A combined pseudouridine biomanufacturing platform enabled by a streamlined designer pathway

Yu, Le; Chen, Ruyi; Zhang, Chenyue; Wang, Zhengyuan; Wang, Zhuqing; Zeng, Xinyue; Liang, Han; He, Yuanyuan; She, Yixuan; Wang, Yifei; Gong, Rong; Song, Xuemin; Deng, Zixin; Fei, Qiang; Chen, Wenqing

doi:10.1038/s41467-025-63906-0

Download PDF

Article
Open access
Published: 06 October 2025

A combined pseudouridine biomanufacturing platform enabled by a streamlined designer pathway

Nature Communications volume 16, Article number: 8866 (2025) Cite this article

Subjects

Abstract

mRNA vaccines, featured by incorporated pseudouridine (Ψ), represent a milestone in combating diseases, thus highlighting Ψ importance in drug development. However, economic and environmental challenges have persisted in sustainable Ψ production. Here, we formulate a streamlined designer Ψ pathway, comprising UMP nucleosidase, ΨMP glycosidase, and ΨMP phosphatase, and realize its gram-scale production by targeted discovery of a prominent UMP-preferred nucleosidase (NmYgdH). The optimized pathway, containing NmYgdH, RjPsuG (ΨMP glycosidase), and HDHD1 (ΨMP-specific phosphatase) is cloned into E. coli and systematic evaluation of multiple strategies achieves a Ψ titer of 44.8 g·L⁻¹. Moreover, a thyA-dependent, tunable, and eco-friendly strategy for sustainable Ψ production is demonstrated in a 5 L bioreactor achieving titer of 45.3 g·L⁻¹. Finally, we establish a simplified-strategy for rapid Ψ purification with a recovery-rate of 71%, and techno-economic analysis is employed to validate the feasibility and advantages of this fermentation platform for Ψ biomanufacturing. Therefore, this study provides a blueprint for industrial-production of nucleoside-related molecules.

A selective and atom-economic rearrangement of uridine by cascade biocatalysis for production of pseudouridine

Article Open access 20 April 2023

Quantitative profiling of pseudouridylation landscape in the human transcriptome

Article 30 March 2023

Promoted read-through and mutation against pseudouridine-CMC by an evolved reverse transcriptase

Article Open access 11 January 2025

Introduction

Pseudouridine (Ψ), an isomer of uridine, is widely present in almost all types of RNA including mRNA, tRNA, rRNA, and SnRNA^1,2,3 (Fig. 1a). As the most prevalent RNA alteration in organisms, Ψ influences RNA’s biological functions, and therefore it is referred to as the fifth nucleoside^2,4. Notably, Ψ has been extensively utilized in mRNA vaccines, as it can significantly increase the stability of mRNA^5,6, and decrease its immunogenicity, while promoting its translation efficiency in vivo⁷. Currently, the mRNA vaccines have already been developed for treating COVID-19 and other clinically-important diseases^8,9. Furthermore, mRNA has also demonstrated its potential in the treatment of hereditary metabolic disorders¹⁰, thereof resulting in a significant surge in the development of mRNA-related therapies¹¹.

Fig. 1: Chemical structures of Ψ as well as related modified nucleosides and the designer artificial biosynthetic strategies for Ψ production in vitro and in *E. coli.*

Over the past decades, chemical strategy for Ψ synthesis, though began in 1961, has been confirmed to be inapplicable due to related flaws, including excessive by-products, and low-stereoselectivity¹². In previous studies, considerable efforts have been paid to employ multi-enzyme cascades to produce Ψ. Relatively, the biocatalytic approach can effectively address the challenges of low stereoselectivity and potential environmental issues, meanwhile maintaining the synthetic efficiency. Alice et al. found that PsuG (EC: 4.2.1.70) selectively catalyzes the linkage and cleavage of the C-C glycosidic bond in ΨMP, with the reaction equilibrium lies far away from the product side¹³, and Riley et al. reported a semi-enzymatic synthesis of Ψ using PsuG and an alkaline phosphatase, with the starting material, ribose-5P, synthesized by depurination reaction of AMP¹⁴ (Supplementary Fig. 1a). Particularly, the Nidetzky group conducted pioneering studies on ΨMP/Ψ synthesis. They achieved ΨMP synthesis by designing an enzymatic cascade from D-ribose and uracil (Supplementary Fig. 1b)¹⁵, and they also developed an elegant four-enzyme cascade, initiating with uridine, for one-pot and atom-efficient production of Ψ (Supplementary Fig. 1c)¹⁶. These biocatalytic strategies indicated promising industrial potentials for Ψ synthesis, with achieving considerable yields under mild conditions.

Recently, related reports have detailed the ways utilizing Escherichia coli (E. coli) to synthesize Ψ. The artificial pathway, as developed by Riley et al., was introduced into E. coli for Ψ production (7.2 g·L⁻¹) by fed-batch fermentation¹⁷ (Supplementary Fig. 2a), moreover, Ψ production (27.5 g·L⁻¹) was also realized by Zhang et al. using E. coli as cell factory with employing a combined fermentation-feeding strategy¹⁸. Nonetheless, the external addition of uracil affects its application for industrial manufacturing (Supplementary Fig. 2b). All of these mean that a robust, and eco-friendly biomanufacturing platform is urgently needed for efficient Ψ production¹⁹.

In this study, we formulate a streamlined designer Ψ pathway and show gram-scale Ψ production enhanced by a UMP-preferring nucleosidase (NmYgdH). We deploy the full-optimized NmYgdH-RjPsuG-HDHD1 cascade in an engineered E. coli chassis, achieving thyA-dependent, tunable, and antibiotic-free fermentation for sustainable Ψ production (45.3 g·L⁻¹) in a 5 L bioreactor. Furthermore, we develop a rapid and scalable purification process for Ψ and use techno-economic analysis (TEA) to assess the economic viability of our biomanufacturing approach. This study opens the way for future development of a powerful platform towards efficient synthesis of more modified nucleoside molecules.

Results and Discussion

In vitro synthesis of Ψ is realized by the designer multi-enzyme cascades

To achieve the goal for efficient synthesis of Ψ in vitro, we first tested the well-documented cascade using uracil and ribose-5P as substrates (Fig. 1b). As anticipated, this cascade, consisting of the ΨMP glycosidase EcPsuG and the alkaline phosphatase PhoA (EC: 3.1.3.2), is workable, but inapplicable, for Ψ synthesis (Supplementary Fig. 3), as it needs expensive ribose-5P as substrate, which promotes us to come up with more economical cascades for the purpose of industrial Ψ synthesis. Inspired by the biosynthesis of angustmycin, whose pathway features an enzyme (AgmA) capable of hydrolyzing AMP, leading to the formation of adenine and ribose-5P (Supplementary Fig. 4a)²⁰, we thus designed an updated cascade using AMP and uracil as substrates (Fig. 1c). The results showed that Ψ is capable of being efficiently synthesized with complete consumption of uracil (Supplementary Fig. 4), while we also noticed that the concomitant production of adenine as by-product is undesirable in that it will increase the cost for the subsequent Ψ separation.

To avoid the potential interference, we then designed a new cascade using CMP as the initial substrate, based on the CMP nucleosidase BlsM (from blasticidin S pathway)^21,22,23 or GouE (from gougerotin pathway)²⁴ (Supplementary Fig. 5), in which cytosine can be easily deaminated by CodA (deaminase)²⁵ to form uracil. To optimize the cascade, we attempted to screen an efficient CMP nucleosidase, with employing phylogenetic analysis of BlsM/GouE against its homologs (Supplementary Fig. 6a, b)²⁶, and two group of enzymes were subsequently selected for further enzymatic assessment according to their phylogenetic distribution sites. Of all enzymes evaluated, we found that SgGouE indicates the most superior performance (Supplementary Fig. 6c, d, e).

Accordingly, we set out to evaluate the cascade comprising SgGouE, CodA, EcPsuG, and PhoA. Conforming to our expectation, the results showed that Ψ can be efficiently synthesized, and we also noticed that a small amount of uracil still existed in the reaction system, likely attributing to the instable phosphate group of ribose-5P (Fig. 2a, and Supplementary Fig. 7). Simultaneously, we observed that SgGouE is also capable of recognizing UMP as substrate, but only with efficiency of 50%, implying that we could figure out a more efficient and economical cascade with UMP as the starter substrate (Supplementary Fig. 6e, and Supplementary Fig. 8). We then tested the feasibility of the SgGouE-EcPsuG-PhoA cascade for Ψ synthesis, and unfortunately, the results indicated that, despite that such cascade is workable for Ψ synthesis, the efficiency is much lower (Supplementary Fig. 9), suggesting that it is inapplicable for industrial synthesis of Ψ. However, it motivates us to further hunt for an efficient UMP-preferred nucleosidase for industrial purposes. Altogether, we have developed a cascade consisting of SgGouE, CodA, EcPsuG, and PhoA, which is likely to be extended for gram-scale synthesis of Ψ.

**Fig. 2: Evaluating the designer pathways for efficient Ψ synthesis utilizing CMP/UMP as initial substrate.**

Mining of NmYgdH as a prominent UMP-preferred nucleosidase facilitates efficient synthesis of Ψ

The experimental phenomenon of SgGouE capable of recognizing UMP as substrate suggests that we may also mine a UMP-preferred nucleosidase from the YgdH (PpnN, pyrimidine/purine nucleotide 5′-monophosphate nucleosidase) family enzymes^27,28,29, and we then retrieved all protein sequences of this family from the UniProt database, thereof constructing a phylogenetic tree utilizing MEGA (Supplementary Fig. 10a)²⁶. Subsequently, we selected 11 proteins based on their phylogenetic relationships (Fig. 2c, Supplementary Fig. 10b) for further evaluating their enzymatic activities against UMP (Fig. 2d). Of all proteins examined, we found that NmYgdH exhibits prominent activity (Fig. 2d). Consequently, we redesigned a three-enzyme cascade using NmYgdH, EcPsuG, and PhoA, and the results showed that NmYgdH, converting UMP to uracil and ribose-5P, significantly facilitates the synthesis of Ψ (Fig. 2e, and Supplementary Fig. 11), suggesting that the present cascade is likely applicable for industrial biomanufacturing purpose.

Next, we determined the kinetic parameters of NmYgdH against UMP, showing K_m = 11.78 ± 0.81 mM, k_cat = 81.90 ± 2.60 s⁻¹, and k_cat/K_m = 6.95 ± 0.53 mM⁻¹·s⁻¹, suggesting that this enzyme displays outstanding catalytic efficiency (Supplementary Fig. 12a). Moreover, we tested substrate specificity of NmYgdH against related nucleotides, including UMP, CMP, AMP, and GMP, and the results exhibited that it prefers UMP as substrate in terms of catalytic efficiency (k_cat/K_m) (Supplementary Fig. 12b–d). In addition, NmYgdH also shows superior stability (Supplementary Fig. 12e).

To further explore the molecular mechanism of NmYgdH for efficient hydrolysis of UMP, we have solved the crystal structure of NmYgdH at 1.80 Å resolution. Overall, the NmYgdH monomer forms a Rossmann α/β structure, and is composed of eight α-helices and seven parallel β-strands, which is considerably homologous to the LOG from Streptomyces coelicolor (PDB ID: 5ZI9), with sharing the conserved “PGGxGTxxE” motif (Supplementary Fig. 13a, b, c; Supplementary Table 1)³⁰. Then we performed a molecular docking analysis of NmYgdH with UMP using Autodock 4.2³¹, and the configuration and binding sites of the molecular docking substrate, as indicated, are nearly identical to those of LOG from Pseudomonas aeruginosa (PDB ID: 5ZBK, structure in complex with AMP) (Supplementary Fig. 13d)³².

Structurally, the bulky purine base of AMP/GMP imposes steric hindrance that prevents its efficient entry into the active-site pocket. To elucidate the molecular basis underlying the preferred hydrolysis of UMP over CMP, we conducted comparative analysis of NmYgdH and SgGouE sequences (a CMP preferred homolog), with finding that they share the conserved YgdH family motif PGGxGTxxE (Supplementary Fig. 14a). Detailed inspection of the substrate-binding pocket revealed that a single residue at the pocket periphery (Q97 in NmYgdH, and E106 in SgGouE), presumably affecting local electrostatic environment of the pocket, is likely responsible for modulating UMP specificity (Supplementary Fig. 14a, b). Consistent with this deduction, in vitro assays showed that the NmYgdH_Q97E variant exhibits a threefold increase of hydrolysis activity for CMP but with a 99% decrease of that for UMP. Likewise, SgGouE_E106Q variant also displays a fourfold enhancement of UMP activity and a 99% reduction in CMP activity (Supplementary Fig. 14c–g). These data clarify the pivotal role of the key residue of NmYgdH/SgGouE for the specific recognition of UMP/CMP.

Gram-scale synthesis of Ψ and ΨMP is individually realized with CMP/UMP as substrate

To facilitate gram-scale synthesis of Ψ, we re-estimated the optimal reaction temperature for the enzymes contained in related cascades, and the results showed that all enzymes, except for SgGouE, maintain the maximal activity at 37 °C of the temperatures selected (Supplementary Fig. 15). Subsequently, we attempted to conduct the NmYgdH-EcPsuG-PhoA cascade for Ψ synthesis in a scale of 1 g substrate (Supplementary Fig. 16). In contrast with our expectation, we found that Ψ is merely synthesized with conversion efficiency of 80%, even with reaction time extended to 1,000 min, which is probably attributed to the fact that PhoA (EC:3.1.3.2) is a substrate-inhibited enzyme under the circumstance of high-concentrated substrate (Supplementary Fig. 17a)^33,34. These results suggest that PhoA is not suitable for a larger scale synthesis of Ψ. More luckily, previous studies reported that YjjG (EC:3.1.3.5) possesses distinguished performance¹⁶, particularly in reactions with higher concentration of substrates (Supplementary Fig. 17b). Consequently, we constructed the optimized NmYgdH-EcPsuG-YjjG cascade for in vitro Ψ synthesis, and the results indicated that efficient synthesis of Ψ can be realized in 350 min with conversion rate of 95% (Supplementary Fig. 18a). The identity of Ψ was then unambiguously confirmed by NMR analysis (Supplementary Fig. 19–22).

Subsequently, we further extended the NmYgdH-EcPsuG-YjjG cascade for Ψ synthesis in a larger scale (10 g UMP as substrate), and the results showed that UMP can be finally converted to Ψ with efficiency of 99% within 350 min (Fig. 3a). More excitingly, we also observed that ΨMP (with either 1 g or 10 g UMP as substrate) is capable of being synthesized with a surprising rate of 93% (Fig. 3b; and Supplementary Fig. 18b, 23–28), and this is mainly because that the limited solubility of uracil promotes the reaction balance towards ΨMP synthesis. Meanwhile, we also evaluated the SgGouE-CodA-EcPsuG-YjjG cascade at 30 °C with CMP (1 g/10 g) as substrate, and the data exhibited that the SgGouE-CodA-EcPsuG-YjjG cascade can also achieve high-efficient conversion (98%) of CMP to Ψ (<350 min). (Fig. 3c; Supplementary Fig. 29a). Likewise, ΨMP was also accumulated with a considerable efficiency of 91% (Fig. 3d; Supplementary Fig. 29b). All these data demonstrate that the two cascades are robust and efficient for in vitro synthesis of Ψ and ΨMP.

**Fig. 3: Monitoring of the scale-up enzymatic synthesis of Ψ as well as ΨMP and the convenient process for rapid purification of Ψ.**

Moreover, we set up a separation process for Ψ from the reactions with 10 g UMP as substrate. Relative to UMP/CMP, Ψ indicates lower solubility, the reactions, with removal of enzymes, was thus precipitated at 4 °C overnight, and Ψ, crystallized under such circumstance, can be easily separated, and prepared with yield of 87% and purity of 98% (Fig. 3e; and Supplementary Fig. 30).

As for the SgGouE-CodA-EcPsuG-YjjG cascade, the maximal concentration of CMP is just 0.2 M, which implicates that we need larger reaction system and extra separation-efforts for Ψ biomanufacturing, more than that, the generation of ammonia is also unfavored, consequently, the further application of such cascade is partly limited. As far as the NmYgdH-EcPsuG-YjjG cascade is concerned, it displays unparalleled advantages over the SgGouE-CodA-EcPsuG-YjjG cascade. Comparatively, it is more efficient (higher substrate-concentration, 1 M UMP; higher reaction-temperature, 37 °C), economical (three-step over four-step, and without ammonia emission). This in vitro designer pathway serves primarily as a proof-of-concept for Ψ synthesis, which could be utilized for local-scale synthesis of Ψ. Most importantly, it can be potentially extended into the microbial cell factory for sustainable production of Ψ.

The designer pathway introduction confers E. coli with the capability of Ψ production

Under intracellular environment, simultaneous existence of a series of nucleotides means that the discovered NmYgdH need exhibit substrate preference for UMP. To see if NmYgdH indicates such feature, we first performed the substrate-competition assays against UMP, CMP, AMP, and GMP, and revealed that it significantly prefers UMP, rather than the other nucleotides (Supplementary Fig. 31a). Subsequently, we conducted the further substrate-competition assays against UMP and dUMP, and excitingly, we found that NmYgdH overwhelmingly favors the former as substrate (Supplementary Fig. 31b), all these suggesting that the NmYgdH-EcPsuG-YjjG cascade is very likely to be extended for in vivo production of Ψ. To realize the goal, E. coli MB219, an engineered strain for efficient production of UMP-derived chemicals³⁵, was chosen as a potential cell factory. However, the strain MB219 contains a redundant gene pbs1 (coding for the enzyme specifically dephosphorylate 5-Methyl UMP and dTMP) for the present system, and thus we deleted it in MB219 to generate the CY1 strain (Supplementary Fig. 32a), which was then utilized as the starter strain for the purpose of synthesizing Ψ.

Given that endogenous phosphatases of E. coli can also recognize Ψ as substrate, we thus introduced the co-expression plasmid pRL01, containing two genes EcpsuG and NmygdH driven by a single T7 promoter, into the CY1 strain. HPLC analysis indicated that metabolites of the strain CY1/pRL01 can generate a characteristic peak that co-migrates with the Ψ standard, and the identity of the target metabolite, as confirmed by LC-MS analysis, are well aligned with that of the Ψ authentic standard (Supplementary Fig. 33), demonstrating the identity of the target metabolite as Ψ, while the average Ψ titer of the strain CY1/pRL01 was just 0.14 g·L⁻¹ (Fig. 4a). Moreover, we also discovered that considerable amount of the by-product uracil was concurrently accumulated in metabolites, implicating that a ΨMP-specific phosphatase is needed to promote Ψ biosynthesis. These results suggest that the introduced two-enzyme cascade is workable but inefficient for Ψ synthesis in the present E. coli strain (CY1/pRL01).

**Fig. 4: A combination of the full-optimized designer three-enzyme pathway and the engineered *E. coli* cell factory leads to significant enhancement of Ψ titer.**

Previously, the enzyme PsuK was demonstrated as Ψ kinase in E. coli¹³. Furthermore, the transporters NupC and NupG in E. coli were also claimed to facilitate the transport of extracellular nucleosides into cells. As for PsuT, it was supposed to function as a specific transporter for Ψ import^18,36. All of these suggest that these enzymes are likely to be associated with metabolic flux of Ψ and UMP. To confirm these, we thus sequentially deleted psuK, nupC, nupG, and psuT, using CY1 as the starter strain (Supplementary Fig. 32b–e). As anticipated, HPLC analysis indicated that Ψ production was gradually elevated in the corresponding strains CY2/pRL01, CY3/pRL01, CY4/pRL01, and CY5/pRL01, and the Ψ titer in strain CY5/pRL01 was obviously enhanced with reaching 0.72 g·L⁻¹, a fourfold enhancement compared to that of CY1/pRL01 (Fig. 4a, and Supplementary Fig. 33a). Subsequently, the target metabolite, purified from CY5/pRL01, was further validated using NMR analysis, definitively demonstrating that the identity of the target metabolite as Ψ (Supplementary Figs. 34, 35).

Earlier studies showed that Ψ catabolism related enzyme, as represented by PsuK¹³, plays a negative role on Ψ production, as it decreases the intracellular concentration of Ψ by direct consumption. As for nucleoside transporters (NupC, NupG, and PsuT)^18,36, they effectively transport the extracellular nucleoside into the cell factory for metabolic recycle, thus leading to the reduction of Ψ titer. Accordingly, inactivating the genes associated with Ψ catabolism and nucleoside transport is a feasible strategy for potential Ψ enhancement. Nevertheless, the unexpected presence of uracil in the metabolites of strain CY5/pRL01 implicates that the metabolic flux towards Ψ formation is inadequate, mainly due to the lack of a ΨMP-specific phosphatase. Apart from this, the unbalanced expression level of Ψ biosynthesis genes in vivo would also contribute to the incomplete utilization of uracil.

Engineering the steps for ΨMP/UMP dephosphorylation and uracil-utilization leads to dramatic enhancement of Ψ titer

The ΨMP dephosphorylation constitutes the committed step in the designer artificial pathway for Ψ biosynthesis. Accordingly, identifying an appropriate and specific phosphatase for ΨMP dephosphorylation should be a viable strategy for further enhancing Ψ titer. Considering that E. coli phosphatases are more likely to recognize a variety of nucleotides, we thus conducted competition assays against UMP and ΨMP, with an aim to screen an endogenous phosphatase preferring the latter as substrate, and we found that all related phosphatases selected do not exhibit obvious specificity against ΨMP (Supplementary Fig. 36a, b), suggesting that none of them could be extended for in vivo purpose of Ψ production. As a result, we endeavored to discover a ΨMP-specific phosphatase from diverse metabolic pathways, either primary or secondary. Of all 7 phosphatases tested, the results indicated that 6 of them associated with C-nucleosides biosynthesis, including MinC, SdmB, YcjU, DickC, SapK, and PumD, are not the optimal one for Ψ synthesis due to the relatively lower activity/specificity against ΨMP (Supplementary Fig. 36c–g)^37,38,39. Surprisingly, a HAD family protein HDHD1, a previously reported ΨMP phosphatase from human red blood cells⁴⁰, was confirmed to possess excellent activity and specificity over ΨMP. In accordance with our expectation, further in vitro competition assays against UMP and ΨMP showed that HDHD1 has prominent specificity and activity against the latter substrate (Supplementary Fig. 36f, g, and Supplementary Fig. 37).

To test if the HDHD1 introduction is beneficial to Ψ titer improvement, it was cloned downstream of NmygdH in pRL01 driven by T7 promoter, obtaining the co-expression vector pRL02. In contrast with our expectation, both Ψ and other metabolites produced by the E. coli CY5/pRL02 exhibited an evident decrease (Fig. 4b), showing that overexpression of HDHD1 is likely to influence the production of related nucleotides required for primary metabolism. Therefore, we set out to solve the issue by integrative-expression of HDHD1 at ΔargR locus (Supplementary Fig. 38a). Subsequently, the resultant strain CY6 with pRL01 (CY6/pRL01) was fermented for metabolites analysis. As anticipated, HPLC analysis indicated that the Ψ titer has a significant enhancement, achieving 1.5 g·L⁻¹, which is 2.1-fold higher than that of CY5/pRL01. These results demonstrate that maintaining appropriate expression level of HDHD1 is an effective strategy for the further enhancement of Ψ titer (Fig. 4b, and Supplementary Fig. 40a).

As UMP is the direct precursor for Ψ biosynthesis in our designer pathway, and in fact, it can also be efficiently recognized by endogenous phosphatases, including PhoA, YjjG, and UshA (Supplementary Fig. 36b), implicating that mutation of related genes is more likely to increase Ψ titer. For these reasons, the three genes were sequentially in-frame deleted to improve the UMP metabolic flux, and HPLC results indicated that sequential mutation of related genes renders a stepwise increase of Ψ production in related strains (CY7/pRL01, 1.8 g·L⁻¹; CY8/pRL01, 2.1 g·L⁻¹; CY9/pRL01, 2.4 g·L⁻¹) (Fig. 4b, and Supplementary Fig. 38b–d, Supplementary Fig. 40a). These data establish that engineering the dephosphorylation step of ΨMP and UMP is an applicable strategy for Ψ titer enhancement.

Under the intricate intracellular environment of E. coli, uracil can be utilized as well by the Rut pathway, which consumes uracil to synthesize the product 3-hydroxypropionic acid⁴¹. In addition, uracil is also capable of being utilized by uracil phosphoribosyl transferase (UPRT, encoded by upp) to form UMP as a futile cycle⁴². Therefore, we conducted successive inactivation of rutA (encoding pyrimidine monooxygenase, which catalyzes the first committed step in Rut pathway, Supplementary Fig. 39a) and upp, and the results showed that Ψ titer of the counterpart strains CY10/pRL01 (CY9ΔrutA/pRL01) and CY11/pRL01 (CY10Δupp/pRL01) attains 2.6 g·L⁻¹ and 2.9 g·L⁻¹, respectively (Fig. 4b, and Supplementary Fig. 39b, c, Supplementary Fig. 40b). Indeed, uracil acts as a pivotal intermediate in Ψ biosynthesis, and consumption by degradation or other recycle pathway would lead to the direct decrease of its intracellular concentration, and accordingly affect Ψ biosynthetic efficiency. Taken together, these data demonstrate that engineering the steps for ΨMP/UMP dephosphorylation and uracil-utilization is an appropriate approach for the enhancement of Ψ titer.

A combined constitutive expression-system exploitation and glucose-utilization pathway reinforcement renders substantial increase of Ψ titer

The single copy of HDHD1 on the chromosome of engineered E. coli strains denotes that promoter activity decides the expression level of the HDHD1 protein. Accordingly, we attempted to test a series of constitutive promoters (comprising two different RBSs)⁴³ and an inducible T7 promoter to replace the previous trc promoter governing HDHD1 in CY11, correspondingly resulting in the strains CY12, CY13, CY14, CY15, and CY16 (Supplementary Figs. 41, 42). The fermentation results indicated that the Ψ titer exhibited a 1.3-fold increase (3.7 g·L⁻¹) in CY15/pRL01 (HDHD1 was controlled by J23119 (RBS2) promoter), while the substitution with T7 promoter contributed to a 1.1-fold increase of Ψ titer, which is somewhat lower than that achieved with J23119 promoter substitution (Fig. 4d). Motivated by such outcome, we therefore hypothesized that the genes EcpsuG and NmygdH can also be effectively driven by a robust constitutive promoter, hence eliminating the IPTG addition. Accordingly, we constructed a new EcpsuG and NmygdH co-expression plasmid, pRL03, which includes a J23119 (RBS2) promoter cloned individually upstream of each gene (Fig. 4c). Of all strains (with individual introduction of pRL03) tested, the strain CY15/pRL03 exhibited the highest Ψ productivity (2.2 g·L⁻¹) when using glucose as carbon source; nevertheless, the Ψ titer of related strains was generally lower than that of the strains with inducible expression system, regardless of glycerol or glucose used as the feeding medium (Fig. 4d, and Supplementary Fig. 43a).

To comprehensively evaluate the intracellular performances of Ψ biosynthetic pathway distinctly driven by inducible and constitutive promoters, a 96 h fed-batch fermentation of CY15/pRL01 and CY15/pRL03 was conducted in a 5 L bioreactor. Furthermore, glycerol and glucose feeding mediums were independently used to evaluate the impact of carbon source on Ψ production (Fig. 4e, and Supplementary Fig. 43b). Notably, CY15/pRL03, utilizing glucose as carbon source, exhibited more optimal performance with a 2.9-fold increase of Ψ titer (28.7 g·L⁻¹) and superior growth status (23.3 g·L⁻¹, DCW), suggesting that the combination of constitutive expression system and supplying glucose as carbon source can significantly enhance Ψ production efficiency (Fig. 4f).

Given that glucose is the ideal carbon source, elevating the expression level of related genes in glucose utilization pathway is likely applicable to boost the metabolic flux of ribose-5P and PRPP. For E. coli cell factory, moderately enhancing the expression level of zwf (coding for glucose-6-phosphate dehydrogenase) and gnd (coding for 6-phosphogluconate dehydrogenase) has been verified to exert positive influence on cell growth, glucose consumption, and pyrimidine biosynthesis⁴⁴. Consequently, a two-gene cassette, including zwf and gnd under different promoters, was integrated into the ΔphoA locus of CY15, generating the strains CY17, CY18, and CY19 (Supplementary Fig. 44). Subsequent fed-batch fermentation was performed with the three recombinant strains (with independent introduction of pRL03), and the results showed that the Ψ titer reaches 38.7 g·L⁻¹ for CY18/pRL03 with a 1.35-fold increase compared to CY15/pRL03 (with zwf and gnd driven by the J23110 promoter). Meantime, we also noticed that excessively strong/weak promoter activity cannot maximize the improvement of Ψ titer (Fig. 5a, b; Supplementary Fig. 45). Likely, overexpressing related genes would incur cellular metabolic disorders, thereof undermining the contribution to Ψ titer enhancement, while relatively-weak expression level of them cannot supply adequate metabolic flux towards Ψ biosynthesis.

Fig. 5: Further improvement of Ψ titer via reinforcing the glucose utilization pathway and evaluating the optimal *psuG.*

Noteworthily, carbon source selection plays key role on large-scale industrial fermentation, and glucose, relative to glycerol, can be directly used for the synthesis of nucleosides and related intermediates. Nevertheless, the glucose effect, for the inducible expression system, seriously diminishes the glucose utilization efficiency. Indeed, it is exactly the case for Ψ production by the inducible expression system. Collectively, we demonstrate the combined constitutive expression-system exploitation and glucose-utilization pathway reinforcement is an applicable strategy for substantial increase of Ψ titer.

Screening of RjPsuG from Rhizobium johnstonii further promotes the biosynthetic efficiency of Ψ

The reaction catalyzed by EcPsuG is reversible, and although the addition of phosphatase can promote the reaction toward Ψ synthesis, the expression level and catalytic efficiency of EcPsuG may also act as the rate-limiting factor which affects the synthesis efficiency, implicating that it is necessary to explore a homologous protein with higher expression level or catalytic efficiency. Accordingly, we conducted homology analysis of ΨMP glycosidases using EcPsuG as query sequence, and 10 EcPsuG homologs were then recognized as candidates for further evaluations (Fig. 5). In vitro assays showed that, of all enzymes tested, EcPsuG and RjPsuG, display greater efficiency towards ΨMP formation in the 2 h reaction, and surprisingly, RjPsuG indicates more desirable feature of soluble expression level (Fig. 5d, and Supplementary Fig. 46). The data suggest that RjPsuG is more likely to perform the improved properties for Ψ biosynthesis in the present designer pathway.

To further assess the application potential of the candidate gene RjpsuG, we constructed the RjpsuG and NmygdH co-expression plasmid pRL04. Subsequently, this plasmid was introduced into CY18, and the resultant strain (CY18/pRL04) was fermented in a 5 L bioreactor. Conforming to our anticipation, CY18/pRL04, in comparison to the strain CY18/pRL03, presents outstanding performance with Ψ titer of 44.8 g·L⁻¹, cellular density of OD₆₀₀ 58.4, and overall Ψ yield of 0.379 g/g glucose (Fig. 5e). All of these establish that discovery and utilization of RjpsuG considerably promotes the increase of Ψ titer.

Interestingly, a relevant ΨMP glycosidase RsPsuG has also been previously screened by Zhou et al., for microbial synthesis of Ψ¹⁷. Despite the high homology of RjPsuG and RsPsuG, the former, as indicated by our in vivo and in vitro assays, indicates slightly better performance for Ψ synthesis (Supplementary Fig. 47).

A tunable, and antibiotic-free strategy based on thyA achieves robust Ψ production

To avoid antibiotic-use during fermentation to maintain stable inheritance of the expression plasmid, the CY18ΔthyA strain, following the technique outlined in earlier work³⁵, was constructed prior to introduction of pRL04-dervied plasmid, in which an alternative thyA homolog (derived from Dickeya fangzhongdai) was cloned. Considering that the plasmid replication mechanism would induce excessive expression of thyA, potentially interfering with normal cell growth, a series of promoters with varying activities were utilized individually for tunable modulation of thyA expression level (Fig. 6a, and Supplementary Fig. 48). Consequently, the CY20ΔthyA strains harboring different expression plasmids (pRL05, pRL06, or pRL07) were generated using the antibiotic-free strategy (Supplementary Fig. 49). Among the strains evaluated, CY20/pRL06 performs superior Ψ biomanufacturing capability, in which expression level of thyA homolog has almost little impact on cell growth. Following fed-batch fermentation results indicated that the final average Ψ titer of the CY20/pRL06 strain arrives at 45.3 g·L⁻¹, with final biomass of 23.0 g·L⁻¹ DCW, and overall Ψ yield of 0.328 g/g glucose (Fig. 6b, and Supplementary Fig. 50). More excitingly, the entire fermentation process is eco-friendly and efficient with nearly no by-products yielded (Supplementary Fig. 51a). Notably, we also evaluated the cell growth of the starter strain MB219 and the engineered strains. The growth curve analysis revealed that the expression of exogenous genes redirects metabolic flux from biomass accumulation to Ψ biosynthesis, which reversely resulting in a reduction in growth rate. However, we also found that the strain engineering at early stage can partially alleviate metabolic stress, thereby enhancing the growth rate. These data indicate that the relationship between growth and production is properly balanced while engineering the E. coli cell factory (Supplementary Fig. 52).

**Fig. 6: Efficient and sustainable production of Ψ by a *thyA*-dependent *E. coli* system and rapid Ψ purification via a simplified separation process.**

Next, we explored the optimized process for rapid purification of Ψ from fermentation cultures, and found that the Ψ product, following the successive steps of decolorization, filtration, concentration, methanol-washing, and redissolution, was conveniently purified from the fermentation cultures of CY20/pRL06, with a recovery rate of 71% and a purity of 98% (Fig. 6c, Supplementary Fig. 51b). More than that, the identity of the product was further confirmed as Ψ using NMR analysis (Supplementary Figs. 53, 54). Collectively, these results demonstrate that our present E. coli platform, employing the tunable, and antibiotic-free strategy, is robust and efficient for sustainable production of Ψ.

Normally, promoter activity of thyA would affect its in vivo expression level, thereby further influencing nucleic acid metabolism and Ψ synthesis. Consequently, in contrast with the antibiotic-containing cell system, dynamic modulation of thyA expression level in response to on-demand of cell growth results in the relatively-maximal final biomass of the antibiotic-free cell system. Nonetheless, we also noticed that cell proliferation shares priority over Ψ production during the initial fermentation phase of CY20-related strains, probably due to the alleviation of antibiotic stress. Among the three recombinant strains with varying thyA expression intensities, the CY20/pRL06 strain exhibits a more balanced carbon source allocation between cell growth and Ψ production during the fermentation process. These results demonstrate that the growth-production conflict has been properly reconciled in the present Ψ biomanufacturing process.

Techno-economic analysis of the sustainable Ψ biomanufacturing

To further evaluate the techno-economic feasibility of the current Ψ biomanufacturing, we developed a process utilizing our Ψ biomanufacturing technique by established commercial software, based on experimental results and assumptions related to commercialization (Supplementary Tables 2, 3), including fermentation, purification, wastewater treatment, and utility (Fig. 7a). According to the simulation model developed by the National Renewable Energy Laboratory (NREL)^45,46,47, the capital expense (CAPEX) and operating expense (OPEX) for this project are illustrated in Fig. 7b, and Supplementary Fig. 55–57, and Supplementary Tables 4, 5, revealing that the fermentation (A100) constituted the majority of costs associated with bioreactors ($12.19 million) and carbon sources ($0.52 million annually), consistent with related case studies^48,49,50. The projected minimum Ψ selling price (MSP) is $40.96 kg⁻¹ (Fig. 7c), substantially lower than the prevailing market pricing of chemically synthesised Ψ ($500 kg⁻¹ to $1000 kg⁻¹ in China).

**Fig. 7: Processes design and techno-economic analysis of Ψ biomanufacturing.**

Conventional Ψ synthesis comprises 4–8 reaction steps with a low carbon conversion rate of ~15% and depends on intricate, expensive, and dangerous chemicals¹⁶. Conversely, our optimized biosynthetic process attained a superior yield (up to 36.28%) and facilitated efficient crystallization, achieving ≥71% recovery and ≥98% purity. In contrast to chemical synthesis, which necessitates cryogenic settings (−20 °C), our biosynthetic approach functions at 37 °C, thereby diminishing energy requirements and operational expenses^12,16,17,51. Therefore, our all-in-one biomanufacturing approach for Ψ production by E. coli offers a promising alternative and reinforces its suitability for industrial-scale implementation.

From an environmental perspective, the efforts of current chemical plants are hindered by tough reaction conditions and the reliance on complex feedstocks, leading to pollutant emissions and resource depletion^52,53. Transitioning from traditional chemical processes to biomanufacturing to produce equivalent products offers a pathway toward environmentally sustainable development^{45,46,54,55,56}. Our de novo Ψ biomanufacturing approach effectively produces target products without the addition of intricate and hazardous substances (e.g., 2,4-dimethoxy-5-bromopyrimidine and 2,4-dimethoxy-5-lithium) and operates under relatively mild conditions, resulting in lower energy consumption compared to traditional chemical pathways, which supports green production by lowering the carbon footprint and mitigating other environmental impacts associated with pollutant emissions in this hard-to-abate sector¹².

In summary, we have developed a combined biomanufacturing platform consisting of three-enzyme pathway for eco-friendly, economic, and efficient production of Ψ. Regarding the in vitro biomanufacturing platform, we have screened a UMP-preferred nucleosidase, which possesses prominent specificity, activity, and stability, and as a proof of concept, we have realized the local-scale production of Ψ via the in vitro platform (While this paper was under review, a two-enzyme cascade to synthesize ΨMP was reported by Pfeiffer et al. ⁵⁷). Encouragingly, we have extended the further-optimized pathway, containing NmYgdH, RjPsuG (ΨMP glycosidase), and HDHD1 (ΨMP-specific phosphatase) into E. coli, thereby achieving Ψ production (44.8 g·L⁻¹) by comprehensive evaluation of multiple strategies. Moreover, we have established a tunable, and antibiotic-free strategy for efficient and sustainable production of Ψ, with a titer of 45.3 g·L⁻¹, and finally, we have developed a streamlined methodology for rapid separation and purification of reaction mixture and fermentation broth to obtain Ψ with high recovery rate and purity. Furthermore, we have demonstrated the effectiveness of the present Ψ biomanufacturing strategy with a systematic techno-economic analysis for upgrading inexpensive, low-polluting, and readily available feedstocks, and the in vitro system is used as a proof of concept for Ψ synthesis. We anticipate that the present advancement in Ψ biomanufacturing will lower the expenses of its downstream products, including RNA-oligonucleotides and mRNA vaccine precursors, and potentially revolutionize their applications across several industries.

Methods

General methods

Strains and plasmids used in this study are described in Supplementary Data 1, and primers are listed in Supplementary Data 2. Sequencing analysis of related mutant strains are shown in Supplementary Data 3. Codon-optimized sequences are shown in Supplementary Data 4. Accession codes and source information of the proteins are shown in Supplementary Table 6. General methods employed in this work were according to the standard protocols by Maniatis Laboratory manual⁵⁸.

Enzymes, chemicals, and reagents

The enzymes used in this study were purchased from New England Biolabs, Yeasen Biotechnology (Shanghai) Co., Ltd., or Takara Bio.; The chemicals and reagents were the products of Sigma-Aldrich, Thermo Scientific, OMEGA, J&K Scientific, Sinopharm Chemical Reagent, or Angel Yeast Co., Ltd.

General detection methods used in this study

Cell growth was monitored by measuring the absorbance at 600 nm (OD₆₀₀) by ultra-micro ultraviolet-visible spectrophotometer (ND-100C, Hangzhou Miu Instruments Co., Ltd.). Glucose content was determined by biosensor analyzer (S-10, Shenzhen Sieman Technology Co., Ltd.). The cell growth curve measurement was monitored by Micro-GCM (BMG LABTECH), with 10 min a cycle (216 cycles, 36 h). The monitoring temperature was 37 °C, and the initial OD₆₀₀ value of each strain was 0.05.

For HPLC analysis, the fermentation broth was processed by adding oxalic acid till pH 3.0-4.0, after centrifugation (13,800 × g, 20 min), the supernatant sample was filtered by 0.22 μm filter for HPLC analysis. The obtained filtrate (5 μL) was analyzed by Shimadzu HPLC (LC-20A) equipped with a reverse-phase C18 column (Diamonsil, 5 μm, 4.6 × 250 mm). The isocratic elution, performed with 97% solvent A (0.15% aqueous trifluoroacetic acid) and 3% solvent B (methanol) at a flow rate of 0.3 mL·min⁻¹ over 30 min, exhibited the characteristic absorption at λ_max = 254 nm. LC-HRMS analysis was carried out on a Thermo Fisher Scientific Q EXACTIVE (Scientific Inc.) instrument controlled by Xcalibur in the positive mode. The parameters for HRMS analysis are as follows: capillary temperature, 275 °C; capillary voltage, 35 V.

Additional detection methods, including overexpression and purification of target proteins, are detailed in Supplementary Method 1. In vitro assays and kinetic analysis of related enzymes are described in Supplementary Methods 2 and 3, respectively. Protein crystallization, structural elucidation, and molecular docking of NmYgdH are outlined in Supplementary Methods 4 and 5. Phylogenetic analysis of related proteins is provided in Supplementary Method 6, and NMR analysis is described in Supplementary Method 7.

Synthesis of ΨMP and Ψ by multi-enzyme cascades

For the cascade with uracil and ribose-5P as starter substrates, 50 µL reaction (50 mM Tris-HCl buffer, pH 8.0) for ΨMP synthesis, containing 1 mM uracil, 1 mM ribose-5P, 1 mM MgCl₂, and 10 µM EcPsuG, was incubated at 30 °C for 2 h. For Ψ synthesis, PhoA (10 µM, final concentration) was extra added when the reaction reaches the equilibrium state for ΨMP synthesis.

For the cascade with uracil and AMP as starter substrates, 50 µL reaction (50 mM Tris-HCl buffer, pH 8.0) for ΨMP synthesis, containing 1 mM uracil, 1 mM AMP, 1 mM MgCl₂, 10 µM EcPsuG, and 10 µM AgmA, was incubated at 30 °C for 2 h. For Ψ synthesis, PhoA (10 µM, final concentration) was extra added when the reaction reaches the equilibrium state for ΨMP synthesis.

For the cascade with CMP as starter substrate, 50 µL reaction (50 mM Tris-HCl buffer, pH 8.0) for ΨMP synthesis, containing 1 mM CMP, 1 mM MgCl₂, 10 µM EcPsuG, 10 µM SgGouE and 1 µM CodA, was incubated at 30 °C for 2 h. For Ψ synthesis, PhoA (10 µM, final concentration) was added when the reaction reaches the equilibrium state for ΨMP synthesis.

For the cascade with UMP as starter substrate, 50 µL reaction (50 mM Tris-HCl buffer, pH 8.0) for ΨMP synthesis, containing 1 mM UMP, 1 mM MgCl₂, 10 µM EcPsuG, and 1 µM NmYgdH, was incubated at 30 °C for 2 h. For Ψ synthesis, 10 µM PhoA was extra added when the reaction reaches the equilibrium state for ΨMP synthesis.

Synthesis of ΨMP and Ψ in 1 g/10 g scale

For the synthesis of ΨMP and Ψ using UMP as the substrate in 10 g scale, 28 mL reaction (20 mM potassium phosphate buffer, pH 8.0), containing 10 g UMP (1 M), 1 mM MgCl₂, 1.75 mg·mL⁻¹ EcPsuG (50 µM), and 0.25 mg·mL⁻¹ NmYgdH (10 µM), was performed using a bottom flask placed in a water bath (37 °C). For Ψ synthesis, YjjG (50 µM, final concertation) was extra added for the dephosphorylation of ΨMP.

For the synthesis of ΨMP and Ψ using CMP as substrate in 10 g scale, 136 mL reaction (20 mM potassium phosphate buffer, pH 8.0), containing 10 g CMP (0.2 M), 1 mM MgCl₂, 0.7 mg·mL⁻¹ EcPsuG (20 µM), 0.14 mg·mL⁻¹ CodA (0.3 µM) and 1.35 mg·mL⁻¹ SgGouE (50 µM) were performed using a bottom flask placed in a water bath (30 °C). For Ψ synthesis, YjjG (20 µM, final concertation) was extra added for the dephosphorylation of ΨMP. Synthesis of ΨMP and Ψ in 1 g scale is described in Supplementary Method 8.

General method for the construction of in-frame deletion mutants and chromosomal integration of the target gene

The homologous recombination technique was used to construct the mutants, and the detailed process was described as follows using CY1ΔpsuK (CY2) as an example. First, the homologous regions were individually amplified from E. coli BL21(DE3) using the primers ΔpsuK-LarmF/R (left arm) and ΔpsuK-RarmF/R (right arm). The two amplified segments were purified and then ligated into the SalI-BamHI sites of pKOV-Kan to generate the plasmid pCY02, which was then introduced into the E. coli CY1. After that, the target psuK in-frame deletion mutant (CY2), confirmed by both PCR and sequencing analysis, was screened according to the method by Maria et al. (Table S1 and S2)⁵⁹. Following the similar protocols, a series of mutants, as shown in Table S3, were also constructed, and validated. Chromosomal integration of the target gene is detailed in Supplementary Method 9. The construction of the expression vector is described in Supplementary Method 10.

Ψ production by shake flask fermentation

The recombinant strains harboring related expression vectors were cultivated in Erlenmeyer flasks at 37 °C in Luria-Bertani medium for about 10 h to achieve comparable OD₆₀₀ values. Followed by inoculating 5 mL of fresh cells into 50 mL (50/500) fermentation medium⁶⁰ (per liter containing glycerol 60 g, MgSO₄ 0.4 g, CaCO₃ 10 g, yeast extract 10 g, soy peptone 14.8 g, and 10 mL of 100 × trace elements) supplemented with 0.3 mM IPTG, 50 mg·L⁻¹ antibiotics, the fermentation cultures were then incubated on a shaker at 37 °C for 72 h. As for the culture utilizing glucose as carbon source, 5 mL of fresh cells were inoculated into 50 mL (50/500) fermentation medium⁶¹ (per liter containing glucose 20 g, yeast extract 4 g, peptone 3 g, sodium citrate 2 g, KH₂PO₄ 2 g, MgSO₄·7H₂O 2 g, and trace elements), which was supplemented with 50 mg·L⁻¹ antibiotics, and then the fermentation cultures were incubated on a shaker at 37 °C for 72 h.

Ψ production by fed-batch fermentation in a 5 L bioreactor

For the fermentation in a 5 L bioreactor, the seed culture (Luria-Bertani medium) and fed-batch culture medium were identical to those used in the shaking flask fermentation. For fed-batch fermentation, 250 mL of seed culture was inoculated into a 5 L bioreactor (Bailun bio, Shanghai, China) containing 2.5 L fermentation medium. The temperature was maintained at 37 ± 0.2 °C. Dissolved oxygen (DO) was detected using a DO analyzer (Hamilton Co., Ltd.), and regulated to 30% ± 5% by adjusting the stirring speed. The pH of the culture broths, measured using a pH meter (Hamilton Co., Ltd.), was maintained at around 7.0 ± 0.2, which was adjusted by addition of NH₄OH (40%, v/v) or citric acid (10%, m/v). In the fermentation process (96 h), the feeding nutrition was batch-added to the culture at an appropriate rate when the dissolved oxygen level exceeds 30%.

For the glycerol feeding medium, each liter comprises 600 mL of glycerol, 25 g of yeast extract, and 40 g of soy peptone. For the glucose feeding medium, each liter consists of 600 g of glucose, 20 g of yeast extract, and 20 g of peptone.

Purification of ΨMP and Ψ from the reaction

For purifying ΨMP from the reaction mixtures, the enzymes were removed by Amicon Ultra filters (10 kDa). The filtrate was lyophilized by freeze drier to give the ΨMP solid with technical-grade purity (>95%), which was analyzed by HPLC and NMR. For Ψ purification, the supernatant was cooled to 4 °C overnight for maximizing Ψ crystallization. The Ψ crystallized product, lyophilized by freeze drier, was further analyzed by HPLC and NMR. Purification of Ψ from the fermentation cultures is detailed in Supplementary Methods 11 and 12.

Process design for Ψ biomanufacturing

The schematics depicting the Ψ production from glucose, and the schematic process design is shown in Fig. 7a (detailed process design can be found in Supplementary Methods). This investigation necessitated the development of a production pathway for Ψ to perform mass and energy balance estimations, facilitated by Aspen Plus version 14 software (Supplementary Tables 7–12). The Ψ biomanufacturing process has five subprocesses: Ψ production (A100), Ψ purification (A200), wastewater treatment (A300), and utilities (A400). The principal assumptions of whole process are illustrated in Supplementary Table 3. The inlet flow rate of glucose was 110 kg·h⁻¹, with corresponding additions of nitrogen sources, specifically yeast extract and peptone, at ~6.37 kg·h⁻¹ and 5.67 kg·h⁻¹, respectively. More details are shown in Supplementary Method 13.

Techno-economic analysis for Ψ biomanufacturing

Techno-economic analysis is an extensive methodology employed to assess the viability of Ψ generation from glucose, considering both economic and technical aspects. CAPEX and OPEX estimations were conducted utilizing an internal model established by NREL⁶², grounded in process design and rigorous materials and energy balance calculations. The Supplementary Methods include a comprehensive explanation of the CAPEX and OPEX calculation methodologies. The inside battery limit (ISBL) is commonly employed in TEA for essential factors concerning warehouse investments, site development, and additional piping. This study demonstrates that the ISBL includes the expenses related to equipment and installation for Ψ production and Ψ purification (Fig. 7a). Supplementary Table 4 illustrates the essential assumptions and key parameters for the process scale-up, whereas the Supplementary Note presents a detailed explanation of Supplementary Table 4, including plant financials and construction timeline, depreciation and economic boundaries, operating conditions, cost breakdowns, fixed operating costs, and financial metrics^45,46,63.

The biorefinery is anticipated to achieve an annual Ψ production capacity of 200 tonnes with a 30-year operational lifespan, considering numerous affecting elements such as equipment dimensions, raw material charges, and associated costs. The expenses related to equipment and raw materials employed in the assessment of CAPEX and OPEX were derived from public literature and official sources (Supplementary Table 13). The expenses related to equipment fluctuate based on the scale of implementation and adhere to a power law in relation to the initial prices⁴⁶. The TEA operates as a “n^th-plant” model, proficient in calculating the investment necessary for the pre-commercial process as well⁴⁶. The discounted cash flow method facilitates the identification of the MSP of Ψ at the point where the NPV (net present value) is zero and the IRR (internal rate of return) is 10%. More details are provided in Supplementary Method 14 and 15.

Statistical analysis

The number of biological replicates is specified in the figure legends. Data analysis was performed using OriginPro statistical software (version 2025, OriginLab Corporation, Northampton, MA, USA). One-way analysis of variance (ANOVA) was used for comparison, and p ≤ 0.05 was considered statistically significant.

Reporting summary

Further information on research design is available in the Nature Portfolio Reporting Summary linked to this article.

Data availability

The data supporting the results are provided in the article and its Supplementary Information. Atomic coordinates and structure factors of NmYgdH have been deposited in the Protein Data Bank (www.rscb.org) under the accession code 9US2. Source data are provided with this paper.

References

Spenkuch, F., Motorin, Y. & Helm, M. Pseudouridine: Still mysterious, but never a fake (uridine)!. RNA Biol. 11, 1540–1554 (2014).
Article PubMed Google Scholar
Liu, W.-W. et al. RNA modifications in cellular metabolism: Implications for metabolism-targeted therapy and immunotherapy. Signal Transduct. Target. Ther. 9, 70 (2024).
Article PubMed PubMed Central Google Scholar
Charette, M. & Gray, M. W. Pseudouridine in RNA: What, where, how, and why. IUBMB Life 49, 341–351 (2000).
Article CAS PubMed Google Scholar
Durairaj, A. & Limbach, P. A. Mass spectrometry of the fifth nucleoside. Anal. Chim. Acta 623, 117–125 (2008).
Article CAS PubMed PubMed Central Google Scholar
Anderson, B. R. et al. Incorporation of pseudouridine into mRNA enhances translation by diminishing PKR activation. Nucleic Acids Res. 38, 5884–5892 (2010).
Article CAS PubMed PubMed Central Google Scholar
Kierzek, E. et al. The contribution of pseudouridine to stabilities and structure of RNAs. Nucleic Acids Res. 42, 3492–3501 (2014).
Article CAS PubMed Google Scholar
Karikó, K., Buckstein, M., Ni, H. & Weissman, D. Suppression of RNA recognition by toll-like receptors: the impact of nucleoside modification and the evolutionary origin of RNA. Immunity 23, 165–175 (2005).
Article PubMed Google Scholar
Lamb, Y. N. BNT162b2 mRNA COVID-19 vaccine: first approval. Drugs 81, 495–501 (2021).
Article CAS PubMed PubMed Central Google Scholar
Baden, L. R. et al. Efficacy and safety of the mRNA-1273 SARS-CoV-2 vaccine. N. Engl. J. Med. https://doi.org/10.1056/NEJMoa2035389 (2020).
Koeberl, D. et al. Interim analyses of a first-in-human phase 1/2 mRNA trial for propionic acidaemia. Nature 628, 872–877 (2024).
Article CAS PubMed PubMed Central Google Scholar
Dolgin, E. The tangled history of mRNA vaccines. Nature 597, 318–324 (2021).
Article CAS PubMed Google Scholar
Shapiro, R. & Chambers, R. W. Synthesis of pseudouridine. J. Am. Chem. Soc. 83, 3920–3921 (1961).
Article CAS Google Scholar
Preumont, A., Snoussi, K., Stroobant, V., Collet, J.-F. & Schaftingen, E. V. Molecular identification of pseudouridine-metabolizing enzymes*. J. Biol. Chem. 283, 25238–25246 (2008).
Article CAS PubMed Google Scholar
Riley, A. T., Sanford, T. C., Woodard, A. M., Clerc, E. P. & Sumita, M. Semi-enzymatic synthesis of pseudouridine. Bioorg. Med. Chem. Lett. 44, 128105 (2021).
Article CAS PubMed Google Scholar
Pfeiffer, M. & Nidetzky, B. Reverse C-glycosidase reaction provides C-nucleotide building blocks of xenobiotic nucleic acids. Nat. Commun. 11, 6270 (2020).
Article CAS PubMed PubMed Central Google Scholar
Pfeiffer, M., Ribar, A. & Nidetzky, B. A selective and atom-economic rearrangement of uridine by cascade biocatalysis for production of pseudouridine. Nat. Commun. 14, 2261 (2023).
Article CAS PubMed PubMed Central Google Scholar
Zhou, M., Tang, R., Wei, L., Wang, J. & Qi, H. Metabolic engineering of Escherichia coli for efficient production of pseudouridine. ACS Omega 8, 36386–36392 (2023).
Article CAS PubMed PubMed Central Google Scholar
Zhang, C. et al. Systematic engineering of Escherichia coli for efficient production of pseudouridine from glucose and uracil. ACS Synth. Biol. 13, 1303–1311 (2024).
Article CAS PubMed Google Scholar
Pfeiffer, M. & Nidetzky, B. C-ribosylating enzymes in the (bio)synthesis of C-nucleosides and C-glycosylated natural products. ACS Catal. 13, 15910–15938 (2023).
Article CAS Google Scholar
Yu, L. et al. Efficient biosynthesis of nucleoside cytokinin angustmycin a containing an unusual sugar system. Nat. Commun. 12, 6633 (2021).
Article CAS PubMed PubMed Central Google Scholar
Grochowski, L. L. & Zabriskie, T. M. Characterization of BlsM, a nucleotide hydrolase involved in cytosine production for the biosynthesis of blasticidin S. ChemBioChem 7, 957–964 (2006).
Article CAS PubMed Google Scholar
Kang, M. et al. Solution structure of the nucleotide hydrolase BlsM: Implication of its substrate specificity. Protein Sci. 29, 1760–1773 (2020).
Article CAS PubMed PubMed Central Google Scholar
Cone, M. C., Yin, X., Grochowski, L. L., Parker, M. R. & Zabriskie, T. M. The blasticidin S biosynthesis gene cluster from Streptomyces griseochromogenes: Sequence analysis, organization, and initial characterization. ChemBioChem 4, 821–828 (2003).
Article CAS PubMed Google Scholar
Niu, G., Li, L., Wei, J. & Tan, H. Cloning, heterologous expression, and characterization of the gene cluster required for gougerotin biosynthesis. Chem. Biol. 20, 34–44 (2013).
Article CAS PubMed Google Scholar
Porter, D. J. & Austin, E. A. Cytosine deaminase. The roles of divalent metal ions in catalysis. J. Biol. Chem. 268, 24005–24011 (1993).
Article CAS PubMed Google Scholar
Tamura, K., Stecher, G. & Kumar, S. MEGA11: Molecular evolutionary genetics analysis version 11. Mol. Biol. Evol. 38, 3022–3027 (2021).
Article CAS PubMed PubMed Central Google Scholar
Bærentsen, R. L., Brodersen, D. E. & Zhang, Y. E. Evolution of the bacterial nucleosidase PpnN and its relation to the stringent response. Microb. Cell Graz Austria 6, 450–453 (2019).
Article Google Scholar
Sévin, D. C., Fuhrer, T., Zamboni, N. & Sauer, U. Nontargeted in vitro metabolomics for high-throughput identification of novel enzymes in Escherichia coli. Nat. Methods 14, 187–194 (2017).
Article PubMed Google Scholar
Zhang, Y. E. et al. (p)ppGpp regulates a bacterial nucleosidase by an allosteric two-domain switch. Mol. Cell 74, 1239–1249.e4 (2019).
Article CAS PubMed Google Scholar
Seo, H. & Kim, K.-J. Structural and biochemical characterization of the type-II LOG protein from Streptomyces coelicolor A3. Biochem. Biophys. Res. Commun. 499, 577–583 (2018).
Article CAS PubMed Google Scholar
Morris, G. M. et al. AutoDock4 and AutoDockTools4: Automated docking with selective receptor flexibility. J. Comput. Chem. 30, 2785–2791 (2009).
Article CAS PubMed PubMed Central Google Scholar
Seo, H. & Kim, K.-J. Structural insight into molecular mechanism of cytokinin activating protein from Pseudomonas aeruginosa PAO1. Environ. Microbiol. 20, 3214–3223 (2018).
Article CAS PubMed Google Scholar
Orhanović, S. & Pavela-Vrančič, M. Dimer asymmetry and the catalytic cycle of alkaline phosphatase from Escherichia coli. Eur. J. Biochem. 270, 4356–4364 (2003).
Article PubMed Google Scholar
Orhanović, S., Bučević-Popović, V., Pavela-Vrančič, M., Vujaklija, D. & Gamulin, V. Effect of a T81A mutation at the subunit interface on catalytic properties of alkaline phosphatase from Escherichia coli. Int. J. Biol. Macromol. 40, 54–58 (2006).
Article PubMed Google Scholar
Yu, L. et al. Developing the E. coli platform for efficient production of UMP-derived chemicals. Metab. Eng. 83, 61–74 (2024).
Article CAS PubMed Google Scholar
Patching, S. G. et al. The nucleoside transport proteins, NupC and NupG, from Escherichia coli: Specific structural motifs necessary for the binding of ligands. Org. Biomol. Chem. 3, 462–470 (2005).
Article CAS PubMed Google Scholar
Kong, L. et al. Divergent biosynthesis of C-nucleoside minimycin and indigoidine in bacteria. iScience 22, 430–440 (2019).
Article CAS PubMed PubMed Central Google Scholar
Palmu, K. et al. Discovery of the showdomycin gene cluster from Streptomyces showdoensis ATCC 15227 yields insight into the biosynthetic logic of C-nucleoside antibiotics. ACS Chem. Biol. 12, 1472–1477 (2017).
Article CAS PubMed Google Scholar
Maffioli, S. I., Sosio, M., Ebright, R. H. & Donadio, S. Discovery, properties, and biosynthesis of pseudouridimycin, an antibacterial nucleoside-analog inhibitor of bacterial RNA polymerase. J. Ind. Microbiol. Biotechnol. 46, 335–343 (2019).
Article CAS PubMed Google Scholar
Preumont, A., Rzem, R., Vertommen, D. & Van Schaftingen, E. HDHD1, which is often deleted in X-linked ichthyosis, encodes a pseudouridine-5′-phosphatase. Biochem. J. 431, 237–244 (2010).
Article CAS PubMed Google Scholar
Kim, K.-S. et al. The Rut pathway for pyrimidine degradation: Novel chemistry and toxicity problems. J. Bacteriol. 192, 4089–4102 (2010).
Article CAS PubMed PubMed Central Google Scholar
Andersen, P. S., Smith, J. M. & Mygind, B. Characterization of the upp gene encoding uracil phosphoribosyltransferase of Escherichia coli K12. Eur. J. Biochem. 204, 51–56 (1992).
Article CAS PubMed Google Scholar
Yan, Q. & Fong, S. S. Study of in vitro transcriptional binding effects and noise using constitutive promoters combined with UP element sequences in Escherichia coli. J. Biol. Eng. 11, 33 (2017).
Article PubMed PubMed Central Google Scholar
Schuster, S., Fell, D. A. & Dandekar, T. A general definition of metabolic pathways useful for systematic organization and analysis of complex metabolic networks. Nat. Biotechnol. 18, 326–332 (2000).
Article CAS PubMed Google Scholar
Yan, X. et al. Paradigm of engineering recalcitrant non-model microorganism with dominant metabolic pathway as a biorefinery chassis. Nat. Commun. 15, 10441 (2024).
Article CAS PubMed PubMed Central Google Scholar
Zhang, C. et al. An upcycling bioprocess for sustainable aviation fuel production from food waste-derived greenhouse gases: Life cycle assessment and techno-economic analysis. Chem. Eng. J. 486, 150242 (2024).
Article CAS Google Scholar
Davis, R., Wiatrowski, M., Kinchin, C. & Humbird, D. Conceptual basis and techno-economic modeling for integrated algal biorefinery conversion of microalgae to fuels and products (2019 NREL TEA update: Highlighting paths to future cost goals via a new pathway for combined algal processing). https://doi.org/10.2172/1665822 (2020).
Choi, K. R., Luo, Z. W., Kim, G. B., Xu, H. & Lee, S. Y. A microbial process for the production of benzyl acetate. Nat. Chem. Eng. 1, 216–228 (2024).
Article Google Scholar
El Abbadi, S. H., Sherwin, E. D., Brandt, A. R., Luby, S. P. & Criddle, C. S. Displacing fishmeal with protein derived from stranded methane. Nat. Sustain. 5, 47–56 (2022).
Article Google Scholar
Wiranarongkorn, K., Im-orb, K., Patcharavorachot, Y., Maréchal, F. & Arpornwichanop, A. Comparative techno-economic and energy analyses of integrated biorefinery processes of furfural and 5-hydroxymethylfurfural from biomass residue. Renew. Sustain. Energy Rev. 175, 113146 (2023).
Article CAS Google Scholar
Yu, C.-P., Chang, H.-Y. & Chien, T.-C. Total synthesis of pseudouridine via heck-type C-glycosylation. N. J. Chem. 43, 8796–8803 (2019).
Article CAS Google Scholar
Mitchell, S., Martín, A. J. & Pérez-Ramírez, J. Transcending scales in catalysis for sustainable development. Nat. Chem. Eng. 1, 13–15 (2024).
Article Google Scholar
Mou, J.-H. et al. Key techno-economic barriers in the valorization of organic solid wastes for chemical production. One Earth 7, 742–746 (2024).
Article Google Scholar
Guillén-Gosálbez, G. & Francisco dos Santos, L. Integrating industrial sectors in the transition to more sustainable chemicals. Nat. Chem. Eng. 1, 336–337 (2024).
Article Google Scholar
Guo, S. et al. Scalable electro-biosynthesis of ectoine from greenhouse gases. Angew. Chem. Int. Ed. 64, e202415445 (2025).
Article CAS Google Scholar
Zuiderveen, E. A. R. et al. The potential of emerging bio-based products to reduce environmental impacts. Nat. Commun. 14, 8521 (2023).
Article CAS PubMed PubMed Central Google Scholar
Pfeiffer, M., Guld, F. & Nidetzky, B. N- to C-glycoside rearrangement of uridine 5′-phosphate in two enzymatic steps for the production of pseudouridine 5′-phosphate. Biotechnol. Bioeng. https://doi.org/10.1002/bit.29037 (2025).
Sambrook, J., Fritsch, E. F. & Maniatis, T. Molecular Cloning: A Laboratory Manual. (Cold Spring Harbor Laboratory, 1989).
Lalioti, M. D. & Heath, J. K. A new method for generating point mutations in bacterial artificial chromosomes by homologous recombination in Escherichia coli. Nucleic Acids Res. 29, e14 (2001).
Article CAS PubMed PubMed Central Google Scholar
Lee, H. C., Kim, J. H., Kim, J. S., Jang, W. & Kim, S. Y. Fermentative production of thymidine by a metabolically engineered Escherichia coli strain. Appl. Environ. Microbiol. 75, 2423–2432 (2009).
Article CAS PubMed PubMed Central Google Scholar
Wu, H. et al. Metabolic engineering of Escherichia coli for high-yield uridine production. Metab. Eng. 49, 248–256 (2018).
Article CAS PubMed Google Scholar
Kobos, P. et al. Techno-economic analysis: Best practices and assessment tools. https://doi.org/10.2172/1738878 (Sandia National Lab, 2020).
Zhang, C. et al. Economic and sustainable revolution to facilitate one-carbon biomanufacturing. Nat. Commun. 16, 4896 (2025).
Article CAS PubMed PubMed Central Google Scholar

Download references

Acknowledgements

We are sincerely grateful to Prof. Shihui Yang, Prof. Zhengyu Zhang, and Prof. You-Sheng Cai for helpful discussion and related assistance with structural analysis. We’d also like to acknowledge the Core Facility of Wuhan University for LC-HRMS and NMR data collection. This work was supported by grants, the National Natural Science Foundation of China (32170026 and 32470047) to W.C., the Hubei Province’s Outstanding Medical Academic Leader Program (Chutian Talent Plan) to W.C., the Science Fund for Distinguished Young Scholars of Shaanxi Province (2022JC-09) to Q.F., and the China Postdoctoral Science Foundation (2025T056HB and GZC20251865) to L.Y.

Author information

These authors contributed equally: Le Yu, Ruyi Chen, Chenyue Zhang.

Authors and Affiliations

Department of Anesthesiology, Zhongnan Hospital of Wuhan University, TaiKang Center for Life and Medical Sciences, School of Pharmaceutical Sciences, Wuhan University, Wuhan, China
Le Yu, Ruyi Chen, Zhengyuan Wang, Zhuqing Wang, Xinyue Zeng, Han Liang, Yuanyuan He, Yixuan She, Yifei Wang, Rong Gong, Xuemin Song, Zixin Deng & Wenqing Chen
School of Chemical Engineering and Technology, Xi’an Jiaotong University, Xi’an, China
Chenyue Zhang & Qiang Fei
Key Laboratory of Combinatorial Biosynthesis and Drug Discovery, Ministry of Education, Wuhan University, Wuhan, China
Zixin Deng & Wenqing Chen

Authors

Le Yu
View author publications
Search author on:PubMed Google Scholar
Ruyi Chen
View author publications
Search author on:PubMed Google Scholar
Chenyue Zhang
View author publications
Search author on:PubMed Google Scholar
Zhengyuan Wang
View author publications
Search author on:PubMed Google Scholar
Zhuqing Wang
View author publications
Search author on:PubMed Google Scholar
Xinyue Zeng
View author publications
Search author on:PubMed Google Scholar
Han Liang
View author publications
Search author on:PubMed Google Scholar
Yuanyuan He
View author publications
Search author on:PubMed Google Scholar
Yixuan She
View author publications
Search author on:PubMed Google Scholar
Yifei Wang
View author publications
Search author on:PubMed Google Scholar
Rong Gong
View author publications
Search author on:PubMed Google Scholar
Xuemin Song
View author publications
Search author on:PubMed Google Scholar
Zixin Deng
View author publications
Search author on:PubMed Google Scholar
Qiang Fei
View author publications
Search author on:PubMed Google Scholar
Wenqing Chen
View author publications
Search author on:PubMed Google Scholar

Contributions

L.Y., R.C., and C.Z. contributed equally to this work. W.C. conceived the project and directed the research. L.Y., R.C., Z.W., Y.H., and Y.S. performed the genetic, biochemical, and fermentation experiments. C.Z. and Q.F. performed the techno-economic analysis. Z.W., X.Z., H.L., Y.W., and R.G. participated in the research. W.C., L.Y., R.C., C.Z., and Q.F. wrote the manuscript. X.S. and Z.D. made a critical reading of the manuscript.

Corresponding author

Correspondence to Wenqing Chen.

Ethics declarations

Competing interests

W.C., L.Y., R.C., Z.D. and Zheng.W. have submitted two patent applications filed by Wuhan University with the China National Intellectual Property Administration (application numbers CN202510665627.7 and CN202510665631.3), as well as one related PCT patent application (international application number PCT/CN2025/105204), all of which pertain to the NmYgdH application and pseudouridine production described in this work. The remaining authors declare no competing interests.

Peer review

Peer review information

Nature Communications thanks Cong Gao, Robert Schnell and the other, anonymous, reviewer(s) for their contribution to the peer review of this work. A peer review file is available.

Additional information

Publisher’s note Springer Nature remains neutral with regard to jurisdictional claims in published maps and institutional affiliations.

Supplementary information

Supplementary Information

Peer Review File

Description of Additional Supplementary Files

Supplementary Data 1

Supplementary Data 2

Supplementary Data 3

Supplementary Data 4

Reporting Summary

Source data

Source Data

Rights and permissions

Open Access This article is licensed under a Creative Commons Attribution-NonCommercial-NoDerivatives 4.0 International License, which permits any non-commercial use, sharing, distribution and reproduction in any medium or format, as long as you give appropriate credit to the original author(s) and the source, provide a link to the Creative Commons licence, and indicate if you modified the licensed material. You do not have permission under this licence to share adapted material derived from this article or parts of it. The images or other third party material in this article are included in the article’s Creative Commons licence, unless indicated otherwise in a credit line to the material. If material is not included in the article’s Creative Commons licence and your intended use is not permitted by statutory regulation or exceeds the permitted use, you will need to obtain permission directly from the copyright holder. To view a copy of this licence, visit http://creativecommons.org/licenses/by-nc-nd/4.0/.

Reprints and permissions

About this article

Cite this article

Yu, L., Chen, R., Zhang, C. et al. A combined pseudouridine biomanufacturing platform enabled by a streamlined designer pathway. Nat Commun 16, 8866 (2025). https://doi.org/10.1038/s41467-025-63906-0

Download citation

Received: 08 March 2025
Accepted: 01 September 2025
Published: 06 October 2025
DOI: https://doi.org/10.1038/s41467-025-63906-0