptualization. Wei Yi: Supervision, Resources, Project administration. Siming Huang: Writing – review & editing, Writing – original draft, Validation, Supervision, Software, Resources, Project administration, Methodology, Funding acquisition, Conceptualization. Gangfeng Ouyang: Writing – review & editing, Project administration, Investigation.

  Acknowledgments

  We acknowledge financial support from projects of the National Natural Science Foundation of China (Nos. 22104159, 22174164), and Guangdong Basic and Applied Basic Research Foundation (Nos. 2023A1515011632, 2024B1515020070).

  Supplementary materials

  Supplementary material associated with this article can be found, in the online version, at doi:10.1016/j.cclet.2025.111223.

  
  
• 1. [1]

     H.E. Schoemaker, D. Mink, M.G. Wubbolts, Science 299 (2003) 1694–1697. doi: 10.1126/science.1079237

  2. [2]

     U.T. Bornscheuer, G.W. Huisman, R.J. Kazlauskas, et al., Nature 485 (2012) 185–194. doi: 10.1038/nature11117

  3. [3]

     M. Pfefferkorn, S. Böhm, T. Schott, et al., Gut 67 (2018) 2045–2053. doi: 10.1136/gutjnl-2017-313811

  4. [4]

     R. Jiang, G. Luo, G. Chen, et al., Sci. Adv. 10 (2024) eadp1796. doi: 10.1126/sciadv.adp1796

  5. [5]

     S. Singh, P.K. Sharma, S. Chaturvedi, et al., Food Chem. 435 (2024) 137217. doi: 10.1016/j.foodchem.2023.137217

  6. [6]

     M.M. Kristjánsson, J.E. Kinsella, Adv. Food Nutri. Res. 35 (1991) 237–316.

  7. [7]

     C. Selvaraj, O. Rudhra, A.S. Alothaim, M. Alkhanani, S.K. Singh, Adv. Protein Chem. Str. 130 (2022) 59–83.

  8. [8]

     M. Bilal, S.A. Qamar, D. Carballares, Á. Berenguer-Murcia, R. Fernandez-Lafuente, Biotechnol. Adv. 70 (2024) 108304. doi: 10.1016/j.biotechadv.2023.108304

  9. [9]

     W. Liang, P. Wied, F. Carraro, et al., Chem. Rev. 121 (2021) 1077–1129. doi: 10.1021/acs.chemrev.0c01029

  10. [10]

      Q. Zhu, Y. Zheng, Z. Zhang, Y. Chen, Nat. Protoc. 18 (2023) 3080–3125. doi: 10.1038/s41596-023-00868-x

  11. [11]

      R.B. Lin, B. Chen, Chem 8 (2022) 2114–2135. doi: 10.1016/j.chempr.2022.06.015

  12. [12]

      S. Huang, G. Chen, G. Ouyang, Chem. Soc. Rev. 51 (2022) 6824–6863. doi: 10.1039/D1CS01011E

  13. [13]

      Z. Chen, K.O. Kirlikovali, P. Li, O.K. Farha, Acc. Chem. Res. 55 (2022) 579–591. doi: 10.1021/acs.accounts.1c00707

  14. [14]

      K. Liang, R. Ricco, C.M. Doherty, et al., Nat. Commun. 6 (2015) 7240. doi: 10.1038/ncomms8240

  15. [15]

      W. Liang, H. Xu, F. Carraro, et al., J. Am. Chem. Soc. 141 (2019) 2348–2355. doi: 10.1021/jacs.8b10302

  16. [16]

      W. Liang, K. Flint, Y. Yao, et al., J. Am. Chem. Soc. 145 (2023) 20365–20374. doi: 10.1021/jacs.3c05488

  17. [17]

      J.C. Díaz, B. Lozano-Torres, M. Giménez-Marqués, Chem. Mater. 34 (2022) 7817–7827. doi: 10.1021/acs.chemmater.2c01338

  18. [18]

      I. Akpinar, X. Wang, K. Fahy, et al., J. Am. Chem. Soc. 146 (2024) 5108–5117. doi: 10.1021/jacs.3c07785

  19. [19]

      T.H. Wei, S.H. Wu, Y.D. Huang, et al., Nat. Commun. 10 (2019) 5002.

  20. [20]

      H.Y. Guan, R.J. LeBlanc, S.Y. Xie, Y. Yue, Coordin. Chem. Rev. 369 (2018) 76–90.

  21. [21]

      Q. Liu, Y. Song, Y. Ma, et al., J. Am. Chem. Soc. 141 (2019) 488–496. doi: 10.1021/jacs.8b11230

  22. [22]

      Z. Zhang, Y. Ye, S. Xiang, B. Chen, Acc. Chem. Res. 55 (2022) 3752–3766. doi: 10.1021/acs.accounts.2c00686

  23. [23]

      Z.J. Lin, S.A.R. Mahammed, T.F. Liu, R. Cao, ACS Cent. Sci. 8 (2022) 1589–1608.

  24. [24]

      I. Hisaki, C. Xin, K. Takahashi, T. Nakamura, Angew. Chem. Int. Ed. 58 (2019) 11160–11170.

  25. [25]

      C. Huang, C. Zhao, Q. Deng, et al., Nat. Catal. 6 (2023) 729–739.

  26. [26]

      S. Huang, J. Li, Y. Lin, et al., J. Am. Chem. Soc. 146 (2024) 1967–1976.

  27. [27]

      G. Chen, S. Huang, X. Ma, R. He, G. Ouyang, Nat. Protoc. 18 (2023) 2032–2050.

  28. [28]

      A. Huang, H. Yang, S. Huang, G. Chen, G. Ouyang, Matter 6 (2023) 2635–2646.

  29. [29]

      W. Huang, H. Yuan, H. Yang, et al., Nat. Commun. 14 (2023) 3644.

  30. [30]

      G. Chen, L. Tong, S. Huang, et al., Nat. Commun. 13 (2023) 4816.

  31. [31]

      W. Huang, H. Yuan, H. Yang, et al., JACS Au 2 (2022) 2048–2058.

  32. [32]

      G. Chen, S. Huang, Y. Shen, et al., Chem 7 (2021) 2722–2742.

  33. [33]

      Z. Tang, X. Li, L. Tong, et al., Angew. Chem. Int. Ed. 60 (2021) 23608– 23613.

  34. [34]

      H. Yang, J. Fu, W. Huang, et al., Small Struct. 4 (2023) 2200346.

  35. [35]

      S. Hayashi, S. Nakamura, BBA–Enzymology 657 (1981) 40–51.

  36. [36]

      N.K. Maddigan, A. Tarzia, D.M. Huang, et al., Chem. Sci. 9 (2018) 4217–4223.

  37. [37]

      F.C. Church, D.H. Porter, G.L. Catignani, H.E. Swaisgood, Anal. Biochem. 146 (1985) 343–348.

  38. [38]

      B.B. Pinheiro, N.S. Rios, G. Zanatta, B.C. Pessela, L.R.B. Gonçalves, Process Biochem. 133 (2023) 292–302.

  39. [39]

      W.H. Chen, M. Vázquez-González, A. Zoabi, R. Abu-Reziq, I. Willner, Nat. Catal. 1 (2018) 689–695.

  40. [40]

      Z. Xu, Y.K. Cen, S.P. Zou, Y.P. Xue, Y.G. Zheng, Crit. Rev. Biotechnol. 40 (2019) 83–98.

  41. [41]

      N.G. Nezhad, R.N.Z.R.A. Rahman, Y.M. Normi, et al., Int. J. Biol. Macromol. 232 (2023) 123440.

  42. [42]

      N. Özer, M. Müftüoglu, D. Ataman, A. Ercan, I.H. Ögüs, J. Biochem. Bioph. Meth. 39 (1999) 153–159.

  43. [43]

      S. O'Connell, G. Walsh, Appl. Microbiol. Biotechnol. 86 (2010) 517–524.

  44. [44]

      Z. Li, J.R. Askim, K.S. Suslick, Chem. Rev. 119 (2019) 231–292.

  45. [45]

      H. Chen, H. Huang, H. Xu, et al., Small 20 (2024) 2308716.

  46. [46]

      L.A. Pérez-Márquez, M.D. Perretti, R. García-Rodríguez, F. Lahoz, R. Carrillo, Angew. Chem. Int. Ed. 61 (2022) e202205403.

  47. [47]

      Y. Wang, B. Li, T. Tian, et al., TrAC-Trends Anal. Chem. 149 (2022) 116565.

  48. [48]

      N.E. Azmi, N.I. Ramli, J. Abdullah, et al., Biosens. Bioelectron. 67 (2015) 129–133.

  49. [49]

      L. Lin, Y. Wen, Y. Liang, N. Zhang, D. Xiao, Anal. Methods 5 (2013) 457–464.

  50. [50]

      A. Singh, A.S.K. Sinha, Appl. Surface Sci. 430 (2018) 184–197.

  51. [51]

      X. Li, C. Li, L. Chen, New J. Chem. 39 (2015) 9976–9982.

• 

  Scheme 1  Schematic of the unsuccessful encapsulation of low-pI enzyme and the facilitated capsulation of the aminated enzyme. (a) Schematic of unsuccessful encapsulation of low-pI enzyme into the HOF framework. (b) Schematic of the facilitated encapsulation of the enzyme upon amination via EDA molecules.

  下载: 全尺寸图片 幻灯片
  

  Figure 1  Principles of the failed encapsulation of the pristine low-pI enzyme and the facilitated encapsulation of modified enzyme. (a) Schematic of the partially deprotonated H₄TBAPy linker and low-pI enzyme with negative charge in water solution, existing a charge repulsion (ⅰ) and a proton transfer that facilitated the deprotonation of the linker (ⅱ), both of which hinder the efficient enzyme encapsulation. (b) The binding mode and binding energy between the partially deprotonated linker and the amino acid residue with -NH₃⁺ group.

  下载: 全尺寸图片 幻灯片
  

  Figure 2  Enzyme surface amination and structure and activity characterization. (a) Schematic of GOx amination by site-specifically reacting to the carboxyl groups in the enzyme via EDC activation, increasing the -NH₃⁺ groups of the enzyme in water solution. (b) Zeta potentials of GOx, GOx-NH₂ and GOx/EDA. (c) IEF electrophoresis photographs of GOx and GOx-NH₂, respectively. (d) Catalytic curves of GOx and GOx-NH₂ based on a typical GOx-HRP cascade reaction. Conditions: enzyme, 10 µg/mL; ABTS, 0.2 mmol/L; glucose, 0.4 mmol/L; HRP, 1 µg/mL. (e) FT-IR spectra of GOx and GOx-NH₂ that the scanning ranges from 1800 cm^-1 to 1200 cm^-1. (f) CD spectra of GOx and GOx-NH₂, in which the bands around 210 nm and 218 nm were assigned to the α-helix and β-sheet structure of the enzyme, respectively.

  下载: 全尺寸图片 幻灯片
  

  Figure 3  Biomimetic encapsulation of enzyme within porous HOF and structure, activity and stability characterization. (a) Photographs of the generation of GOx@HBF-1 and GOx-NH₂@HBF-1 biocomposites. (b) PXRD patterns of GOx@HBF-1 and GOx-NH₂@HBF-1 as compared to the simulated PFC-1 crystal. (c) CLSM images of the GOx-NH₂@HBF-1, where the GOx-NH₂ was pre-labeled by AF350. (d) N₂ adsorption/desorption isotherms of the PFC-1 and GOx-NH₂@HBF-1 composite. Inset was the curve of pore volume against pore size distribution. (e) Catalytic curves of free GOx, GOx@HBF-1 and GOx-NH₂@HBF-1. Conditions: The enzyme concentration in the GOx or GOx-NH₂@HBF-1 was 50 µg/mL. The GOx@HBF-1 and GOx-NH₂@HBF-1 amounts were kept at 89 µg/mL. Other conditions: 3.0 mmol/L glucose, 0.2 mmol/L ABTS, 8 µg/mL HRP. (f) The kinetic curve of GOx-NH₂@HBF-1 fitted by Michaelis-Menten equation under a series of glucose concentrations. Conditions: 50 µg/mL GOx-NH₂, 0.2 mmol/L ABTS, 8 µg/mL HRP. (g) The retained activity of GOx, GOx-NH₂ and GOx-NH₂@HBF-1 after different treatments. The treatment time was kept at 30 min.

  下载: 全尺寸图片 幻灯片
  

  Figure 4  Fabrication of QDs+GOx-NH₂@HBF-1 glucose senser and sensing performance. (a) Photographs of QDs, GOx-NH₂@HBF-1 and the QDs+GOx-NH₂@HBF-1 mixture solution under a UV lamp (365 nm), respectively. (b) The fluorescence spectra of QDs, GOx-NH₂@HBF-1 and the QDs+GOx-NH₂@HBF-1 mixture solution, respectively. Excitation wavelength was set at 365 nm. (c) Schematic of the QDs+GOx-NH₂@HBF-1 fluorescence sensor for glucose detection. (d) The linear range for glucose detection based on the QDs+GOx-NH₂@HBF-1 sensor.

  下载: 全尺寸图片 幻灯片
•