Multi-omics-guided metabolic engineering of Limosilactobacillus reuteri for high-level 3-hydroxypropionaldehyde production from glucose Authors: Zuojun Liu, Qiang Yin, Liang Li, Hang Wu, Yong Fang, Fuqiang Cheng, Ting Guo Published: 2026 ABSTRACT BackgroundLimosilactobacillus reuteri is an important probiotic chassis for producing 3-hydroxypropionaldehyde (3-HPA), a broad-spectrum antimicrobial and biochemical. However, its production is constrained by a metabolic trilemma comprising coenzyme B₁₂ auxotrophy, redox imbalance, and carbon catabolite repression (CCR). The main bottleneck currently limiting cost-effective glucose-based production is the inability to simultaneously resolve these interconnected constraints.ResultsThrough integrated multi-omics analyses of differential carbon source performance, an unexpected regulatory mechanism was obtained. Galactose orchestrates a tripartite metabolic program in which UDP-galactose may act as a signaling molecule potentially involved in alleviating CCR; redirects carbon flux through the pentose phosphate pathway to balance ATP/NADH, and transcriptionally co-activates dhaB (glycerol dehydratase) and cobA/Q (B₁₂ biosynthesis) for coordinated holoenzyme assembly. Strikingly, this native regulatory network enables 8.51 g/L 3-HPA from galactose despite its poor support for cell growth, versus 1.67 g/L from glucose which readily supports robust growth—a 5.1-fold advantage that highlights galactose’s metabolic priming efficacy. Guided by these insights, an engineered strain overexpressing glycerol dehydratase was combined with process optimization and a rationally designed four-factor targeted supplementation. Synergistic supplementation with coenzyme B₁₂, ATP, KCl, and MgCl₂ achieved 28.96 g/L 3-HPA—a 17.34-fold improvement over native glucose-based performance.ConclusionsTo our knowledge, this is the first systems-level strategy for compensating the metabolic trilemma in L. reuteri by engineering carbon source-dependent regulatory networks and cofactor dynamics in a cost-effective glucose-based system, achieving the highest 3-HPA titer reported in lactobacilli in shake-flask fermentation to date. This work establishes a scalable platform for industrial 3-HPA production and provides a systems-level framework for reprogramming central carbon metabolism in probiotic chassis to address persistent bioproduction bottlenecks. Background As antimicrobial resistance increases, interest in sustainable alternatives to conventional antibiotics has intensified, and probiotic-derived antimicrobials have gained attention within “One Health” strategies. Limosilactobacillus reuteri, a commensal bacterium widely distributed in the gastrointestinal tracts of vertebrates, produces reuterin, a broad-spectrum antimicrobial agent comprising hydrated, non-hydrated, and dimeric forms of 3-hydroxypropionaldehyde (3-HPA). This compound exhibits activity against both Gram-positive and Gram-negative bacteria, fungi, and protozoa, providing L. reuteri with a competitive advantage in its ecological niches [1]. Beyond its antimicrobial properties, 3-HPA also holds considerable promise as a biochemical precursor for the synthesis of valuable compounds such as acrylic acid and functional polymers, highlighting its industrial potential [2]. Despite this considerable translational potential, scaling 3-HPA biosynthesis in L. reuteri remains difficult because of fundamental metabolic constraints that limit industrial viability. In this organism, 3-HPA production is intrinsically linked to glycerol metabolism, a pathway constrained by several interconnected limitations [3–6]. First, the process is strictly dependent on coenzyme B₁₂ because the catalytic activity of glycerol dehydratase (GDHt), the key enzyme in the pathway, requires this cofactor for optimal function [6]. However, L. reuteri possesses only limited endogenous capacity for B₁₂ synthesis, resulting in suboptimal enzyme activation [7]. Second, pronounced redox imbalance arises because NADH accumulates during glycerol metabolism, which promotes the diversion of 3-HPA toward 1,3-propanediol via NADH-dependent propanediol oxidoreductase (PDOR), thereby substantially reducing 3-HPA yield [8]. Third, carbon catabolite repression (CCR) suppresses glycerol metabolic genes in the presence of preferred carbon sources such as glucose, drastically limiting carbon flux into the 3-HPA biosynthetic pathway [9]. These three limitations, coenzyme B₁₂ scarcity, redox stress, and transcriptional repression, collectively form a tightly interconnected “metabolic trilemma” that complicates engineering strategies for enhanced 3-HPA production. Current research specifically targeting 3-HPA enhancement in L. reuteri remains limited. Although one study used carboxyl-functionalized superparamagnetic nanoparticles attached to the bacterial surface to improve conversion efficiency, this strategy increases operational complexity and cost [10]. Carbon source selection represents an underexplored yet potentially transformative lever for metabolic optimization. Carbon sources are pivotal determinants of microbial metabolic profiles because they dictate the distribution of carbon flux across pathways [11]. Glucose, a preferred carbon source for most microorganisms, rapidly activates glycolysis but often triggers CCR, thereby suppressing genes involved in alternative metabolic routes [12]. Heterologous expression of 3-HPA biosynthesis pathways in Escherichia coli has also been attempted to circumvent native regulation in L. reuteri, but these approaches sacrifice the probiotic compatibility required for food and pharmaceutical applications [13]. Translating 3-HPA production to an industrially viable scale requires improvements in both titer and cost. In other microorganisms such as Klebsiella pneumoniae, metabolic engineering for 1,3-propanediol production has achieved notable success, validating key strategies including cofactor rebalancing and endogenous cofactor synthesis [14]. Zhang et al. engineered K. pneumoniae to efficiently convert glycerol to 1,3-propanediol by reprogramming central metabolism to minimize byproducts and autonomously supply critical cofactors, thereby eliminating the need for exogenous vitamin B₁₂ through enhanced endogenous synthesis [15]. These advances demonstrate the potential of systematic metabolic engineering. Similarly, the use of low-cost carbon sources such as glucose presents a compelling route for L. reuteri, but its potential to boost 3-HPA yield remains largely untapped because of CCR [16]. In this study, an integrated approach combining multi-omics technologies and metabolic engineering was used to decipher the carbon source-dependent metabolic trilemma in L. reuteri LR618. The metabolic responses to different carbon sources were systematically investigated. The results suggest that galactose orchestrates a tripartite mechanism: UDP-galactose accumulation may contribute to alleviating CCR, pentose phosphate pathway flux rebalances ATP/NADH pools, and transcriptional co-activation of dhaB (GDHt) and cob genes coordinates holoenzyme assembly. Guided by these mechanistic insights, a rationally designed engineering strategy was implemented in a glucose-based fermentation system by combining GDHt overexpression with targeted supplementation of coenzyme B₁₂, ATP, K⁺, and Mg²⁺. This approach achieved a 17.34-fold increase in 3-HPA yield (28.96 g/L) compared with the native glucose-based system, substantially exceeding previously reported levels. This work provides a systems-level framework for leveraging native metabolic networks in probiotic chassis and indicates that coordinated multi-target interventions are important for overcoming the metabolic trilemma in sustainable antimicrobial production. Materials and methods Bacterial strain and cultivation conditions L. reuteri LR618, isolated from pig feed, was used as the experimental strain. To investigate the effects of different carbon sources on strain growth and subsequent 3-hydroxypropionaldehyde (3-HPA) production, pre-cultures were established in de Man-Rogosa-Sharpe (MRS) broth. The medium was supplemented with one of the following sole carbon sources: glucose, lactose, galactose, sucrose, dextrin, soluble starch, maltose, fructose, deoxyribose, xylose, or glycerol, each at 20 g/L, consistent with the carbohydrate concentration in the standard MRS formulation. Static cultures were incubated at 37 °C under 5% CO₂ for 26 h. Growth was monitored by measuring the optical density at 600 nm (OD₆₀₀), and cultures were harvested at the stationary phase to ensure a consistent physiological state for subsequent experiments. Glycerol conversion reactions After 26 h of cultivation, cells were harvested by centrifugation at 8,000 × g for 10 min at 4 °C. The pellets were washed twice with sterile phosphate-buffered saline (pH 6.0) and resuspended in glycerol conversion medium consisting of 50 mM potassium phosphate buffer (pH 6.0) supplemented with 600 mM glycerol. The final inoculum size was adjusted to a wet cell weight (WCW) of 10 g/L. Bioconversion reactions were carried out in 250 mL Erlenmeyer flasks at 37 °C under static conditions to produce 3-HPA. Samples were collected every 0.5 h. A 5 mL aliquot of the reaction mixture was rapidly chilled on ice to halt metabolic activity and then centrifuged at 12,000 × g for 5 min at 4 °C to separate cells from the supernatant. Supernatants were stored at 4 °C for biochemical analyses to quantify 3-HPA and other metabolites. Cell pellets were frozen in liquid nitrogen and stored at − 80 °C for metabolomic and transcriptomic profiling. The 3-HPA concentration was determined spectrophotometrically. A standard curve was established using purified 3-HPA (0.05–1.5 mg/L) dissolved in ultrapure water. For sample analysis, 1 mL culture supernatant was mixed with 0.75 mL 10 mM tryptophan reagent and 3 mL concentrated HCl, followed by vortexing and incubation at 37 °C for 20 min. Absorbance at 560 nm was measured using a UV-Vis spectrophotometer (Shimadzu, Japan). The 3-HPA concentration was then calculated from the standard curve. The values reported in this study represent total aldehyde content expressed as 3‑HPA equivalents. Metabolomic profiling of LR618 Metabolite extraction from cells: A 100 µL aliquot of cell pellets was mixed with 400 µL extraction solution (MeOH: ACN, 1:1, v/v). The extraction solution contained deuterated internal standards. The mixture was vortexed for 30 s, sonicated for 10 min in a 4 °C water bath, and incubated for 1 h at − 40 °C to precipitate proteins. Samples were then centrifuged at 12,000 rpm for 15 min at 4 °C. The supernatant was transferred to a fresh glass vial for analysis. The quality control sample was prepared by pooling equal aliquots of the supernatants from all samples. LC-MS analysis: LC-MS/MS analyses were performed using a UHPLC system (Vanquish, Thermo Fisher Scientific) equipped with a Waters ACQUITY UPLC BEH Amide column (2.1 mm × 50 mm, 1.7 μm) coupled to an Orbitrap Exploris 120 mass spectrometer (Thermo). The mobile phase consisted of 25 mmol/L ammonium acetate and 25 mmol/L ammonium hydroxide in water (pH = 9.75) (A) and acetonitrile (B). The autosampler temperature was 4 °C, and the injection volume was 2 µL. The Orbitrap Exploris 120 mass spectrometer was operated in information-dependent acquisition mode under the control of Xcalibur software (Thermo) to acquire MS/MS spectra. In this mode, the acquisition software continuously evaluates the full-scan MS spectrum. The electrospray ionization source conditions were set as follows: sheath gas flow rate, 50 Arb; auxiliary gas flow rate, 15 Arb; capillary temperature, 320 °C; full MS resolution, 60,000; MS/MS resolution, 15,000; collision energy, stepped normalized collision energy 20/30/40; and spray voltage, 3.8 kV in positive mode and − 3.4 kV in negative mode. Transcriptomic profiling of LR618 Triplicate biological replicates were harvested by rapid centrifugation (8,000 × g, 4 °C, 5 min). Cell pellets were immediately flash-frozen in liquid nitrogen and stored at − 80 °C until RNA extraction. Total RNA was isolated using the RNeasy Mini Kit (Qiagen, Germany) with on-column DNase I treatment to remove genomic DNA. RNA integrity was verified using an Agilent 2100 Bioanalyzer (RNA Integrity Number, RIN ≥ 9.0), and RNA quantity was measured using a Qubit 4.0 Fluorometer (Thermo Fisher Scientific, USA). Ribosomal RNA was depleted using the MICROBExpress™ Bacterial mRNA Enrichment Kit (Ambion, USA) to enrich mRNA. Strand-specific RNA-seq libraries were prepared using the NEBNext® Ultra™ II Directional RNA Library Prep Kit (NEB, USA) according to the manufacturer’s protocol. Briefly, fragmented RNA (200–300 nt) underwent cDNA synthesis, end repair, adapter ligation, and PCR amplification (12 cycles). Library quality was assessed using an Agilent 2100 Bioanalyzer (average insert size, 370–420 bp) and quantified by qPCR (Kapa Biosystems, USA). Paired-end sequencing (2 × 150 bp) was performed on the Illumina NovaSeq 6000 platform (Novogene, China). Raw reads were preprocessed using Trimmomatic v0.39 to remove adapters and low-quality bases (Phred score < 20). Clean reads were aligned to the L. reuteri DSM 17,938 reference genome (NCBI Assembly ASM2698v1) using Bowtie2 v2.4.2 with default parameters. Transcript abundance was quantified using feature counts v2.0.3, and differentially expressed genes (DEGs) were identified using DESeq2 v1.34.0 (|log₂FC| > 2, FDR-adjusted p < 0.05). Functional annotation of DEGs was performed using the KEGG Orthology (KO) and Gene Ontology (GO) databases through clusterProfiler v4.4.4. Validation of glycerol dehydratase overexpression The dhaB operon was amplified from LR618 genomic DNA using primers GDHt-F/R under the following PCR conditions: 98 °C for 3 min; 30 cycles of 98 °C for 10 s, 55 °C for 15 s, and 72 °C for 3 min; followed by 72 °C for 5 min. The purified amplicon (TransGen Gel Kit) was cloned into pLEM415 by homologous recombination and electroporated into LR618 at 2.0 kV for 5 ms. The transformant (LR620) was selected on erythromycin (10 µg/mL). 3-HPA yields of LR620 and the controls, wild-type LR618 and empty-plasmid LR618/pLEM415, were assayed as described above. For qRT-PCR, RNA from the pre-culture (26 h) and bioconversion (2 h) phases was reverse-transcribed, and dhaB expression was quantified using GDHt-Q-F/R primers with 16 S ribosomal RNA (16 S-Q-F/R) gene as the reference gene. The qPCR program consisted of 40 cycles at 95 °C for 5 s and 60 °C for 30 s. Fold changes were calculated using the 2^−ΔΔCt method. Exogenous metabolite supplementation The engineered strain LR620 was precultured anaerobically in MRS broth. Single-factor experiments were performed to examine the effects of biotransformation time (0.5–5 h), temperature (23–51 °C), inoculum size (10–130 g/L, WCW), pH (4–9), and glycerol concentration (100–800 mM). An orthogonal design was then used to evaluate interactions among these factors. Based on the orthogonal analysis, the most significant factors were further optimized by response surface methodology using a Box-Behnken design. To investigate glycerol dehydratase reactivation, exogenous additives were supplemented into the glycerol solution under the optimized conditions. The individual additives tested were KCl (1–5%), ATP (0.1–0.9%), coenzyme B₁₂ (0.001–0.009%), and MgCl₂ (0.03–0.15%). Based on the optimal individual concentrations, binary, ternary, and quaternary combinations were then tested. All experiments were performed in triplicate. Coenzyme B₁₂ stock solutions (0.1%, w/v) were prepared in sterile deionized water, filter-sterilized, and stored at 4 °C in amber vials. Working concentrations were obtained by serial dilution, and all additions were performed using calibrated micropipettes. Statistical analyses All data in this study represent the mean (± standard deviation) of three independent experiments. Statistical analyses (p < 0.05) were performed using SPSS 20.0 to analyze the relationship between metabolomic and transcriptomic data. Origin software was used for graphical analysis. Results and discussion Carbon source screening identifies galactose as superior for 3-HPA production In the two-step fermentation process using L. reuteri LR618, 3-HPA production varied significantly across carbon sources (Fig. 1). Galactose supported the highest titer (8.51 g/L), followed by lactose (7.01 g/L), whereas glucose yielded only 1.67 g/L. Deoxyribose (1.37 g/L) and xylose (1.28 g/L) showed moderate production, whereas maltose, dextrin, soluble starch, and sucrose produced negligible amounts (< 1 g/L). Strikingly, growth curve analysis revealed that glucose supported rapid proliferation, with biomass reaching OD600 5.1, whereas galactose-cultured cells grew slowly, attaining a maximum OD600 of only 3.8. This dramatic disparity suggests fundamental differences in how these carbon sources prime cellular metabolism for subsequent glycerol bioconversion. Galactose and lactose, both metabolized through the Leloir pathway [17], appear to enhance the metabolic capacity for 3-HPA synthesis, whereas glucose and fructose, which preferentially enter glycolysis, do not establish a similarly favorable state. Polysaccharides (starch and dextrin) and disaccharides (sucrose and maltose) likely suffer from inefficient hydrolysis or transport under the experimental conditions, limiting their utility [18]. Based on these results, galactose and glucose were selected as representative high-performance and baseline carbon sources, respectively, for subsequent multi-omics dissection of the underlying mechanisms. Fig. 13-HPA production from different carbon sources in the two-step fermentation system using LR618. The error bars represent the standard deviation 3-HPA production from different carbon sources in the two-step fermentation system using LR618. The error bars represent the standard deviation Metabolomic profiling reveals carbon source-dependent metabolic reprogramming To dissect the metabolic basis of the differential 3-HPA production, untargeted metabolomics was performed on samples from four experimental conditions: cells cultured in glucose for 26 h (PLM26), glucose-cultured cells transferred to glycerol for 2 h (PLG2), cells cultured in galactose for 26 h (BAM26), and galactose-cultured cells transferred to glycerol for 2 h (BAG2). This design enabled separate interrogation of the pre-culture phase (carbon source priming) and the bioconversion phase (3-HPA production). Metabolomic profiling revealed 1554 differentially abundant metabolites in the pre-culture phase (Fig. 2A) and 1650 in the bioconversion phase (Fig. 2B). Principal component analysis revealed clear separation between the glucose- and galactose-fed groups, indicating substantial metabolic reprogramming induced by carbon source selection. Pathway enrichment analysis identified coenzyme B₁₂ biosynthesis, the Leloir pathway, glycerol metabolism, and redox homeostasis as the key pathways differentiating the two carbon source regimes. The findings indicate that galactose enhances cofactor synthesis and stress resilience in lactic acid bacteria, whereas glucose-driven glycolytic overflow is associated with redox imbalance. Fig. 2Volcano plots showing differential metabolites in LR618 fermentation between glucose and galactose conditions. A Pre-culture phase; B Bioconversion phase. Red: significantly upregulated; blue: significantly downregulated (|log₂FC| > 1, p < 0.05); gray: not significant. Dashed lines indicate significance thresholds Volcano plots showing differential metabolites in LR618 fermentation between glucose and galactose conditions. A Pre-culture phase; B Bioconversion phase. Red: significantly upregulated; blue: significantly downregulated (|log₂FC| > 1, p < 0.05); gray: not significant. Dashed lines indicate significance thresholds In the pre-culture phase, the most significantly enriched pathways were concentrated in amino acid metabolism (alanine, aspartate and glutamate metabolism, beta-alanine metabolism), cofactor and vitamin metabolism (nicotinate and nicotinamide metabolism, pantothenate and CoA biosynthesis), with the above pathways showing high enrichment significance (p < 0.05) and abundant differential molecules (Fig. 3A). In galactose-preconditioned cells (BAM26), metabolic priming was characterized by synergistic upregulation of the Leloir pathway and coenzyme B₁₂ biosynthesis. Elevated levels of UDP-galactose, a key Leloir intermediate, were observed (2.3-fold vs. PLM26, p < 0.01), accompanied by accumulation of 5′-deoxyadenosylcobalamin, a coenzyme B₁₂ precursor. Concurrent enrichment of the methionine salvage and selenocompound metabolism pathways supplied ATP and selenocysteine, both essential for coenzyme B₁₂ synthesis. This coordinated metabolic configuration ensured that, upon transfer to glycerol, both GDHt and its essential cofactor were readily available [19]. Fig. 3KEGG pathway enrichment treemaps of differentially expressed metabolites in LR618 fermentation between glucose and galactose conditions. A Pre-culture phase; B Bioconversion phase. The area of each rectangle represents the number of differential genes/metabolites enriched in the corresponding pathway, and the color intensity indicates the enrichment significance (–lnP-value), with darker colors reflecting higher significance KEGG pathway enrichment treemaps of differentially expressed metabolites in LR618 fermentation between glucose and galactose conditions. A Pre-culture phase; B Bioconversion phase. The area of each rectangle represents the number of differential genes/metabolites enriched in the corresponding pathway, and the color intensity indicates the enrichment significance (–lnP-value), with darker colors reflecting higher significance In contrast, glucose-preconditioned cells (PLM26) exhibited metabolic suppression characterized by accumulation of acidic byproducts (acetate and lactate) and depletion of nucleotide sugars. This acidification may trigger stress responses that downregulate nonessential pathways, whereas CCR actively silences genes involved in alternative carbon utilization [20]. The absence of UDP-galactose and coenzyme B₁₂ precursors in PLM26 indicates that glucose metabolism fails to establish the metabolic infrastructure required for efficient 3-HPA synthesis. In sharp contrast, the strain exhibited profound metabolic network remodeling during glycerol bioconversion for 3-HPA production, with core enriched pathways shifting to those directly associated with 3-HPA biosynthesis—alanine, aspartate and glutamate metabolism remained highly significant, alongside robust enrichment of glycerol catabolism, coenzyme regeneration and energy supply pathways, and sustained significant enrichment of stress resistance-related pathways (Fig. 3B). Upon transfer to glycerol, galactose-fed cells (BAG2) achieved 5.1-fold higher 3-HPA titers than glucose-fed cells (PLG2), driven by different metabolic configurations. In BAG2, glycerol was preferentially channeled through the reductive pathway, as indicated by elevated GDHt activity markers and reduced glycerol-3-phosphate accumulation (1.2-fold downregulation of glycerol kinase flux). Critically, redox analysis revealed a more oxidative NAD⁺/NADH ratio in BAG2 (2.3) than in PLG2 (1.6, p < 0.05), which suppressed PDOR-mediated diversion of 3-HPA to 1,3-propanediol. This favorable redox state originated from Leloir pathway activity during pre-culture, which generates NADH in a controlled manner without the glycolytic overflow characteristic of glucose metabolism. Glucose-fed cells (PLG2) displayed metabolic dysregulation characterized by ATP depletion and NADH overload. Excess NADH activated PDOR, redirecting 3-HPA toward 1,3-propanediol as a redox sink. At the same time, acid stress caused by organic acids produced during pre-culture persisted and further compromised metabolic efficiency [21]. UDP-galactose, a key intermediate of the Leloir pathway, accumulated to 2.3-fold higher levels in galactose-fed cells than in glucose-fed controls (p < 0.01). Similarly, S-adenosylmethionine (SAM) reserves were 2.1-fold higher in galactose-preconditioned cells (p < 0.05) and persisted during bioconversion. The NAD⁺/NADH ratio was also significantly elevated in galactose-fed bioconversion cultures (2.3) compared with glucose-fed controls (1.6, p < 0.05). These metabolite signatures—UDP-galactose accumulation, SAM enrichment, and a more oxidized redox state—collectively distinguished the metabolic configuration established by galactose from that of glucose. The mechanistic implications of these metabolic differences are further explored through transcriptomic analysis in the following section. Transcriptomic reprogramming of glycerol metabolism by galactose To complement the metabolomic findings and establish the transcriptional basis of the metabolic advantages conferred by galactose, comparative transcriptomics was performed on the same four experimental conditions. RNA-seq analysis identified 346 differentially expressed genes between the galactose- and glucose-fed groups (|log₂FC| > 1, FDR-adjusted p < 0.05), Hierarchical clustering showed clear transcriptional specialization under galactose conditions, with BAM26 and BAG2 forming a cohesive cluster. Transcriptomic profiling revealed 374 differentially expressed metabolism-associated transcripts across the experimental conditions in the pre-culture phase and 331 in the bioconversion phase (Fig. 4A, B). Fig. 4Transcriptomic volcano plots of transcriptomic changes in LR618 fermentation between glucose and galactose conditions. A Pre-culture phase; B Bioconversion phase. The x-axis shows log₂(fold change) in gene expression, and the y-axis shows -log₁₀(adjusted P-value, padj). Red: up-regulated; blue: down-regulated (|log₂FC| > 1, p < 0.05); gray: not significant. Selected differentially expressed genes are annotated Transcriptomic volcano plots of transcriptomic changes in LR618 fermentation between glucose and galactose conditions. A Pre-culture phase; B Bioconversion phase. The x-axis shows log₂(fold change) in gene expression, and the y-axis shows -log₁₀(adjusted P-value, padj). Red: up-regulated; blue: down-regulated (|log₂FC| > 1, p < 0.05); gray: not significant. Selected differentially expressed genes are annotated During the bioconversion phase, galactose feeding resulted in 3.1-fold upregulation of GDHt (EC 4.2.1.30) gene in BAG2 relative to PLG2 (p < 0.05) (Fig. 5). This transcriptional elevation directly correlated with the enhanced 3-HPA synthesis capacity observed in the metabolomic analysis. Importantly, this upregulation was not limited to GDHt alone. Genes encoding enzymes required for coenzyme B₁₂ biosynthesis were also co-induced during the pre-culture phase, with cobA (C6H63_RS00365) and cobQ (C6H63_RS00370) upregulated 2.6- and 3.3-fold, respectively, in BAM26 relative to PLM26 (p < 0.05). This coordinated transcriptional program indicates that glycerol exposure significantly enhances the availability of both the GDHt and its essential cofactor coenzyme B₁₂, a key regulatory axis driving glycerol-dependent 3-HPA biosynthesis in Limosilactobacillus reuteri that was comprehensively characterized in the previous work [22]. Fig. 5KEGG glycerolipid metabolism pathway map of LR618, showing transcriptomic changes in the glycerol-to-3-HPA biosynthetic pathway. Colored boxes indicate differentially expressed genes between galactose and glucose carbon source cultures during 3-HPA fermentation, with green for upregulation and red for downregulation in galactose-fed groups. EC numbers are labeled in boxes, and arrows represent metabolic reactions KEGG glycerolipid metabolism pathway map of LR618, showing transcriptomic changes in the glycerol-to-3-HPA biosynthetic pathway. Colored boxes indicate differentially expressed genes between galactose and glucose carbon source cultures during 3-HPA fermentation, with green for upregulation and red for downregulation in galactose-fed groups. EC numbers are labeled in boxes, and arrows represent metabolic reactions Galactose metabolism was driven by 4.2-fold induction of Leloir pathway genes (galK, C6H63_RS00285; galT, C6H63_RS00290), enabling efficient galactose catabolism to generate ATP and NADH without the glycolytic overflow characteristic of glucose metabolism. At the same time, genes encoding the oxidative glycerol pathway (glpK, C6H63_RS00450; glpD, C6H63_RS00455) were downregulated 2.5-fold in galactose-fed cells, thereby minimizing carbon diversion toward glycolysis and ensuring that glycerol entering the cell was preferentially directed toward 3-HPA synthesis. This transcriptional configuration, activation of the desired pathway coupled with suppression of competing routes, represents a more favorable metabolic state that glucose metabolism fails to establish. The co-upregulation of dhaB and cob genes suggests a transcriptionally coupled mechanism that optimizes GDHt function. This dual regulation differs from previous L. reuteri studies in which coenzyme B₁₂ limitation constrained GDHt activity despite adequate enzyme expression [23], highlighting the unique ability of galactose to bypass this bottleneck. The transcriptional coupling likely arises from shared regulatory elements responsive to UDP-galactose or related metabolites, thereby linking carbon source availability to both pathway enzyme production and cofactor synthesis. In glucose-grown cells, CCR actively represses genes related to alternative carbon utilization. In galactose-fed cells, transcript levels of genes associated with CCR regulation were reduced 1.5-fold (p < 0.05), consistent with alleviation of CCR. The accumulation of UDP-galactose, which does not occur during glucose metabolism, correlated with this derepression, suggesting a potential role for this metabolite or one of its derivatives in modulating the CCR machinery (Fig. 6). Although direct molecular evidence is currently lacking, it is hypothesized that a UDP-galactose-mediated signaling cascade may contribute to CCR alleviation under galactose conditions. Fig. 6KEGG galactose metabolism pathway map of LR618. Colored boxes show differentially expressed genes/metabolites between galactose and glucose carbon source cultures, with green for upregulation and red for downregulation in galactose-fed groups. Numbers indicate enzyme commission (EC) numbers, and arrows represent metabolic reactions KEGG galactose metabolism pathway map of LR618. Colored boxes show differentially expressed genes/metabolites between galactose and glucose carbon source cultures, with green for upregulation and red for downregulation in galactose-fed groups. Numbers indicate enzyme commission (EC) numbers, and arrows represent metabolic reactions The transcriptomic data provide a mechanistic explanation for the metabolomic observations. The elevated GDHt activity inferred from metabolite profiles in BAG2 directly reflects transcriptional upregulation of dhaB. The favorable NAD⁺/NADH ratio in galactose-fed cells arises from induction of the Leloir pathway, which generates NADH without the excessive glycolytic flux characteristic of glucose metabolism. Sustained coenzyme B₁₂ availability during bioconversion results from induction of cob genes during pre-culture, combined with SAM-mediated stabilization. Together, these data indicate that galactose orchestrates a multilayered metabolic program, transcriptional, enzymatic, and metabolic, that collectively optimizes 3-HPA synthesis. Hierarchical metabolic constraints govern 3-HPA biosynthetic efficiency The multi-omics framework described above predicted that efficient 3-HPA production requires coordinated resolution of all three components of the metabolic trilemma: coenzyme B₁₂ availability, redox balance, and transcriptional derepression. To test this prediction and translate mechanistic insights into engineering strategies, systematic phenotypic analyses were performed across engineered and supplemented L. reuteri strains. To validate the rate-limiting role of GDHt predicted by transcriptomics, strain LR620 was constructed by overexpressing the dhaB operon in LR618 using plasmid pLEM415. qRT-PCR confirmed 4.5-fold upregulation of dhaB in LR620 relative to the wild type (p < 0.05) (Fig. 7A). In the glucose-based two-step fermentation system, LR620 produced 5.77 g/L 3-HPA, a 3.45-fold increase over wild-type LR618 (1.67 g/L, p < 0.05), confirming GDHt as a key bottleneck (Fig. 7B). Fig. 7Transcriptional levels of key genes and 3-HPA production in Lactobacillus reuteri strains. A Relative transcription levels of dhaB and cobA in LR620 and LR618. B 3-HPA production titers of LR618, LR620, and LR618/pLEM415. The error bars represent the standard deviation Transcriptional levels of key genes and 3-HPA production in Lactobacillus reuteri strains. A Relative transcription levels of dhaB and cobA in LR620 and LR618. B 3-HPA production titers of LR618, LR620, and LR618/pLEM415. The error bars represent the standard deviation However, two observations indicated that GDHt overexpression alone was insufficient to achieve galactose-level performance. First, the empty-plasmid control (LR618/pLEM415) yielded only 1.42 g/L, which was 15% lower than the wild type (p < 0.05), directly demonstrating a measurable ATP burden caused by plasmid maintenance. This 15% deficit is consistent with the metabolomic observation of energy depletion in glucose-fed systems, in which ATP scarcity constrains multiple metabolic processes. Second, despite elevated GDHt expression, LR620 showed no increase in endogenous coenzyme B₁₂ synthesis, with qRT-PCR showing unchanged cobA expression relative to the wild type (1.02-fold, p < 0.05). This uncoupling of enzyme overexpression from cofactor biosynthesis, a feature inherent to galactose metabolism but absent in glucose-based engineering, explains why the LR620 yield (5.77 g/L) remained below that achieved with galactose (8.51 g/L) despite higher GDHt transcript levels [24]. After establishing that enzyme availability is necessary but insufficient, systematic optimization of biotransformation parameters was carried out using LR620. Initial single-factor experiments established the baseline ranges of biotransformation time (0.5–5 h), temperature (23–51 °C), inoculum size (10–130 g/L, WCW), pH (4–9), and glycerol concentration (100–800 mM). To capture interactions among factors, an L₁₆(3⁵) orthogonal design was implemented. ANOVA revealed that inoculum size and glycerol concentration exerted the strongest effects on 3-HPA production, followed by temperature and pH, whereas biotransformation time within the 1.5–3.0 h range was not significant. The optimal combination identified by range analysis, 110 g/L inoculum size, 700 mM glycerol, 30 °C, pH 6.0, and 2.0 h, increased the titer to 18.54 g/L. Response surface methodology using a Box-Behnken design was then applied to refine the three most significant factors: inoculum size, glycerol concentration, and temperature. The quadratic model showed an excellent fit (R² = 0.9957, adjusted R² = 0.9901) and a non-significant lack of fit (p < 0.05), validating its predictive capacity. Model coefficients revealed significant positive linear effects of inoculum size (p < 0.05) and glycerol concentration (p < 0.05), as well as significant negative quadratic effects for all three factors (p < 0.05), indicating the existence of optimal intermediate values. The interaction term inoculum size with temperature was significant (p < 0.05), demonstrating that the effect of inoculum size depends on temperature, likely reflecting a trade-off between enzyme availability at high cell density and temperature-sensitive mass transfer efficiency. The model predicted a maximum 3-HPA production of 23.5 g/L at 102 g/L inoculum size (WCW), 724 mM glycerol, and 28.9 °C. Experimental validation under these conditions yielded 22.79 g/L, closely matching the prediction and representing a 3.95-fold improvement over the initial unoptimized condition (5.77 g/L for LR620 without optimization). This progression, from 5.77 to 18.54 to 22.79 g/L, demonstrates the value of hierarchical optimization that progressively incorporates additional layers of system complexity. Despite achieving 22.79 g/L through parameter optimization, 3-HPA accumulation remained transient, peaking at 2.0 h and declining thereafter. This pattern reflects the well-characterized suicide inactivation of GDHt. Upon binding glycerol, the enzyme induces a conformational change in coenzyme B₁₂, leading to irreversible cofactor-enzyme complexation and loss of catalytic activity [25]. Reactivation requires removal of the damaged cofactor and replacement with fresh coenzyme B₁₂, an ATP-dependent process facilitated by Mg²⁺ ions, whereas K⁺ is essential for maintaining active-site conformation and substrate positioning [26]. To overcome this limitation, exogenous additives were tested individually under the optimized biotransformation conditions. KCl supplementation, as a K⁺ source, enhanced 3-HPA production in a dose-dependent manner, with an optimum at 3% (w/v), yielding 24.33 g/L, a 6.76% increase over the response surface methodology-optimized baseline (Fig. 8A). Exogenous ATP supplementation also enhanced 3-HPA production (Fig. 8B). The maximum 3-HPA titer (25.11 g/L, representing a 10% increase over the unsupplemented control) was achieved at an ATP concentration of 0.3% (w/v), whereas higher concentrations resulted in reduced titers. The presence of a defined optimal concentration indicates that the effect of ATP on 3-HPA biosynthesis is concentration-dependent [27]. Fig. 8Effect of single-factor supplementation on 3-HPA production by LR620 under optimized biotransformation conditions. A KCl, B ATP, C coenzyme B₁₂, D MgCl₂. The error bars represent the standard deviation Effect of single-factor supplementation on 3-HPA production by LR620 under optimized biotransformation conditions. A KCl, B ATP, C coenzyme B₁₂, D MgCl₂. The error bars represent the standard deviation Exogenous coenzyme B₁₂ supplementation also enhanced 3-HPA production (Fig. 8C). The highest 3‑HPA titer (25.56 g/L) was achieved at 0.005% (w/v), representing a 12% improvement over the unsupplemented control (22.79 g/L). At lower concentrations (0.001–0.003%), titers ranged from 23.19 to 23.69 g/L, while at higher concentrations (0.007–0.009%), titers decreased to 24.5 and 24.1 g/L, respectively. This pattern is consistent with a typical cofactor saturation effect: GDHt is a coenzyme B₁₂-dependent enzyme, and moderate B₁₂ supplementation facilitates its catalytic reaction, whereas excessive B₁₂ may competitively inhibit the reactivation process or lead to non-productive binding [19]. MgCl₂ supplementation showed a narrower optimum, with peak production (23.21 g/L) at 0.06% (w/v) and sharp declines at higher concentrations, likely because of osmotic stress (Fig. 8D). Based on the optimal individual concentrations of each component, various combinatorial supplementations were then investigated to identify potential synergistic effects. The quaternary combination (ATP + KCl + B₁₂ + MgCl₂) produced the highest titer, 28.96 g/L, corresponding to a 27% improvement over the response surface methodology-optimized baseline (22.79 g/L) and a 17.34-fold increase over the initial unoptimized condition (1.67 g/L for wild-type LR618) (Table 1). Table 13-HPA production of Limosilactobacillus reuteri under different treatmentsRegionTreatments3-HPALR618glucose as the carbon source1.67 g/LLR618galactose as the carbon source8.51 g/LLR618/pLEM415with the pLEM4151.42 g/LLR620overexpression of GDHt5.77 g/LLR620optimization of fermentation engineering22.79 g/LLR620with 0.3% ATP25.11 g/LLR620with 3% KCl24.33 g/LLR620with 0.005% coenzyme B₁₂25.56 g/LLR620with 0.06% MgCl₂23.21 g/LLR620with ATP, KCl, coenzyme B₁₂, MgCl₂28.96 g/L Table 13-HPA production of Limosilactobacillus reuteri under different treatmentsRegionTreatments3-HPALR618glucose as the carbon source1.67 g/LLR618galactose as the carbon source8.51 g/LLR618/pLEM415with the pLEM4151.42 g/LLR620overexpression of GDHt5.77 g/LLR620optimization of fermentation engineering22.79 g/LLR620with 0.3% ATP25.11 g/LLR620with 3% KCl24.33 g/LLR620with 0.005% coenzyme B₁₂25.56 g/LLR620with 0.06% MgCl₂23.21 g/LLR620with ATP, KCl, coenzyme B₁₂, MgCl₂28.96 g/L 3-HPA production of Limosilactobacillus reuteri under different treatments The use of exogenous ATP and coenzyme B₁₂ inevitably adds to production costs. Nevertheless, several considerations suggest that the overall economic potential of this glucose‑based platform remains attractive. First, glucose is considerably cheaper than galactose, which already provides a substantial saving. Second, the optimal concentrations of ATP (0.3%) and B₁₂ (0.005%) are relatively low, and the KCl and MgCl₂ supplements are inexpensive and widely used in industrial fermentation. Third, the current work serves as a mechanistic proof of concept, and future metabolic engineering, including chromosomal integration of the dhaB-cobA operon or enhancement of intracellular ATP, could further reduce or eliminate the reliance on exogenous supplements. Mechanistically, these results establish a hierarchical framework for GDHt function. The reactivation process itself, removal of damaged B₁₂ and replacement with fresh cofactor, requires ATP and Mg²⁺ and is enabled by exogenous B₁₂ supplementation. However, once reactivated, the enzyme must rapidly bind glycerol and convert it to 3-HPA, a catalytic step critically dependent on K⁺ for active-site integrity. Thus, K⁺ acts as a gatekeeper that translates reactivation events into productive catalysis [25]. In its absence, even successfully reactivated enzyme molecules show suboptimal activity, thereby limiting total 3-HPA production. The additive concentrations (3% KCl, 0.3% ATP, 0.005% B₁₂, 0.06% MgCl₂) represent economically viable levels compatible with downstream processing. The systematic phenotypic analyses provide quantitative validation of the multi-omics framework and explain why single-gene engineering alone cannot achieve galactose-level performance. The performance gap between LR620 (22.79 g/L after process optimization) and the fully supplemented system (28.96 g/L) reflects three fundamental limitations. First, cellular energy status imposes a measurable constraint. The empty-plasmid control produced 15% less 3-HPA than wild-type LR618 (1.42 vs. 1.67 g/L), directly demonstrating that ATP diversion to plasmid replication and selection marker expression restricts energy-intensive processes. The presence of a defined optimal ATP concentration (0.3%) further indicates that the effect of ATP on 3-HPA production is concentration-dependent, and that appropriate supplementation within this narrow range contributes to enhanced yields when combined with other cofactors [28, 29]. Second, enzyme overexpression without cofactor enhancement creates an unbalanced metabolic state. Despite 4.5-fold higher dhaB expression, LR620 showed no increase in endogenous coenzyme B₁₂ synthesis, as cobA expression remained unchanged. This uncoupling means that, although more GDHt apoenzyme is available, the supply of its essential cofactor remains limiting. This is a classic example of hierarchical metabolic constraint in which downstream capacity cannot compensate for an upstream limitation [30]. The 3.45-fold improvement from GDHt overexpression (5.77 vs. 1.67 g/L) was therefore substantially lower than the 5.1-fold advantage achieved through galactose’s coordinated program (8.51 vs. 1.67 g/L). Third, the ionic and cofactor environment imposes its own constraints. The narrow optimal range for coenzyme B₁₂, ATP and Mg²⁺, the indispensable role of K⁺ in translating reactivation into catalysis collectively reveal that each component of the GDHt functional cycle must be precisely balanced. Together with the ATP burden from plasmid maintenance and the uncoupling of enzyme overexpression from cofactor synthesis, these observations establish that hierarchical metabolic constraints—spanning energy supply, cofactor availability, and ionic environment—cannot be overcome by single-gene engineering alone [28]. Finally, The multi-omics analyses and phenotypic validations presented above collectively demonstrate that galactose’s native superiority arises from an integrated regulatory program that simultaneously addresses coenzyme B₁₂ availability, redox balance, and carbon catabolite repression. This coordinated tripartite mechanism achieves what reductionist engineering strategies cannot: a self-reinforcing metabolic state in which enzyme supply, cofactor availability, and ionic environment are optimized concurrently. These findings shift the paradigm of microbial aldehyde production away from an “enzyme-centric” approach—focused on overexpressing a single rate-limiting step—toward a “systems-centric” framework in which cofactor dynamics, energy management, and ionic homeostasis dominate yield optimization. Future efforts should prioritize chromosomal integration of GDHt-cobA operons coupled with redox-balancing modules, such as NADH oxidase expression or Leloir pathway reconstruction in glucose-grown cells, to eliminate plasmid burden and recapitulate the integrated advantages of galactose in cost-effective glucose-based processes. Conclusion This study hypothesizes a cross-phase regulatory axis in L. reuteri, in which galactose may optimize 3-HPA biosynthesis through an integrated tripartite mechanism that alleviates CCR, balances ATP/NADH pools, and coordinates holoenzyme assembly—together achieving a 5.1-fold increase over glucose-grown cultures. Translating these mechanistic insights into engineering strategies demonstrated that single-gene overexpression alone cannot overcome hierarchical metabolic constraints. Instead, synergistic supplementation with ATP, K⁺, coenzyme B₁₂, and Mg²⁺, which addresses enzyme reactivation, cofactor availability, and catalytic efficiency, enabled a 17.34-fold improvement, reaching a final titer of 28.96 g/L. To our knowledge, this is the highest 3-HPA titer reported in Limosilactobacillus reuteri in shake-flask fermentation to date. The final optimized process consisted of cell harvest at 24 h, biotransformation at 28.9 °C for 2.0 h with 102 g/L inoculum size (WCW) and 724 mM glycerol (pH 6.0), and supplementation with 3% KCl, 0.3% ATP, 0.005% coenzyme B₁₂, and 0.06% MgCl₂. While achieved at shake-flask scale, this titer provides a strong foundation for bioreactor scale-up studies. Although the current process relies on exogenous supplements, the cost analysis suggests economic viability, and future engineering efforts aim to minimize or eliminate these additions. This work establishes a systems-level blueprint for reprogramming probiotic chassis and demonstrates that resolving the metabolic trilemma of cofactor dependence, redox imbalance, and CCR is essential for next-generation sustainable antimicrobial production.