To achieve dynamic regulation of the cadaverine synthesis pathway, we constructed a Temperature-responsive system. The temperature-responsive repressor protein TlpA39 is expressed under the control of the constitutive promoter J23100. At temperatures below 39°C, TlpA39 forms a dimer that can bind to the promoter PtlpA, thereby repressing the expression of downstream genes; at temperatures above 39°C, TlpA39 exists as a monomer, loses the ability to bind to PtlpA, and thus activates the expression of downstream genes.

We placed the reporter gene of mWasabi (a green fluorescent protein) and the lysine decarboxylase gene cadA downstream of the PtlpA promoter, respectively. The recombinant plasmids pET28a(ΔI7)-PJ23100-RBS-TlpA39-PtlpA-mWasabi and pET28a(ΔI7)-PJ23100-RBS-TlpA39-PtlpA-CadA were respectively transformed into Escherichia coli NT1003. The two engineered strains were cultured aerobically at 33, 35, 37, 39, and 42°C, respectively. By determining fluorescence intensity, the mWasabi reporter strain showed low fluorescence at ≤37°C but strong fluorescence at ≥39°C (Figure 1-1A). For the CadA group, lysine accumulated to 4.27 g/L, 2.6 g/L, and 2.4 g/L at 33°C, 35°C, and 37°C, respectively (Figure 1-1 B, C, D), but was not detected at 39°C and 42°C (Figure 1-1 E, F). The results indicated the temperature-sensitive switch operated effectively. Low fluorescence with lysine accumulation at ≤37°C confirmed the "OFF" state, while high fluorescence with no lysine accumulation at ≥39°C confirmed the "ON" state. It defined a clear activation threshold near 39°C, successfully enabling the induction of downstream gene through a simple temperature shift.

To achieve dynamic regulation of lysine decarboxylase (CadA), we attempted to construct a protein encapsulation system based on artificial intrinsically disordered proteins (A-IDPs). We designed A-IDPs of different sizes according to literature, including 0.5A-IDP, 1.0A-IDP, and 2.0A-IDP, and inserted TEV/TVMV cleavage sites between CadA and A-IDPs to achieve controllable protein release. To investigated the encapsulation function in E. coli NT1003, we applied this encapsulation system to mWasabi (a green fluorescent protein) and lysine decarboxylase, respectively.

Although the fluorescence results indicated that the 1.0A-IDP group was lower intensity than the 0.5A-IDP group, both A-IDP groups did not accumulate lysine (Figure 1-2). The current protein encapsulation system failed to prevent lysine consumption by CadA in both 0.5A-IDP and 1.0A-IDP, suggesting insufficient enzyme encapsulation. Further increasing the size to 2.0A-IDP, lysine still did not accumulate (Figure 1-3).

Thus, it suggested that 2.0A-IDP failed to perfectly encapsulate the CadA protein to block its activity during the aerobic phase. Therefore, we decided to adopt the temperature-sensitive switch in subsequent experiments.

To optimize the stringency and sensitivity of the temperature-sensitive switch for more accurate control of CadA expression, promoters with different strengths (PJ23100, PJ23106, and PJ23119) were used to regulate the expression of TlpA39. The three engineered strains harboring different promoters were subjected to aerobic culture at 33°C for 23 h, then shifted to anaerobic culture at 42°C. The titers of lysine and glucose during the aerobic and anaerobic phases were measured (Figure 2-1). Both PJ23106 and PJ23119 led to more lysine accumulation compared with PJ23100.Among them, PJ23106 showed the most significant effect, accumulating 5.1 g/L during the aerobic phase. Subsequently, during the high-temperature anaerobic phase, lysine was gradually completely consumed..

This demonstrated that replacing the promoter with PJ23106 could effectively reduce CadA leaky expression, and that anaerobic high temperature allows normal expression of CadA to catalyze lysine consumption.

RNA thermometers U7 and U8 can cause RBS (ribosome binding site) and a segment of sequence to form a stem-loop structure under low-temperature conditions, preventing ribosomes from binding normally to RBS and thus inhibiting the translation of downstream proteins. At high temperatures, they unlock the stem-loop structure, allowing normal translation of downstream proteins. Therefore, we designed to link them into the PJ23106-RBS-TlpA39-PtlpA-CadA construct as a translational-level regulation. The original temperature-sensitive switch, combined with RNA thermometer, formed a dual switch that is controlled at transcriptional and translational levels via temperature.

The different recombinant plasmids pET28a(ΔI7)-PJ23106-RBS-TlpA39-PtlpA-U7-CadA and pET28a(ΔI7)-PJ23106-RBS-TlpA39-PtlpA-U8-CadA were separately transformed into E. coli NT1003 to evaluate the effect of RNA thermometers. The strains were aerobically cultured at 33°C, 35°C, and 37°C for 23 hours, then transferred to an anaerobic environment and cultured at an elevated temperature of 42°C. As shown in Figures 3-1, the engineered E. coli with the dual-switch exhibited higher lysine accumulation at temperatures of 33°C, 35°C, and 37°C. It indicated that lysine was hardly consumed during the aerobic stage. Moreover, after increasing the temperature, lysine began to be gradually consumed, producing cadaverine simultaneously conducting anaerobic fermentation to synthesize succinate. These demonstrated that the dual switchs inserting RNA thermometers can achieve more precise temperature-responsive regulation. In addition, it has been initially proven that we can achieve the co-production of cadaverine and succinate through its regulation.

Lysine decarboxylase (CadA) is a key enzyme for cadaverine synthesis. To improve its catalytic efficiency, we performed multiple rounds of engineering modification on CadA.

We first employed a computationally guided strategy to perform molecular docking simulations using the structure of CadA (PDB No. 6YN6). Hydrogen bonds, salt bridges, and hydrophobic interactions between the enzyme and its substrate were systematically analyzed. Based on the interaction analysis, we identified some residues located near the active pocket that might make key contributions to substrate binding and catalysis: I182, H245, K246, W333, E526, K527 and Y652. These candidate sites were subsequently evaluated using a site-directed mutagenesis prediction model for virtual amino acid substitutions, and their catalytic efficiencies (K_cat) were predicted. The model recommended several pentuple mutants with predicted high enzymatic activity (Table1).

We successfully constructed plasmids encoding the five predicted quintuple mutants. However, whole-cell catalytic experiments showed that all of these mutants completely lost enzymatic activity, with no detectable consumption of the substrate lysine. It was likely due to structural alterations that disrupt the active site and/or cofactor binding.

Since multi-site mutations resulted in a complete loss of enzymatic activity, we conducted a systematic single-site mutation analysis. Based on autodock result anlysis, the key sites (I182, H245, K246, W333, E526, K527 and Y652) were identified to assess their individual functional importance(Figure4-1). We first performed alanine-scanning, followed by further substitutions at positions that exhibited less severe activity loss (I182, H245, K246) with amino acids of diverse physicochemical properties (E, L, S, F, V) to explore their functional roles.

We successfully constructed plasmids for most of the single-site mutants. Enzymatic activity assay results showed that mutations W333A, E526A, K527A, and Y652A led to an almost complete loss of enzymatic activity, indicating that these residues are directly involved in substrate binding or the catalytic active site, and are functionally essential (Figure 4-2(A)). In contrast, alanine substitutions and other mutations at positions I182, H245, and K246 did not significantly enhance activity, but several mutants retained partial activity. These findings provide important clues for the catalytic mechanism of CadA and the functions of its residues (Figure 4-2(B)).

Apart from the aforementioned key amino acid residues, we shifted our modification strategy to enhancing the stability of CadA's decameric structure, as its decameric form is essential for efficiently exerting decarboxylation function. Through computer simulation analysis combined with literature reports, we predicted several pairs of double-site mutations (F14C/K44C, V12C/K41C, F14C/L45C, L93C/E445C, Y13C/P36C, E104K) or single-site mutations (E104K) that may stabilize the decameric structure by forming disulfide bonds or other interactions. We also referenced mutation pairs from the literature (V12C/K44C, F14C/D41C, L89C/E445C).

Whole-cell catalytic experiments showed that multiple mutants exhibited higher enzymatic activity than WT. Among them, the mutant Y13C/P36C showed the most significant effect, with its relative enzymatic activity increasing by approximately 113.26% compared with the wild-type. In addition, the enzymatic activities of mutants V12C/D41C, F14C/D41C, and L93C/E445C also increased to varying degrees compared with the wild-type. These results indicate that stabilizing the decameric structure of CadA through strategies such as introducing disulfide bonds is an effective approach to improve its catalytic efficiency (Figure4-3).

To further enhance enzymatic activity, we aimed to perform iterative mutations. First, we selected the two best-performing mutants from the previous round, Y13C/P36C and L93C/E445C, for iteration, constructing the mutant Y13C/P36C/L93C/E445C. Considering that multiple disulfide bonds might lead to excessively poor structural flexibility, we speculated a combination of one set of disulfide bonds and a single-site mutation more favorable. Since our current experiments did not screen out single-site mutations with improved enzymatic activity, we selected literature-reported mutations E445Q and K477R, constructing the mutants Y13C/P36C/E445Q and Y13C/P36C/K477R.

Whole-cell catalytic experiments showed that the effect of iterative mutations was not ideal. The mutant Y13C/P36C/E445Q exhibited a significant decrease in enzymatic activity, and the enzymatic activities of the other mutants were also significantly lower than that of the best mutant Y13C/P36C (Figure 4-4).

To comprehensively evaluate the catalytic performance of the Y13C/P36C mutant, we compared its catalytic efficiency with that of the wild-type under different substrate concentrations (20 g/L, 50 g/L, 100 g/L, 200 g/L) and different pH conditions (6, 7, 8, 9).

The results showed that CadA(Y13C/P36C) exhibited superior catalytic performance to the wild-type under all tested conditions, with the advantage being more pronounced especially at higher substrate concentrations. In contrast to the wild-type, Y13C/P36C mutant demonstrated superior lysine consumption kinetics across a substrate concentration of 20-200 g/L. More significantly, Y13C/P36C mutant maintained persistent catalytic activity at high substrate. This indicates that the mutant has better industrial application potential. These results indicate that CadA(Y13C/P36C) efficiently improved the substrate tolerance.

Under pH conditions from 6-9, Y13C/P36C mutant exhibited superior catalytic performance. Its relative activities were respectively 1.59, 2.01and 2.07-fold higher than that of WT at pH 6, 7 and 8. Furthermore, the relative activity of WT at pH 9 was 37.07 %, while the relative activities of Y13C/P36C mutant was 74.14% . This indicated that CadA(Y13C/P36C) efficiently improved the pH tolerance and has better application potential.

For our other co-produced product, succinate, based on the metabolic flux analysis, we chose to overexpress glyceraldehyde-3-phosphate dehydrogenase (GAP) to enhance the supply of reducing power. Succinate is a reductive product that requires the consumption of excess reducing power during its synthesis. To boost succinate synthesis, we screened four NADP⁺-dependent glyceraldehyde-3-phosphate dehydrogenases (GAP) originated from Corynebacterium glutamicum ATCC 13032 (CggapC), Bacillus subtilis (BsrocG), Clostridium saccharobutylicum DSM 13864 (CsgapC), and Streptococcus mutans (SmgapN). These enzymes catalyze the conversion of D-glyceraldehyde 3-phosphate to 3-phospho-D-glycerate and simultaneously generate NADPH, thereby increasing intracellular reducing power.

The aforementioned strains were cultured in a fermentation medium supplemented with 20 g/L glucose, and the titers of lysine and succinate were measured. As shown in Figure 5-2 (B),the results showed that the expression of CggapC could effectively increase the titer of succinate by 24.5%. However, we also found that it exerted an negative effect on lysine during the aerobic phase, with a 40% decrease compared with the control. It might be due to excessive reducing power led to an imbalance in intracellular redox levels and imposed the metabolic burden on the cells.

To mitigate the effects of excess reducing power on cellular metabolism and lysine accumulation during the aerobic growth phase, the expression of glyceraldehyde-3-phosphate dehydrogenase was further optimized using a low-copy-number plasmid pACYCDuet.

The results showed that the expression of four dehydrogenase using a low-copy-number plasmid led to an increase in succinate. Among them, CggapC still exhibited the highest succinate titer, and no decrease in lysine titer (Figure 5-3). Therefore, the CggapC was selected in subsequent work.

To further enhance succinate production and maintain cell vitality, cofactor distribution and flow were rationally engineered. For the NADH required in the succinate synthesis, the soluble pyridine nucleotide transhydrogenase (sthA) from E. coli MG1655 catalyzes the conversion of NADPH to NADH, which was selected to balance different types of cofactor. Co-expression of sthA with CggapC enabled optimization of the intracellular cofactor pool, thereby improving reducing power allocation. CggapC facilitated additional NADPH synthesis, thereby providing the necessary reducing power and then SthA catalyzed the conversion of NADPH to NADH for succinate production. The results showed that the succinate titer of the strain harboring pACYCDuet-CggapC-sthA increased by 28% relative to control; meanwhile, the lysine titer reached 14.9 g/L. These results indicated that proper regulation of cofactors could effectively enhance succinate synthesis.

Through preliminary experiments, we integrated the thermo-responsive switch to control the expression of CadA and the cofactor regulation system. A single-cell factory was constructed to simultaneously produce cadaverine and succinate. According to the fermentation results, the recombinant strain successfully achieved the co-synthesis of cadaverine and succinate, with the highest titer reaching 5.7 g/L and 5.4 g/L from glucose. It laid a robust foundation for the efficient and green production of the PA54.