Reconstitution of adrenergic signaling restores catecholamine-responsive transcription in E. coli Nissle 1917

To determine whether adrenergic signaling could be reconstituted in E. coli Nissle 1917 (EcN), we heterologously expressed the enterohemorrhagic E. coli (EHEC) qseBC operon, which encodes the catecholamine-responsive sensor kinase QseC and its cognate response regulator QseB (Fig 1A). Although EcN harbors homologs of qseBC with high sequence similarity to those of EHEC (~97% identity), bioinformatic analysis revealed naturally occurring mutations in both qseC and qseB that are predicted to impair adrenergic sensing (S1 and S2 Figs). To restore catecholamine responsiveness, we expressed the EHEC qseBC cassette under a synthetic constitutive promoter (J23111) on a low-copy-number plasmid (pSC101) (Fig 1B).

Download:

• PNG
  larger image
• TIFF
  original image

Fig 1. Transcriptomic profiling and characterization of adrenergic signaling in EcN.

(A) Schematic overview of the QseBC-mediated adrenergic signaling pathway in EHEC. QseC detects NE and Epi and activates the response regulator QseB to drive transcriptional responses. (B) Genetic strategy for heterologous expression of the EHEC qseBC operon in EcN using a low copy pSC101 plasmid under a constitutive promoter (J23111), followed by NE exposure. RNA was isolated from untreated EcN_WT, NE-treated EcN_WT, and NE-treated EcN expressing qseBC_EHEC. (C) Hierarchical clustering heatmap of gene expression profiles highlighting NE-dependent transcriptional changes across the indicated treatment groups. (D) Volcano plot highlighting the selected upregulated and downregulated genes in EcN_{qseBC-EHEC-NE} vs EcN_WT-NE, emphasizing the specific genes regulated by the QseB in response to NE exposure. RNA-seq–derived differentially expressed genes (DEGs), selected using thresholds of padj < 0.05 and |log₂(fold change) > 0.5. Differential expression analysis was performed in R using the DESeq2 package. (E) Schematic representation of potential regulators interacting with the flhDC_EHEC operator that may either constrain or facilitate activation of the flhDC_EHEC operon (F) Bubble plot summarizing transcriptional changes of regulatory factors associated with the flhDC_EHEC regulon following NE induction in the engineered EcN_qseBC-EHEC strains compared to EcN_WT-NE. (G) Evaluation of NE sensitivity using the flhDC_EHEC promoter fused to RFP in EcN, with fluorescence intensity normalized to OD₆₀₀ after exposure to 100 µM NE. Data are presented as mean ± SEM (n = 3). The data underlying this Figure can be found in S1 Data. Raw transcriptomic data can be found under BioProject accession number PRJNA1467173. Created in BioRender. Aggarwal, N. (2026) https://BioRender.com/z371w74.

https://doi.org/10.1371/journal.pbio.3003926.g001

To assess transcriptional responses to norepinephrine (NE), we performed RNA sequencing on four conditions: untreated wild-type EcN (EcN_WT), NE-treated EcN_WT, EcN expressing the qseBC operon (EcN_qseBC–EHEC) and NE-treated EcN_qseBC–EHEC. NE exposure elicited minimal transcriptional changes in wild-type EcN, indicating weak endogenous adrenergic responsiveness (S3A Fig). In contrast, EcNqseBC–EHEC exhibited robust NE-dependent transcriptional reprogramming, with distinct expression profiles evident in both heatmap and volcano plot analyses, when compared to untreated EcN expressing the qseBC operon (Figs 1C and S3A).

To distinguish NE-dependent effects from basal qseBC expression, we additionally compared uninduced EcNqseBC–EHEC with the corresponding NE-induced condition, which revealed widespread transcriptional reprogramming and substantial changes in gene expression (S4 Fig). Overall, differential expression analysis showed that NE-induced QseBC signaling preferentially regulated genes associated with flagellar biogenesis, quorum sensing, stress adaptation, and biofilm formation [13,22] (Figs 1D and S3B). These pathways mirror transcriptional programs previously reported for adrenergic signaling in EHEC, confirming that the heterologously expressed qseBC operon is functional in the EcN background and capable of transmitting catecholamine-dependent signals.

Among the NE-responsive pathways, genes within the flagellar biosynthesis regulon were prominently represented. This regulon is hierarchically controlled by the master regulator operon flhDC, which integrates environmental inputs to coordinate motility and stress-adaptive responses in EHEC (S3C Fig). However, despite broad activation of downstream flagellar genes, transcription of the flhDC operon itself was only weakly induced by NE in EcN_qseBC–EHEC (S3B Fig).

The limited responsiveness of flhDC in this heterologous context is consistent with interference from global transcriptional regulators and sigma-factor–dependent control mechanisms known to act on the flhDC promoter, including H-NS, RcsB, EcpR, HdfR, and FliZ [22] (Fig 1E and 1F). In addition, sequence divergence in the endogenous EcN flagellar promoters may further reduce QseB binding affinity relative to their EHEC counterparts (S3D Fig).

To identify adrenergic-responsive promoter elements suitable for biosensor construction, we therefore evaluated flagellar gene promoters derived from EHEC rather than EcN. The flhDC, flgBCDEF, and flgGHIJ promoters from EHEC were fused to a red fluorescent protein (RFP) reporter and co-expressed with qseBC_EHEC in EcN. Promoter activity was assessed under microaerobic conditions sealed tube or sealed plate in Dulbecco’s Modified Eagle Medium (DMEM), which enhances QseBC-dependent signaling in EHEC [23]. Among the promoters tested, only the flhDC_EHEC promoter exhibited measurable NE-dependent induction (Fig 1G), whereas the flgBCDEF_EHEC and flgGHIJ_EHEC promoters showed no detectable responsiveness (S3E and S3F Fig).

Together, these results establish that heterologous expression of the EHEC qseBC operon restores catecholamine-responsive transcription in EcN and identify flhDC_EHEC as a QseBC-regulated promoter with intrinsic but constrained NE responsiveness. This finding provided the basis for subsequent promoter engineering to enhance adrenergic signal transduction.

Promoter truncation reduces regulatory interference and enhances adrenergic responsiveness of flhDC

To determine whether regulatory complexity limits adrenergic responsiveness of the flhDC_EHEC promoter, we examined its native architecture in light of prior studies [14,20,22]. The flhDC_EHEC promoter contains a σ²⁸-binding site in its proximal region, while the central and distal regions harbor multiple binding sites for global transcriptional regulators as well as phosphorylated QseB, including both high- and low-affinity motifs [20,22,24] (Fig 2A and 2B).

Download:

• PNG
  larger image
• TIFF
  original image

Fig 2. Truncation of flhDC_EHEC promoter and synthetic rewiring enhances QseBC-mediated adrenergic sensing.

(A) Schematic showing the truncation of the wild-type flhDC_EHEC promoter to generate flhDC_EHEC-300 and flhDC_EHEC-600 variants enabling reduced regulatory complexity and increased sensitivity to adrenergic signaling. (B) WebLogo representation of conserved high- and low-affinity QseB-binding motifs together with the genomic organization of proximal and distal QseB-binding regions within the native flhDC _EHEC promoter. (C-E) Quantification of NE-responsive transcriptional activity of full-length and truncated flhDC_EHEC variants fused to RFP following exposure to 50 μM and 100 μM NE (F) Heatmap comparing relative activities of Anderson synthetic promoters in EcN and E. coli MG1655, highlighting strain-dependent differences in promoter strength. (G-J) Effect of qseBC_EHEC expression stoichiometry on NE signal transduction. Expression of the qseBC_EHEC operon was tuned using synthetic promoters of defined strengths, and NE dependent activation of the flhDC_EHEC-300 promoter was quantified by normalized RFP fluorescence following exposure to 50 μM and 100 μM NE. (K) Schematic representation of flhDC₋₃₀₀ promoter variants in which the native σ²⁸-dependent region of flhDC₋₃₀₀ was replaced with synthetic σ70 promoters of defined strengths (J23117, J23105, J23111) to overcome environmental regulation and enhance transcriptional stability. (L–O) Fluorescence-based reporter assay showing the activity of engineered flhDC-_EHEC promoter variants in response to 50 μM NE or Epi. The J23117 variant showed the highest induction and lowest variability. (All Data are presented as mean ± SEM (n = 3). One-way ANOVA followed by Tukey’s multiple comparisons test was used for the statistical analysis; *p < 0.05, **p < 0.01, ***p < 0.001, ****p < 0.0001 vs. uninduced (0µM NE). The data underlying this Figure can be found in S1 Data. Created in BioRender. Aggarwal, N. (2026) https://BioRender.com/z371w74.

https://doi.org/10.1371/journal.pbio.3003926.g002

To reduce potential interference from competing regulatory inputs while preserving QseB responsiveness, we generated truncated flhDC_EHEC promoter variants. The flhDC_EHEC-600 variant retains several regulatory binding motifs in addition to the low-affinity QseB-binding site, whereas the flhDC_EHEC-300 variant preserves only the low-affinity QseB-binding site together with the core σ²⁸-dependent regulatory region (Fig 2A). Each promoter variant, along with the full-length flhDC_EHEC promoter, was fused to a RFP reporter on a low-copy-number plasmid (pSC101).

Promoter activity was assessed under microaerobic conditions following exposure to NE. Compared to the full-length and flhDC_EHEC-600 promoters, the flhDC_EHEC-300 variant exhibited significantly enhanced NE responsiveness at both 50 and 100 μM NE (Fig 2C–2E). These results indicate that removal of distal regulatory elements increases adrenergic sensitivity, consistent with the presence of competing regulators that constrain QseB-mediated transcriptional activation in the native promoter context.

QseBC expression level and translational coupling govern catecholamine sensing

To determine how QseBC abundance influences catecholamine sensing, we examined the effect of qseBC expression level on NE-dependent transcriptional output. Because promoter activity can vary substantially between E. coli strains, we first assessed a panel of Anderson synthetic promoters in both EcN and the laboratory strain MG1655 (Figs 2F and S5). Despite high genomic similarity, promoter strengths differed markedly between strains, underscoring the strain-specific nature of transcriptional control.

Based on promoter activity in EcN, we selected synthetic promoters spanning a range of strengths to drive the expression of EHEC qseBC operon, together with an engineered RBS containing a canonical AGGAG Shine–Dalgarno motif while preserving native operon architecture. The engineered RBS was designed using the Salis Lab RBS Calculator and yielded a predicted translational initiation rate of ~10,000, nearly 10-fold higher than the native qseBC RBS. These constructs were combined with the flhDC_EHEC-300-RFP reporter. Increasing qseBC expression enhanced NE responsiveness, with J23104 producing the highest reporter induction among the promoter configurations tested, whereas weaker promoters yielded diminished responses (Fig 2G–2J). These results indicate that QseBC expression level is a key determinant of sensor sensitivity.

In native EHEC, qseB and qseC are encoded within a single operon and are translationally coupled via overlapping stop and start codons (S6A Fig). To assess whether this organization influences sensing performance, we decoupled qseB and qseC expression by placing each gene under independent promoters (S6B Fig). While QseB and QseC protein expression was detectable by SDS–PAGE (S6C Fig), decoupling their expression resulted in a marked reduction in NE responsiveness compared to the operon-encoded configuration (S6D and S6E Fig).

The reduced responsiveness observed upon decoupling is consistent with prior reports that translational coupling ensures coordinated expression and functional stoichiometry of two-component system components. Together, these results demonstrate that both QseBC expression level and translational coupling are critical for efficient catecholamine sensing and identify the flhDC_EHEC-300 promoter combined with operon-encoded qseBC driven by J23104 as the optimal configuration for subsequent analyses.

Sigma-factor replacement and QseB motif engineering yield a synthetic catecholamine-responsive promoter

To reduce context-dependent variability in adrenergic signaling, we next examined the role of sigma-factor dependence in regulating flhDC transcription. In EHEC, transcriptional initiation at the flhDC_EHEC promoter is mediated in part by σ²⁸, whose activity is modulated by the anti-sigma factor FlgM under non-permissive conditions [25]. The truncated flhDC_EHEC-300 promoter retains this native σ²⁸-binding region which may constrain promoter activity under conditions where σ²⁸ availability is limited.

We therefore decoupled adrenergic responsiveness from σ²⁸-dependent regulation by replacing the native σ²⁸-binding site in the flhDC_EHEC-300 promoter with synthetic σ⁷⁰-dependent promoters of defined strengths (Fig 2K). Among the variants tested, substitution with the weakest σ⁷⁰ promoter (J23117) produced the most robust and reproducible induction in response to NE and epinephrine (Epi), whereas stronger σ⁷⁰ promoters exhibited reduced dynamic range and increased variability (Fig 2L–2O). Synthetic promoter activity is highly context dependent, with local sequence architecture and spacing constraints substantially influencing signaling output [26]. Consistent with this structural sensitivity, replacing the native σ²⁸ region within the truncated flhDC_-300 promoter altered promoter behavior and deviated from canonical Anderson promoter rankings.

Catecholamine responsiveness was further enhanced by introducing both high- and low-affinity phosphorylated QseB-binding sites at defined positions upstream of the core promoter (Fig 3A). This design was intended to promote cooperative QseB binding and enable graded transcriptional activation. To assess the contribution of motif architecture, we additionally generated a construct containing only the low-affinity QseB-binding site (S7A Fig). Compared with the dual-motif design, the low-affinity-only construct exhibited limited responsiveness across 25–100 μM NE and Epi concentrations (S7B and S7C Fig), supporting a role for combined high- and low-affinity motifs in enhancing catecholamine-responsive transcriptional activation.

Download:

• PNG
  larger image
• TIFF
  original image

Fig 3. A synthetic QseB-responsive promoter enables dose-dependent catecholamine sensing and controlled secretion of a CRF antagonist.

(A) Design of a synthetic promoter (SynflhDC) incorporating both high and low-affinity QseB binding sites to enhance adrenergic responsiveness. (B, C) Quantification of normalized RFP fluorescence demonstrated a clear dose-dependent activation pattern with NE and Epi (25, 50, and 100 μM). (D) Flow cytometry analysis showing population-level fluorescence shifts corresponding to graded NE/Epi responsive promoter activation. (E) Illustration depicting engineered EcN detecting NE and Epi via the QseBC two-component system, resulting in conditional expression and secretion of the defined biological output rCRF(9–41) (F) Schematic of an NE/Epi-responsive genetic circuit coupled to a YebF–EK–rCRF (9–41) gene cassette. (G) Western blot analysis of extracellular fractions showing NE dependent secretion of the YebF–EK–CRF(9–41) fusion protein, detected using an anti-His HRP–conjugated antibody. SDS–PAGE analysis showing EK cleavage of the purified YebF–EK–CRF(9–41) fusion protein to release the rCRF(9–41) peptide. (I, J) qRT-PCR analysis showing modulation of CRF (100 nM)–mediated TLR4 expression in THP-1 macrophages and Caco-2 cells following pretreatment with either recombinant [rCRF(9–41)] or synthetic [synCRF(9-41)] antagonist (500 nM). (K, L) Quantification of IL-6 and IL-1β in differentiated THP-1 macrophages treated with CRF (100 nM) and LPS (1 µg/mL), alone or in combination, following pretreatment with either rCRF(9-41) or synCRF(9-41) antagonist (500nM). Data are presented as mean ± SEM (n = 3) and were analyzed by one-way ANOVA with Tukey’s multiple-comparison test (*p < 0.05, **p < 0.01, ***p < 0.001, ****p < 0.0001). The data underlying this Figure can be found in S1 Data and S1 Raw Images. Created in BioRender. Aggarwal, N. (2026) https://BioRender.com/z371w74.

https://doi.org/10.1371/journal.pbio.3003926.g003

The dual-motif engineered promoter, designated SynflhDC, exhibited detectable activation at 25 μM NE and showed progressively enhanced activity at higher concentrations of NE and Epi (Fig 3B and 3C). Additionally, the engineered strain exhibited a similar growth profile to the wild-type EcN strain (S8A Fig). Flow cytometry analysis confirmed population-wide fluorescence shifts consistent with enhanced transcriptional output (Fig 3D). Dose–response analysis further revealed concentration-dependent activation of the engineered sensor, with measurable induction observed from approximately 25 μM, followed by progressive signal enhancement and saturation at higher ligand concentrations (S8B and S8C Fig). This response range is consistent with previously reported QseBC-based catecholamine sensing systems, which typically operate in the low-to-mid micromolar range under in vitro conditions [14,27]. We also evaluated SynflhDC responsiveness to the quorum-sensing signal AI-3 and compared its activation with NE across a 25–100 μM concentration range. Although the engineered system responded to both ligands, NE consistently elicited stronger activation than AI-3, indicating preferential responsiveness to catecholamine signaling under the tested conditions (S8D Fig).

To investigate the molecular basis of catecholamine-responsive signaling by the QseC sensor kinase, we performed structure-guided mutagenesis focused on the predicted periplasmic sensing domain. As no experimentally resolved structure of QseC is currently available, ligand docking analyses were performed using AlphaFold-predicted structural models restricted to the putative sensor domain (residues 34–156). Docking of norepinephrine (NE) and epinephrine (Epi) identified a shared set of candidate interaction residues, including Thr72, Asn74, Asp83, and Asp101, predicted to contribute to ligand-responsive signaling (S9A and S9B Fig).

To evaluate their functional relevance, candidate residues were subjected to site-directed mutagenesis followed by reporter-based analysis of catecholamine responsiveness. Among the variants tested, substitution of Asn74 with alanine (N74A) markedly reduced catecholamine-responsive signaling, whereas D83A and D101A produced partial reductions in activity and T72A showed no detectable effect (S9C–S9G Fig). Given that QseBC signaling can also be activated by the quorum-sensing autoinducer AI-3, we additionally assessed AI-3–responsive signaling in wild-type QseC and the N74A mutant. Reduced activation of the N74A variant following AI-3 stimulation suggests that Asn74 contributes to QseC-mediated signal transduction rather than acting as a selective catecholamine-sensing determinant (S9D Fig).

To further exclude the possibility that the observed signaling defects resulted from altered protein stability or defective membrane localization, we performed preliminary biochemical analysis of membrane-associated QseC variants. Comparable membrane-associated expression of QseC WT and mutant proteins suggested that the reduced signaling phenotypes are unlikely to arise from gross protein misfolding or impaired membrane localization (S10 Fig). However, definitive elucidation of the molecular mechanism underlying QseC-mediated catecholamine sensing will require future structural and direct ligand-binding studies, which are beyond the scope of the present study.

Catecholamine sensing can be coupled to regulated secretion of a bioactive peptide

To test whether adrenergic sensing could be coupled to a defined biological output, we linked the optimized catecholamine-responsive promoter to a secretion cassette encoding the CRF receptor antagonist peptide rCRF(9–41), used here as a model bioactive output. CRF(9–41) is a well-characterized competitive antagonist of CRF1 receptors and is widely used to probe CRF-mediated signaling pathways [18,28].

To implement a sense-and-respond circuit, we engineered EcN to detect NE/Epi and conditionally secrete the CRFR1 antagonist rCRF(9–41) (Fig 3E). To enable extracellular peptide release, rCRF(9–41) was fused to the E. coli secretion carrier YebF via an enterokinase cleavage site (Fig 3F). Engineered EcN strains harboring the optimized sensing module and the YebF–rCRF(9–41) cassette were exposed to NE under microaerobic conditions at concentrations of 25 μM and 50 μM, which fall within the upper physiological range reported for the gut during stress and enteric infection [11,29]. Western blot analysis of concentrated culture supernatants confirmed NE-dependent secretion of the fusion protein following induction with 25 μM and 50 μM NE (Fig 3G). To further quantify peptide production under non-concentrated conditions, secreted rCRF(9–41)-HisTag levels were measured directly from culture supernatants using a HisTag ELISA assay, revealing secretion levels of approximately 65–84 nM under the tested conditions (S8E Fig). Following purification and enterokinase treatment, the expected rCRF(9–41) fragment was released (Fig 3H), confirming correct processing of the secreted product.

NE concentrations in the gut microenvironment are difficult to quantify and may vary substantially under conditions of stress and inflammation [5,12,30,31]. Under these conditions, rCRF (9–41) secretion was reproducibly induced by NE exposure, demonstrating that the engineered genetic circuit converts catecholamine detection into regulated peptide secretion.

Secreted rCRF(9–41) modulates CRF-dependent inflammatory and barrier responses in vitro

To assess whether the secreted bioactive peptide produced by the engineered sensing module retained functional activity, we performed in vitro analyses using purified recombinant rCRF(9–41) derived from engineered EcN. Direct co-culture of engineered bacteria with mammalian cells was not pursued due to instability of catecholamines under standard cell culture conditions and their rapid degradation by catechol-O-methyltransferase and monoamine oxidases expressed by mammalian cells [32]. Use of purified peptide enabled controlled evaluation of bioactivity independent of these constraints.

Purified YebF–rCRF(9–41) fusion protein was processed by enterokinase to release the active peptide and applied to human THP-1 macrophages and Caco-2 intestinal epithelial cells. Because CRF has been shown to prime innate immune responses through modulation of Toll-like receptor 4 (TLR4) signaling [33], we first examined the effect of rCRF(9–41) on CRF-induced TLR4 expression. CRF treatment increased TLR4 mRNA levels in both cell types, whereas pretreatment with rCRF(9–41) or a synthetic CRF antagonist [synCRF(9-41)] significantly attenuated this response (Fig 3I and 3J). To exclude the possibility that the observed effects resulted from YebF itself or purification-associated contaminants, purified YebF lacking the rCRF(9–41) cargo was processed under identical secretion, purification, and enterokinase-treatment conditions. Treatment with purified YebF alone did not suppress CRF-induced TLR4 expression in either THP-1-derived macrophages or Caco-2 cells (S11A and S11B Fig), supporting that the observed effects are specifically attributable to rCRF(9–41).

We next examined whether rCRF(9–41) could modulate inflammatory responses under conditions of combined neuroendocrine and microbial stimulation. In differentiated THP-1 macrophages, CRF alone had minimal effect on IL-6 and IL-1β expression, whereas co-stimulation with CRF and lipopolysaccharide (LPS) increased pro-inflammatory cytokine expression (Figs 3K, 3L, and S12). Pretreatment with rCRF(9–41) reduced cytokine induction to levels comparable to those observed with a synthetic CRF antagonist, indicating effective antagonism of CRF-dependent inflammatory priming.

To assess effects on epithelial barrier-associated responses, Caco-2 monolayers were treated with CRF, LPS, or CRF + LPS in the presence or absence of CRF antagonists. CRF + LPS co-stimulation markedly increased expression of the tight junction protein claudin-2 (Cldn2), whereas pretreatment with rCRF(9–41) or synthetic antagonist reduced Cldn2 accumulation (Fig 4A). Consistent with these observations, CRF + LPS increased paracellular permeability in FITC–dextran trans well assays, an effect that was attenuated by CRF antagonist pretreatment (Fig 4B).

Download:

• PNG
  larger image
• TIFF
  original image

Fig 4. Attenuation of gut epithelial barrier disruption and inflammatory signalling by secreted rCRF(9–41).

(A) Immunofluorescence analysis of Cldn2 expression in Caco-2 monolayers treated with CRF (100 nM) alone or in combination with LPS (2.5 µg/mL), following pretreatment with either the recombinant [rCRF(9–41)] or synthetic [synCRF(9–41)] antagonist (500 nM). The expression of Cldn2 was detected using a polyclonal primary antibody against Cldn2 and an Alexa Fluor 594 (red)-conjugated secondary antibody. Cell nuclei are counterstained with DAPI (blue) (B) FITC-dextran (FD10) concentration in the basal chamber of a transwell permeability assay with Caco2 cell monolayer. The cells were treated with CRF (100 nM) alone or in combination with LPS (2.5 µg/mL), following pretreatment with either rCRF(9-41) or synCRF(9-41) antagonist (500nM). Data are presented as mean ± SEM (n = 3) and One-way ANOVA followed by Tukey’s multiple comparisons test was used for the statistical analysis; *p < 0.05, **p < 0.01, ***p < 0.001, ****p < 0.0001vs. control (C) Heatmap of differentially expressed inflammatory genes in Caco-2 cells stimulated with CRF and LPS, in the presence or absence of the recombinant CRF antagonist rCRF(9–41), relative to untreated control groups. Differential gene expression analysis was performed utilizing the DESeq2 package in R. Data are visualized as row Z-scores computed from normalized transcript counts to illustrate relative up- or down-regulation across experimental treatments (n = 3). The data underlying this Figure can be found in S1 Data. Raw transcriptomic data can be found under BioProject accession number PRJNA1468903.

https://doi.org/10.1371/journal.pbio.3003926.g004

To further characterize transcriptional changes associated with CRF antagonism, we performed RNA sequencing of Caco-2 cells subjected to CRF + LPS stimulation with or without rCRF(9–41) pretreatment. CRF + LPS exposure induced expression of multiple inflammatory mediators, including IL-6, TNF-α, IL-18, NOD2, and COX-2, while reducing expression of TGFB1, a regulator of epithelial homeostasis (Fig 4C). Pretreatment with rCRF(9–41) attenuated these transcriptional changes, resulting in reduced inflammatory gene expression and partial preservation of TGFB1 levels.

Together, these results demonstrate that rCRF(9–41) secreted by the engineered sensing module retains biological activity in vitro and can modulate CRF-dependent inflammatory and barrier-associated responses under conditions of combined CRF and LPS stimulation.

Construction and expression of the QseBC catecholamine-sensing module

The qseBC gene cassette from EHEC was synthesized as a double-stranded DNA fragment (gBlock; Integrated DNA Technologies). The cassette was PCR-amplified using primers listed in S4 Table containing constitutive promoters (J23104, J23106, J23108, J23111, and J23114) from the Anderson promoter library (S3 Table) to enable promoter-strength–dependent tuning of qseBC expression. Amplified fragments were cloned into the pBbS8A plasmid backbone using EcoRI and BamHI restriction sites and transformed into the E. coli Top10 (S1 Table). All constructs were sequence verified prior to downstream use (S2 and S5 Tables).

Transcript profiling of Escherichia coli Nissle 1917 expressing EHEC QseBC

To analyze transcriptional changes associated with catecholamine sensing, transcriptomic profiling was performed on EcN expressing the EHEC-derived QseBC system following (−)-norepinephrine (Cat# A7257, Sigma-Aldrich) 100 μM induction. Three independent colonies of wild-type EcN and engineered EcN harboring pBbS8A-J111-qseBC (EHEC) were each inoculated into 5 mL of Dulbecco’s modified Eagle medium (11965092, DMEM) supplemented with 25 mM glucose, 0.6% (w/v) casamino acids, and Trace Metal Mix A5 (Cat# 9249, Sigma-Aldrich). Cultures were grown overnight at 37 °C with shaking at 220 rpm, then sub-cultured into freshly supplemented DMEM (Cat# 11965092, DMEM) and grown to mid-log phase (OD₆₀₀ ≈ 0.6). Cells were harvested by centrifugation at 4,000g, and total RNA was extracted from cell pellets using the RNeasy Mini Kit (Cat# 74106, Qiagen) according to the manufacturer’s instructions. RNA quality and concentration were assessed prior to sequencing. High-quality RNA samples were submitted to Novogene (NovogeneAIT Genomics) for library preparation and sequencing. Directional mRNA libraries were prepared and sequenced on an Illumina NovaSeq platform using paired-end 150-bp reads (PE150), generating approximately 12 million reads per sample.

Differential gene expression analysis was performed using the DESeq2 package in R. Genes with an adjusted p value (padj) < 0.05 and an absolute log₂fold change ≥ 0.5 were considered significantly differentially expressed.

Construction and characterization of EHEC flagellar promoter-based catecholamine biosensors

To evaluate the sensitivity of EHEC flagellar promoters to NE/Epi signaling in EcN, selected promoters (flgBCDEF, flgGHIJ, and flhDC) were synthesized as double-stranded gBlocks (Integrated DNA Technologies). Each promoter was cloned upstream of a RFP reporter in the pBbS8A plasmid backbone containing the EHEC-derived qseBC regulatory cassette.

To assess the impact of promoter length on transcriptional activity, truncated versions of the flhDC promoter (300 and 600 bp upstream of the transcription start site) were generated using primers listed in S4 Table, yielding flhDC_EHEC-300 and flhDC_EHEC-600 variants. These fragments were cloned upstream of the RFP reporter using the same cloning strategy. In addition, a synthetic promoter (SynflhDC) containing both high- and low-affinity QseB binding motifs were designed, synthesized as a gBlock and cloned upstream of the RFP reporter. To further tune promoter responsiveness, RNA polymerase σ⁷⁰/σ²⁸ binding motifs were modified using Anderson constitutive promoter variants (J23105, J23111, and J23117), and the corresponding promoter sequences were synthesized as gBlocks incorporating these motifs.

For biosensor characterization, single colonies were grown overnight in LB broth and subsequently sub-cultured into phenol red–free DMEM (Cat# 31053028, DMEM) supplemented with Trace Metal Mix A5 (Cat# 9249, Sigma-Aldrich). Cultures were induced with varying concentrations of norepinephrine (NE), epinephrine (Epi) (25 μM, 50 μM, 100 μM) and AI-3 [(3,5 dimethylpyrazin-2-ol-) 25 μM, 50μM, 100 μM)] and incubated under microaerobic conditions in sealed plate or sealed tube at 37 °C for 16 h. Fluorescence was quantified using a BioTek Synergy H1 plate reader (Agilent) with excitation/emission wavelengths of 535/600 nm. For flow cytometric analysis, cells were washed with 1× phosphate-buffered saline (PBS), resuspended, and analyzed using a BD Accuri C6 flow cytometer (BD Biosciences). All experiments were performed in biological triplicates, and fluorescence values were normalized to optical density at 600 nm (OD₆₀₀).

For decoupling studies, qseB_EHEC and qseC_EHEC were individually cloned under the control of the constitutive promoters J23104 and J23111, respectively. Plasmid constructs were introduced into electrocompetent EcN by electroporation. For analysis of QseB and QseC protein expression, EcN cultures were harvested by centrifugation at 4,000g, and cell pellets were lysed by heat treatment at 95 °C for 5 min. Crude lysates were resolved by SDS–PAGE and stained overnight with InstantBlue protein stain (Cat# ab119211, Abcam). Gels were washed twice with Milli-Q water and imaged to confirm protein expression. To assess sensing activity, a flhDC_EHEC-_{300–RBS–RFP} reporter cassette was cloned upstream of qseB_EHEC and qseC_EHEC in the decoupled constructs. Sensor sensitivity was then evaluated in response to norepinephrine at concentrations of 50 µM and 100 µM, as described above.

QseC mutant design and functional characterization

To investigate the structural determinants of ligand recognition by the QseC periplasmic sensor domain, a structural model of QseC was generated using AlphaFold2. Molecular docking was then performed to identify potential interactions between sensor domain QseC (34-156) and the small-molecule ligands NE and Epi. Docking simulations were carried out using AutoDock 4.5 software [34]. For each ligand, 100 docking poses were generated using default parameters, and the conformation with the lowest predicted binding energy was selected for further interaction analysis. The docking results identified residues within the QseC that are predicted to contribute to catecholamine binding, including Thr72, Asn74, Asp83, and Asp101. These residues were selected for experimental validation based on their predicted proximity to the ligand and conservation across QseC homologs. Site-directed mutagenesis was carried out to individually substitute the selected residues using the QuikChange Lightning Multi Site-Directed Mutagenesis Kit (Cat# 210513, Agilent), following the manufacturer’s instructions. All mutant constructs were verified prior to functional analysis. The biosensor performance of QseC variants was evaluated using the catecholamine-responsive genetic circuit described above, enabling assessment of the contribution of individual residues to NE and Epi sensing.

Membrane fractionation and analysis of QseC variants

Membrane fractionation was performed to assess localization of QseC variants and exclude gross protein misfolding or defective membrane association. Engineered E. coli Nissle 1917 (EcN) strains expressing wild-type QseBC, point mutants (N74A, D83A, D101A), or an empty-vector control (EcN-EV) were grown to mid-log phase (OD ₆₀₀ ~0.6) and harvested (5,000g, 10 min, 4 °C). Cell pellets were resuspended in lysis buffer (50 mM Tris-HCl pH 8.0, 150 mM NaCl, 1 mM EDTA, protease inhibitor cocktail) and lysed by probe sonication on ice (40% amplitude; 10 cycles of 10 s ON/30 s OFF). Cell debris was removed and crude membrane fractions were isolated by ultracentrifugation, for 30 min, 4 °C. Membrane pellets were solubilized in lysis buffer containing 1% Triton X-100 for 30 min at room temperature and clarified by ultracentrifugation, for 30 min, 4 °C. Solubilized membrane fractions were analyzed by SDS–PAGE and InstantBlue staining. Comparable abundance of membrane-associated QseC (~50.2 kDa) across wild-type and mutant strains suggested that the reduced signaling phenotypes were unlikely to arise from major defects in protein stability or membrane localization.

Protein expression, purification and enterokinase digestion

To generate the NE and Epi sensing and response construct, the CRF receptor antagonist peptide rCRF(9–41) encoded gene was synthesized from GenScript and placed under the control of the engineered catecholamine-responsive promoter SynflhDC. To assemble the gene cassette, the rCRF(9–41) peptide was fused to YebF, an Escherichia coli K-12 protein known for efficient secretion into the extracellular environment [35]. An enterokinase (EK) recognition site (Asp-Asp-Asp-Asp-Lys) was inserted between YebF and rCRF(9–41) to allow proteolytic cleavage and activation of the peptide. DNA fragments encoding the SynflhDC promoter, the YebF–EK–rCRF(9–41) fusion sequence, and the pBbS8A–J104–qseBC plasmid backbone were assembled using Gibson Assembly with the HiFi DNA Assembly Master Mix (Cat# E2621L, New England Biolabs), following the manufacturer’s instructions. The resulting construct was transformed into EcN to generate a strain capable of sensing NE/Epi and secreting rCRF(9–41). To evaluate secretion of rCRF(9–41), EcN harboring the SynflhDC–YebF–EK–rCRF(9–41) construct was cultured in Dulbecco’s Modified Eagle Medium (Cat# 11965092, DMEM) supplemented with Trace Metal Mix A5. Norepinephrine was added to the culture medium at final concentrations of 25 and 50 μM. Cultures were incubated overnight at 37 °C under microaerophilic conditions in a sealed tube. Cells were subsequently pelleted by centrifugation, and the culture supernatant was concentrated using Vivaspin 500 centrifugal concentrators (Cat# VS15T92, 3 kDa MWCO; Sartorius). Expression of the fusion protein and release of the recombinant CRF antagonist [rCRF(9–41)] were assessed by Western blotting using an anti-His tag antibody. Protein samples were separated by SDS–PAGE and transferred onto a 0.45 μm nitrocellulose membrane (Bio-Rad). Membranes were blocked with 3% bovine serum albumin (BSA) for 1 h at room temperature and incubated overnight at 4 °C with HRP-conjugated anti-His antibodies (MA121315, Invitrogen). After washing with Tris-buffered saline containing 0.1% Tween-20 (TBST), signals were detected using SignalFire Elite ECL reagent (Cat# 12757, Cell Signaling Technology) according to the manufacturer’s instructions. Blots were visualized using the iBright CL1500 Imaging System (Invitrogen).

For bioassay studies, the secreted fusion peptide [YebF–EK–CRF(9–41)] was purified from the culture supernatant using Ni Sepharose Excel resin (Cytiva). The purified protein was buffer-exchanged into 20 mM Tris-HCl and digested with recombinant enterokinase (Merck) at a ratio of 1 U per 50 μg of protein. Digestion was performed either overnight at room temperature or for 1 h at 37 °C.

Quantification of secreted rCRF(9–41)-HisTag

Single colonies of the NE-responsive SynflhDC-based EcN biosensor strain carrying the rCRF(9–41)-HisTag output cassette were grown overnight in LB broth at 37 °C. Overnight cultures were sub-cultured into phenol red–free DMEM (Cat. No. 31053028, Gibco/Thermo Fisher Scientific) supplemented with Trace Metal Mix A5 (Cat. No. 92949, Sigma-Aldrich). Cultures were induced with 50 μM norepinephrine (NE) and incubated under microaerobic conditions in sealed tubes at 37 °C for 16 hours. Following induction, cultures were centrifuged, and clarified supernatants were collected.

Secreted rCRF(9–41)-HisTag levels were quantified using a His-tag ELISA Kit (Abcam, ab322565) according to the manufacturer’s instructions. Peptide concentrations were calculated from a standard curve and expressed as nanomolar values.

Cell culture and differentiation THP-1 cell lines

THP-1 (human monocytic) cells were cultured in RPMI-1640 (Cat# SH40007.01, Cytiva) medium supplemented with 10% fetal bovine serum (Cat# S1810, Atlantis Bioscience) and maintained in suspension at 37 °C with 5% CO₂, with medium changes every 2–3 days. Differentiation into macrophage-like cells was induced by treating THP-1 cells with 100 ng/mL phorbol 12-myristate 13-acetate (P8139, Sigma-Aldrich) for 48–72 hours. To examine the effects of the CRF antagonist, the medium was replaced with fresh RPMI-1640 without FBS prior to stimulation. Differentiated THP-1 macrophages were then stimulated with 100 nM CRF, 1.0 µg/mL lipopolysaccharide (Cat# L4391, LPS-Sigma-Aldrich), or a combination of CRF (100 nM) and LPS (1.0 µg/mL). In parallel experiments, macrophages were pretreated for 1 hr with either the enterokinase digested recombinant enterokinase-digested fusion peptide [YebF-EK-CRF (9−41)-500nM] or the synthetic CRF antagonist [synCRF(9-41)-500 nM] prior to CRF + LPS stimulation. Levels of IL-6 and IL-1β in the culture supernatants were quantified using Human IL-6 and IL-1β ELISA Kit (Cat# ab178013 and ab46052, Abcam) according to the manufacturer’s instruction. Synthetic CRF and synCRF(9–41) peptides used in these experiments were obtained from GenScript.

Cell culture of Caco-2 and RNA-Seq analysis of treatment groups

Caco-2 cells were cultured in DMEM supplemented with 15% Fetal Bovine Serum (FBS) and 1× Penicillin–Streptomycin (PS) solution. Cells were seeded in 6-well plates and maintained at 37°C in a 5% CO₂, with regular media changes until confluence is reached. To investigate the effect of the [rCRF(9-41)] antagonist, the culture medium was replaced with fresh FBS-free DMEM. Confluent Caco-2 cells were stimulated with a combination of CRF (100 nM) and LPS (2.5 μg/mL). In parallel, before CRF and LPS stimulation, an additional set of Caco-2 cells was pretreated for 1 hour with either enterokinase-digested fusion peptide [YebF-EK-CRF (9-41)-500nM] or synthetic CRF antagonist [synCRF(9-41)-500nM]. Following antagonist treatment, Caco-2 cells were washed once with 1× PBS and RNA was extracted using TRIzol Reagent (Cat# 15596018, Invitrogen) according to an optimized protocol [36]. Briefly, 200 μl of TRIzol Reagent was added to each well and adherent cells were removed by pipetting. Fifty μl of chloroform was then added, and samples were vigorously vortexed for 30 s and placed at room temperature for 5 min. Following centrifugation at 12,000g at 4 °C, the aqueous layer containing RNA was carefully transferred to a clean tube and RNA was precipitated using standard isopropanol methods. RNA pellets were reconstituted in DEPC-treated water and stored at −20 °C. RNA from three independent replicates from each group: (1) Control, (2) CRF+ LPS, and (3) CRF+ LPS+ rCRF(9-41) pretreatment was sent for directional mRNA library preparation (NovogeneAIT Genomics). Poly-A enrichment was performed, and libraries were sequenced on the NovaSeq platform. The sequencing was conducted with paired-end 150 bp (PE150) reads, generating approximately 40 million reads (12 Gb) per sample, providing high-depth transcriptome data for downstream analyses. Data analysis was performed using NovoMagic, a cloud-based bioinformatics platform developed by Novogene. In addition, differential gene expression analysis was conducted using the DESeq2 R package. DESeq2 and Z-score normalization was conducted to normalize between the groups.

Transwell and permeability assay

To perform the FITC-dextran permeability assay, Caco-2 cells were seeded at a density of 2.5 × 10⁴ cells per well in 24-well trans wells (Cat# PIHP01250, Millipore). Thereafter, cells were subjected to polarization with FBS-free DMEM on the apical compartment and FBS-supplement DMEM on the basal compartment. Media change was conducted every 2 days for 18 days. On the final day, a media change was conducted, and cells were stimulated with 100 nM CRF, 2.5 μg/mL LPS, or a combination of CRF (100 nM) and LPS (2.5 μg/mL) on apical chamber and incubated for 24 hours. In another group, cells were pretreated with recombinant CRF antagonist peptide [rCRF(9-41)] or synthetic CRF antagonist [synCRF (9-41)] for 1 hour before stimulation with the CRF and LPS combination and incubated 24 hours.

On the day of the experiment, the culture medium was removed from the apical side and washed with PBS. Thereafter, 100 μg/mL solution of FITC-dextran 10 kDa (Cat# FD10S-100 MG, Sigma-Aldrich) was prepared in PBS and added to the apical compartment (200 μL), while phenol red-free DMEM was added to the basolateral compartment (600 μL). The transwell plates were incubated at 37° C for 2 hours. After incubation, 100 μL samples were collected from the basolateral compartment and replaced with fresh DMEM media to maintain volume consistency. The fluorescence of the samples was measured using a fluorescence plate reader at an excitation wavelength of 485 nm and emission wavelength of 520 nm. We made a standard curve of FITC-Dextran to calculate the FITC-dextran flux. Permeability was calculated based on the fluorescence intensity of FITC-Dextran in the basolateral compartment.

qRT-PCR for TLR4 expression in THP-1 and Caco-2

Total RNA was extracted from THP-1 and Caco-2 cells using the TRIzol method, as described previously [37,38]. Complementary DNA (cDNA) was synthesized using the HiScript III cDNA Synthesis Kit (Cat# R312-02, Vazyme) according to the manufacturer’s instructions. Briefly, 1 μg of total RNA was treated to remove genomic DNA prior to reverse transcription using the HiScript III enzyme mix at 37 °C. The resulting cDNA was diluted, and quantitative PCR was performed using Luna Universal qPCR Master Mix (Cat# M3003, New England Biolabs) on a CFX Connect Real-Time PCR Detection System (Bio-Rad Laboratories). Relative RNA expression levels were calculated using the 2^−ΔΔCt method, with β-actin used as the internal housekeeping gene for normalization [39].

Immunofluorescence staining

Caco-2 cells were seeded onto sterile glass coverslips placed in 6-well culture plates and grown to the desired confluence. For immunofluorescence staining, culture medium was removed and cells were washed three times with 1× PBS to remove residual medium. Cells were then fixed with 4% paraformaldehyde (Cat# 16005, Sigma-Aldrich) for 10 min at room temperature or overnight at 4 °C. Following fixation, cells were washed three times with PBS and blocked with 1% BSA in PBS for 1 hour at room temperature to reduce nonspecific antibody binding. Cells were subsequently incubated overnight at 4 °C with gentle agitation with a rabbit anti–claudin-2 polyclonal primary antibody (Cat# 51-6100, Thermo Fisher Scientific). After primary antibody incubation, cells were washed three times with PBS and incubated with an Alexa Fluor 594–conjugated anti-rabbit IgG secondary antibody (Cat# 8889, Cell Signaling Technology) for 1 hour at room temperature in the dark. Nuclei were counterstained with 4′,6-diamidino-2-phenylindole (DAPI; NucBlue, Cat# R37605, Invitrogen). Coverslips were mounted onto glass slides using a mounting medium and sealed. Fluorescent images were acquired using a Leica DM4000 B fluorescence microscope under identical exposure settings for all experimental conditions.

Statistical analysis

All statistical analyses were performed using GraphPad Prism software (GraphPad Software, San Diego, CA). Data are presented as mean ± standard error of the mean (SEM), unless otherwise indicated. Statistical significance was assessed using one-way analysis of variance (ANOVA) followed by Tukey’s multiple-comparison test to compare differences among experimental groups. A p value of < 0.05 was considered statistically significant, with significance thresholds defined as *p < 0.05, **p < 0.01, ***p < 0.001, ****p < 0.0001.