Construction of L-proline chassis strain
To enable the engineered strain of E. coli to de novo synthesize T-4-HYP from glucose, significant accumulation of the precursor proline (L-Pro) is necessary. In E. coli, proline is catabolized to glutamate by proline dehydrogenase (PutA) [14]. Previous studies [15] have shown that the deletion of putA not only increases the accumulation of L-Pro but also accelerates the rate of T-4-HYP synthesis. Therefore, we deleted the putA gene in the E. coli W3110 genome, inhibiting the pathway for L-Pro degradation. The resulting strain, Pro-1, accumulated 1.39 g/L of L-Pro after 26 h of shake flask fermentation(Fig. 2a).
Next, we deleted lacI from the Pro-1 strain to ensure constitutive expression of genes controlled by the Ptrc promoter. Feedback inhibition and overexpression of key metabolic nodes in the pathway are important and effective strategies for overproducing L-Pro [16, 17]. Proline synthesis from glutamate involves three enzymatic reactions. The first step involves the conversion of glutamate to γ-glutamyl phosphate by γ-glutamyl kinase, and allosteric inhibition ofγ-glutamyl kinase activity leads to the breakdown of proline into glutamate. This reaction is catalyzed by the proB gene encoding γ-glutamyl kinase, which is the rate-limiting enzyme in the synthesis [16]. It has been reported [18] that the mutant gene proB74 effectively relieved feedback inhibition by replacing the 319th G in proB with an A, resulting in the substitution of asparagine for aspartic acid at position 107 of the predicted protein. Therefore, we integrated the proB74 gene into the Pro-1 genome under the control of the Ptrc promoter to obtain Pro-2, which produces 5.9 g/L of L-Pro. This increased the yield 4.2 times higher than that of the control strain Pro-1 (1.39 g/L), confirming that the mutant gene proB74 alleviated feedback inhibition caused by proline accumulation, which is consistent with previous research[18]. ProB74 was inserted into the rpH locus of Pro-2 to generate Pro-3 while being controlled by the Ptrc promoter. The shake flask data for E. coli are shown in Fig. 2a, with Pro-3 accumulating 10.6 g/L of L-Pro.
Construction of trans-4-hydroxyproline expression element and its introduction into the L-proline chassis strain
The key determinant for converting L-proline (L-Pro) into T-4-HYP is the activity of proline hydroxylase, which plays a pivotal role in enhancing hydroxylation efficiency and subsequently increasing T-4-HYP production rates. Therefore, we chose a previously reported trans-Proline-4-hydroxylase gene (DaP4H) from Dactylosporangium sp. [6]. To ensure its successful expression in E. coli, we subjected DaP4H to codon optimization and subsequently integrated it into plasmid Ptrc99a, resulting in the Ptrc-DaP4H construct. Upon successful transformation into the wild-type E. coli strain W3110, the engineered strain H-1 was generated. In parallel, an blank Ptrc99a plasmid was introduced into the wild-type E. coli W3110 strain to generate the control strain, H-A. Both strains were subjected to shake-flask fermentation using a substrate mixture of 10.0 g/L proline and 2.0 g/L α-ketoglutarate for 26 h. Notably, the H-1 strain exhibited a considerable T-4-HYP yield of 8.9 g/L, whereas the control strain H-A did not produce any T-4-HYP (Fig. 2b). However, the fact that both strains exhibited nearly identical optical densities at 600 nm (OD600), confirmed the efficient expression of the heterologous proline hydroxylation element Ptrc-DaP4H in E. coli W3110, with no discernible impact on cellular growth.
Furthermore, we introduced the Ptrc-DaP4H plasmid into the Pro-5 strain, resulting in the recombinant strain HYP-1. Concurrently, an blank Ptrc99a plasmid was introduced into the Pro-5 strain to generate the control strain, HYP-A. The results of 24-hour shake flask fermentation are presented in Fig. 2c. Remarkably, strain HYP-1 produced a substantial amount of T-4-HYP (4.98 g/L), whereas strain HYP-A did not produce T-4-HYP but accumulated 16.9 g/L of L-Pro. Importantly, both strains displayed comparable OD values, confirming the successful transformation of endogenously produced L-Pro into T-4-HYP and the non-interference of the introduced T-4-HYP expression element with cellular growth. Intriguingly, strain HYP-1 also produced 11.4 g/L of L-Pro, which was not effectively converted into T-4-HYP. Chen et al. [19] suggested that this observation might be attributed to an inadequate α-ketoglutarate precursor availability. α-ketoglutarate plays a pivotal role in the T-4-HYP biosynthetic pathway, serving as an intermediate in the TCA cycle and as a precursor for T-4-HYP biosynthesis. Therefore, to drive the efficient hydroxylation of α-ketoglutarate, downregulation of the transcriptional level of the SucAB gene is required, facilitating the redistribution of metabolic flux at the α-ketoglutarate node. Additionally, 1.1 g/L of acetate and 0.83 g/L of lactate were produced, signifying the concurrent diversion of carbon metabolic flux towards acetate biosynthesis during the T-4-HYP synthesis pathway.
The genotype of strains: Pro-1, E. coli W3110, ∆putA ;Pro-2, Pro-1 ∆lacI ,yghX::Ptrc- proB*; Pro-3, Pro-2 rpH::Ptrc- proB*; Pro-4, Pro-3 yjiT::Ptrc- proA; Pro-5, Pro-4 yciQ::Ptrc- proC; H-1, E. coli W3110 Ptrc-DaP4H; H-A, E. coli W3110 Ptrc99a; HYP-A, Pro-5 Ptrc99a; HYP-1, Pro-5, Ptrc99a-DaP4H; HYP-2, HYP-1, ∆ldhA; HYP-3, HYP-2, ∆poxB; HYP-4, HYP-3 ∆ackA; HYP-5, HYP-4, ∆aceA;
Data are expressed as mean and error bars are expressed as standard deviation ( n = 3 independent experiments ). * means 0.01 < P < 0.05, * * means P < 0.01.
Enhancing the metabolic flux of the precursor to increase the production of trans-4-hydroxyproline
In E. coli, acetyl-CoA and α-ketoglutarate serve as crucial central intermediates in the T-4-HYP biosynthetic pathway; however, they are also key bottlenecks in efficient T-4-HYP production. To address this, we strategically inhibited the side pathways of pyruvate to eliminate byproduct formation and enhance pyruvate availability. Specifically, we disrupted the lactate dehydrogenase gene (ldhA) to impede lactate synthesis, resulting in strain HYP-2 (Fig. 2c), which exhibits a modest 6.2% increase in pyruvate accumulation (5.29 g/L) compared to the control strain HYP-1. Additionally, Pox B can convert pyruvate into acetic acid and CO2,we deleted pyruvate oxidase (PoxB) to generate the strain HYP-3 in order to prevent acetate production. HYP-3 yielded 0.35 g/L of acetate and achieved a T-4-HYP titer of 5.76 g/L, representing a 9.0% improvement over HYP-2. However, pyruvate-to-acetate conversion catalyzed by acetyl-CoA via acetate kinase A (AckA) leads to carbon loss, which is detrimental to fermentation [20]. Thus, ackA was knocked out to obtain the strain HYP-4. As shown in Fig. 2c, HYP-4 accumulated merely 0.22 g/L of acetate, 59% less than HYP-3. The augmented acetyl-CoA pool in HYP-4 contributed to an enhanced T-4-HYP production of 6.62 g/L, a 14.9% increase compared to HYP-3 (Fig. 2c). These results unequivocally demonstrate that ackA knockout significantly reduces acetate byproduct formation. Isocitrate lyase is encoded by the aceA gene, which can directly convert isocitrate into succinic and malic acids through the glyoxylic acid cycle and is a competitive pathway for the synthesis of α-ketoglutarate. Therefore, aceA (an isocitrate lyase-encoding gene) was deleted to prevent the loss of isocitrate and improve the metabolic flow to α-ketoglutarate. Deletion of the aceA gene produced the strain HYP-5, and the yield of T-4-HYP reached 7.97 g/L (Fig. 2d), which was 17.6% higher than that of strain HYP-4.
Considering the pivotal role of central carbon metabolism in both cell growth and T-4-HYP biosynthesis, the precise carbon flux at the α-ketoglutarate node must be tailored to achieve a harmonious equilibrium between these two processes. To achieve this, we introduced weak promoters of varying strengths upstream of sucAB into strain HYP-5 to modulate sucAB transcription. The weak promoters BBa-j23109, BBa-j23115, and BBa-j23114 resulted in strains HYP-6-1, HYP-6-2, and HYP-6-3, respectively. After 26 hours of shake flask fermentation (Fig. 3a), all three engineered strains HYP-6-1, HYP-6-2, and HYP-6-3 displayed elevated T-4-HYP titers, reaching 9.56 g/L, 8.75 g/L, and 8.27 g/L, respectively, compared to the control strain HYP-5. Although these strains exhibited slower growth than HYP-5 during the initial stages, their later growth converged to similar levels with no significant differences in biomass. As L-Pro was converted to T-4-HYP, α-ketoglutarate was directed towards succinate through the glyoxylate bypass, thus facilitating cellular growth restoration, consistent with prior research findings [2]. Analysis of the relative transcription levels of SucAB at different fermentation time points (Fig. 3b) corroborated the successful downregulation of SucAB gene expression using different weak promoters. BBa-j23109 exhibited the most pronounced effect. By comparing Figs. 3a and b, it was evident that fine-tuning the expression of sucAB effectively improved T-4-HYP biosynthesis without significantly compromising E. coli growth.
These genetic modifications effectively alleviate the metabolic flux bottleneck in T-4-HYP biosynthesis. Nonetheless, a residual amount of 6.5 g of L-Pro remained (Fig. 2d). To address this, we explored the addition of 2.0 g/L of α-ketoglutarate to the existing culture medium to investigate whether exogenous α-ketoglutarate supplementation could enhance T-4-HYP synthesis. After 26 h of shake flask fermentation, the results (Fig. 3c) demonstrated that the addition of α-ketoglutarate led to a remarkable increase in T-4-HYP production, reaching 11.64 g/L, signifying a 21.7% increase over the control group. This outcome underscores the need for further reinforcement of α-ketoglutarate metabolism, and the subsequent introduction of the NOG pathway is anticipated to further intensify carbon flux towards α-ketoglutarate.
The genotype of strains: HYP-6-1, HYP-5 PBBa-j23109- SucAB; HYP-6-2, HYP-5 PBBa-j231115- SucAB; HYP-6-3, HYP-5 PBBa-j23114- SucAB; HYP-7-1, HYP-6-1 ycdN::Ptrc- Ba-xfp; HYP-7-2, HYP-6-1 ycdN::Ptrc- Ac-xfp; HYP-7-3, HYP-6-1 ycdN::Ptrc- Bl-xfp
Data are expressed as mean and error bars are expressed as standard deviation (n = 3 independent experiments). * means 0.01 < P < 0.05, * * means P < 0.01.
Introducing the NOG pathway to reduce carbon loss and increase yield
The glycolytic pathway is the basic pathway of glucose catabolism in cells and converts external glucose into a carbon source for use in the body. Pyruvate decarboxylase catalyzes the conversion of pyruvate to acetyl-CoA via a series of catalytic reactions using glucose as the starting material. This process loses CO2, resulting in carbon loss and limiting the yield of the target product [21] (Fig. 4b). However, phosphoketolase (encoded by xfp), the key enzyme in the NOG pathway, is mainly present in the phosphoketolase pathway of heterolactic fermentation and shunt metabolism in Bifidobacterium. [22] As shown in Fig. 4a, xfp weakens the Embden-Meyerhof-Parnas (EMP) pathway during expression, converting one molecule of glucose into three molecules of acetyl-CoA without CO2 emissions. The theoretical carbon yield of acetyl-CoA increased from 0.66 mol mol mol− 1 to 1.00 mol mol mol− 1 [23], and carbon is totally preserved in glycolytic metabolism, similar to acetyl-CoA. Enhanced carbon flux in the TCA cycle. Heterologous introduction of this pathway into other engineered bacteria can significantly increase product yield using acetyl-CoA as a precursor. [ 22 ] Kanokarn et al. [ 24 ] used strains with the phosphoenolpyruvate pathway to increase the supply of acetyl-CoA as a precursor, resulting in the maximum PHB(poly-β-hydroxybutyrate) yield after 100 h. Acetyl-CoA is an important intermediate in the TCA cycle [ 25 ]. Lee et al. [ 26 ] replaced the initial stage of EMP in Corynebacterium glutamicum with NOG pathway to construct a hybrid EMP, which reduced CO2 emissions by 10% and increased acetyl-CoA by 19%. Based on the above findings, we introduced the NOG pathway into strain HYP-1, which enhanced the carbon flux flowing into the TCA cycle. According to reports, when the NOG pathway was supplemented to supply the T-4-HYP synthesis precursors acetyl-CoA and α-ketoglutaric acid, the yield of hydroxyproline was greatly improved [27]. The phosphoketolase pathway exists in fungi and several bacteria and shows different substrate specificities and expression intensities among different species [ 28 ]. Therefore, we selected three reported xfps for comparative experiments. These were isolated from Bifidobacterium adolescentis (Gene ID:56674845), Clostridium acetobutylicum (GenBank: KHD36088.1), and Bifidobacterium longum (NCBI-Gene ID:56674845). The three genes, Baxfp, Acxfp, and Blxfp, which are controlled by the Ptrc promoter, were codon optimized and integrated into the ycdN locus of the HYP-6 genome to evaluate the feasibility of this pathway in E. coli and strains HYP-7-1, HYP-7-2, and HYP-7-3 were obtained, respectively. Strain HYP-6 was used as the control for shake-flask fermentation. Data from the E. coli shake flask fermentation are shown in Fig. 3d. After 26 h of fermentation, the T-4-HYP yields in the HYP-7-1, HYP-7-2, and HYP-7-3 groups were 13.21, 10.52, and 11.42 g/L, respectively. The yield from strain HYP-7-1 was 30. % higher than that of HYP-6. This demonstrates that engineered E. coli cells can convert glucose through the NOG pathway with sufficient biotransformation activity and that Ba-xfp is more suitable for the expression of T-4-HYP in E. coli. The production performance of strains HYP-6 and HYP-7-1 was tested in a 5 L fermenter. As shown in Fig. 3e, the yields of T-4-HYP were 45.5 and 62.4 g/L, respectively, and in strain HYP-7-1, the yield of T-4-HYP from glucose was 0.28 g/g. Introduction of the NOG pathway increased carbon flux to the TCA cycle, and the yield of T-4-HYP was greatly improved. This finding is consistent with those of the previous studies. Strain HYP-7-1 uses glucose as a carbon source to supply acetyl-CoA for high T-4-HYP yields.
Enhancing NADPH supply
NADPH provides a large amount of reducing power for the growth and metabolism of various organisms. Fuhrer et al. [ 29 ] reported that the in vivo demand for NADPH in biosynthesis far exceeds the amount of NADPH produced. To maintain growth, the catabolism of NADPH should be balanced with the demand for the anabolism of the target products. In this study, 3 mol of NADPH was required for the synthesis of 1 mol of L-Pro by acetyl-CoA (Fig. 4c), but only 1 mol of NADP+ was reduced to NADPH during this process. Therefore, an increase in the NADPH/NADP+ ratio in cells promotes the synthesis of T-4-HYP. Thus, additional methods are required to synthesize NADPH for cellular use. The E. coli membrane-bound pyridyl nucleotide transhydrogenase encodes PntAB, which is oxidized by NADH to NAD+ to drive the reduction of NADP+ to NADPH [30]. The strategy of maintaining intracellular NAD(P)H balance by enhancing the expression of PntAB to promote the synthesis of target products has proven to be simple and effective [31, 32]. In this study, the natural promoter of pntAB was replaced with the Ptrc promoter to improve its expression and obtain enhanced HYP-8 yields. The T-4-HYP yield of strain HYP-8 reached 15.24 g/L, which was approximately 15.3% (Fig. 5a) higher than that of the control HYP-7 strain. In addition, we introduced a second pntAB gene controlled by the Ptrc promoter to the yeeL site of HYP-8 and obtained the strain HYP-9, which produced 16.85 g/L of T-4-HYP (Fig. 5a) and was 10.5% higher than that of strain HYP-8. Intracellular reduced cofactor levels in the two strains were detected, and the total amount of NAD(P)H in HYP-9 was significantly higher than that in HYP-8 (Fig. 5b). These results indicate that pntAB overexpression has a positive effect on T-4-HYP production and cell growth.
Enhanced cofactor supply to improve T-4-HYP synthesis
Enhanced dissolved oxygen concentration during fermentation
Proline hydroxylase is a kind of α-ketoglutarate-dependent dioxygenase that requires molecular oxygen as a donor. However, as fermentation progresses, the volume of bacteria and products in the fermentation tank increases [10], which results in an insufficient supply of dissolved oxygen and limited bacterial growth. Oxygen supply is generally considered the rate limiting factor in the production of T-4-HYP by recombinant strains [10]. According to previous studies [3, 30, 33], there are four methods to improve the oxygen content during the bacterial fermentation process: (1) increasing the stirring rate, ventilation, and tank pressure, (2) introducing the videocall hemoglobin gene (vHb), (3) adding protease, and (4) adding an oxygen carrier. However, increasing stirring can cause serious damage to the cells, which in turn leads to a decrease in the product synthesis rate [34]. Liu and Ciobanu [33, 34] added separately protease and n-dodecane to increase the dissolved oxygen content during the fermentation process, which greatly improved the yield of the target product. However, this not only increased the production cost, but also made the subsequent extraction and purification processes more challenging. It has been reported that Vitreoscilla hemoglobin (vHb) can efficiently transport oxygen [35]. The presence of vHb accelerates oxygen transfer to enhance ATP production and oxidative metabolism, and has been widely used in aerobic fermentation [36]. Therefore, in this study, vHb was first expressed under the control of the Ptrc promoter and integrated into the HYP-9 genome to obtain HYP-10. After 26 h of shake-flask fermentation, the yield of T-4-HYP was 16.97 g/L, which was not significantly higher than that of HYP-9. During the production of T-4-HYP in a 5L fermenter, excessive cell growth resulted in an insufficient supply of dissolved oxygen at 6 h of fermentation. Monitoring of the fermenter data showed that the dissolved oxygen in the fermentation broth was zero at this time, and finally at 36 h of fermentation. OD and T-4-HYP production stabilized at 122.1 and 72.2, respectively (Fig. 6b). The strain HYP-9 produced 70.5 g/L of T-4-HYP (Fig. 6a). The yield of the two was not much different, and both showed L-Pro accumulations of 12.1 and 14.3 g/L, respectively. Therefore, alternative approaches must be used to solve the problem of insufficient dissolved oxygen in the fermentation broth.
During the production process in 5L fermentor, the dissolved oxygen can be regulated to a certain extent, by sugar supplementation. The glucose addition strategy was changed from batch feeding to continuous feeding. The dissolved oxygen concentration in the fermentation broth was controlled by regulating the glucose addition rate. When the glucose concentration in the fermentation broth was controlled at 0.3 g/L, the dissolved oxygen in the fermentation broth was maintained at approximately 35% until the end of fermentation. It can be seen from Figs. 6c and d that the OD of the experimental group was 130.4, which was 6.7% higher than it was before optimization. The yield of T-4-HYP reached the highest value of 80.1 g/L, which was 10.9% higher than that before optimization, and L-Pro did not accumulate. Based on the comparison between Fig. 6-d and e, the maximum sugar consumption rate without this feeding strategy was 14.3 g/(h·L). Using this strategy, the overall sugar consumption rate was relatively stable. At 28 h, the sugar consumption rate reached its highest levels which was 12.1 g /(h·L). The sugar consumption rate decreased during the later stages of fermentation. Although this strategy prolonged the fermentation time of the strain, the slow sugar feeding rate reduced sugar consumption by the bacteria, thereby increasing the yield of T-4-HYP. After 44 h, the final yield of 0.33 g/g was achieved.
The effect of Fe2+ on the fermentation process
Fe2+ is an essential trace element in cell growth and production, and plays a role in electron-transfer reactions, gene regulation, oxygen binding, and transport [37]. It is also a cofactor of the key enzyme (proline hydroxylase) in the process of T-4-HYP fermentation and participates in the hydroxylation of proline. Therefore, a practical strategy to increase the yield of T-4-HYP is to maintain Fe2+ in the medium at the minimum concentration required for T-4-HYP fermentation. According to Vasylkivska et al. [38], in C. acetobutylicum, iron is a component of the redox protein, which is an iron-sulfur protein involved in electron transfer. They participate in the formation of hydrogen by catalyzing proton reduction, and play an important role in this pathway. Alsaker et al. [39] found that Fe2+ plays a role in spore formation, resulting in a significant increase in the expression of the ferrous absorption system X276_13995-X276_13990-X276_13985 in the stationary phase. To ensure a sufficient supply of Fe2+, a key regulatory factor during the fermentation process of E. coli producing T-4-HYP, we chose a continuous fed-batch fermentation strategy in a 5L bioreactor. This ensured that Fe2+ was maintained at a relatively stable concentration during the fermentation process, providing sufficient power for proline hydroxylase. FeSO4 (20 mg/L) was added to the initial medium, and FeSO4 was continuously fed at a rate of 30 mg/L/h from 10 h. The results of fermentation showed that the cells grew in the logarithmic phase from the beginning of fermentation and entered the stable phase at 28 h. The maximum yield of 89.4 g/L was obtained after 44 h of fermentation.with the highest OD600 of 75.3 at 44 h.