De Novo Biosynthesis of Valinomycin From Glucose Using In Vitro Reconstituted Hybrid Pathways

Abstract

In vitro biotransformation mediated by cell‐free biosynthetic systems provides a flexible biomanufacturing platform that enables the reconstitution of hybrid metabolic pathways for complex natural product biosynthesis from simple substrates. Here, we report the total biosynthesis of valinomycin from glucose using in vitro reconstituted hybrid pathways consisting of three catalytic modules. First, a four‐enzyme short glycolytic pathway was integrated to convert glucose to pyruvate, one of the key precursors of valinomycin. Subsequently, a second pathway for L‐valine biosynthesis catalyzes the conversion of pyruvate to α‐ketoisovalerate and L‐valine, which serve as the other two precursors. Finally, de novo biosynthesis of valinomycin is achieved through the third module, valinomycin synthetase, which assembles the three precursors generated from the first two modules. Overall, we demonstrate the successful reconstitution of in vitro hybrid pathways for valinomycin biosynthesis using glucose as the sole input substrate. This synthetic strategy provides a modular framework for designing easy‐to‐use enzymatic pathways to produce value‐added complex natural products from simple and inexpensive substrates.

A modular cell‐free biosynthetic system enables de novo production of the nonribosomal peptide valinomycin from glucose. By integrating a four‐enzyme short glycolytic pathway, an L‐valine biosynthetic module, and valinomycin synthetase (Vlm1/Vlm2), three key precursors are generated and assembled in vitro, demonstrating hybrid pathway design for complex natural product biosynthesis from simple substrates.

1. Introduction

Valinomycin is a nonribosomal peptide (NRP) natural product produced mainly by various Streptomyces strains and exhibits a wide range of biological activities, including antibacterial, antifungal, antiviral, and antitumor properties [1, 5]. Structurally, valinomycin is a 36‐membered cyclododecadepsipeptide synthesized by two large nonribosomal peptide synthetases (NRPSs), Vlm1 (370 kDa) and Vlm2 (284 kDa) [1, 6, 7]. Similar to other NRPS systems [8, 9], Vlm1 and Vlm2 function sequentially as assembly lines that incorporate three precursors—pyruvate, α‐ketoisovalerate (α‐Kiv), and L‐valine—to construct the valinomycin molecule [6]. Because these precursors originate from the primary metabolism (i.e., the pathway from glucose to L‐valine), we previously achieved heterologous biosynthesis of valinomycin by reconstituting the vlm1 and vlm2 genes in E. coli [5, 10]. Subsequently, valinomycin production in E. coli was systematically improved through enzyme engineering, strain engineering, and bioprocess optimization [11, 13]. However, due to the cellular toxicity of valinomycin, the overall production yield in E. coli remained modest. To overcome this limitation, we previously reconstituted the valinomycin biosynthetic pathway in vitro by coupling cell‐free protein synthesis (CFPS) and cell‐free metabolic engineering (CFME), enabling high‐yield production of valinomycin [14].

Although the total in vitro biosynthesis of valinomycin using native E. coli metabolites has been successfully demonstrated, generation of the precursor pyruvate from glucose requires 10 enzymatic steps through glycolysis. In addition, four more enzymatic reactions are required to produce the other two precursors (α‐Kiv and L‐valine), resulting in a total of 14 catalytic steps to generate the three building blocks needed for valinomycin biosynthesis. Such pathway complexity presents a significant challenge for efficient precursor generation and final product formation in vitro. Therefore, we ask whether this long pathway could be shortened.

Previous studies have shown that a non‐phosphorylative Entner–Doudoroff (np‐ED) pathway derived from hyperthermophilic archaea provides a shortened glycolytic route consisting of only four enzymes [15, 16]. Due to their thermophilic nature, these enzymes can be readily purified from cell lysates by heat treatment and have been used to construct in vitro synthetic enzymatic systems for pyruvate production from glucose. However, to our knowledge, this four‐enzyme short glycolytic pathway has not yet been applied to the in vitro biosynthesis of complex natural products such as valinomycin.

To address this opportunity, we integrated this nature‐derived short glycolytic pathway into the valinomycin biosynthetic route by coupling three metabolic modules (Figure 1). In Module 1, pyruvate is generated from glucose via the short glycolytic pathway consisting of four thermophilic enzymes. Module 2 is derived from the native E. coli L‐valine biosynthetic pathway and converts pyruvate into α‐Kiv and L‐valine through four enzymatic reactions. The final Module 3 contains two large NRPS enzymes, Vlm1 and Vlm2, responsible for assembling the three precursors into valinomycin. By combining these 10 enzymes, we successfully reconstituted an in vitro hybrid pathway for the de novo biosynthesis of valinomycin using glucose as the sole input substrate. Furthermore, valinomycin synthesis can be achieved using different reaction formats, demonstrating the flexibility and efficiency of modular enzymatic pathways for complex natural product biosynthesis. This study provides a promising example for the in vitro production of value‐added compounds starting from glucose or pyruvate‐derived metabolites.

FIGURE 1.

Schematic representation of in vitro hybrid enzymatic pathways for valinomycin biosynthesis from glucose. In Module 1, glucose is converted to pyruvates through a non‐phosphorylative Entner–Doudoroff (np‐ED) pathway, consisting of four enzymes. Module 2 is the L‐valine biosynthetic pathway that converts pyruvate to α‐ketoisovalerate and L‐valine by four‐enzyme‐based sequential reactions. In Module 3, Vlm1 and Vlm2 catalyze the assembly of three precursors into valinomycin. AHAIR, acetohydroxy acid isomeroreductase (encoded by ilvC); AHAS I, acetohydroxy acid synthase I (encoded by ilvB/N); ALDH, glyceraldehyde dehydrogenase; BCAT, branched‐chain amino acid aminotransferase (encoded by ilvE); DHAD, dihydroxy acid dehydratase (encoded by ilvD); DHT, dehydratase; GDH, glucose dehydrogenase; KDGA, 2‐keto‐3‐deoxygluconate aldolase; Vlm1, valinomycin synthetase 1; Vlm2, valinomycin synthetase 2.

2. Results and Discussion

2.1. Design and Reconstitution of an In Vitro Hybrid Pathway for Valinomycin Biosynthesis

To achieve de novo biosynthesis of valinomycin from glucose in vitro, a key prerequisite is the generation of three precursors (pyruvate, α‐Kiv, and L‐valine) for the valinomycin synthetase enzymes Vlm1 and Vlm 2. We addressed this requirement by designing a hybrid metabolic route to generate these precursors. To minimize the number of catalytic steps required for glycolysis, we selected a short four‐enzyme pathway to convert glucose into pyruvate (Module 1, Figure 1). The enzymes involved include glucose dehydrogenase (GDH) from Sulfolobus solfataricus [15], dehydratase (DHT) from Paralcaligenes ureilyticus [17], 2‐keto‐3‐deoxygluconate aldolase (KDGA) from Sulfolobus acidocaldarius [15], and glyceraldehyde dehydrogenase (ALDH) from Thermoplasma acidophilum [15]. Notably, these enzymes are thermophilic, allowing their isolation from cell lysates through simple heat precipitation.

To generate α‐Kiv and L‐valine, we employed the L‐valine biosynthetic pathway from E. coli, which also consists of four enzymes: acetohydroxy acid synthase I (AHAS I), acetohydroxy acid isomeroreductase (AHAIR), dihydroxy acid dehydratase (DHAD), and branched‐chain amino acid aminotransferase (BCAT) [18]. In this pathway, AHAS I, AHAIR, and DHAD sequentially convert pyruvate to α‐Kiv, which is then transformed into L‐valine by BCAT (Module 2, Figure 1).

Once the three precursors are generated in the reaction system, Vlm1 and Vlm2 from Streptomyces tsusimaensis assemble them into the final product valinomycin (Module 3, Figure 1). Overall, this in vitro enzymatic biosystem comprises 10 enzymes derived from six microbial species and reconstitutes a hybrid biosynthetic pathway for the de novo production of valinomycin using glucose as a single input substrate.

2.2. Enzyme Preparation and Activity Assay

All enzymes were expressed and purified according to their biochemical properties. Enzymes used in Modules 1 and 2 were expressed in E. coli BL21 Star (DE3), whereas Vlm1 and Vlm2 were expressed in E. coli BAP1 as previously described [14]. The genome of BAP1 contains an integrated sfp gene encoding a phosphopantetheinyl transferase responsible for post‐translational phosphopantetheinylation and activation of apo‐NRPS enzymes (here are Vlm1 and Vlm2) [19]. After expression, the thermally stable enzymes in Module 1 were purified by heat treatment to remove endogenous proteins. The remaining enzymes were purified using nickel affinity chromatography followed by size exclusion chromatography and analyzed by sodium dodecyl sulfate‐polyacrylamide gel electrophoresis (SDS‐PAGE).

The results showed that enzymes involved in Modules 1 and 2 were purified to near homogeneity with the expected molecular weights (Figure 2). Vlm1 and Vlm2 were also successfully purified, although their purity was relatively lower due to their large molecular sizes. Enzyme activity assays confirmed that all enzymes in Modules 1 and 2 were active toward their respective substrates (Table 1 and Figure S1). Notably, the thermophilic enzymes GDH, DHT, KDGA, and ALDH involved in glycolysis were tested at 50°C. The specific activities of DHT toward the substrates gluconate and glycerate were determined to be 1.21 and 3.14 U/mg, respectively. Enzymatic reactions from the L‐valine biosynthetic pathway were conducted at 30°C. AHAS consists of two subunits (ilvB and ilvN), both of which are active toward pyruvate. The activities of Vlm1 and Vlm2 were not measured individually because they function cooperatively to assemble the three precursors into valinomycin. Instead, their activity was verified by detecting valinomycin production in subsequent experiments.

FIGURE 2.

SDS‐PAGE analysis of purified enzymes. Analysis of GDH (42 kDa), DHT (62 kDa), KDGA (33 kDa), ALDH (57 kDa), AHAS I‐ilvB (60 kDa), AHAS I‐ilvN (11 kDa), AHAIR (54 kDa), DHAD (65 kDa), and BCAT (34 kDa) was performed with a 10% separation gel. Analysis of Vlm1 (370 kDa) and Vlm2 (284 kDa) was performed with a 4%–12% precast separation gel. Marker, protein marker (kDa).

TABLE 1.

Specific activity of purified enzymes.

Enzyme	Abbreviation	GenBank no.	Specific activity, U/mg
Glucose dehydrogenase	GDH	AAK43106.1	0.65
Dehydratase	DHT	WP_132585145.1	1.21/3.14a
2‐Keto‐3‐desoxy gluconate aldolase	KDGA	AAY79644.1	7.29
Glyceraldehyde dehydrogenase	ALDH	CAC11938.1	3.14
Acetohydroxy acid synthase I (ilvB)	AHAS I	CAQ34014.1	0.43
Acetohydroxy acid synthase I (ilvN)		CAQ34013.1	2.38
Acetohydroxy acid isomeroreductase	AHAIR	CAQ34118.1	0.36
Dihydroxy acid dehydratase	DHAD	CAQ34115.1	0.33
Branched‐chain amino acid aminotransferase	BCAT	CAQ34114.1	1.02

^a DHT exhibits specific activities of 1.21 and 3.14 U/mg toward gluconate and glycerate, respectively.

2.3. Evaluation of Valinomycin Precursor Synthesis

Having confirmed the enzymatic activities, we next sought to construct the metabolic pathways for glycolysis (Module 1) and L‐valine biosynthesis (Module 2). In the short glycolytic pathway (Figure 1), the enzyme dehydratase catalyzes two reactions: the conversion of gluconate to 2‐keto‐3‐deoxygluconate and the dehydration of glycerate to pyruvate. Previous studies have reported that a dehydratase from Sulfolobus solfataricus can catalyze both reactions, yet its dehydration activity is relatively low [15]. Therefore, we selected the dehydratase (DHT) from Paralcaligenes ureilyticus [17], which has been reported to exhibit high catalytic activity toward the conversion of glycerate to pyruvate. Given the key role of DHT in the glycolytic pathway, we first evaluated the effect of DHT loading on pyruvate production in Module 1. To do this, 30 mM glucose was used as the initial substrate, and 1 U/mL each of GDH, KDGA, and ALDH was added to the reaction mixture, while the concentration of DHT was varied from 0 to 4 U/mL. The results indicated that the addition of 2 U/mL DHT provided optimal pyruvate production under the tested conditions (Figure S2).

Next, we examined the time course of glycolysis at 50°C by monitoring glucose consumption and pyruvate formation. As shown in Figure 3a, the glucose concentration gradually decreased over time, with approximately 50% of the glucose consumed after 8 h of reaction. After 20 h, about 70% of the initial glucose had been converted into downstream metabolites, leaving approximately 9 mM glucose in the reaction mixture. Meanwhile, pyruvate—the end product of Module 1—was detected, reaching a final concentration of 4.6 mM after 20 h. These results suggest that a portion of the glucose was converted into intermediates rather than fully transformed into pyruvate.

FIGURE 3.

In vitro synthesis of valinomycin precursors from glucose. (a) Time courses of glucose consumption and pyruvate production. The reaction was performed at 50°C in 100 mM HEPES‐NaOH buffer (pH 7.0) containing 30 mM glucose, 5 mM NAD⁺, 5 mM MgCl₂, 1 U/mL GDH, 2 U/mL DHT, 1 U/mL KDGA, and 1 U/mL ALDH. (b) Time courses of pyruvate consumption and the formation α‐Kiv and L‐valine. The reaction was performed at 30°C in 100 mM HEPES‐NaOH buffer (pH 7.0) containing 20 mM pyruvate, 5 mM NADPH, 5 mM L‐glutamate, 5 mM MgCl₂, 1 U/mL AHAS I, 1 U/mL AHAIR, 1 U/mL DHAD, and 1 U/mL BCAT. Values show means with error bars representing standard deviations (S.D.) of at least three independent experiments.

We then constructed Module 2 by using 20 mM pyruvate as the starting substrate and adding 1 U/mL of each enzyme involved in the L‐valine biosynthetic pathway, followed by incubation at 30°C for 20 h. The results showed that approximately 50% of pyruvate was consumed after 12 h reaction, and ultimately, about 93% was converted into downstream products (Figure 3b). At the end of the reaction (20 h), the concentrations of α‐Kiv and L‐valine were determined to be 2.9 and 0.7 mM, respectively. Taken together, these results demonstrate that the three valinomycin precursors (pyruvate, α‐Kiv, and L‐valine) can be successfully produced from Modules 1 and 2, providing a key foundation for subsequent coupling with Module 3 for valinomycin biosynthesis.

2.4. De Novo Biosynthesis of Valinomycin From Glucose

After verifying the generation of all three precursors from Modules 1 and 2, we next coupled these modules with Module 3 to synthesize valinomycin. Because the four glycolytic enzymes are thermophilic with an optimal reaction temperature at 50°C, whereas the remaining enzymes are mesophilic with optimal activity at 30°C, we designed two reaction formats: (i) a one‐pot two‐step reaction and (ii) a one‐pot three‐step reaction (Figure 4a). In both formats, the first step involved glycolysis (Module 1), in which GDH, DHT, KDGA, and ALDH converted glucose into pyruvate at 50°C for 8 h. After completion of this step, enzymes from Modules 2 and 3 were added to the reaction mixture and incubated at 30°C. In the one‐pot two‐step format, enzymes from Modules 2 and 3 were introduced simultaneously, allowing concurrent formation of α‐Kiv and L‐valine as well as valinomycin synthesis. Alternatively, in the one‐pot three‐step format, enzymes from Module 2 were first added to generate α‐Kiv and L‐valine, followed by the addition of Vlm1 and Vlm2 from Module 3 to catalyze valinomycin biosynthesis. The total reaction time for both formats was 20 h. After completion of the reactions, valinomycin was extracted and analyzed by liquid chromatography‐mass spectrometry (LC‐MS).

FIGURE 4.

Total in vitro biosynthesis of valinomycin from glucose. (a) Design of two reaction formats for valinomycin biosynthesis. (b) Representative chromatogram of de novo biosynthesized valinomycin analyzed by LC‐MS/MS. Insert, MS/MS spectrum of valinomycin peak; EIC, extracted ion chromatogram. (c) Biosynthesis of valinomycin from two different reaction formats. Values show means with error bars representing standard deviations (S.D.) of at least three independent experiments.

The results showed that valinomycin was successfully synthesized using both reaction formats (Figure 4b,c). The one‐pot three‐step reaction produced valinomycin at a titer of 3 μg/L, which was approximately 1.5‐fold higher than that obtained from the one‐pot two‐step reaction format. This improvement is likely due to the more efficient conversion catalyzed by Module 2 when it operates independently, resulting in higher accumulation of α‐Kiv and L‐valine prior to the addition of Module 3. When sufficient amounts of all three precursors are available, the subsequent addition of Vlm1 and Vlm2 enables more efficient valinomycin assembly. Although the de novo biosynthesis of valinomycin from glucose using in vitro reconstituted hybrid pathways was successfully achieved, the overall yield obtained in this study remains lower than that reported for previous cell‐based and CFPS/CFME production systems [10, 14]. Therefore, further efforts are needed to improve valinomycin production, for example, through engineering key enzymes such as DHT to enhance catalytic activity, optimizing enzyme concentrations within individual modules, and balancing metabolic flux across the entire biosynthetic pathway.

Nevertheless, we demonstrate a feasible approach for in vitro valinomycin synthesis. Beyond such feasibility, this work highlights the potential of modular hybrid enzymatic systems for constructing streamlined metabolic routes to complex natural products. By integrating a shortened glycolytic module derived from thermophilic microorganisms with a native L‐valine biosynthetic pathway and a large NRPS assembly system, we significantly reduced the enzymatic complexity required to generate key precursors from glucose. While not investigated in the current study, the L‐valine biosynthetic pathway could potentially be replaced by thermophilic homologs or engineered thermostable variants [20]. Such modifications may further simplify enzyme purification through heat treatment and enable Modules 1 and 2 to operate under a common high‐temperature condition, thereby facilitating precursor generation for subsequent valinomycin biosynthesis in Module 3. The present modular design not only simplifies pathway construction but also enables flexible control over individual reaction steps, allowing pathway optimization without the constraints imposed by cellular metabolism. Importantly, the successful coupling of enzymes originating from multiple organisms further demonstrates the compatibility and versatility of hybrid in vitro biosynthetic systems. Such strategies provide a promising framework for the bottom‐up assembly of synthetic metabolic networks and may facilitate the production of diverse nonribosomal peptides and other complex natural products directly from simple and inexpensive substrates.

3. Conclusion

In this study, we reconstituted a hybrid enzymatic framework consisting of three biosynthetic pathway modules for the de novo biosynthesis of valinomycin from glucose. Using a four‐enzyme short glycolytic pathway, approximately 70% of the supplied glucose was converted into downstream intermediates and the end product pyruvate. The second module efficiently transformed more than 90% of pyruvate into α‐Kiv and L‐valine through the L‐valine biosynthetic pathway. Finally, valinomycin synthetase (Vlm1 and Vlm2) assembled the three precursors (pyruvate, α‐Kiv, and L‐valine) into valinomycin, resulting in a final titer of 3 μg/L using the one‐pot three‐step reaction format. Overall, we successfully demonstrated the in vitro reconstitution of hybrid metabolic pathways for valinomycin biosynthesis using glucose as the sole input substrate, thereby achieving the main objective of this work. This study not only provides an alternative strategy for valinomycin production but also offers a general framework for designing in vitro modular hybrid biosynthetic pathways to produce other value‐added compounds of chemical and pharmaceutical significance, as demonstrated recently [21, 30].

4. Methods

4.1. Plasmid Construction

The plasmids pET28a‐Ssgdh, pET28a‐Sackdga, and pET28a‐Taaldh were kindly provided by Dr. Chun You [9]. The dht gene of Paralcaligenes ureilyticus was synthesized by GENEWIZ (Suzhou, China) and cloned into pET28a, yielding pET28a‐Pudht. The dht gene sequence is listed in Table S1. The genes of AHAS I (ilvB/N), AHAIR (ilvC), DHAD (ilvD), and BCAT (ilvE) were PCR amplified from the genome of E. coli BL21(DE3) and subsequently were cloned into pET28a, generating pET28a‐ilvB, pET28a‐ilvN, pET28a‐ilvC, pET28a‐ilvD, and pET28a‐ilvE, respectively. To express Vlm1 and Vlm2, pCTUT7‐vlm1 and pKS01‐vlm2 were used as described in our previous work [13, 14].

4.2. Strains and Culture Media

Plasmid construction and propagation were performed in E. coli DH5α. E. coli BAP1 (kindly provided by Dr. Yong Wang) was used for the expression of Vlm1 and Vlm2. E. coli BL21 Star (DE3) was used for all other protein expression. Luria–Bertani (LB) medium (10 g/L NaCl, 5 g/L yeast extract, and 10 g/L tryptone) with 50 μg/mL kanamycin was used for cell growth. Terrific broth (TB) medium (12 g/L tryptone, 24 g/L yeast extract, 0.4% glycerol, 2.31 g/L KH₂PO₄, and 12.54 g/L K₂HPO₄, pH 7.4) and 2 × YTPG (yeast extract 10 g/L, tryptone 16 g/L, NaCl 5 g/L, K₂HPO₄ 7 g/L, KH₂PO₄ 3 g/L, and glucose 18 g/L, pH 7.2) with 50 μg/mL kanamycin, 34 μg/mL chloramphenicol, or 100 μg/mL ampicillin were used for recombinant protein expression.

4.3. Enzyme Expression and Purification

The overnight‐precultured strains containing pCTUT7‐vlm1 or pKS01‐vlm2 were transferred into 1 L 2 × YPTG medium (with 34 μg/mL chloramphenicol for Vlm1 and 100 μg/mL ampicillin for Vlm2) for overexpression of Vlm1 and Vlm2 with an initial OD₆₀₀ of 0.05 at 30°C and 250 rpm. When OD₆₀₀ reached 0.6–0.8, cells were induced with 20 µM isopropyl‐β‐D‐thiogalactoside (IPTG) for 8  at 30°C and 250 rpm for protein expression. Similarly, the overnight‐precultured strains containing the other plasmids were transferred into 500 mL LB medium (100 μg/mL kanamycin) with an initial OD₆₀₀ of 0.05 at 37°C and 250 rpm. When OD₆₀₀ reached 0.6–0.8, cells were induced with 100 µM IPTG for 8 h at 30°C and 250 rpm for protein expression.

The cell pellets were collected after centrifugation at 5000 g and 4°C. After washing with 50 mM HEPES‐NaOH buffer (pH 7.5) twice, the pellets were then resuspended in the same buffer (1 g cell pellet per 10 mL buffer) for ultrasonication. To purify thermophilic enzymes (GDH, DHT, KDGA, and ALDH), the cell lysates were heated at 70°C (note that ALDH was heated at 60°C) for 20 min. For the other enzymes (AHAS I, AHAIR, DHAH, BCAT, Vlm1, and Vlm2), cell lysates were first centrifugated at 21,000 g and 4°C for 20 min. Then, each supernatant with target 6 × His‐tagged proteins was purified by Histrap column FF Crude 5 mL (GE Healthcares, USA), followed by AKTA (Cytiva, USA) purification if needed. The purity of the purified enzymes was examined by SDS‐PAGE.

4.4. Enzyme Activity Assay

The activity of GDH was assayed at 50°C by measuring the increase of NADH at 340 nm. The reaction system contained 50 mM glucose, 4 mM NAD⁺, and 5 mM MgCl₂ in 100 mM HEPES‐NaOH buffer (pH 7.0).

The activity of DHT was assayed at 50°C by measuring the reduction of gluconate. The reaction system contained 20 mM gluconate and 5 mM MgCl₂ in 100 mM HEPES‐NaOH buffer (pH 7.0).

The activity of DHT for pyruvate was assayed at 50°C by measuring the reduction of glyceraldehyde. The reaction system contained 20 mM glyceraldehyde, 4 mM NAD⁺, 5 mM MgCl₂, and 2 U/mL ALDH in 100 mM HEPES‐NaOH buffer (pH 7.0).

The activity of KDGA was assayed at 50°C by measuring the reduction of gluconate. The reaction system contained 20 mM D‐gluconate, 5 mM MgCl₂, and 2 U/mL DHT in 100 mM HEPES‐NaOH buffer (pH 7.0).

The activity of ALDH was assayed at 50°C by measuring the reduction of glyceraldehyde. The reaction system contained 20 mM glyceraldehyde, 4 mM NAD⁺, and 5 mM MgCl₂ in 100 mM HEPES‐NaOH buffer (pH 7.0).

The activity of AHAS I was assayed at 30°C by measuring the reduction of pyruvate. The reaction system contained 20 mM pyruvate and 5 mM MgCl₂ in 100 mM HEPES‐NaOH buffer (pH 7.0).

The activity of AHAIR was assayed at 30°C by measuring the reduction of NADPH at 450 nm by NADP⁺/NADPH assay kit (Beyotime, Shanghai, China). The reaction system contained 50 mM pyruvate, 2 mM NADPH, 2 U/L AHAS I, and 5 mM MgCl₂ in 100 mM HEPES‐NaOH buffer (pH 7.0).

The activity of DHAD was assayed at 30°C by measuring the increase of α‐ketoiosvalerate. The reaction system contained 50 mM pyruvate, 2 mM NADPH, 2 U/L AHAS I, 2 U/L AHAIR, and 5 mM MgCl₂ in 100 mM HEPES‐NaOH buffer (pH 7.0).

The activity of BCAT was assayed at 30°C by measuring the reduction of α‐ketoiosvalerate. The reaction system contained 20 mM α‐ketoiosvalerate, 10 mM L‐glutamate, and 5 mM MgCl₂ in 100 mM HEPES‐NaOH buffer (pH 7.0).

One unit of enzyme activity was defined as the amount of enzyme that released 1 μmol of product (or consumed 1 μmol of product) per minute. Each measurement was conducted in triplicate.

4.5. Valinomycin Synthesis and Intermediate Assay

The one‐pot reaction mixture of glycolysis pathway contained 100 mM HEPES‐NaOH buffer (pH 7.0), 30 mM glucose, 5 mM NAD⁺, and 5 mM MgCl₂, together with GDH, DHT, KDGA, and ALDH (each enzyme added 1 U/mL except DHT for 2 U/mL) at 50°C and 600 rpm. The one‐pot reaction mixture of L‐valine biosynthetic pathway consisted of 100 mM HEPES‐NaOH (pH 7.0), 5 mM L‐glutamate, 5 mM NADPH, and 5 mM MgCl₂ as well as AHAS I, AHAIR, DHAD, and BCAT (each enzyme added 1 U/mL). The one‐pot reaction mixture of valinomycin biosynthesis consisted of 100 mM HEPES‐NaOH (pH 7.0), 5 mM MgCl₂, 5 mM ATP, 1 mM CoA, 25 mM PEP, 4 µM Sfp (to further enhance the catalytic activity of the enzymes), and Vlm1 and Vlm2 (0.1 µM of each) at 30°C and 600 rpm.

After reaction, the mixture was transferred to a 10 kDa filter (Millipore, Germany) and centrifuged at 10,000 g for 5 min to terminate the reaction. The intermediates gluconate, glyceraldehyde, and α‐ketoiosvalerate were determined by high‐performance liquid chromatography (HPLC) (Agilent, USA) coupled to a RID detector with a Hi‐Plex H column (Agilent, USA) and 5.0 mM H₂SO₄ as the mobile phase. The intermediates pyruvate, L‐valine, and substrate L‐glutamate were detected via HPLC with DAD detector. A C18 column (Agilent, USA) was used for the amino acid detection after precolumn derivatization by 2,4‐dinitrofluorobenzene (DNFB) with a mobile phase of 70% A (50 mM C₂H₃NaO₂ with 1% N,N‐dimethylformamide) and 30% B (acetonitrile) at a flow rate of 1 mL/min. The concentration of glucose was determined by glucose assay kit (Beyotime, Shanghai, China). Valinomycin was prepared and detected as described in our previous work [8]. The one‐pot reaction mixture was extracted with threefold volumes of ethyl acetate and then centrifugated at 16,000 g for 5 min. The organic fraction with valinomycin was dried and resuspended in methanol for analysis by Agilent 6470 Triple Quadrupole LC‐MS System equipped with an Eclipse Plus C18 column. The separation was performed using mobile phase A (water + 0.1% formic acid) and mobile pahse B (acetonitrile + 0.1% formic acid) with a linear gradient from 80% to 100% B over 2.5 min, followed by a 7.5 min wash at 100% B and a 10 min re‐equilibration at 80% B at a flow rate of 0.4 mL/min. Each measurement was conducted in triplicate.

Funding

This work was supported by the National Key Research and Development Program of China (grant no. 2023YFA0914000), the National Natural Science Foundation of China (grant no. 32571664), the Shanghai Science and Technology Committee (grant no. 24ZR1451000), and the School of Physical Science and Technology of ShanghaiTech University (grant no. SPST‐YSFZ‐2024‐01).

Conflicts of Interest

The authors declare no conflicts of interest.

Supporting information

Supplementary Material

Data Availability Statement

The data that support the findings of this study are available on request from the corresponding author.