Fabrication of a 3D nano-bienzyme cascade reactor based on a Dps protein scaffold for chiral amine synthesis

TL;DR

Researchers built a 3D nano-enzyme reactor using a Dps protein scaffold, which efficiently and sustainably produces chiral amines with high yield and selectivity.

Contribution

A novel 3D nano-bienzyme cascade reactor was fabricated using a Dps protein scaffold for efficient chiral amine synthesis.

Findings

The 3DNECR showed significantly higher catalytic efficiency due to optimized spatial organization.

The reactor maintained over 80% activity after 9 days and high stability across pH and temperature ranges.

The 3DNECR achieved 99.9% yield and enantioselectivity in synthesizing R-BPA with 92% activity after six reuse cycles.

Abstract

Based on the self-assembling properties of the SpyCatcher/SpyTag system and the structural advantages of Dps protein, this study successfully constructed a three-dimensional nano-enzyme cascade reactor (3DNECR) through the covalent self-assembly of SpyTag-ADH and SpyCatcher-Dps-ATA117 fusion proteins. The 3DNECR exhibited significantly enhanced catalytic efficiency compared to the two-dimensional control, attributed to optimized spatial organization promoting substrate channeling. The reactor exhibited remarkable storage, pH, and thermal stability. It maintained over 80% activity after 9 days of storage, showed superior pH tolerance across pH 8–10, and remained stable in the temperature range of 4–40 ℃. Molecular docking confirmed strong interfacial binding (− 17.3 kcal/mol) between assembly components and favorable substrate binding (− 7.4 kcal/mol) within the active site. Furthermore, the 3DNECR was applied to the asymmetric synthesis of (R)-1-[3,5-bis(trifluoromethyl)phenyl]ethanamine (R-BPA) in an oil-water biphasic system. Under optimized conditions, the 3DNECR consistently achieved high yields (99.9%) and excellent enantioselectivity (99.9%). The 3DNECR maintained a relative enzyme activity as high as 92% even after six cycles of reuse. This integrated platform showcases substantial potential for efficient and sustainable biocatalytic applications.

The online version contains supplementary material available at 10.1186/s40643-026-01022-7.

Figures

Schematic diagram of the molecular docking. A Binding mode of the SpyTag-ADH/SpyCatcher-Dps-ATA117 complex. SpyTag-ADH is shown in pink and the SpyCatcher-Dps-ATA117 protein is shown in blue. Hydrogen bonds are indicated by yellow dashed lines. B Binding mode of 3D nano dual-enzyme cascade reactor (3DNECR) with 3,5-bis(trifluoromethyl)acetophenone. Protein residues are depicted in blue-grey sticks. The R molecule is shown in yellow sticks. Hydrogen bonds are represented by solid blue lines. Hydrophobic interactions are indicated by grey dashed lines. π-π stacking interactions are shown as gree

Effect of substrate concentration (A) and pH (B) on the transamination reaction

Effect of number of reuses on relative enzyme activity

Coupling mechanism of SpyCatcher/SpyTag

Schematic diagram of the construction of the SpyTag-ADH (A) and SpyCatcher-Dps-ATA117 (B) fusion proteins

Hydrophobicity (A) and signal peptide (B) analysis of SpyCatcher-Dps-ATA117

Schematic diagram of the recombinant plasmid construction

SDS-PAGE analysis of SpyTag-ADH, SpyCatcher-Dps-ATA117, and 3D nano dual-enzyme cascade reactor (3DNECR)

Schematic illustration of the preparation of the 3D nano dual-enzyme cascade reactor (3DNECR)

Optimization of self-assembly time (A) and concentration (B) for the 3D nano dual-enzyme cascade reactor (3DNECR)

FE-SEM and TEM images. A FE-SEM image of SpyTag-ADH. B FE-SEM image of SpyCatcher-Dps-ATA117. C FE-SEM image of 3D nano dual-enzyme cascade reactor (3DNECR). D TEM image of 3DNECR

Stability of 3D nano dual-enzyme cascade reactor (3DNECR). A Storage stability. B pH stability. C Thermal stability

Full text

Introduction

In nature, highly organized multi-enzyme complexes drive diverse metabolic reactions with remarkable specificity (Lim et al. 2020; Röder et al. 2017). These complexes catalyze cascade reactions through the synergistic action of multiple enzymes, leveraging their distinct substrate specificities to make the reaction pathway more efficient and environmentally benign (Cai et al. 2025; Yan et al. 2025). The spatial proximity of enzymes facilitates substrate channeling, which protects unstable intermediates from degradation, avoids unfavorable reaction equilibria, and ultimately enhances synergistic effects and catalytic efficiency (Schmid-Dannert and López-Gallego 2019). Researchers have employed various strategies to assemble multiple enzymes with different structures and catalytic properties into multi-enzyme complexes for cascade reactions (Li et al. 2025; Ma et al. 2025). These constructed complexes often demonstrate superior stability and catalytic efficiency compared to simple mixtures of free enzymes (Luo et al. 2024; Patil et al. 2025). Current methods for constructing artificial multi-enzyme complexes primarily include co-immobilization techniques, protein fusion technology, and scaffold-mediated self-assembly (Aer et al. 2024; Dang et al. 2025; Zhong et al. 2022). Scaffold systems provide an effective platform for immobilizing multiple proteins and enhance the catalytic efficiency of multi-enzyme systems through substrate channeling and synergistic mechanisms, representing an ideal model for orchestrating cascade reactions. Common scaffolds used in these systems include DNA, RNA, and proteins (Duarte et al. 2025; Lai et al. 2025). While DNA and RNA scaffolds offer superior precision in controlling inter-enzyme distances, their high cost often limits large-scale application (Lin et al. 2023). Therefore, designing protein scaffolds that mimic and draw inspiration from natural catalytic systems is crucial for constructing highly efficient multi-enzyme cascade reactors.

Chiral amines serve as indispensable building blocks for active pharmaceutical ingredients, agrochemicals, and bioactive natural products, playing a vital role in the national economy (Jongkind et al. 2025; Wang et al. 2024). ω-Transaminase (ATA117) and alcohol dehydrogenase (ADH) have been employed as biocatalysts in cascade reactions for chiral amine synthesis (Lu et al. 2022). The SpyCatcher/SpyTag system originates from the split CnaB2 domain of the Streptococcus pyogenes surface protein FbaB. This system spontaneously forms a stable isopeptide bond under diverse conditions between a lysine (Lys) residue on SpyCatcher and an aspartic acid (Asp) residue on SpyTag (Fig. 1). SpyTag can be positioned at the N-terminus, C-terminus, or within internal sites of a target protein, offering greater flexibility than labeling methods that rely on split inteins or sortases. The reaction between SpyTag and SpyCatcher proceeds rapidly, with a reported half-life of approximately 74 s at 10 µM (Peng et al. 2021). This robust covalent intermolecular isopeptide linkage makes the system a powerful tool for constructing and controlling various functional proteins (Gentili et al. 2025; Si et al. 2016). However, practical applications face challenges including limited spatial organization capability and suboptimal stability, which adversely affect enzymatic synergy and reusability (Che et al. 2024). The miniferritin (Dps) derived from Listeria innocus consists of 12 subunits that self-assemble into a symmetrical nanocage-like structure (Zhu et al. 2021). This architecture provides an excellent scaffold material for spatially extending multi-enzyme metabolic networks, enabling efficient spatial organization and aggregation of dual enzymes, thereby facilitating the preparation of a three-dimensional nanoscale dual-enzyme cascade reactor (3DNECR) with controllable spatial configuration. Based on this principle, this study utilized the self-reactive coupling capability of the SpyCatcher/SpyTag system to co-assemble ATA117 and ADH onto a Dps scaffold with three-dimensional topological structure, designing and constructing an efficient 3DNECR. This 3DNECR was applied in an oil-water biphasic system to catalyze the asymmetric transamination of 3,5-bis(trifluoromethyl)acetophenone (BPO) for (R)-1-[3,5-bis(trifluoromethyl)phenyl]ethanamine (R-BPA) production. This study contributes to the sustainable development goals, particularly SDG 9 (industry, innovation and infrastructure) and SDG 12 (responsible consumption and production), by developing a highly efficient, stable, and reusable 3DNECR that enhances biocatalytic efficiency, reduces resource and energy consumption, and supports greener, more sustainable chemical manufacturing processes (Fegade 2024a, c).

Fig. 1. Coupling mechanism of SpyCatcher/SpyTag

Materials and methods

Materials and reagents

3,5-Bis(trifluoromethyl)acetophenone (BPO), isopropylamine hydrochloride, ethanol, pyridoxal 5’-phosphate (PLP), isopropylamine hydrochloride (IPA), NADH, n-hexane, dibutyl phthalate, benzene, toluene, n-heptane, agar, tryptone, yeast extract, sodium chloride, tris(hydroxymethyl)aminomethane (Tris), concentrated hydrochloric acid, and sodium phosphate buffer. Escherichia coli BL21(DE3) host strain was purchased from Tsingke Biotechnology Co., Ltd.

Instruments

The instruments used in this study included an ultrasonic cell disruptor, an electrophoresis system, a gel imaging system, a centrifuge, a constant temperature incubator, a clean bench, an orbital shaker, a Malvern Zeta potential analyzer and particle size analyzer (ZSU3200, United Kingdom), a microplate reader (Ensight, Singapore), and a field-emission scanning electron microscope (FE-SEM, SU8600, Hitachi).

Preparation and purification of the SpyTag-ADH and SpyCatcher-Dps-ATA117

The ADH gene (NP_014555.1) used in this study was obtained from the National Center for Biotechnology Information (NCBI) database. In earlier work, our group successfully constructed the recombinant plasmid pET28a-SpyTag-ADH by modifying ADH with the SpyTag (Lu et al. 2023). The sequences of both ADH and SpyTag-ADH are provided in the Supplementary Material. A 10 µL aliquot of the laboratory-constructed and preserved glycerol stock of E. coli BL21(DE3)/pET28a-SpyTag-ADH was inoculated into 50 mL of Luria-Bertani (LB) liquid medium (containing 5 g/L yeast extract, 10 g/L peptone, 10 g/L sodium chloride, pH 7.0) supplemented with 50 µg/mL kanamycin. Meanwhile, the SpyCatcher-Dps-ATA117 fusion gene was commercially synthesized and inserted between the BamH I and Xho I sites of a pET28a plasmid for preservation in E. coli DH5α. The gene sequence of ATA117 (GenBank: JA717225) was obtained from the NCBI. The sequences for both ATA117 and SpyCatcher-Dps-ATA117 are provided in the Supplementary Material. The pET28a-SpyCatcher-Dps-ATA117 plasmid was extracted and transformed into E. coli BL21(DE3) competent cells to generate the recombinant strain BL21(DE3)/pET28a-SpyCatcher-Dps-ATA117. A single colony of this strain was inoculated into 50 mL of LB medium supplemented with 50 µg/mL kanamycin. Both cultures were incubated overnight at 37 ℃ with orbital shaking at 180 rpm to obtain the seed cultures. Subsequently, each of the aforementioned seed cultures was transferred at a 3% (v/v) inoculum into 150 mL of fresh LB medium and cultivated under identical conditions until the optical density at 600 nm (OD_600_) reached 0.8. Following the addition of 1 mM isopropyl β-D-1-thiogalactopyranoside (IPTG) to cultures of both engineered strains, the cultures were transferred to a 20 ℃ shaker at 180 rpm for 16 h to induce protein expression. The induced cultures were then centrifuged at 8000 rpm for 10 min at 4 ℃ to harvest the cell pellets. The pellets were washed three times with 0.9% (w/v) saline. The harvested wet cells were resuspended in 100 mM sodium phosphate buffer (pH 8.0) at a concentration of 50 g/L and disrupted by ultrasonication for 10 min. The lysates were centrifuged at 10,000 rpm for 10 min at 4 ℃, and the resulting supernatants were collected and filtered through a water-based microporous membrane (0.22 μm) to obtain crude enzyme extracts of SpyTag-ADH and SpyCatcher-Dps-ATA117, respectively. SpyTag-ADH and SpyCatcher-Dps-ATA117 were purified via affinity chromatography using a BeyoGold His-tag Purification Resin (Beyotime, Shanghai, China). The eluate for each protein was collected and then transferred to a dialysis membrane with a molecular weight cutoff of 8000–14,000 Da for overnight dialysis against a 10-fold volume of 0.01 M phosphate buffer (PB, pH 8.0) at 4 ℃. The retentate was pre-frozen at -80 ℃ for 6 h and then lyophilized for 12 h. The resulting dried powder represented the lyophilized enzyme powders of the fusion proteins SpyTag-ADH and SpyCatcher-Dps-ATA117. These powders were thoroughly dissolved in deionized water to obtain the final SpyTag-ADH and SpyCatcher-Dps-ATA117 enzyme solutions.

Preparation of 3DNECR

Protein concentrations were measured using the Bradford method with bovine serum albumin as a standard (Wang et al. 2023). The lyophilized enzyme powders of the fusion proteins SpyCatcher-Dps-ATA117 and SpyTag-ADH were separately reconstituted with deionized water to prepare enzyme concentration of 5 µM. Then, 100 µL aliquot of the freshly prepared SpyCatcher-Dps-ATA117 solution was mixed with 100 µL of the SpyTag-ADH solution, and the total volume was adjusted to 500 µL with deionized water to form the assembly system. The mixture was incubated at 4 °C for 4 h, with gentle shaking every 30 min to facilitate the assembly process. The success of self-assembly was preliminarily assessed by sodium dodecyl sulfate-polyacrylamide gel electrophoresis (SDS-PAGE). The resulting solution was pre-frozen at − 80 °C for 6 h and then lyophilized for 12 h to obtain a dry powder, which was denoted as the 3DNECR. Under identical conditions, the fusion protein SpyCatcher-Dps-ATA117 was replaced with SpyCatcher-ATA117 to prepare the 2D nano dual-enzyme cascade reactor (2DNECR). The SpyCatcher-ATA117 fusion protein was obtained by inducing expression in the previously constructed engineered strain E. coli BL21(DE3)/pET28a-SpyCatcher-ATA117, followed by isolation and purification. Detailed information on BL21(DE3)/pET28a-SpyCatcher-ATA117 is provided in Table 1.

Table 1. Details of the genetically engineered strains used in this studyStrainsEnzymes/ProteinsAmino acid sequenceSourceBL21(DE3)/pET28a-ATA117ATA117MAFSADTPEIVYTHDTGLDYITYSDYELDPANPLAGGAAWIEGAFVPPSEARIPIFDQGFYTSDATYTTFHVWNGNAFRLGDHIERLFSNAESIRLIPPLTQDEVKEIALELVAKTELREAMVTVTITRGYSSTPFERDITKHRPQVYMSASPYQWIVPFDRIRDGVHLMVAQSVRRTPRSSIDPQVKNFQWGDLIRAIQETHARGFELPLLLDCDNLLAEGPGFNVVVIKDGVVRSPGRAALPGITRKTVLEIAESLGHEAILADITPAELYDADEVLGCSTGGGVWPFVSVDGNSISDGVPGPVTQSIIRRYWELNVEPSSLLTPVQYOur laboratoryBL21(DE3)/pET28a-SpyCatcher-ATA117SpyCatcher-ATA117GAMVDTLSGLSSEQGQSGDMTIEEDSATHIKFSKRDEDGKELAGATMELRDSSGKTISTWISDGQVKDFYLYPGKYTFVETAAPDGYEVATAITFTVNEQGQVTVNGKATKGDAHIGGGGSGGGGSMAFSADTPEIVYTHDTGLDYITYSDYELDPANPLAGGAAWIEGAFVPPSEARIPIFDQGFYTSDATYTTFHVWNGNAFRLGDHIERLFSNAESIRLIPPLTQDEVKEIALELVAKTELREAMVTVTITRGYSSTPFERDITKHRPQVYMSASPYQWIVPFDRIRDGVHLMVAQSVRRTPRSSIDPQVKNFQWGDLIRAIQETHARGFELPLLLDCDNLLAEGPGFNVVVIKDGVVRSPGRAALPGITRKTVLEIAESLGHEAILADITPAELYDADEVLGCSTGGGVWPFVSVDGNSISDGVPGPVTQSIIRRYWELNVEPSSLLTPVQYOur laboratoryBL21(DE3)/pET28a-SpyTag-ADHSpyTagAHIVMVDAYKPTKOur laboratoryADHMSIPETQKGVIFYESHGKLEYKDIPVPKPKANELLINVKYSGVCHTDLHAWHGDWPLPVKLPLVGGHEGAGVVVGMGENVKGWKIGDYAGIKWLNGSCMACEYCELGNESNCPHADLSGYTHDGSFQQYATADAVQAAHIPQGTDLAQVAPILCAGITVYKALKSANLMAGHWVAISGAAGGLGSLAVQYAKAMGYRVLGIDGGEGKEELFRSIGGEVFIDFTKEKDIVGAVLKATDGGAHGVINVSVSEAAIEASTRYVRANGTTVLVGMPAGAKCCSDVFNQVVKSISIVGSYVGNRADTREALDFFARGLVKSPIKVVGLSTLPEIYEKMEKGQIVGRYVVDTSKLinker 1GGGGSGGGGSBL21(DE3)/pET28a-SpyCatcher-Dps-ATA117SpyCatcherGAMVDTLSGLSSEQGQSGDMTIEEDSATHIKFSKRDEDGKELAGATMELRDSSGKTISTWISDGQVKDFYLYPGKYTFVETAAPDGYEVATAITFTVNEQGQVTVNGKATKGDAHIGene synthesisDpsMSTAKLVKSKATNLLYTRNDVSDSEKKATVELLNRQVIQFIDLSLITKQAHWNMRGANFIAVHEMLDGFRTALIDHLDTMAERAVQLGGVALGTTQVINSKTPLKSYPLDIHNVQDHLKELADRYAIVANDVRKAIGEAKDDDTADILTAASRDLDKFLWFIESNIEATA117MAFSADTPEIVYTHDTGLDYITYSDYELDPANPLAGGAAWIEGAFVPPSEARIPIFDQGFYTSDATYTTFHVWNGNAFRLGDHIERLFSNAESIRLIPPLTQDEVKEIALELVAKTELREAMVTVTITRGYSSTPFERDITKHRPQVYMSASPYQWIVPFDRIRDGVHLMVAQSVRRTPRSSIDPQVKNFQWGDLIRAIQETHARGFELPLLLDCDNLLAEGPGFNVVVIKDGVVRSPGRAALPGITRKTVLEIAESLGHEAILADITPAELYDADEVLGCSTGGGVWPFVSVDGNSISDGVPGPVTQSIIRRYWELNVEPSSLLTPVQYLinker 2AGAGAGPEGAGAGAGPEGAGAGAGPEGAGAGAGPEGAGAGAGPEGLinker 3GGGGSGGGGS

Enzyme activity assay

The enzyme activity assay for ATA117, SpyCatcher-ATA117, SpyCatcher-Dps-ATA117, and 3DNECR was carried out in a 2 mL reaction mixture containing 832.8 mM IPA, 1 mM PLP, 0.3 mL of DMSO with 166.6 mM BPO, 1 mL of enzyme solution (1 mg/mL), and 100 mM Tris-HCl buffer (pH 9). The reaction proceeded at 45 °C with shaking at 180 rpm for 2 h and was terminated by adding 1 µL of 1% trifluoroacetate. The production of R-BPA was quantified using a gas chromatography (GC) method. The method is described in the Supplementary Material. Under these conditions, one unit (U) of enzyme activity was defined as the amount of enzyme that catalyzes the formation of 1 µmol of R-BPA per minute.

Effect of self-assembly time on the 3DNECR

Solutions of SpyTag-ADH (5 µM) and SpyCatcher-Dps-ATA117 (5 µM) were mixed in equal volumes and incubated continuously. Aliquots were collected at time points of 1, 2, 3, 4, 5, and 6 h and transferred into the corresponding sample vials of a Malvern Zeta potential analyzer and particle size analyzer. Each sample was sonicated for 5 min to ensure homogeneous particle dispersion. All measurements were performed in triplicate to ensure data accuracy.

Effect of self-assembly concentration on the 3DNECR

To investigate the effect of self-assembly concentration on the self-assembly of 3DNECR, purified SpyTag-ADH and SpyCatcher-Dps-ATA117 were mixed at an equimolar ratio and allowed to self-assemble at 4 °C at different concentrations (2, 5, and 10 µM). The self-assembly process was monitored by dynamic light scattering (DLS). The 3DNECR prepared at different concentrations were transferred to corresponding sample vials, followed by sonication for 5 min to ensure homogeneous dispersion of particles. The samples were then loaded into a cuvette, and all measurements were performed in triplicate to ensure data accuracy.

FE-SEM analysis

The structure and morphology of SpyTag-ADH, SpyCatcher-Dps-ATA117, and the 3DNECR were characterized using FE-SEM (SU8600, Hitachi, Japan) operated at 2 kV. A small amount of the SpyTag-ADH, SpyCatcher-Dps-ATA117, and SpyTag-ADH/SpyCatcher-Dps-ATA117 assembly was transferred onto conductive adhesive tape mounted on an FE-SEM stub using a toothpick. Loosely adhered particles were removed by gentle blowing. To minimize charging effects, the sample was sputter-coated with a thin layer of gold prior to FE-SEM observation.

Kinetic analysis of 3DNECR

The enzyme kinetic parameters for the 3DNECR was determined spectrophotometrically by monitoring NADH consumption at 340 nm. The reactions were conducted in a 4 mL system containing 0.1 M Tris-HCl buffer (pH 9), 100 mM IPA, 10 mg 3DNECR or 2DNECR (equivalent to 22.9 or 26.6 µM, respectively), 1 mM PLP, and 1 × 10^− 3^ mol/L NADH, with BPO concentrations varying from 10 to 50 mM. Each substrate concentration was tested in triplicate. Absorbance measurements were taken at two-minute intervals, and the apparent kinetic constants Km​and Kcat were derived by fitting the data to the Michaelis–Menten equation.

Storage stability assay

The 3DNECR was constructed by incubating SpyTag-ADH (5 µM) and SpyCatcher-Dps-ATA117 (5 µM) in 2 mL of Tris-HCl buffer (0.1 M, pH 9) at 4 °C for 4 h. Both the assembled 3DNECR and the 2DNECR, assembled from 5 µM SpyTag-ADH and 5 µM SpyCatcher-ATA117) were stored in Tris-HCl buffer (100 mM, pH 9) at 4 °C for varying durations (0–9 days). Enzyme activities are reported relative to the initial activity of the 2DNECR (measured at pH 9 on day 0), which was set as 100%.

pH stability assay

The 3DNECR was prepared by incubating SpyTag-ADH (5 µM) and SpyCatcher-Dps-ATA117 (5 µM) in 2 mL of Tris-HCl buffer (0.1 M, pH 9) at 4 °C for 4 h. Both the resulting 3DNECR and the 2DNECR (assembled from 5 µM SpyTag-ADH and 5 µM SpyCatcher-ATA117) were subsequently incubated for 24 h at 4 °C in buffers with pH values ranging from 5 to 10. The buffer systems included Na_2_HPO_4_–NaH_2_PO_4_ for pH 5–8, Tris–HCl for pH 8–9, and glycine–NaOH for pH 9–10. The initial enzyme activity of each reactor before storage in pH buffers was set as 100%. After 24 h storage, the relative enzyme activity at each pH was calculated as the percentage of residual activity compared to the initial value.

Thermal stability assay

To evaluate the thermal stability of 3DNECR, it was prepared by incubating 5 µM SpyTag-ADH with 5 µM SpyCatcher-Dps-ATA117 in 2 mL of Tris-HCl buffer (0.1 M, pH 9) at 4 °C for 4 h. Similarly, 2DNECR was prepared via the self-assembly of 5 µM SpyTag-ADH and 5 µM SpyCatcher-ATA117. Both 3DNECR and 2DNECR were incubated in pH 9 buffer at different temperatures (4, 25, 30, 35, and 40 °C) for 24 h, after which the residual enzyme activity was measured. For each reactor, the activity before incubation at each temperature was defined as 100%. After 24 h, the relative activity at each temperature was calculated as a percentage of this initial value.

Molecular docking

The binding mode between the SpyCatcher-Dps-ATA117 fusion protein and the SpyTag-ADH fusion protein was predicted using Chai from the Chai Discovery Lab (https://github.com/chaidiscovery/chai-lab). The prediction was performed de novo based on the amino acid sequences of the two fusion proteins, without employing homologous templates or experimentally resolved structures. The top-ranked predicted complex (top 1) was subsequently subjected to energy minimization with the AMBER18 software package using the ff14SB force field. The energy-minimized structure was then evaluated for its binding energy using the online tool Prodigy (https://wenmr.science.uu.nl/prodigy). Finally, the complex structure with the most favorable binding energy was visualized and rendered using PyMOL version 2.5.3. Furthermore, molecular docking between the constructed protein complex 3DNECR and the small-molecule drug BPO was performed using AutoDock Vina version 1.2.5. A global docking approach was employed for the protein. The number of docking runs was set to 32, and the grid spacing was set to 0.375 Å, with all other parameters kept at their default values. Similarly, the complex structure with the most favorable binding energy was visualized and rendered using PyMOL version 2.5.3.

Effect of substrate concentration on the amination catalyzed by the 3DNECR

To investigate the effect of substrate concentration on the amination reaction catalyzed by 3DNECR in a biphasic system, a biphasic system with a 1:1 volume ratio of organic to aqueous phases was constructed, with 2DNECR at an equivalent mass used as a control. Specifically, 10 mg of the prepared 3DNECR was dispersed in 0.7 mL of 0.1 M Tris-HCl buffer (pH 9). To this dispersion were added the amino donor IPA at final concentrations of 1000, 1050, 1100, 1150, 1200, and 1250 mM, the cofactor pyridoxal phosphate PLP at 1 mM, and the cofactor NADH at 1 × 10^− 3^ mol/L. Subsequently, 0.3 mL of an ethanolic solution containing BPO at concentrations of 200, 210, 220, 230, 240, and 250 mM was introduced. The aqueous phase was then brought to a final volume of 1 mL with 0.1 M Tris-HCl buffer (pH 9), followed by the addition of 1 mL of the organic phase, dibutyl phthalate. The reaction vessels were incubated at 45 °C for 24 h. After the reaction, the mixtures were centrifuged at 8000 rpm for 10 min at 4 °C. The supernatant was collected and extracted three times with an equal volume of ethyl acetate. The combined organic extracts were dried over anhydrous sodium sulfate, and the ethyl acetate was evaporated at room temperature. The yield and enantiomeric excess (ee) of R-BPA were subsequently determined by GC as described in the Supplementary Material.

Effect of pH on the amination catalyzed by the 3DNECR

Biphasic reaction systems with a 1:1 aqueous-to-organic phase volume ratio were constructed using 0.1 M sodium phosphate buffer (pH 6), sodium phosphate buffer (pH 7), Tris-HCl (pH 8), Tris-HCl (pH 9), or glycine-NaOH (pH 10) as the aqueous medium, and dibutyl phthalate as the organic phase. To each aqueous buffer at the specified pH was added 0.3 mL of an ethanol solution containing 230 mM BPO, 1150 mM IPA, 1 mM PLP, 1 × 10^− 3^ mol/L NADH, and 10 mg of the prepared 3DNECR. The aqueous phase was then brought to a final volume of 1 mL with the corresponding buffer, followed by the addition of 1 mL of dibutyl phthalate. The reaction vessels were incubated at 45 °C with shaking at 180 rpm for 24 h. After the reaction, the mixtures were centrifuged at 8000 rpm for 10 min at 4 °C. The supernatant was collected and extracted three times with an equal volume of ethyl acetate. The combined organic extracts were dried over anhydrous sodium sulfate, and the ethyl acetate was evaporated at room temperature. The product yield and ee of R-BPA were subsequently determined.

Effect of recycling cycles of the 3DNECR on the amination reaction

The 3DNECR was prepared by mixing the two fusion enzymes, SpyTag-ADH and SpyCatcher-Dps-ATA117 (each at 5 µM), at a 1:1 molar ratio and incubating at 4 °C for 4 h. The resulting 3DNECR (10 mg) were introduced into 2 mL of 0.1 M Tris-HCl buffer (pH 9) containing 0.3 mL of a DMSO solution with 230 mM BPO, 1150 mM IPA, 1 mM PLP, and 1 × 10^− 3^ mol/L NADH. The cascade reaction proceeded at 45 °C with shaking at 180 rpm for 24 h. After reaction, the entire solution was transferred to a 30 kDa ultrafiltration device, and the 3DNECR was collected by centrifugation at 3000 × g and 4 °C for 30 min for reuse in the next cycle. This recovery and reuse procedure was repeated five times. The relative enzyme activity after each cycle was expressed relative to that of the 0th cycle, which was set to 100%. All cascade reactions were performed in triplicate.

Results and discussion

Design of SpyCatcher-Dps-ATA117

In this study, pET28a-SpyCatcher-Dps-ATA117 was designed as illustrated in Fig. 2. Following the design of the SpyCatcher-Dps-ATA117 fusion protein, theoretical analyses were performed to assess its potential for stable soluble expression in the expression host E. coli BL21(DE3). The amino acid sequence and composition analysis of SpyCatcher-Dps-ATA117 are presented in Tables 1 and 2, respectively. As indicated in Table 2, the content of sulfur-containing amino acids is markedly low. This suggests that the protein may contain very few or no disulfide bonds. Accordingly, it can be inferred that the stabilization of its three-dimensional structure likely does not rely on disulfide bridges. This characteristic is consistent with an intracellular localization, given that the reducing nature of the intracellular environment is unfavorable for disulfide bond formation. Furthermore, the notably higher proline content implies the presence of an increased number of turns or loop regions within the protein structure, which could influence both the rigidity and the folding mechanism of the protein. Hydrophobicity analysis of the fusion protein revealed a balanced distribution of hydrophilic and hydrophobic residues in its amino acid sequence, resulting in an overall amphiphilic character without strong hydrophilic or hydrophobic tendencies (Fig. 3A). The presence of hydrophobic amino acids facilitates the formation of an internal hydrophobic core during protein folding, which serves as a key driving force for the structural stability of globular proteins. Concurrently, sufficient hydrophilic residues enable the formation of a hydrated surface, allowing favorable interactions with aqueous solvents and promoting stable dissolution in the cytoplasm or physiological fluids. This moderate hydrophobicity suggests that the protein adopts a compact, soluble, and typical globular conformation. When expressed in E. coli, it is highly likely to accumulate in the cytoplasm in a soluble form rather than forming insoluble inclusion bodies. A signal peptide is a short amino acid sequence located at the N-terminus of a protein, whose primary function is to direct the newly synthesized protein across cellular membranes to specific locations. Once the protein reaches its target destination, the signal peptide is typically cleaved off by specific enzymes. Prediction results for the protein designed in this study indicate the absence of a signal peptide (Fig. 3B), suggesting that it can be efficiently expressed and accumulated in the cytoplasm. This facilitates effective purification through cell disruption and affinity chromatography, which is consistent with our design strategy.

Fig. 2. Schematic diagram of the construction of the SpyTag-ADH (A) and SpyCatcher-Dps-ATA117 (B) fusion proteins

Table 2. Amino acid composition analysis of SpyCatcher-Dps-ATA117Amino acidNumberPercentage (%)Sulfur-containing amino acidsCysteine (C)20.3Methionine (M)121.8Other amino acidsProline (P)334.9

Fig. 3. Hydrophobicity (A) and signal peptide (B) analysis of SpyCatcher-Dps-ATA117

Construction, expression, and purification of SpyTag-ADH and SpyCatcher-Dps-ATA117

The recombinant plasmids pET28a-ATA117, pET28a-SpyTag-ADH, and pET28a-SpyCatcher-ATA117 were previously constructed by our group. Schematic diagrams of the construction of these four plasmids are presented in Fig. 4, with detailed information on the corresponding genetically engineered strains listed in Table 1. The verified recombinant plasmid pET28a-SpyCatcher-Dps-ATA117 was transformed into the expression host E. coli BL21(DE3). Protein expression was induced with 1 mM IPTG when the OD_600_ reached 0.8, and successful production and purification were confirmed by SDS-PAGE. As shown in Fig. 5, His-tag affinity chromatography purification of SpyTag-ADH yielded a single discrete band with high clarity at the expected position corresponding to its theoretical molecular weight of ~ 43.0 kDa. In contrast, the purification of SpyCatcher-Dps-ATA117 resulted in discrete bands dominated by the distinct target band, though with additional minor bands present. The theoretical molecular weight of the SpyCatcher-Dps-ATA117 fusion protein is approximately 66.0 kDa, and a prominent band was observed at the corresponding position. These results demonstrate that the expected SpyTag-ADH and SpyCatcher-Dps-ATA117 were obtained following induced expression and purification of the genetically engineered strain in this study.

Fig. 4. Schematic diagram of the recombinant plasmid construction

Fig. 5SDS-PAGE analysis of SpyTag-ADH, SpyCatcher-Dps-ATA117, and 3D nano dual-enzyme cascade reactor (3DNECR)

Design and preparation of 3DNECR

The SpyCatcher/SpyTag system is derived from the engineered splitting of the CnaB2 domain within the FbaB protein of Streptococcus pyyogenes (Hosseini et al. 2025). This system exhibits notable advantages, including the spontaneous formation of a stable isopeptide bond across a broad spectrum of reaction conditions (Veron et al. 2025). Such site-specific covalent linkage enables the precise assembly and regulation of multifunctional protein complexes, establishing it as a robust platform for protein engineering. Nevertheless, certain limitations have been observed in practical implementations. For example, in the context of multi-enzyme cascade reactors, the SpyCatcher/SpyTag system offers limited capacity for spatial organization, thereby constraining synergistic interactions among enzymes and leading to suboptimal catalytic efficiency. Moreover, under harsh operational conditions or prolonged use, its structural integrity may be compromised, adversely impacting recyclability and long-term functional performance. The mini-ferritin (Dps) derived from Listeria spp. belongs to the ferritin superfamily and features a symmetrical nanocage structure formed through the self-assembly of 12 subunits. This cage-like architecture renders Dps an ideal platform for nano-bioreactors, enabling applications in catalysis and materials synthesis. The compact size of Dps minimizes steric hindrance, while its dodecameric nanocage structure facilitates the formation of substrate channels. Meanwhile, the SpyCatcher/SpyTag system further promotes the covalent self-assembly between SpyTag-ADH and SpyCatcher-Dps-ATA117, resulting in a stable and efficient 3DNECR (Fig. 6). As described in Sect.  “Preparation of 3DNECR”, SpyTag-ADH and SpyCatcher-Dps-ATA117 were mixed to carry out covalent coupling, and the products were analyzed by SDS-PAGE. As shown in Fig. 5, co‑incubation of SpyTag‑ADH (~ 43.0 kDa) and SpyCatcher‑Dps‑ATA117 (~ 66.0 kDa) resulted in the appearance of a distinct new high‑molecular‑weight band close to the theoretical molecular weight of ~ 109 kDa. This confirms the efficient and specific covalent conjugation via the SpyCatcher/SpyTag system, leading to the successful assembly of the 3DNECR.

Fig. 6. Schematic illustration of the preparation of the 3D nano dual-enzyme cascade reactor (3DNECR)

Effect of self-assembly time on the 3DNECR

The covalent self-assembly process between SpyTag-ADH and SpyCatcher-Dps-ATA117 was analyzed by measuring the particle size distribution of individual fusion proteins and their mixtures. As shown in Fig. 7A, although SpyCatcher-Dps-ATA117 is slightly larger than SpyTag-ADH, both proteins exhibit particle sizes below 100 nm. Prior to mixing, the solution primarily contained the above two distinct fusion protein components. A minor peak around 30 nm likely corresponds to unassembled individual fusion enzymes or early-stage simple complexes. As self-assembly progressed, this peak consistently shifted toward larger sizes, indicating the consumption of free components or simple complexes for the formation of larger assemblies. A prominent peak around 1000 nm is attributed to the successfully formed large 3DNECR. With the self-assembly time increasing from 1 h to 6 h, the intensity of this peak increased significantly and monotonically, directly demonstrating that the SpyCatcher/SpyTag system effectively drives the formation of large supramolecular structures through a slow kinetic process lasting several hours. Although the assembly process initiated within the first hour, it continued progressively over 6 h without reaching complete equilibrium, suggesting that the reaction rate may be limited by effective macromolecular collisions or the intrinsic kinetics of isopeptide bond formation. The broad and polydisperse size distribution of the final 3DNECR products indicates limited uniformity in size and morphology, likely due to nonspecific cross-linking or aggregation of multiple Dps nanocages mediated by surface-conjugated enzymes. These larger assemblies may include non-specifically aggregated or over-cross-linked species, which could compromise the structural stability and catalytic efficiency of the reactor. Based on these findings, a self-assembly time of 4 h was selected as the standard condition for subsequent studies.

Fig. 7. Optimization of self-assembly time (A) and concentration (B) for the 3D nano dual-enzyme cascade reactor (3DNECR)

Effect of self-assembly concentration on the 3DNECR

Under a fixed self-assembly duration, the influence of three distinct reactant concentrations (2 µM, 5 µM, and 10 µM) on the assembly products was evaluated by comparing the particle size distributions of the resulting products with those of the individual components, namely SpyCatcher-Dps-ATA117 and SpyTag-ADH (Fig. 7B). Prior to mixing, both SpyCatcher-Dps-ATA117 and SpyTag-ADH at 5 µM displayed size distributions in the 20–100 nm range, each profile consisting of two distinct peaks corresponding to relatively smaller and larger particle sizes, thereby establishing a standard for the unassembled fusion proteins. At a fusion protein concentration of 2 µM, the intensity distribution still exhibited small peaks resembling those of the unassembled proteins, suggesting that a considerable proportion of proteins remained unreacted. A weakly intense peak around 250 nm was also observed, indicative of low 3DNECR formation efficiency, likely resulting from insufficient molecular collision frequency and incomplete assembly under dilute conditions. When the concentration of both fusion proteins was increased to 5 µM, the intensity of the peak corresponding to smaller particles was significantly reduced, reflecting more efficient consumption of free components. Concurrently, a pronounced intensity increase was detected for the peak near 300 nm, demonstrating enhanced assembly efficiency and the generation of more fully assembled 3DNECR complexes. These observations confirm that higher protein concentrations promote 3DNECR formation, consistent with chemical reaction kinetics in which elevated reactant concentrations increase the effective collision frequency between SpyCatcher and SpyTag, thereby accelerating covalent self-assembly. However, further increasing the assembly concentration to 10 µM led to a noticeable rightward shift of the peak corresponding to larger assemblies, implying the onset of nonspecific over-crosslinking at high concentrations. In contrast to the 5 µM system, the small peak in the 10 µM sample shifted toward smaller sizes, which may be attributed to altered molecular interaction kinetics under crowded conditions. Within the 2–5 µM range, increasing concentration contributed positively to assembly efficiency. Therefore, 5 µM is identified as the optimal concentration, offering a balanced combination of high assembly yield and superior product uniformity.

FE-SEM analysis

The morphology and structure of SpyTag-ADH (A), SpyCatcher-Dps-ATA117 (B), and the final assembly product 3DNECR (C) were characterized and compared using FE-SEM. As shown in Fig. 8A, SpyTag-ADH appeared as relatively uniform and fine particulate structures. Their small size and homogeneous distribution indicate that the protein remained well-dispersed without significant aggregation under the sample preparation conditions. Figure 8B reveals that SpyCatcher-Dps-ATA117 formed well-defined, dispersed spherical or near-spherical nanoparticles. This morphology is highly consistent with the known dodecameric cage-like structure of Dps proteins. The distinct spherical shape and uniform size suggest that the native quaternary structure of Dps remained intact after fusion with SpyCatcher and the ATA117 enzyme. In contrast to panels A and B, Fig. 8C shows the emergence of large, three-dimensional network-like or aggregated architectures. These structures, formed through the interconnection and cross-linking of numerous nanoparticles, reached scales up to the micrometer range. This distinct morphological transformation serves as direct evidence for the successful self-assembly among the protein components. The resulting 3DNECR possesses a high specific surface area and a complex porous three-dimensional structure, which is highly conducive to substrate diffusion and enzyme-substrate interactions, structurally explaining its potential efficiency as a cascade reactor. This was also confirmed by the transmission electron microscopy (TEM) image of 3DNECR (Fig. 8D). In summary, these images strongly demonstrate the successful use of the SpyCatcher/SpyTag system to achieve spatially controlled co-localization of ADH and ATA117 enzymes on the Dps nanocage, constructing a well-defined and sophisticated three-dimensional multi-enzyme cascade reactor.

Fig. 8FE-SEM and TEM images.** A** FE-SEM image of SpyTag-ADH.** B** FE-SEM image of SpyCatcher-Dps-ATA117.** C** FE-SEM image of 3D nano dual-enzyme cascade reactor (3DNECR).** D** TEM image of 3DNECR

The reaction kinetic analysis of 3DNECR

To analyze the reaction kinetics of 3DNECR, we compared the catalytic efficiency of the 2DNECR and 3DNECR. The Vmax value reflects the maximum reaction rate when the enzyme is saturated with substrate. As shown in Table 3, the Vmax of 3DNECR is approximately 15% higher than that of 2DNECR. This indicates that under sufficient substrate concentration, the 3DNECR reactor can convert more substrate per unit time, demonstrating a higher catalytic throughput. The Michaelis constant Km is defined as the substrate concentration at half-maximal velocity, and a lower Km corresponds to higher enzyme-substrate affinity. The calculated Km values for both systems are very close, with that of 3DNECR being slightly higher. This suggests that the transition from a 2D to a 3D structure did not significantly alter the intrinsic substrate affinity of the enzymes. Therefore, the enhancement in catalytic efficiency is likely not due to changes in the enzymes themselves but can be attributed to advantages conferred by the three-dimensional nanostructure, which is consistent with the three-dimensional network observed by FE-SEM. The Kcat value, representing the maximum number of substrate molecules converted per enzyme molecule or per active site per unit time, is approximately 34% higher for 3DNECR than for 2DNECR. This indicates a significantly enhanced intrinsic catalytic activity per enzyme molecule within the 3DNECR assembly. The Kcat/Km ratio, considered the gold standard for evaluating catalytic efficiency as it accounts for both catalytic rate and substrate affinity, is about 28% higher for 3DNECR than for 2DNECR. This clearly demonstrates that 3DNECR is a reactor with superior overall catalytic efficiency. Together, these kinetic data strongly support that upgrading the multi-enzyme system from a two-dimensional assembly to a three-dimensional network structure significantly enhances the efficiency of the cascade reaction by optimizing the spatial arrangement of enzymes and facilitating substrate channeling. This provides solid kinetic evidence for 3DNECR as an efficient multi-enzyme catalytic platform.

Table 3. The kinetic parameters of the 2DNECR and 3DNECRRection systemsVmax (µmol/min)Km (mM)Kcat (s^− 1^)Kcat/Km (s^− 1^·mM^− 1^)2DNECR0.49622.78119.4270.8533DNECR0.57223.88625.9791.088Vmax was calculated based on the 4 mL reaction system

Analysis of the enzymatic activities of ATA117, SpyCatcher-ATA117, and SpyCatcher-Dps-ATA117

To investigate the impact of structural modifications on enzyme function, we analyzed the trend in specific activity of ATA117 throughout its stepwise integration into the SpyCatcher/SpyTag system and eventual incorporation into the final 3DNECR (Table 4). The native activity of the free, unmodified ATA117 enzyme served as the baseline reference, set at 100%. Fusion of ATA117 with SpyCatcher resulted in no significant decrease in enzymatic activity compared to the free enzyme. This indicates that the fusion process itself had a minimal impact on the catalytic domain of ATA117, confirming the successful construction of the fusion protein. Subsequent assembly of the fusion protein with the Dps nanocage led to an approximately 15% decrease in activity, which may be attributed to subtle influences from the Dps microenvironment. Notably, the final covalent self-assembly step, forming the complex 3D structure by combining SpyTag-ADH with SpyCatcher-Dps-ATA117, did not cause severe additional impairment to ATA117 activity. This demonstrates the mildness and effectiveness of the employed SpyCatcher/SpyTag system and Dps nanocage as a platform for protein engineering.

Table 4. Enzyme activity assayEnzymesEnzyme activity (U/mg)Relative activity (with ATA117 set as 100%)Yield (%)Enantiomeric excess (ee, %)ATA1170.232 ± 0.062100%55.70 ± 14.8999.6 ± 0.3SpyCatcher-ATA1170.226 ± 0.08397.4%54.26 ± 19.9399.4 ± 0.5SpyCatcher-Dps-ATA1170.198 ± 0.06785.3%47.54 ± 16.0999.7 ± 0.23DNECR0.187 ± 0.04580.6%44.90 ± 10.8199.8 ± 0.1One unit (U) of enzyme activity is defined as the amount of enzyme that catalyzes the conversion of 1 µmol of substrate per minute

Stability analysis

To evaluate the storage, pH, and thermal stability of the 3DNECR, we compared its relative enzyme activity with that of the 2DNECR. For the storage stability assessment, the enzyme activity of the 2D assembly stored in pH 9 buffer for 0 days was defined as 100%. The relative activity under other conditions was calculated as the ratio of the activity measured under those conditions to the activity at day 0 in pH 9 buffer. Figure 9A shows the changes in relative enzyme activity of the 3DNECR and 2DNECR after storage at 4 °C for different durations. Overall, the relative activity of both reactors decreased to varying degrees over time. However, the 3DNECR consistently maintained a higher level of relative activity than the 2DNECR. This can be attributed to the stable and protective microenvironment provided by the 3DNECR architecture. The nanocage structure formed by Dps proteins may effectively reduce enzyme unfolding, denaturation, or inactivating aggregation, thereby significantly slowing the decay of enzymatic activity. These results indicate that enzyme immobilization within a three-dimensional nanostructure helps extend the operational lifetime of biocatalysts.

Fig. 9. Stability of 3D nano dual-enzyme cascade reactor (3DNECR).** A** Storage stability.** B** pH stability.** C** Thermal stability

Analysis of the pH stability revealed that both reactors exhibited maximum activity around pH 9, indicating their optimal performance under mildly alkaline conditions (Fig. 9B). In the range of pH 5–7, the 3DNECR retained significantly higher relative activity than the 2DNECR. As the pH increased further, the difference in relative activity between the two assemblies gradually narrowed, with the 2DNECR showing slightly higher activity than the 3DNECR at pH 9. Nevertheless, at pH 10, the 3DNECR still demonstrated superior residual activity. Taken together, these findings indicate that the three-dimensional structure constructed from Dps nanocages serves not only as a scaffold but also provides a relatively stable internal microenvironment. This structure mitigates the effects of external environmental fluctuations on the enzymatic active sites, thereby maintaining a higher catalytic activity in the 3DNECR over a wider pH range and longer storage period.

In the thermal stability assay, the initial activity of each reactor at each temperature was defined as 100%. After 24 h, the relative activity at each temperature was calculated as a percentage of this initial value. As shown in Fig. 9C, the relative activity of both the 3DNECR and the 2DNECR decreased with increasing temperature. Notably, the 3DNECR exhibited a more gradual decline in relative enzymatic activity. This trend suggests that both reactors are better suited for low‑temperature storage. It was found that over a broad temperature range (25 °C to 40 °C), the 3DNECR consistently maintained a higher level of relative enzymatic activity than the 2DNECR, indicating its superior thermal stability. After 24 h of storage at a low temperature (4 °C), the relative activities of the two reactors showed almost no significant difference. However, as the temperature increased further, a clear disparity in relative activity between them emerged. Notably, at a higher temperature (40 °C), the relative activity of the 3DNECR was approximately 20% higher than that of the 2DNECR. These data provide strong evidence that the three‑dimensional confined space constructed based on the Dps scaffold can effectively enhance the thermal stability of the immobilized enzymes. This finding lays a solid foundation for its potential application in industrial biocatalytic processes that require mild heating or are subject to thermal effects.

Molecular docking

The binding affinity between the SpyCatcher-Dps-ATA117 and SpyTag-ADH proteins was evaluated based on the Prodigy docking score. Five models were generated through de novo prediction using the Chai web server, and the affinity scores for each model are listed in Table S1. All computed binding affinity values were negative, indicating favorable binding. More negative values correspond to stronger binding. The binding energy between SpyCatcher-Dps-ATA117 and SpyTag-ADH was − 17.3 kcal/mol, reflecting a strong binding effect(Table 5). Their binding mode and interactions were further analyzed (Fig. 10A). According to the molecular docking and interaction analysis, multiple stable interfacial interactions were formed between SpyCatcher-Dps-ATA117 and SpyTag-ADH, demonstrating strong molecular recognition and stable binding. In terms of hydrogen bonding, several residues on SpyCatcher-Dps-ATA117 (K31, S33, R35, E37, D36, D38, D85, G86, E88, Y87, and R580) formed multiple hydrogen bonds with corresponding residues on SpyTag-ADH (A1, I3, M5, H2, Y9, K10, D112, K109, D332, Y113, and Q167). The bond distances ranged from 1.7 to 2.2 Å, falling within the typical range of strong hydrogen bonds and indicating tight interfacial complementarity. Additionally, several salt bridges were observed, such as those between SpyCatcher-Dps-ATA117 residues R580, K40, D36, D38, and D85 and SpyTag-ADH residues D332, D112, K109, and K10. These charge-based interactions further enhanced the electrostatic stability of the complex. Most critically, a covalent amide bond was formed between K34 of SpyCatcher-Dps-ATA117 and D7 of SpyTag-ADH, enabling covalent fusion and resulting in a stable fusion protein system. The characteristic covalent linkage of the SpyCatcher/SpyTag system ensures high stability and irreversible conjugation under physiological conditions, laying a molecular foundation for subsequent functional assembly and applications.

Table 5. Binding energies from molecular dockingSystemPredicted binding affinity (kcal/mol)SpyCatcher-Dps-ATA117/SpyTag-ADH (3DNECR)− 17.33DNECR/BPO− 7.4

Fig. 10. Schematic diagram of the molecular docking.** A** Binding mode of the SpyTag-ADH/SpyCatcher-Dps-ATA117 complex. SpyTag-ADH is shown in pink and the SpyCatcher-Dps-ATA117 protein is shown in blue. Hydrogen bonds are indicated by yellow dashed lines.** B** Binding mode of 3D nano dual-enzyme cascade reactor (3DNECR) with 3,5-bis(trifluoromethyl)acetophenone. Protein residues are depicted in blue-grey sticks. The R molecule is shown in yellow sticks. Hydrogen bonds are represented by solid blue lines. Hydrophobic interactions are indicated by grey dashed lines. π-π stacking interactions are shown as green dashed lines

The binding energy between the 3DNECR and BPO was evaluated using AutoDock Vina. In AutoDock Vina, more negative docking scores indicate higher binding affinity. A docking energy of -5 kcal/mol is generally considered to represent a moderate to high likelihood of binding. As shown in Table 5, the calculated binding energy for the 3DNECR/BPO complex was − 7.4 kcal/mol, indicating a strong potential for spontaneous binding between the ligand and the protein. Furthermore, we analyzed the specific interactions between 3DNECR and BPO. Figure 10B reveals that residues A277 and F247 of 3DNECR form hydrophobic interactions with the BPO molecule. Hydrophobic interactions arise from the tendency of non-polar molecules or groups to associate in an aqueous environment to minimize their contact with water. This association serves as a major driving force that contributes significantly to binding stability. Additionally, a hydrogen bond is formed between S272 and BPO. Hydrogen bonds are highly directional and specific interactions, which are crucial for precisely positioning and anchoring the substrate within the active site of 3DNECR. Moreover, a π-π stacking interaction occurs between the benzyl side chain of F247 and the benzene ring of BPO. This type of interaction, which takes place between two aromatic ring systems, can greatly enhance the stability of the complex and is a key stabilizing force in the binding of aromatic substrates. The presence of these diverse interactions provides a structural explanation for the high catalytic efficiency observed in 3DNECR.

Effect of reaction conditions on the amination catalyzed by 3DNECR

As shown in Fig. 11A and Table S2, both the 3DNECR and 2DNECR achieved high yields (≥ 99.9%) at substrate concentrations of 200 and 210 mM, while their ee values were also comparable at these levels. As the substrate concentration increased further, a noticeable gap in yield emerged between the two systems, with the 3DNECR consistently exhibiting higher yields than the 2DNECR. Overall, the yield of the 3DNECR remained above 95%, whereas that of the 2DNECR varied between 86% and 100%. In terms of enantioselectivity, both systems maintained high and stable ee values, indicating that each was capable of producing the target product with high optical purity. These results clearly demonstrate that 3DNECR has a higher catalytic efficiency than 2DNECR, with an optimal substrate concentration of 230 mM. This advantage becomes particularly evident when compared to conventional free-enzyme or immobilized enzyme systems documented in the literature (Lu et al. 2022). Traditional approaches are often constrained by limitations such as mass transfer inefficiencies and inadequate operational stability. In contrast, the spatially controlled self-assembly strategy employed in this work effectively addresses these challenges, offering a novel pathway toward efficient cascade catalysis.

Fig. 11. Effect of substrate concentration (A) and pH (B) on the transamination reaction

To evaluate the effect of buffer pH on the amination reaction catalyzed by the 3DNECR, both the yield and ee of the reaction system were measured at different pH levels (Fig. 11B and Table S3). It was observed that the yield of the 3DNECR was consistently and significantly higher than that of the 2DNECR across the entire pH range tested. The yields of both reactors increased sharply between pH 6 and 8, reached a maximum at pH 9, and then plateaued. The 3DNECR maintained a high yield (≥ 94%) over a broad pH range from 8 to 10, achieving a near-quantitative yield (~ 100%) at pH 9, indicating its exceptional and stable catalytic efficiency under alkaline conditions. Furthermore, the ee of the 3DNECR remained consistently high under all pH conditions and was stably superior to that of the 2DNECR, demonstrating that the 3DNECR not only exhibits high efficiency but also consistently produces the product with excellent optical purity. Therefore, the 3DNECR outperforms the 2DNECR in both catalytic efficiency and enantioselectivity, particularly within its optimal pH range of 8–10.

Reusability of 3DNECR

To evaluate the operational stability of the 3DNECR upon repeated use, we measured the relative enzyme activity after six consecutive reaction cycles, with the relative activity of the 2DNECR over the same number of reuses serving as a control. As shown in Fig. 12, the relative activity of both reactors decreased with increasing number of reuses, which is commonly observed for immobilized enzymes and can be attributed to partial inactivation of enzyme active sites and enzyme loss during recycling. Throughout the test cycles, the relative activity of the 3DNECR remained consistently slightly higher than that of the 2DNECR, indicating superior operational stability. This can likely be ascribed to the stable microenvironment provided by the Dps protein cage, which helps shield the enzymes from structural damage caused by external factors. Moreover, compared with the 2DNECR, the 3DNECR exhibited a more gradual decline in relative activity, demonstrating better reusability. This result may be explained by the stable anchoring of enzymes to the Dps nanocage via covalent bonds formed by the SpyCatcher/SpyTag system, which effectively reduces enzyme leaching during reuse and recovery.

Fig. 12. Effect of number of reuses on relative enzyme activity

Conclusion

This study adopts a green approach for the asymmetric synthesis of R-BPA by using a biocatalytic enzyme cascade reactor termed 3DNECR in an oil-water biphasic system under mild temperature and pH conditions. This strategy enables high selectivity and recyclability while reducing hazardous reagents and waste, thereby applying green engineering principles of benign solvents and conditions, energy efficiency, waste prevention, and design for reuse and efficiency (Fegade 2015, 2024b). The 3DNECR was successfully constructed through the rational integration of the SpyCatcher/SpyTag covalent conjugation system with the Dps protein nanocage scaffold. The foundational design of the SpyCatcher-Dps-ATA117 fusion protein was informed by robust in silico analyses. The absence of a signal peptide and a balanced amphiphilic amino acid composition correctly predicted its successful soluble expression in the cytoplasm of E. coli BL21(DE3). Subsequent experimental validation confirmed the high-yield expression and purification of both SpyTag-ADH and SpyCatcher-Dps-ATA117 building blocks. The self-assembly process, driven by the specific covalent isopeptide bond formation between SpyTag and SpyCatcher, was rigorously optimized. DLS analysis revealed that the assembly is a concentration-dependent and time-evolving process, with optimal conditions identified at 5 µM reactant concentration and a 4 h incubation period. This process efficiently consumed free protein components to form large, supramolecular assemblies. FE-SEM provided direct visual evidence of the successful construction, showing a distinct morphological transition from discrete, spherical SpyCatcher-Dps-ATA117 nanoparticles to a complex, interconnected three-dimensional network architecture for the 3DNECR. This unique structure inherently provides a high specific surface area and a porous matrix, which are crucial for substrate diffusion and enzyme-substrate interactions.

The catalytic superiority of the 3DNECR was unequivocally demonstrated through detailed enzyme kinetics. The significant increases in Vmax and Kcat for the 3DNECR, compared to the 2DNECR, indicate a higher maximum reaction rate and enhanced turnover number per enzyme molecule. Most importantly, the catalytic efficiency, represented by the Kcat/Km ratio, was approximately 28% higher for the 3DNECR. This enhancement is not attributable to a change in intrinsic substrate affinity but is a direct consequence of the advantageous spatial organization within the 3D nanostructure. The close proximity of ADH and ATA117 enzymes facilitates substrate channeling, minimizing the diffusion loss of intermediates and thereby accelerating the overall cascade reaction rate.

In practical application, the 3DNECR consistently achieves high product yields across varying substrate concentrations while maintaining excellent enantioselectivity, confirming that the 3D architecture enhances both productivity and stereochemical precision. This work establishes an effective strategy for constructing advanced multi-enzyme systems by integrating protein nanocages with peptide-mediated assembly. The 3DNECR platform demonstrates compelling potential for applications requiring efficient, stable, and stereoselective biocatalysis. By exchanging the enzyme components fused to the SpyCatcher/SpyTag system, the reactor can be tailored for various application scenarios. For continuous-flow production, the 3DNECR can be retained by a membrane, enabling continuous substrate feeding and simultaneous product separation—a configuration particularly suitable for synthesizing high-value chemicals such as the chiral pharmaceutical intermediates described in this study. Moreover, microreactor technology offers an ideal platform for numbering-up these nano‑reactor units, thereby achieving high-throughput and well‑controlled scaled production. Future work will focus on evaluating its performance in complex media and advancing toward industrial‑scale implementation.

Supplementary Information

Below is the link to the electronic supplementary material.

Supplementary Material 1.