3-methyl-1,2,4-triazole (Adamas-beta, Hmtz), zinc nitrate hexahydrate (Sinopharm Chemical Reagent Co., Ltd., Zn (NO₃)₂·6H₂O), NH₃·H₂O (25%, Macklin), sodium phenylpyruvate (PPA, Aladdin), Fe₃O₄ (Aladdin), Bradford protein detection kit (TaKaRa). None of the chemicals were further purified before usage.

The expression and purification of D-LDH (hereinafter referred to as LDH) refer to the previous research methods of our laboratory (Luo et al., 2020a). The LDH expression strain was grown and induced expression in Luria-Bertani (LB) medium containing 50 μg mL⁻¹ kanamycin. Cells were collected by centrifugation and LDH crude enzyme solution was obtained from lysed cells. The LDH crude enzyme was purified using Ni²⁺-nitrilotriacetic acid column (1.6 cm × 10 cm, BioRad, United States). And then dialyzed overnight against 20 mM potassium phosphate buffer (pH 8.0) to obtain purified LDH. See support information for specific experimental steps (S 1.1 Expression and purification of d-lactate dehydrogenase).

LDH/MNPs@MAF-7 was synthesized in an aqueous system (total volume was 4 ml) comprising Zn (NO₃)₂·6H₂O (40 mM), Hmtz (120 mM), Fe₃O₄ (20 mM), NH₃·H₂O (10%, 60 μL) and the indicated concentration of LDH at 25°C under stirring (200 rpm) for 24 h. The obtained material was recovered with magnets, washed thoroughly in distilled water and repeated three times to remove loosely adsorbed LDH.

The enzyme activity was measured by measuring the concentration of D-PLA catalyzed, using PPA (20 mM) and NADH (20 mM) as substrate. HPLC was used for detecting the concentration of D-PLA (Luo et al., 2020a). Specific test conditions refer to the laboratory published literature (Wang Y. et al., 2022). The standard curve of enzyme activity assay is shown in Supplementary Figure S3.

LDH/MNPs@MAF-7 was repeated eight times and the residual activity was measured after each use. After each cycle, it was collected with magnets, then added with a new substrate solution (20 mM PPA and NADH) to start a new catalytic reaction.

By adopting a one-pot encapsulation strategy, we encapsulated the LDH and MNPs into MAF-7 frame. The performances of the LDH/MNPs@MAF-7 composite were evaluated under various conditions using the preparation of D-PLA as a model reaction (Figure 1).

The ultrastructural characteristics of MAF-7 (Figure 2A), MNPs@MAF-7 (Figure 2B), and LDH/MNPs@MAF-7 (Figure 2C) were observed via SEM. The MAF-7 particles displayed a typical dodecahedral structure with a diameter of approximately 2 –10 μm.

As shown in Figure 2C, LDH/MNPs@MAF-7 exhibited a roughly spherical shape. This may be due to a number of the enzyme is encapsulated and some is exposed. According to the SEM images, the particle sizes of LDH/MNPs@MAF-7 were smaller than those of MAF-7 and MNPs@MAF-7, which might be due to the enzyme immobilization. Ji et al. (2021) observed a similar phenomenon when encapsulating lipase and Fe₃O₄ in ZIF-8. Additionally, elemental mapping (Figures 2D–G) demonstrated that the sample contains nitrogen, zinc, iron, and phosphorus, which is in good agreement with the expected chemical composition of LDH/MNPs@MAF-7.

The FT-IR bands corresponding to the different functional groups of MAF-7, MNPs@MAF-7, and LDH/MNPs@MAF-7 are indicated in Figure 3. In the FT-IR spectrum of MAF-7, the strong peaks around 424, 1,640 and 2,930 cm⁻¹ could correspond to the stretching vibrations of Zn-N, C-N, and imidazole C-H, respectively (Tian et al., 2020; Wang Q. et al., 2022). Several bands were observed in the broad region of 2,500–3,500 cm⁻¹, which corresponded to the stretching vibrations of N-H, O-H, and C-H. In the FT-IR spectra of Fe₃O₄ particles, a well-defined band near 600 cm⁻¹ indicated the presence of Fe-O bonds. The FI-IR spectrum of MAF-7 was consistent with that of MNPs@MAF-7 and LDH/MNPs@MAF-7, barring the peak around 600 cm⁻¹ (Fe-O). Thus, the FT-IR spectra indicated that both LDH and Fe₃O₄ were successfully immobilized in the MAF-7.

XRD analysis was performed using pure MAF-7, Fe₃O₄, MNPs@MAF-7, and LDH/MNPs@MAF-7 (Figure 4). Compare to MAF-7, it showed a slightly decreased in the intensity of the characteristic diffraction peaks of LDH/MNPs@MAF-7. It is possibly because that part of MAF-7 sites is occupied by enzyme molecules, which partly disorganized the crystalline structure of the MOF. In addition, the XRD pattern of Fe₃O₄ nanoparticles showed six characteristic peaks at 2θ values of 30.26°, 35.74°, 43.22°, 53.32°, 57.24° and 62.8°, respectively corresponding to planes (220) (311) (400) (422) (511) and (440) of the face-centered cubic spinel Fe₃O₄ structure (Nosike et al., 2020). The XRD patterns collected for LDH/MNPs@MAF-7 indicated that the crystal structure of MAF-7 was not influenced by enzyme and magnetic nanoparticles immobilization.

The VSM with the corresponding field of −20 KOe ∼20 KOe was employed to study the magnetic properties of prepared MNPs@MAF-7 and LDH/MNPs@MAF-7 at 300 K. As shown in Figure 5, the appeared magnetism increased with the external magnetic field until the saturation magnetization of the MNPs@MAF-7 and LDH/MNPs@MAF-7 was reached at Ms = 13.65 and 8.98 emu/g. This value was much smaller than that of bulk Fe₃O₄ (76.2 emu/g), and the magnetization of Fe₃O₄ was basically consistent with that reported in the literature (81.3 emu/g) (Ge et al., 2020). This indicates that the MAF-7 and LDH shell does not contribute to the magnetization, resulting in a smaller magnetic moment per unit mass. However, a strong response to an external magnetic field was still observed (inset of Figure 5A). After the significant amplification, the coercivity (defined as the field magnitude necessary to obtain M = 0) was as low as 48 and 59.5 Oe, respectively, indicating that the microspheres can be easily dispersed in the absence of an external magnetic field. This directly demonstrates that the LDH/MNPs@MAF-7 possess strong magnetic properties. It will be easy and efficient to separate LDH/MNPs@MAF-7 particles from the reaction system.

The MAF-7, MNPs@MAF-7 and LDH/MNPs@MAF-7 were analyzed by X-ray photoelectron spectroscopy to investigate the composition as well as the chemical state. In Figure 6A, similar coordination of Zn and N was present in MNPs@MAF-7 and LDH/MNPs@MAF-7 as in MAF-7. The peak at 1,022.48 eV corresponded to Zn 2p peak (Figure 6B), slightly higher than the standard binding energy of Zn²⁺ (1,022 eV), which also indicated that the chemical environment of Zn is dominated by N-coordination (Wang J. et al., 2020). As illustrated in Figure 6C, there are two characteristic peaks at 710.92 and 724.68 eV, which were interpreted as the Fe 2p peaks of Fe₃O₄. As demonstrated in Supplementary Table S2, we calculated the chemical composition of the elements (C, N, Zn, P, Fe) through the integration of the peak areas. The existence of Fe and P suggested the successful encapsulation of MNP and LDH.

The TGA curves exhibited two platforms, meaning two obvious mass loss phases (Figure 7). The initial weight loss phases (start from 30 to 400°C) may corresponded to the removal of water molecules and unreacted materials from the pores (Wang Y. et al., 2022).

The deformation of frame probably accounted for the second mass loss phase (from 400 to 850°C), which was more conspicuous in LDH/MNPs@MAF-7 (Wang L. et al., 2020). After heating to 850°C, the remaining weight of LDH/MNPs@MAF-7 and MNPs@MAF-7 was 32.6 and 45.7%, respectively. This can be explained by the loss of protein, which illustrates that LDH is successfully encapsulated in MAF-7 framework (Shieh et al., 2013; Shieh et al., 2015). According to the results of TGA, the loading efficiency of the enzyme in LDH/MNPs@MAF-7 was 12.1%.

In order to investigate the BET surface area and pore structure of MAF-7 (Figure 8A), MNPs@MAF-7 (Figure 8B) and LDH/MNPs@MAF-7 (Figure 8C), the N2 absorption and desorption experiments were conducted. The isotherms were categorized as a type-Ⅱ hysteresis loop attributed to the microporous structure of LDH/MNPs@MAF-7, which is similar to the isotherms of MAF-7. The specific surface area of LDH/MNPs@MAF-7 was ∼511 m²/g, which was smaller than MNPs@MAF-7 (∼755 m²/g) and MAF-7 (∼1,115 m²/g), due to the encapsulation of LDH and MNPs (Schejn et al., 2015; Wang L. et al., 2020). Further comparison of the pore diameters of LDH/MNPs@MAF-7, MNPs@MAF-7 and MAF-7 (Figure 8D) suggested that LDH/MNPs@MAF-7 and MNPs@MAF-7 had similar pore structure with MAF-7, illuminating good structural preservation during the encapsulation process.

The enzymatic properties of LDH/MNPs@MAF-7 were investigated using pyruvic acid as the substrate. To evaluate the ability of MAF-7 to protect LDH at high temperatures, LDH/MNPs@MAF-7 was incubated at 30°C–80°C for 20 min. As displayed in Figure 9A, the relative enzyme activity retention rate of LDH/MNPs@MAF-7 decreased with increasing temperature, but the catalyst retained more than 10% of the maximum activity even at 80°C. By contrast, the free LDH was practically completely inactivated at 80°C. The enhanced thermal stability could be illustrated that LDH was protected by the rigid structure of MAF-7. This can minimize the unfolding of the enzyme (Ji et al., 2022).

To evaluate the influence of pH on the catalytic activity of LDH, the catalyst was incubated in tris-HCl buffers with different pH (4–10). As shown in Figure 9B, LDH/MNPs@MAF-7 demonstrated better tolerance under acidic conditions, and the activity reached the maximum at pH 6. This indicates that the three-dimensional structure of MAF-7 provided a microenvironment, thus helping the LDH retain its activity and preventing the subunits dissociation (Li et al., 2018; Ji et al., 2022).

The solubility of PPA is very poor, and it needs to be solubilized with organic solvents. Thus, LDH/MNPs@MAF-7-catalyzed reactions have been assessed in the presence of a variety of organic solvents (DMSO, DMF, DCM) (Figure 9C). LDH/MNPs@MAF-7 showed excellent tolerance to organic solvents (>58% residual activity). It has been reported that MAF-7 forms a barrier, which effectively reduced exposure to enzymes and organic solvents (Liang et al., 2015; Nadar and Rathod, 2019). The results indicated that the LDH/MNPs@MAF-7 was suitable for the biosynthesis of PLA.

Finally, the reusability of LDH/MNPs@MAF-7 was assessed as an important performance indicator for actual applications. After each cycle, the LDH/MNPs@MAF-7 were recycled and washed completely, then repeated the experiment of enzyme activity assessment. As illustrated in Figure 9D, the residual activity remained above 80% after eight successive cycles. The results demonstrated that LDH/MNPs@MAF-7 had a certain technical dominance in the D-PLA biosynthesis. The decrease of enzyme activity may be attributed to the leakage of enzyme from the supports and denaturation of enzyme during the biocatalytic process over many reuse cycles (Liang et al., 2020).

It is noticed that the intrinsic fragile nature of enzymes makes them prone to denaturation or destabilization in harsh practical conditions, leading to unavoidably shortened lifespan and extremely high cost. Thus, we provided a possible synthesis strategy for application of enzyme immobilization by innovatively magnetic modifications to MAF-7. Herein, LDH/MNPs@MAF-7 exhibited more stable catalytic performances than those of free enzymes, including improved enzyme efficiency, stability and recyclability, due to the protection of enzymes by highly ordered frameworks. In addition, it exhibited excellent resistance to organic reagents, which makes it suitable for the industrial bioconversion of poorly soluble substrates. To summarize, the considerable universality and expansibility of this synthesis strategy would promise great potential in industrial applications of immobilized enzymes.