Product-Responsive Regulation of NADPH and Malonyl-CoA Allocation Enhances 3‑Hydroxypropionic Acid Production in Escherichia coli Authors: Yuchan Kim, Kyung Hyun Cho, Chandran Sathesh-Prabu, Sung Kuk Lee Published: 2026 ABSTRACT Efficient microbial production of platform chemicals requires precise regulation of metabolic fluxes. However, static control strategiesincluding constitutive overexpression and multi-inducer systems requiring manual optimizationcannot adapt in real time to changing intracellular states and therefore hinder fine-tuning of cofactor and precursor pools. Here, we constructed a product-responsive regulatory circuit in Escherichia coli that coordinates two key biosynthetic determinantsNADPH regeneration and malonyl-CoA allocationin response to intracellular 3-hydroxypropionic acid (3-HP) accumulation. Following initial IPTG induction of the 3-HP biosynthesis module, intracellular 3-HP accumulation activates the downstream product-responsive regulatory circuit, enabling coordinated modulation of NADPH regeneration and malonyl-CoA allocation without additional external intervention. The final engineered strain, SBH40, achieved a 3-HP titer of 3.0 g/L and a yield of 0.19 g/g glucose, representing an approximately 5-fold improvement in both metrics relative to a nonregulated control strain (0.64 g/L; 0.04 g/g glucose). These results demonstrate that a single-input product-responsive circuit can autonomously coordinate multiple metabolic nodes after pathway initiation, offering a practical and modular framework for dynamic regulation of cofactor- and precursor-intensive biosynthesis. Introduction Dynamic regulation of metabolic flux is increasingly recognized as a critical requirement for efficient microbial production of biochemicals. , Many biosynthetic pathways impose coupled demands on cellular cofactors and precursor metabolites, which vary with growth phase, pathway activity, and product accumulation. , Static engineering strategiessuch as constitutive overexpression or permanent gene deletionsare unable to adapt to these changing intracellular states, often resulting in metabolic imbalance, growth inhibition, and suboptimal product yields. , These limitations underscore the need for autonomous regulatory systems that dynamically coordinate multiple metabolic nodes in response to cellular and product-derived signals. Recent advances in synthetic biology have enabled dynamic control strategies that couple intracellular metabolites to transcriptional regulation via genetically encoded sensing modules, allowing automatic adjustment of pathway activity in response to physiological state and thereby improving robustness and efficiency. − However, most reported dynamic regulatory circuits modulate only a single metabolic factor at a time, such as cofactor balance, precursor supply, or pathway timing. , These systems typically operate in a binary on/off manner and lack the capacity to proportionally coordinate interconnected pathways. , For biosynthetic processes that require simultaneous balancing of both cofactor regeneration and precursor allocation, single-node dynamic control remains insufficient, highlighting the need for integrated, multinode regulatory architectures. − This challenge is particularly evident in pathways that simultaneously demand high levels of redox cofactors and precursors. One representative example is the microbial production of 3-hydroxypropionic acid (3-HP), a versatile platform chemical. It is used in the synthesis of acrylic acid and 1,3-propanediol and has been identified by the U.S. Department of Energy as a top biomass-derived chemical. − Among the reported biosynthetic routes, the malonyl-CoA (Mal-CoA) pathwayenabled by heterologous expression of Mal-CoA reductase (MCR)offers a high theoretical yield from glucose. However, this pathway imposes two tightly coupled metabolic constraints: (i) a strong NADPH demand, as MCR consumes two moles of NADPH per mole of 3-HP produced, and (ii) competition for Mal-CoA, a key precursor also required for fatty acid synthesis (FAS) and biomass formation. To address these constraints, prior studies have relied largely on static engineering approaches. NADPH availability has been enhanced through overexpression of transhydrogenases (e.g., PntAB) or by redirecting carbon flux from the Embden-Meyerhof-Parnas pathway (EMP) to the Entner-Doudoroff (EDP) and pentose phosphate pathways (PPP) via pfkAB deletion. − Mal-CoA availability has been increased by chemically or genetically suppressing FAS. While effective to some extent, these strategies independently optimize cofactor and precursor pathways and cannot dynamically adjust their relative contributions (Supplementary Table S1). As a result, they frequently lead to cofactor imbalance, impaired growth, or inefficient carbon utilization. , These limitations demonstrate that efficient 3-HP production may benefit from coordinated regulation of NADPH regeneration and Mal-CoA allocation. Such coordination is fundamentally lacking in static or single-factor dynamic control systems. In this study, we developed a product-responsive dynamic regulatory system in which a single regulatory protein orchestrates both cofactor regeneration and precursor allocation. This circuit links intracellular product accumulation to coordinated modulation of multiple metabolic nodes following pathway induction. Rather than regulating cofactor or precursor supply in isolation, the designed circuit integrates control of NADPH regeneration and Mal-CoA allocation in a unified, product-responsive framework. The utility of this multinode dynamic regulation strategy is demonstrated using 3-HP biosynthesis in Escherichia coli as a model system. This approach establishes a generalizable framework for dynamically balancing cofactor and precursor metabolism and improves the production of NADPH- and Mal-CoA-dependent biochemicals. Results and Discussion Design and Operating Principle of a 3-HP-Responsive Regulatory Circuit Efficient production of 3-HP via the Mal-CoA pathway requires simultaneous regulation of cofactor balance and precursor allocation during the production phase. To achieve this coordination, we designed a 3-HP-responsive regulatory circuit that links intracellular 3-HP accumulation to dynamic control of NAD­(P)H regeneration and Mal-CoA allocation. The overall design strategy and circuit architecture are summarized in Figure . Design strategy and architecture of a 3-HP-responsive dual-factor regulatory circuit. The circuit consists of a 3-HP biosynthesis module, a 3-HP sensing module, and GntR-mediated cofactor and precursor control modules. After initial IPTG induction of the 3-HP biosynthesis module, intracellular 3-HP accumulates during the production phase, leading to increased GntR expression, repression of the EDP and FAS, and activation of the PPP. Through this product-responsive regulation, the downstream circuit promotes a transition from growth-associated metabolism toward production-associated metabolism. The logic table summarizes the logical relationship between IPTG input, intracellular 3-HP levels, and the resulting metabolic states, illustrating a product-responsive shift from growth-associated metabolism to production-associated metabolism after pathway initiation. 1, active/high; 0, inactive/low. The circuit is composed of five functionally connected modules: a 3-HP biosynthesis module, a 3-HP sensing module, an NADH regeneration module, an NADPH regeneration module, and a Mal-CoA allocation module. In the 3-HP biosynthesis module, IPTG induces expression of mcr-C and mcr-N, initiating conversion of Mal-CoA to 3-HP. Accumulated intracellular 3-HP is sensed by the transcriptional regulator HpdR, which activates the P hpdH promoter to drive expression of gntR in the 3-HP sensing module. GntR serves as the central regulator linking 3-HP sensing to downstream metabolic control. The EDP primarily supports cell growth by supplying NADH and metabolic energy, particularly during the early growth phase. , In the NADH regeneration module, GntR represses the native edd–eda operon, thereby reducing carbon flux through the EDP. Concurrently, GntR activates expression of gnd through a TetR/P tet inverter implemented using a GntR-responsive promoter. This activation results in PPP activation, thereby enhancing NADPH regeneration in the NADPH regeneration module. Through this dual action, GntR shifts glycolytic flux from the EDP toward the PPP in response to 3-HP accumulation. In parallel, the Mal-CoA allocation module links GntR activity to FAS. The native fabHDG operon is placed under the control of a GntR-responsive promoter, such that increasing GntR levels repress fabHDG expression. This repression limits Mal-CoA consumption by FAS and redirects the precursor toward 3-HP formation. Importantly, Mal-CoA flux redistribution via fabHDG regulation is conditional on intracellular 3-HP accumulation and is therefore minimized during the early growth phase. The phase-dependent logical behavior of the circuit is summarized in the logic table shown in Figure . During the growth phase, low intracellular 3-HP levels result in minimal GntR expression, allowing the EDP and FAS to remain active while PPP activity remains low, thereby providing NADH and fatty acids required for growth. After initial IPTG induction of the 3-HP biosynthesis module, intracellular 3-HP accumulates during the production phase, leading to increased GntR expression, repression of the EDP and FAS, and activation of the PPP. Through this product-responsive regulation, increasing GntR availability autonomously drives the transition from growth phase to production, synchronizing NADPH regeneration and Mal-CoA availability with 3-HP biosynthetic demand. Characterization of the GntR-Dependent Promoters and Derived Biosensors Given GntR’s established role in repressing the edd–eda operon in E. coli, this regulator represents a suitable regulatory node for controlling flux through the EDP. To implement the GntR-mediated control within the product-responsive circuit, GntR-responsive regulatory elements suitable for reliable integration into the circuit architecture were characterized. Specifically, the transcriptional responses of candidate GntR-regulated promoters were examined in the presence and absence of GntR. Their suitability for dynamic control of downstream metabolic pathways was evaluated. To couple a 3-HP sensing module to GntR-mediated control of target-gene expression, strong GntR-responsive promoters appropriate for downstream circuit construction were first identified. RNA-seq data were reanalyzed from our previously engineered strain, Pfk_ALE-1. This E. coli strain was derived via an adaptive laboratory evolution and carries an inactive gntR gene, resulting in preferential utilization of the EDP and upregulation of GntR-regulated transcriptional targets. Comparison of transcriptomes between wild-type MG1655 and Pfk_ALE-1 revealed multiple transcriptional units under the control of GntR. Among these, transcripts of gntT, gntK, gntU, and idnK changed by at least 2-fold. To ensure promoter independence, gntU was excluded because its position within an operon suggests regulation by its corresponding upstream promoter rather than direct GntR control. The promoter regions of gntK (P gntK ), gntT (P gntT ), and idnK (P idnK ) were cloned upstream of an rfp reporter gene on a plasmid. Fluorescence reporter assays were then performed to evaluate promoter activity in the presence or absence of GntR. For each promoter, reporter plasmids were introduced into either a ΔgntR host or a gntR-restored host, such that the two conditions differed only in GntR availability. The following strain pairs were tested: SBH01/SBH02 (pP gntK –RFP in the absence/presence of chromosomal gntR expression), SBH03/SBH04 (pP gntT –RFP in the absence/presence of chromosomal gntR expression), and SBH05/SBH06 (pP idnK –RFP in the absence/presence of chromosomal gntR expression). As expected, reporter expression was high in the absence of GntR and strongly repressed when GntR was present. Fluorescence decreased more than 5-fold under GntR-positive conditions, validating the RNA-seq screen and confirming P gntK , P gntT , and P idnK as suitable regulatory elements for GntR-based control modules (Figure A). Characterization of GntR-dependent promoters and derived genetic modules. (A) Fluorescence analysis of candidate GntR-responsive promoters in the presence and absence of GntR. Each promoter was placed upstream of an rfp reporter gene. (B) Fluorescence analysis of genetic inverter modules derived from GntR-responsive promoters. Candidate promoters were placed upstream of tetR, which controls rfp expression via the downstream P tet promoter. To convert GntR repression into a measurable activation signal, a TetR/P tet -based logic inverter was implemented in which each GntR-responsive promoter (P gntK , P gntT , and P idnK ) was used to drive tetR expression. Binding of GntR to these promoters inhibits tetR transcription, thereby relieving TetR-mediated repression of the downstream P tet -driven reporter and resulting in fluorescence activation. This design effectively transforms a negative GntR input into a positive, easily quantified output signal (Figure B). Functionality was evaluated in the Pfk_ALE-1_ΔgntR before and after chromosomal restoration of the native gntR gene. Among the three variants, the P gntK -based module exhibited more than 1,000-fold increase in fluorescence between GntR-negative and GntR-positive conditions, indicating that this switch module responds robustly to GntR availability. In contrast, modules based on P gntT and P idnK showed no appreciable difference between the two genetic backgrounds, presumably due to residual basal promoter activity that continues to drive tetR expression even in the presence of GntR. These results identify P gntK as the most suitable promoter for constructing a GntR-based inverter. This inverter was next incorporated into a 3-HP sensing module to achieve product-responsive GntR control. The fully modular architecturecomprising a GntR-regulated promoter, TetR/P tet -based logic inverter, and reporter cassettecan be readily combined with other regulatory modules, facilitating broader synthetic biology applications. Integration of a 3-HP Sensing Module with GntR for Autonomous Metabolic Rewiring Having identified P gntK as a robust GntR-responsive promoter, this promoter was integrated with a 3-HP sensing module. This created a product-responsive regulatory circuit that links intracellular 3-HP accumulation to GntR expression and to downstream regulation via P gntK . To enable product-coupled sensing and regulation, gntR was placed under control of the HpdR/P hpdH module, resulting in strain SBH13. To enhance circuit stability and minimize plasmid-dependent variation, the sensing module (HpdR/P hpdH -gntR) was chromosomally integrated at the native gntR locus. To match the dynamic range of P gntK and optimize downstream target expression, the 3-HP-dependent GntR output was systematically tuned by modulating hpdR expression using a series of RBS variants and, subsequently, a strong synthetic promoter–RBS combination. This configuration enabled tuning of GntR output levels in response to intracellular 3-HP and ensured that gntK promoter activity aligned with the desired sensing strength. RBS sequences for hpdR were redesigned using UTR Designer to increase translation efficiency by 10-, 100-, 500-, and 1,000-fold, generating strains SBH14–SBH17. , Following construction of the sensing modules, transcription of gntR was confirmed by quantitative RT-PCR (qRT-PCR). Relatively small differences in gntR transcript levels were observed among strains SBH13–17 (Supplementary Figure S1). To further enhance sensing output and improve control, strain SBH18 was constructed by replacing the native promoter and RBS of hpdR with a strong synthetic promoter (P J23100 ) and RBS (B0034), respectively. Even with this enhanced configuration, qRT-PCR revealed only moderate variation in gntR transcript levels among strains SBH13–18 following supplementation with 0.5 g/L 3-HP (Supplementary Figure S1). As GntR acts as a transcriptional repressor, it cannot directly activate gene expression in response to 3-HP. Although gntR transcript levels differed only modestly, small differences in repressor abundance can be amplified through inverter logic. Therefore, to achieve a positive output signal upon 3-HP sensing, we designed a transcriptional inverter that converts GntR-dependent repression into inducible gene activation. To evaluate this effect, a TetR/P tet -based inverter was employed to convert the repressive GntR input into a positive, quantifiable fluorescence output upon 3-HP sensing. To accelerate regulatory response and minimize residual TetR accumulation, an LAA degradation tag was fused to the C-terminus of TetR in the inverter module. The resulting plasmid (pINV1: P gntK -tetR-LAA, P tet -rfp) was introduced into SBH13–18 strains, generating strains SBH19–24 (each corresponding to its SBH13–18 parent). This allowed us to characterize the inverter’s performance via RFP fluorescence (Figure A). Characterization of a 3-HP-inducible sensing module for GntR-mediated dynamic control. (A) Schematic of the genetic circuit. The 3-HP-sensing module (left) uses HpdR, activated by 3-HP, to drive P hpdH -dependent expression of GntR. The NADPH regeneration module (right) uses GntR to repress P gntK , relieving TetR (LAA degron) repression and activating P tet to induce rfp (pINV1) or gnd (pINV2). (B) RFP fluorescence of strains SBH19–SBH24 carrying pINV1 and 3-HP-sensing module variants with increasing RBS strengths (1× to 1,000×, predicted by UTR Designer), with SBH24 using P J23100 as promoter. Cells were grown with 0 or 0.5 g/L 3-HP. (C) Intracellular NADPH/NADP+ ratios in strains harboring pINV2 (P gntK -tetR-LAA, P tet -gnd). SBH25 (control, no sensing module) and SBH31 (pINV2 coupled with the 3-HP-sensing module) were cultured with or without 0.5 g/L 3-HP. Nonsignificant differences are not indicated. Asterisks denote statistical significance (*p < 0.05, **p < 0.01, ***p < 0.001). RFP expression in SBH24 was at least 11.4-fold higher than in SBH19, consistent with stronger 3-HP-induced gntR expression (Figure B). Although SBH23 and SBH24 were predicted by UTR Designer to exhibit similar RBS strengths (∼1000× to WT), only SBH24 incorporated the strong P J23100 –B0034 promoter–RBS combination, accounting for its superior fluorescence output (Figure B). This dual enhancement explains why SBH24 consistently displayed the highest dynamic response among the tested variants. RFP expression remained at basal levels across all variants in the absence of 3-HP, confirming tight repression and minimal leakiness of the sensing module under uninduced conditions. Time-course fluorescence measurements confirmed that the inverter remains inactive under basal conditions and is activated upon 3-HP addition (Supplementary Figure S2). A dose-dependent increase in RFP fluorescence was observed across the tested 3-HP range (0–1 g/L), with SBH24 exhibiting the strongest concentration-dependent response. In contrast, SBH19–SBH23 showed only modest fluorescence increases even at 1 g/L 3-HP, consistent with their weaker RBS strengths and promoter activities. These results represent characterization of sensor responsiveness within the tested concentration range rather than a complete quantitative transfer-function analysis. To assess metabolic flux redirection, the rfp gene was replaced with gnd, encoding 6-phosphogluconate dehydrogenasea key enzyme in the PPPyielding pINV2 (P gntK -tetR-LAA, P tet -gnd), which constitutes the NADPH regeneration module. This construct was introduced into Pfk_ALE-1_ΔgntR and SBH18 (the parental background of SBH24), generating strains SBH25 and SBH31, respectively. Because SBH24 (SBH18 carrying pINV1) exhibited the strongest 3-HP-induced inverter response (Figure B), SBH18 was selected for construction of SBH31 by replacing pINV1 with pINV2. Intracellular NADPH levels were measured in the presence and absence of 0.5 g/L 3-HP. SBH31 exhibited an approximately 20% increase in the NADPH/NADP+ ratio relative to the control strain SBH25, which lacks the 3-HP sensing module (Figure C). These results indicate that the product-responsive circuit effectively activates gnd expression and enhances NADPH regeneration. Overall, this 3-HP-inducible regulatory circuit enables dynamic cofactor regeneration by rewiring metabolic flux in response to product accumulation. More broadly, it provides a generalizable framework for product-coupled metabolic engineering, where biosynthetic output is directly linked to pathway regulation. Dynamic Regulation of Glycolytic Flux Enhances 3-HP Production With evidence that the circuit boosts NADPH in response to 3-HP, we next examined whether these dynamic changes would improve 3-HP production. To this end, the optimized HpdR/P hpdH -gntR sensing module was coupled with two GntR-mediated cofactor control modules: the NADPH regeneration module that activates the PPP via the TetR/P tet -gnd inverter, and the NADH regeneration module that modulates the EDP. These modules were further integrated with the 3-HP biosynthesis module (mcr) (Figure A). In this configuration, GntR simultaneously represses the EDP and activates the PPP through the cofactor control modules, thereby boosting NADPH availability. This dynamic cofactor regulation enhances metabolic flux toward 3-HP synthesis, resulting in substantially increased product titers. Promoter/RBS-modulated product-responsive regulation enhances 3-HP production through an EDP-to-PPP metabolic switch. (A) Schematic of the engineered genetic circuit for dynamic metabolic rewiring. Intracellular 3-HP activates HpdR, driving P hpdH -dependent GntR expression. GntR represses the EDP (edd–eda) and, via a TetR/P tet inverter, activates gnd expression, redirecting carbon flux toward the PPP and enhancing NADPH regeneration. (B) 3-HP titers of strains SBH32 (no dynamic control) and SBH33–SBH38 carrying RBS variants of increasing strength controlling hpdR expression. Intermediate variants SBH33–SBH37 were not statistically distinguishable from one another, whereas SBH38 showed a significant increase relative to the nonregulated control SBH32. Nonsignificant differences are not indicated. Asterisks denote statistical significance (*p < 0.05 and ***p < 0.001). Building on this assembled circuit, the impact of sensing-module strength on 3-HP production was systematically examined. Strains carrying sensing-modules of increasing strength (SBH33–38) showed a general trend toward higher 3-HP titers, with a pronounced increase observed at the strongest configuration (SBH38; P J23100 –B0034), while intermediate variants produced broadly similar titers (Figure B). Strains harboring weak sensing modules produced less than 1.0 g/L of 3-HP, only marginally higher than the nonregulated control strain SBH32 (0.64 g/L), indicating insufficient GntR output to drive effective flux redirection. Among the strains with intermediate sensing-module strengths (SBH33–37), 3-HP titers ranged from approximately 1.7 to 2.0 g/L, representing a two- to 3-fold increase relative to SBH32, though these differences were not statistically distinguishable from one another. In contrast, SBH38 produced 2.5 g/L of 3-HP, corresponding to an approximately 4-fold increase over the nonregulated control SBH32 (p < 0.001) (Figure B), consistent with a nonbinary regulatory response in which only the strongest sensing-module configuration generates sufficient GntR output to drive effective EDP-to-PPP flux redirection. By coupling cofactor regeneration to intracellular product accumulation, the circuit prevents premature diversion of carbon flux during growth and synchronizes NADPH supply with biosynthetic demand, thereby enhancing overall productivity. This behavior is consistent with the observed increase in NADPH availability (Figure C), in which strains carrying the dynamic regulatory module exhibited an approximately 20% increase in the NADPH/NADP+ ratio compared to the control strains. Collectively, these data demonstrate that the engineered circuit effectively coordinates product sensing with redox regeneration, supporting increased flux through MCR. Importantly, the graded activation of the PPP avoids premature carbon redirection during early exponential growtha common limitation of static or inducer-based strategieswhile ensuring sufficient NADPH supply during the production phase. Overall, these findings show that the GntR-mediated rewiring of glycolysis, guided by the 3-HP sensor, substantially increases 3-HP production by dynamically synchronizing NADPH supply with product accumulation. This study underscores the broader utility of product-responsive control circuits as product-responsive regulatory modules for other NADPH-dependent bioproduction, although the present implementation still relies on IPTG for MCR expression. Additional Precursor Control Contributes to 3-HP Production Having optimized the cofactor regeneration modules, we sought to further enhance 3-HP production by addressing the precursor supply. To this end, an additional control module was introduced to redirect Mal-CoA flux. Specifically, the native fabHDG operon (responsible for FAS) was placed under the control of the GntR-responsive promoter P gntK , thereby creating a precursor control module that suppresses Mal-CoA consumption for FAS as intracellular 3-HP accumulates. When combined with the GntR-mediated cofactor control modules and mcr-driven 3-HP biosynthesis module, this multimodule design enabled coordinated regulation of both cofactor balance and precursor allocation (Figure A). Stepwise comparison of engineered strains demonstrates enhanced 3-HP titers through product-responsive integrated control. (A) Schematic of the integrated regulatory circuit. As intracellular 3-HP accumulates, HpdR activates P hpdH -dependent expression of GntR, which simultaneously represses the EDP (via P edd -mediated suppression of edd–eda) and, through a TetR/P tet inverter driven by P gntK , activates gnd expression to enhance NADPH regeneration. In parallel, GntR represses the fabHDG operon via P gntK , reducing Mal-CoA flux toward FAS and redirecting precursor supply toward 3-HP biosynthesis. (B) Time-course profiles of cell growth (OD600; left axis; squares), 3-HP titer (left axis; diamonds), and residual glucose (right axis; dashed lines) in SBH32 (no dynamic regulation; light green) and SBH40 (dynamic 3-HP-inducible EDP/PPP and FAS control; dark teal) over 48 h following IPTG induction of the 3-HP biosynthesis module. (C) Final 3-HP titers after 48 h of shake-flask cultivation across four engineered strains: SBH32, no dynamic regulation (baseline); SBH39, static IPTG-inducible EDP/PPP control; SBH38, dynamic 3-HP-inducible EDP/PPP control; and SBH40, dynamic 3-HP-inducible EDP/PPP and FAS control. The 3-HP biosynthesis module was induced with 100 μM of IPTG. The SBH38 vs SBH40 difference was not statistically significant. Asterisks denote statistical significance (*p < 0.05 and **p < 0.01). The contribution of each control module to 3-HP production was quantified through a stepwise comparison of four strains (SBH32, SBH38, SBH39, and SBH40). In addition, growth and 3-HP production were monitored over time to examine the dynamic behavior of the regulatory circuit. Figure B compares the baseline strain SBH32 and the fully regulated strain SBH40 and supports a production-associated shift in the integrated dynamic-control strain. The baseline strain SBH32 (mcr under an IPTG-inducible promoter, without cofactor or precursor control modules) produced only minimal 3-HP (∼0.64 g/L). Strain SBH39, which expresses both Gnd and GntR under an IPTG-inducible promoter on a plasmid (mimicking a static cofactor control system) to redirect flux from the EDP to the PPP, increased the 3-HP titer to approximately 2.4 g/L (Figure C). Replacing the static control with a product-responsive sensing module (as in SBH38) further elevated the titer to ∼ 2.8 g/L, highlighting the advantage of dynamic, product-coupled regulation (Figure C). Finally, SBH40which integrates both the cofactor regeneration and precursor supply control modulesachieved the highest 3-HP titer (3.0 g/L); however, this improvement over SBH38 was modest and did not reach strong statistical significance (p > 0.05), indicating that precursor control provides a secondary rather than primary contribution under the tested conditions. Compared with the baseline strain SBH32, the fully regulated strain SBH40 improved the 3-HP yield from 0.04 to 0.19 g/g glucose, corresponding to an approximately 5-fold increase in carbon efficiency. This increase in carbon efficiency reflects the coordinated redirection of metabolic flux by the dynamic circuit. To further support the proposed mechanism, relative Mal-CoA availability was assessed using an RppA-based biosensor, in which flaviolin production (OD340/OD600) serves as an indirect proxy for Mal-CoA flux. The strain harboring the Mal-CoA allocation module (SBH42) exhibited statistically higher relative Mal-CoA availability than the control strain (SBH41) at both 24 h (p < 0.01) and 48 h (p < 0.001) (Supplementary Figure S3). Although indirect, these results provide statistically supported evidence for the Mal-CoA flux under the tested condition; however, absolute intracellular Mal-CoA concentrations were not directly quantified. Together with enhanced NADPH supply and repression of competing pathways, these changes promote greater channeling of glucose-derived carbon toward 3-HP biosynthesis, with product-dependent activation ensuring that flux redistribution occurs primarily during the production phase. This stepwise improvement highlights two key principles. First, coordinated regulation of precursor supply and cofactor regeneration yields greater benefits than optimizing each independently. Although enhancing NADPH regeneration substantially increased titers, the additional redirection of Mal-CoA flux away from FAS provided only a modest and statistically nonsignificant further improvement under the tested conditions, suggesting that cofactor regeneration is the primary driver of production enhancement in the current system. Second, strains equipped with a dynamic, product-responsive regulatory circuit consistently outperformed those relying on static IPTG induction in both final 3-HP titers and metabolic balance. By responding directly to intracellular 3-HP accumulation, the circuit ensured that NADPH regeneration and Mal-CoA reallocation occurred predominantly during the production phase, thereby preserving cell growth while enhancing 3-HP production. The NADPH/NADP+ assay and the RppA-based Mal-CoA proxy assay are consistent with the proposed mechanism, however direct, time-resolved quantification of NADH/NAD+, NADPH/NADP+, or intracellular Mal-CoA during 3-HP production was not performed in this study. Collectively, these findings demonstrate that integrated dynamic feedback control provides a modular framework for dynamic metabolic engineering. By combining NADPH regeneration via EDP-to-PPP redirection (NADH and NADPH regeneration modules) and Mal-CoA redistribution via fabHDG regulation (Mal-CoA allocation module), the engineered system achieved a 5-fold increase in 3-HP titer relative to the baseline strain, primarily driven by enhanced cofactor regeneration with a secondary contribution from precursor modulation. The current system retains IPTG-dependent MCR expression, as constitutive or chromosomally integrated expression resulted in insufficient 3-HP production in preliminary experiments, highlighting the difficulty of balancing pathway flux with cellular fitness in fully autonomous configurations. Accordingly, IPTG induction was retained to ensure sufficient and reproducible MCR activity, enabling rigorous evaluation of the downstream product-responsive circuit. As summarized in Supplementary Table S1, the 3.0 g/L titer achieved in this study remains below the highest reported flask and fed-batch 3-HP titers. However, the present work differs conceptually from prior pathway-output optimization studies by introducing a product-responsive multinode regulatory architecture that coordinates NADPH regeneration and Mal-CoA allocation after pathway initiation. Thus, the main contribution of this study is not maximal titer, but the demonstration of a modular product-coupled control strategy for balancing cofactor and precursor metabolism. Future work will focus on achieving fully autonomous regulation by eliminating external induction, fermentation-scale evaluation, improved sensor sensitivity, and further pathway optimization to enhance productivity. Importantly, the modularity of this architecture, enabled by the graded and nonbinary regulatory properties of HpdR- and GntR-controlled promoters, provides a generalizable framework that can be extended to other NADPH- and Mal-CoA-dependent biosynthetic pathways. , Conclusions A product-responsive regulatory circuit was developed that links intracellular 3-HP accumulation to coordinated dynamic metabolic control in E. coli through a single sensing module. Via GntR, the circuit simultaneously redirects glycolytic flux from the EDP to the PPP to enhance NADPH regeneration and represses the fabHDG operon to divert Mal-CoA from FAS toward 3-HP biosynthesis, dynamically reallocating carbon and cofactor resources as the culture transitions from growth to production phase. Cofactor regeneration was the primary driver of improved production, while fabHDG repression provided a secondary contribution; together, these modules achieved an approximately 5-fold improvement in both 3-HP titer (3.0 g/L) and yield (0.19 g/g glucose) relative to the unregulated baseline. These results establish a generalizable design principle for product-coupled multinode metabolic control, in which a downstream product-response circuit coordinates cofactor supply and precursor allocation after IPTG-mediated pathway initiation. The current system still relies on IPTG-dependent MCR expression and has been validated only under shake-flask conditions. Future work will prioritize achieving fully inducer-independent operation and extending this modular framework to other NADPH- and Mal-CoA-dependent pathways under bioreactor-scale cultivation conditions. Materials and Methods Reagents, Strains, and Media All chemical reagents used in this study were obtained from Sigma-Aldrich (St. Louis, MO, USA) unless otherwise specified. Bacterial strains and plasmids employed in this work are listed in Supplementary Table S2. E. coli DH10B was used for plasmid construction and amplification. For routine cultivation and cloning, Luria–Bertani (LB) medium was prepared with 5 g/L yeast extract, 10 g/L peptone, and 10 g/L NaCl. M9 minimal medium (BioShop Canada Inc., Burlington, Canada) was supplemented with 2 mM MgSO4, 0.1 mM CaCl2, glucose, and appropriate antibiotics as indicated. Antibiotics were added at the following concentrations: ampicillin (100 μg/mL), kanamycin (50 μg/mL), and chloramphenicol (30 μg/mL). Genome Editing A tetA-based dual selection system was employed for genome integration and replacement. Briefly, a tetA cassette was first inserted at the target locus and subsequently replaced with a PCR-amplified DNA fragment encoding either the 3-HP-inducible sensing module or the P gntK for the Mal-CoA allocation module. Each fragment contained 50-bp homologous arms flanking the insertion site (upstream of gntR or upstream of fabHDG) to facilitate homologous recombination. Recombinants were selected using negative selection with 50 μM NiCl2, followed by verification of successful integration via colony PCR and Sanger sequencing. Plasmid Construction All plasmids constructed in this study were derived from BioBrick plasmids. Primers used for cloning are listed in Supplementary Table S3 and were synthesized by Macrogen Inc. (Seoul, Korea). PCR amplification was performed using Q5 High-Fidelity DNA Polymerase (New England Biolabs, USA). Restriction enzymes were purchased from Thermo Scientific (USA). For the construction of pP gntK -RFP, pP gntT -RFP, and pP idnK -RFP reporter plasmids, BioBrick plasmid pBbS2c-RFP was used as the vector backbone. The backbone was digested with AatII and NdeI, and promoter regions of gntK, gntT, and idnK were PCR-amplified from E. coli MG1655 genomic DNA based on NCBI reference sequences. Promoter fragments were assembled into the digested backbone using Gibson assembly, generating the corresponding reporter constructs. For construction of pP gntK -TetR, pP gntT -TetR, and pP idnK -TetR inverter plasmids, pBbS2c-RFP was digested with AatII and NdeI. To replace the native promoter of tetR with a GntR-responsive promoter, the tetR coding sequence, candidate promoter fragments, the P tet promoter, and a double terminator were PCR-amplified and assembled by Gibson assembly. For construction of pINV1, an LAA degron tag was fused to the C-terminus of tetR to enable rapid protein turnover and enhance regulatory responsiveness. This construct was generated following the same cloning strategy as pP gntK -TetR, using Gibson assembly. Fluorescence Analysis Recombinant E. coli strains were grown overnight in LB medium at 37 °C with shaking at 200 rpm. Cells were washed three times with phosphate-buffered saline (PBS, pH 7.0) and resuspended in M9 minimal medium. Cultures were inoculated into a clear-bottom 96-well plate at an initial OD600 of 0.1 and exposed to either 0.5 g/L 3-HP or a range of 3-HP concentrations (0–1 g/L). After incubation for 16 h, RFP fluorescence was measured using an Infinite F200 PRO microplate reader (Tecan, Austria) with excitation at 535 nm and emission at 585 nm. Fluorescence values were normalized to OD600 and are reported as arbitrary units (a.u.). 3-HP Production For 3-HP biosynthesis, single bacterial colonies were inoculated into 3 mL of LB medium and cultured at 37 °C with shaking at 200 rpm. Overnight cultures were diluted 1:20 into M9 minimal medium supplemented with 20 g/L glucose, 1.0 g/L yeast extract, 2 mM MgSO4, 0.1 mM CaCl2, and trace elements (2.4 g FeCl3·6H2O, 0.3 g CoCl2, 0.15 g CuCl2·2H2O, 0.3 g ZnCl2, 0.3 g Na2MoO4·2H2O, 0.075 g H3BO3, and 0.495 g MnCl2·4H2O per liter). After 3 h of cultivation, mcr expression was induced with 100 μM IPTG and cultures were incubated at 37 °C for 48 h. Samples were collected either at the end point or at 12 h intervals for growth, glucose, and 3-HP analyses. Quantification of Intracellular NADPH/NADP+ Ratio Intracellular NADPH levels were quantified using an NADP/NADPH Assay Kit (Colorimetric) (ab65349, Abcam, UK) according to the manufacturer’s protocol. Briefly, cells were harvested by centrifugation and washed with cold PBS. Cell pellets were extracted using NADP/NADPH extraction buffer through repeated freeze–thaw cycles, followed by centrifugation at 14,000 rpm for 5 min to remove cell debris. For NADPH-specific detection, extracts were incubated at 60 °C for 30 min to decompose NADP+ while preserving NADPH. Reaction mixtures were prepared in a 96-well plate by combining the samples with NADP cycling buffer and enzyme mix, followed by incubation at room temperature for 5 min. Developer solution was then added, and absorbance at 450 nm was measured using a microplate reader (Tecan, Austria) after color development. NADPH concentrations were determined using a standard curve generated from the NADPH standards provided with the kit and normalized to cell density (OD600). The NADPH/NADP+ ratio was calculated according to the manufacturer’s protocol. RppA-Based Biosensor Assay for Malonyl-CoA Availability Relative Mal-CoA availability was evaluated using an RppA-based biosensor as previously described. The rppA gene, encoding a type III polyketide synthase that converts Mal-CoA to flaviolin, was introduced into the control strain lacking the Mal-CoA allocation module (SBH41) and the engineered strain harboring the Mal-CoA allocation module (SBH42). For the RppA-based biosensor assay, cells were cultivated under the same conditions as described for 3-HP production, except that the MCR expression plasmid was omitted. RppA expression was induced with 100 μM IPTG, and exogenous 3-HP was supplied at 0.5 g/L to activate the sensing module. At 24 and 48 h postinduction, culture samples were collected and flaviolin production was quantified by measuring absorbance at 340 nm (OD340). OD340 values were normalized to cell density (OD600), and the OD340/OD600 ratio was used as an indirect indicator of relative Mal-CoA availability. Analytical Methods Sample preparation followed previously described methods. , Culture samples were centrifuged at 13,200 × g for 15 min, and the supernatant was collected and heat-treated at 80 °C for 1 h. Samples were centrifuged again at 13,200 × g for 30 min prior to analysis. Supernatants were diluted 10-fold, and 20 μL was injected into the high-performance liquid chromatography (HPLC) system. Glucose and 3-HP concentrations were quantified using an Aminex HPX-87H (Bio-Rad, USA) column (300 mm × 7.8 mm) operated at 35 °C with 5.41 mM H2SO4 as the mobile phase. , Statistical Analysis All data are presented as mean ± standard deviation (SD) from three independent biological replicates unless otherwise indicated. For comparisons among multiple strains, statistical significance was evaluated using one-way analysis of variance (ANOVA) followed by Tukey’s multiple-comparison test. For prespecified pairwise comparisons, two-sided unpaired Student’s t tests were used. Supplementary Material