Cell immobilization and characterization
The scanning electron microscopy (SEM) micrographs in Fig. 2a demonstrate that E. coli particles exhibit a rod-like morphology with smooth surfaces, measuring between 0.5 μm and 2 μm in length. Fig. 2b present SEM images of DE/PEI/GLU-immobilized E. coli, where the immobilized E. coli cells exhibit uniform particle sizes, characterized by membrane-coated surfaces and tightly interconnected cellular junctions, with no noticeable structural damage. This observation suggests that following diatomite adsorption, glutaraldehyde crosslinked with PEI, leading to the changes in cell structure.
FT-IR analysis was performed to confirm the successful immobilization of diatomite and cells using GLU as a crosslinking agent. Fig. 3 illustrates the spectra of free cells, diatomite, and diatomite-immobilized cells. Upon cell immobilization, several distinct peaks emerged in the spectrum, differentiating it from those of E. coli and diatomite alone. Specifically, peaks at 792, 617, 471, and 1090 cm-1 were attributed to Si-O stretching vibrations, Si-O-Si symmetric stretching vibrations, Si-O bending vibrations, and the binding vibration of calcite to CO₃2- in diatomite, respectively. Furthermore the peak at 3305 cm-1 corresponded to the stretching vibrations of -OH and -NH, indicating the presence of these functional groups (Hu et al. 2017, Benayache et al. 2018) . Additionally, the peaks at 1656 cm-1 and 1537 cm-1 represented the bending vibrations within the N-H plane of primary and secondary amines.
Upon treatment with GLU and PEI as cross-linking agent, the stretching vibrations of the C-H bond at 2927 cm-1, bending vibrational peak of CH2 at 1393 cm-1, in-plane rocking absorption peak for CH2 in diatomite immobilized cells at 1230 cm-1, which can be attributed to PEI and GLU (Hu et al. 2017) . Additionally, a new stretch at 1453 cm−1 can be assigned to stretching vibration absorption peaks of -C=N-. These peaks in the spectra (1230, 1393, 1453 and 2937 cm-1) indicate the successful immobilization of diatomite and cells using GLU as a cross-linking agent.
Optimization of parameters for cell immobilization by diatomite
The use of diatomite as an immobilization matrix presents several advantages, including its high resistance to dissolution and exceptional adsorption capacity. However, relying solely on its adsorption properties for cell immobilization can result in challenges such as cell dislodgement. To improve the stability of immobilized cells, we investigated the efficacy of various cross-linking agents in forming covalent bonds with the cells. The effects of different types of cross-linker on enzyme activity recovery is presented in Table 1, while their impact on reusability is illustrated in Fig. 4a.
The enzyme activity recoveries of the immobilized cells were found to be similar across all four cross-linkers. However, when polyvinyl alcohol and p-phenylenedial were employed as cross-linking agents, the reusability of the immobilized cells was considerably poor, with yields of β-alanine amounting to only 25.15 % and 17.55 % after three recycling cycles, respectively. In contrast, the use of GLU as a cross-linking agent resulted in a yield of 75.88 % after six recycling cycles. This enhanced reusability can be attributed to the utilization of free amino groups by GLU for cross-linking, which minimizes the impact on the activity. Consequently, GLU was selected as the preferred cross-linking agent for subsequent experiments.
Table 1Effect of crosslinking agents on enzyme activity.

PEI serves as a flocculant for the immobilization of cells onto diatomite. The primary amine groups in PEI can cross-link with glutaraldehyde, resulting in the formation of a stable mesh structure. Variation in the relative molecular mass of PEI lead to changes in its viscosity and the proportion of secondary to tertiary amines, thereby influencing the immobilization effectiveness. Consequently, we conducted an investigation on the impact of PEI molecular weight on the immobilization effect, as illustrated in Fig. 4b.The viability of immobilized cells exhibited a gradual increase with the molecular weight of PEI up to 10 kDa, followed by a decline upon reaching 70 kDa. This phenomenon can be attributed to the increased viscosity and greater branching of the polymer when the molecular weight of PEI reaches 10 kDa. Consequently, a denser polymer film forms on the cell surface, leading to an enhanced effective specific surface area during interactions with the cross-linking agent, thereby improving the efficiency of immobilization (Zhao et al. 2021) . An increase in the molecular weight of PEI to 70 kDa may result in a reduction in the proportion of secondary amine groups within the amine group, potentially decreasing the efficiency of immobilization. Ultimately, PEI with a molecular mass of 10 kDa was chosen as the optimal flocculant.
Diatomite serves as a carrier for immobilized cells, and its particle size significantly influences the effectiveness of immobilization. Fig. 4c demonstrates that the viability of immobilized cells initially increases and then decreases as the diatomite particle size increases. This phenomenon can be attributed to the enhancement of the pore structure, total pore volume, and micropore volume of diatomite with increasing particle size. As a result, both the loading capacity and loading efficiency of cells are improved. Consequently, diatomite with a particle size of 36.8 μm was chosen as the optimal immobilization carrier.
A quantity of 0.3 g of diatomite was measured and added to varying masses of recombinant E. coli cells, aiming to investigate the impact of cell loading on the viability of immobilized cells, as depicted in Fig. 4d. The viability of immobilized cells exhibited an increase as the cell loading was raised to 5 g, but declined beyond this threshold. This phenomenon can be attributed to the accumulation of a high concentration of bacteria within the immobilized cells, which results in the blockage of channels and a reduced probability of substrate-cell contact. Consequently, the viability of the immobilized cells decreases. Therefore, a cell loading of 5 g was selected.
The optimal volume fraction of PEI for immobilization was determined, as shown in Fig. 4e. The viability of immobilized cells exhibited a gradual increase with an increase as the PEI volume fraction continues to increase, the viability of immobilized cells began to decline. At low concentrations of PEI, the structure formed by the cross-linking of GLU and cells is loose, making the cells prone to detachment (Chen et al. 2008) . Conversely, as the concentration of PEI in the system increases, the molecules of PEI cross-link, resulting in heightened mass transfer resistance and a subsequent decrease in the viability of the immobilized cells (HOU et al. 2016) . Consequently, a volume fraction of 8 % PEI was chosen as the optimal flocculant.
The selection of flocculation time for PEI is illustrated in Fig. 4f. Initially, the viability of immobilized cells increased with longer PEI flocculation time increased, peaking at 2.5 h. However, further increase in flocculation time resulted in a decrease in cell viability. This phenomenon may be attributed to the tightening of the crosslinked structure of the immobilized cells as the flocculation time is extended, which leads to excessive mass transfer resistance and ultimately reducing cell viability. Therefore, a flocculation time of 2.5 h was determined to be optimal.
When immobilizing cells using cross-linking, the choice of crosslinker volume fraction significantly and irreversibly influences cell structure. Therefore, optimal immobilization is often achieved through precise control of the concentration and duration. The efffect of GLU on the viability of immobilized cells is depicted in Fig. 4g, revealing a positive correlation between GLU concentrations and cell viability. Optimal viability was observed at a GLU volume fraction of 5 %. This maximum viability is attributed to the toxicity of GLU to recombinant E. coli; excessive GLU concentration can reduce the viability of immobilized cells. Consequently, a GLU the concentration of 5 % was selected.
The impact of GLU crosslinking duration on the activity of immobilized is illustrated in Fig. 4h. The viability of the immobilized cells increased as the crosslinking time extended, peaking at 1.5 h (The specific activity was 310.73 U/g, with an activity recovery yield of 70.71%.). This enhancement is attributed to the low molecular weight of GLU and the limited availability of aldehyde groups for crosslinking with PEI, a process that is completed within 1.5 h. Beyond this duration, structural changes in the immobilized cells remained minimal, but any residual GLU in the system began to affect viability, resulting in a decline in cell activity. Therefore, a final crosslinking time of 1.5 h for GLU was established.
Properties of immobilized cells
In biocatalysis, the rate of catalytic reactions typically increases with increasing temperatures. However, excessively high temperatures can lead to a decline in cellular activity. The impact of temperature on the specific activity of both immobilized and free cells is illustrated in Fig. 5a. Viability for both types of cells showed an upward trend with increasing temperature, peaking at 50 ℃. Beyond this point, viability for both immobilized and free cells began to decrease. The optimal temperature for cell activity, whether immobilized or free, was determined to be 50 ℃.
The effect of temperature on the stability of free and immobilized cells is depicted in Fig. 5b. Following a 12-hour incubation at 45 ℃ , the activities of immobilized cells and free cells were 90.15 % and 82.50 % of their initial activities, respectively, indicating minimal activity loss and good stability. After the same duration at 50 ℃ , the activities decreased to 81.29 % and 65.59 % for immobilized and free cells, respectively. Subsequent incubation at 55 ℃ resulted in further activity decline, with immobilized and free cells reaching 26.60 % and 12.56 % of their initial activities. Immobilization enhanced thermal stability, with diatomite providing additional cellular protection. Optimal stability for both cell types was observed at 45 ℃ , which was consequently chosen as the preferred reaction temperature.
The effect of temperature on the stability of free and immobilized cells is depicted in Fig. 5b. After a 12 h incubation at 45 ℃ , the activities of immobilized cells and free cells were 90.15 % cells achieving 26.60 % and 12.56 % of their initial activities. Immobilization enhanced thermal stability, with diatomite providing and 82.50 % of their initial activities, respectively, indicating minimal activity loss and good stability. Following the same duration at 50 ℃ , the activities decreased to 81.29 % and 65.59 % for immobilized and free cells, respectively. Subsequent incubation at 55℃ resulted in a further decline in activity, with immobilized and free cells additional cellular protection. Optimal stability for both cell types was observed at 45 ℃ , which was subsequently chosen as the preferred reaction temperature.
In biocatalytic reactions, variations in the pH of the reaction solution lead to the dissociation of specific groups in aspartate lyase to varying degrees, which affects their substrate binding capacities and results in differing activities at different pH levels. The impact of pH on the specific activity of both immobilized and free cells is illustrated in Fig. 5c. Viability for both types of cells exhibited a gradual increase with rising pH. The peak viability for immobilized cells was observed at pH 9.5, followed by a decline with further increases in pH. Conversely, free cell activity reached its maximum at pH 10.0, decreasing beyond this point. The optimal pH was determined to be 9.5 for immobilized cells and 10.0 for free cells.
The effect of pH on cell stability before and after immobilization is shown in Fig. 5d. Following a 12 h storage in Tris-HCl buffer at pH 8.0, both immobilized and free cells exhibited higher viability compared to those stored in other buffers, with free cells demonstrating superior pH stability. Immobilized cells displayed robust stability within the pH range of 8.0 to 9.5, retaining over 85 % of their activity. Overall, pH stability increased following immobilization likely due to the protective cross-linking network formed by PEI and GLU, which shielded the cells and minimized environmental influences.
Kinetic Properties
Investigating kinetic parameters is essential for assessing the efficiency of an immobilization process. The kinetics of the catalytic reaction were analyzed, and the Michaelis-Menten parameters were determined using a Lineweaver-Burk plot. Fig. 6 depicts the resultant fitted curves, revealing a Km of 636.93 mmol/L and Vmax of 99.66 U/g for free cells, and a Km of 748.13 mmol/L and Vmax of 43.55 U/g for immobilized cells. The Km value for immobilized cells increased by a factor of 1.17, primarily due to the heightened mass transfer resistance induced by the mesh structure within the immobilized cells.
Storage stability and reusability
Immobilized and free cells were stored at 4 °C for 60 days to evaluate changes in activity (Fig. 7a). Immobilized cells remained above 93 % for the first 30 days, demonstrating a 64 % higher residual viability compared to free cells. By day 60, immobilized cells retained a viability of 72.29 %, while free cells only reached 14.26 %. Immobilized cells exhibit significantly enhanced storage stability, effectively preserving cell activity during storage.
Systematic evaluation of biocatalyst recyclability revealed that the immobilized cell system maintained excellent operational stability during repeated-batch β-alanine synthesis in Fig. 7b. The initial 6-hour conversion cycle achieved near-quantitative yield (96.6 ± 1.2%), with subsequent cycles showing minimal biocatalyst deactivation (0.89 ± 0.15% per cycle), ultimately retaining 80.06% production yield and 86.26% initial activity after 10 consecutive batches. These results demonstrate exceptional sustainability for potential industrial implementation. In contrast, batches using free cells achieved only around a 10 % yield after four consecutive reactions. Diatomite-immobilized cells demonstrate superior reusability, offering promising prospects for industrial applications.
β-alanine synthesis in fixed bed bioreactor
To achieve continuous production, the immobilized cells were placed in a fixed bed reactor. This method eliminates the need for mechanical stirring and prevents direct contact between the enzyme and acidoid, thereby ensuring the overall stability of the whole cell. The effects of acrylic acid concentration and volumetric flow rate on the continuous production yield of β-alanine were systematically evaluated in a fixed-bed bioreactor packed with immobilized E. coli cells. The experimental results are shown in Fig. 8a. When the concentration of acrylic acid is 2 mol/L, the optimal flow rate is 0.03 mL/min, resulting in a space-time yield of 193.08 g·L-1·h-1. At acrylic acid concentrations of 1.5 mol/L and 2 mol/L, the space-time yield remians the same at the optimal flow rate. Higher concentrations of acrylic acid and aqueous ammonia can cause damage to the cells. Therefore, the subsequent synthesis of β-alanine was carried out in a fixed-bed reactor under the conditions of 1.5 mol/L acrylic acid concentration and a flow rate of 0.04 mL/min. The ability of immobilized cells to continuously synthesize β-alanine was investigated in the fixed-bed reactor under optimal conditions for 360 h, as shown in Fig. 8b. In the fixed-bed reactor, the immobilized cells demonstrated stable synthesis of β-alanine, achieving a yield of over 71.30 % after 360 h of continuous reaction.