Adhesion Module Results
1. Identification and Validation of Lp_0018 as the Optimal OppA Variant
The heatmap displays the docking energies (ΔG) of seven Lactobacillus plantarum OppA variants with candidate host ligands. Darker blue indicates stronger binding affinity. (B) Bubble chart summarizing the docking energies and docking scores between mucin and the seven OppA proteins, where circle size represents binding strength and color indicates different OppA variants. As shown in panel (A), among the tested variants, Lp_0018 exhibited the most favorable binding profile with NaHS and was therefore selected for subsequent experimental validation. Panel (B) reveals potential binding between mucin (present in the intestinal environment) and OppA, further validating the necessity of incorporating a hypoxia-responsive suicide element.
(A--G) Representative binding conformations of NaHS with individual OppA proteins (A: Lp_0018; B: Lp_0020; C: Lp_0783; D: Lp_0092; E: Lp_0201; F: Lp_3686; G: Lp_1261). NaHS molecules are shown in red, OppA binding residues in green. Hydrogen bonds and key interactions are indicated.
We constructed an overexpression plasmid and expressed Lp_0018 in L. plantarum WCFS1. SDS-PAGE analysis revealed a distinct protein band at the expected molecular weight (highlighted in Fig. X), confirming successful expression of Lp_0018.
2. Inducer Optimization for Adhesion Efficiency
To ensure robust expression of the adhesion modules, we first evaluated the effect of inducer concentration on cell--cell adhesion. Overnight cultures of Escherichia coli Nissle 1917 (EcN) harboring Nb3 and Ag3 adhesion pathways were exposed to a gradient of IPTG concentrations, with non-induced samples serving as controls. After 24 h of induction, OD600 values of supernatants were recorded to quantify aggregation. The titration assay demonstrated that adhesion efficiency increased with IPTG concentration up to a plateau. The optimal working concentration was determined to be 0.8 mM, which was subsequently applied in all downstream experiments. This result highlights the necessity of precise inducer tuning to balance gene expression and cellular fitness.
3. Validation of Orthogonality in Adhesin Pairs
To rigorously evaluate the orthogonality of engineered adhesins, we designed a three-step validation strategy.
(i) Quantitative co-culture assay. We first co-cultured EcN strains carrying Nb3 and Ag3, as well as Nb2 and Ag2 adhesion plasmids, under induced and non-induced conditions. After 24 h of induction, OD600 measurements of the supernatants revealed a significant increase in aggregation in the induced groups. This result demonstrated the successful establishment of specific adhesion interactions at the population level.
(ii) Fluorescence microscopy of Nb3--Ag3 interaction. To directly visualize orthogonal adhesion, we introduced fluorescence reporter plasmids (pJUMP-RFP into Nb3 cells and pJUMP-sfGFP into Ag3 cells). Fluorescence microscopy confirmed that red Nb3 cells adhered specifically to green Ag3 cells upon induction, forming distinct aggregates. In contrast, control groups without induction exhibited minimal clustering.
(fig.X Fluorescence microscopy images show EcN cells labeled in red and green. (I) Without inducer, EcN cells were mostly dispersed, with limited cell--cell contact. (J) With inducer, EcN cells exhibited clear clustering and close contacts, indicating successful self-adhesion upon induction. Black boxes highlight representative adhesion events.)
(iii) Construction of an orthogonality matrix. Finally, we extended the validation by mixing strains carrying different adhesins (Nb3, Ag3, Nb2, Ag2) without fluorescent markers. Co-culture experiments showed preferential adhesion within cognate pairs (Nb3--Ag3, Nb2--Ag2), while non-cognate combinations displayed negligible interaction. The orthogonality matrix derived from these assays highlighted the pairwise specificity of adhesins, confirming the modular and programmable nature of the system.
4. Intraspecies and Interspecies Adhesion Assays
We next investigated adhesion efficiency both within EcN strains and between EcN and Lactobacillus plantarum WCFS1.
(i) Quantitative co-culture of EcN and WCFS1. For interspecies assays, Nb2-expressing EcN strains (without fluorescence reporters) were co-cultured with WCFS1. OD600-based quantification revealed that induced groups exhibited significantly higher aggregation compared with non-induced controls, confirming functional interspecies adhesion.
(ii) Fluorescence microscopy of Nb2--WCFS1 interaction. To visualize the adhesion events, we further constructed a red-fluorescent Nb2--EcN strain and co-cultured it with non-fluorescent WCFS1. Fluorescence microscopy images revealed that induced EcN cells adhered tightly to WCFS1 surfaces, while controls showed dispersed cell distribution.
5. A single diagram summarising our adhesive plate
(A) Representative fluorescence microscopy image of Ag3--Nb3 aggregates. Red cells (Nb3) express RFP and green cells (Ag3) express GFP. (B) Quantification of cell--cell adhesion shown in (A). Significant differences were observed between induced and non-induced groups (*p < 0.0001). (C) Orthogonality matrix derived from (B), showing the probability that cells of one color (row) adhere to cells of another color (column). (D) Representative fluorescence microscopy images of interspecies adhesion between Lactobacillus plantarum and Escherichia coli. Red Nb2 cells adhered to L. plantarum cells. (E) Quantification of interspecies adhesion shown in (D). The JC group (joint cultivation) displayed significant differences between induced and non-induced conditions (*p < 0.0001). (F, G) Macroscopic views of Ag2--Nb2 aggregates. (F) Aggregation between EcN strains; (G) aggregation between EcN and L. plantarum WCFS1. (H) IPTG titration curve determining the optimal inducer concentration. The optimal working concentration was identified as 0.8 mM.
Control Module Results
1. AHL system Results
(A) Uninduced blank control. Cells without inducer addition showed negligible fluorescence, confirming the absence of leaky expression.
(B) Endogenously induced samples. Cells harboring the endogenous Lux promoter exhibited faint but detectable red fluorescence, suggesting a low basal promoter activity that enabled limited expression under native conditions. The weak fluorescence and high background are likely due to the intrinsic low strength of the Lux promoter and the contribution of cellular autofluorescence and uneven illumination.
(C) Exogenously induced samples. Upon addition of external inducers, a strong and uniform red fluorescence signal was observed, demonstrating effective activation of the promoter and successful protein expression.
(D) Quantitative fluorescence analysis. Fluorescence intensity was quantified using a microplate reader. The blank showed negligible signal, indicating minimal leakiness. Endogenous induction yielded a weak increase (10.2±1.2-fold vs. blank, CV=5.46%), whereas exogenous induction produced a strong and uniform response (81.1±8.6-fold vs. blank; CV=1.4%).
(E) Lux promoter activity curve. Time‐ and concentration‐dependent induction profiles of the Lux promoter were measured. The fluorescence intensity increased in proportion to inducer concentration and incubation time, indicating a dose--response relationship and confirming the dynamic regulatory capacity of the Lux system.
2. AIP system Results
(A) Uninduced blank control. Cells without inducer addition exhibited negligible background fluorescence, confirming the absence of leaky expression from the PsppA promoter.
(B) Endogenously induced samples. Cells under endogenous induction showed weak but detectable red fluorescence, indicating that the native regulatory elements of PsppA possess basal activity even without exogenous stimulation. The relatively low fluorescence intensity and high background noise may result from weak promoter strength and cellular autofluorescence.
(C) Exogenously induced samples. Upon addition of external inducers, cells displayed a strong and homogeneous red fluorescence signal, suggesting that PsppA is highly responsive to exogenous stimuli and capable of robust activation.
(D) Quantitative fluorescence measurement. The fluorescence intensity of each group was quantified. The blank showed negligible signal, indicating minimal leakiness. Endogenous induction yielded a weak increase (18.6±5.8-fold vs. blank, CV = 31.4%), whereas exogenous induction produced a strong and uniform response (54.1 ± 3.5-fold vs. blank, CV = 6.5%).
(E) Time‐dependent induction curve of PsppA promoter activity. Fluorescence intensity was recorded over time following induction. PsppA reached half-maximal activity at ~1 h (t₅₀ = 1.00 h) and approached a stable plateau by 2--3 h (mean ≈ 0.94 RPU, normalized to P23). The maximal response occurred at 3 h (Y_max = 0.97 RPU; dynamic range = 0.87 RPU), after which the fluorescence signal declined slightly at ~0.11 RPU·h⁻¹.
(F) Concentration‐dependent induction curve of PsppA promoter activity. Fluorescence increased in a dose‐dependent manner with inducer concentration, reaching saturation at higher levels, consistent with a sigmoidal regulatory behavior characteristic of inducible promoter systems.
3. Application Results
(A) Schematic of bidirectional regulation between Lactiplantibacillus plantarum and Escherichia coli. The left panel illustrates L. plantarum--mediated control of the E. coli therapeutic production module; the right panel shows E. coli--mediated regulation of L. plantarum adhesion/function.
(B) SDS--PAGE validation in L. plantarum carrying pSIP403-Pspp-RFP. Cells were induced and harvested after 6 h; whole-cell lysates were analyzed by SDS--PAGE to confirm induction-dependent expression.
(C) Melanin production assay in E. coli carrying pET-J23100-LuxR-Plux-HpaBC. After 24 h of induction, pigment formation was quantified by absorbance at OD400; PC denotes the positive control.
(D) Visual phenotype of E. coli cultures after 24 h induction. Representative color changes across experimental groups following induction of pET-J23100-LuxR-Plux-HpaBC.
Therapeutic Module Results
1. L-dopa production module
This experiment aimed to achieve two objectives: first, to evaluate the influence of the exogenous plasmid and inducer on E. coli growth by comparing the growth curves among the three groups; and second, to confirm the successful construction and functional expression of the production module by measuring L-DOPA yield over time in the induced group.
Analysis of the growth curves (Figure 1) revealed that neither plasmid carriage nor IPTG induction significantly affected the growth of the engineered E. coli strain.
The yield-time curve plotted based on HPLC data (Link 1) (as shown in Figure 2) reveals that the maximum concentration of levodopa in the sampled supernatant reached 36,000 ng/mL. Subsequently, as substantial amounts of levodopa were oxidised into melanin, its concentration in the sample supernatant progressively decreased.
To quantitatively assess the degree of oxidation of levodopa during fermentation, we measured the absorbance of the fermentation supernatant at 400 nm based on literature reports. One OD400 unit corresponds to 0.066 g/L melanin. We calculated its approximate concentration, with specific data and corresponding levodopa content shown in Figure 3. Based on this, we further estimated the total amount of oxidised levodopa.
2. Glutathione production module
Analysis of the growth curve (as shown in Figure 4) indicates that the carriage of the exogenous plasmid, the addition of inducers, or the supplementation of amino acids did not exert a significant effect on the growth of the engineered Escherichia coli.
The yield-time curve plotted based on HPLC data (Link 2) (as shown in Figure 5) reveals that the maximum concentration of reduced glutathione in the sampled supernatant reached 252 ng/mL. Over time, GSH was progressively oxidised to GSSG, with GSH levels gradually decreasing and GSSG levels gradually increasing.
Safety Module Results
1. Verification of the Hypoxia-Responsive Suicide Module
We first verified the function of the hypoxia-responsive suicide module. As shown in Figure A, compared with aerobic culture, the number of colonies formed under hypoxic conditions after 48 h was dramatically reduced, indicating that the constructed hypoxia-responsive suicide element was effective and could significantly inhibit the growth of Escherichia coli.
2. Validation of the Inducible Suicide Modules (Arabinose- and Nisin-Inducible Systems)
We evaluated the inducible suicide modules. As shown in Figure C, compared with the control group, both the engineered E. coli carrying the arabinose-inducible suicide system and the engineered Lactobacillus plantarum harboring the nisin-inducible suicide system exhibited markedly slower OD₆₀₀ increases after the addition of arabinose and nisin, respectively. These results demonstrate that both engineered strains responded effectively to the corresponding inducers.
3. Evaluation of the Temperature-Sensitive Suicide System
Finally, we assessed the temperature-sensitive suicide system. We first examined the effectiveness of the temperature-responsive protein and the downstream promoter. As shown in Figure D, the engineered E. coli harboring the single-plasmid system showed no significant difference in specific fluorescence intensity after preculturing at 30 °C and 37 °C for 18 h followed by 24 h of main culture, indicating that the single-plasmid temperature-sensitive system exhibited poor performance. In contrast, the E. coli carrying the dual-plasmid system displayed a markedly higher specific fluorescence intensity at 30 °C than at 37 °C under the same conditions, suggesting that the dual-plasmid construct possessed stronger temperature sensitivity. Subsequently, we replaced the red fluorescent protein in the dual-plasmid temperature-sensitive--fluorescence system with the toxin gene mazF. As shown in Figure B, the engineered E. coli exhibited a slightly reduced colony number after static incubation at 37 °C for 24 h compared with 30 °C, suggesting a modest temperature-dependent effect.
Figure 19. Validation of the synthetic safety modules.
(A) Colony growth of E. coli harboring plasmid pET-phyb-MazF after 48 h of incubation under aerobic (top row) and hypoxic (bottom row) conditions at 37 °C. The drastic reduction of colonies under hypoxia demonstrates the effectiveness of the hypoxia-responsive suicide element.
(B) Colony morphology of E. coli carrying pETDuet-pRM-CI857 and pJUMP-pR-phIF-pPhIF-MazF after 24 h of static incubation at 30 °C (top) and 37 °C (bottom). The slightly lower colony number at 37 °C suggests a modest temperature-dependent activation of the dual-plasmid thermo-responsive suicide system.
(C) Growth curves of engineered bacteria carrying inducible suicide modules. Top: E. coli harboring pBAD-MazF cultured with or without arabinose induction. Bottom: Lactobacillus plantarum containing Pnis-MazF cultured with or without nisin induction. Both strains showed attenuated growth upon inducer addition, confirming successful activation of the inducible suicide circuits.
(D) Specific fluorescence intensity of E. coli strains transformed with temperature-responsive red-fluorescent protein reporters. Top: single-plasmid construct (pETDuet-J23110-Cl857-pR-phIF-pPhIF-RFP); Bottom: dual-plasmid construct (pETDuet-pRM-Cl857 + pJUMP-pR-phIF-pPhIF-RFP). The dual-plasmid system exhibited markedly higher fluorescence at 30 °C than 37 °C, indicating improved temperature sensitivity.
Co-culture of Two Bacteria
To investigate the co-culture conditions of EcN and Lp, we explicitly established the following three experimental systems (detailed experimental design is provided in the Engineering section): (1) EcN monoculture expressing sfGFP; (2) Lp monoculture expressing mRFP; (3) a co-culture of the two strains mixed in a 1:1 ratio. This study pursued two objectives: firstly, to quantitatively analyse the growth kinetics of both strains under separate and coexisting conditions by comparing growth curves across the three groups; secondly, to elucidate the spatial distribution and relative density of the two strains within the co-culture system via fluorescence microscopy.
Analysis of the growth curves revealed that the growth trends in the single-culture and co-culture systems were largely overlapping. This indicates that, under the experimental conditions employed, co-culture did not significantly inhibit the growth of either strain.
To further validate the coexistence capability of EcN and Lp, we conducted cell counting analysis on fluorescence microscopy images from three experimental systems. By quantifying cell numbers in both single-strain cultures and co-culture systems, we calculated and compared their relative proportions within the mixed culture. Statistical results indicate that under co-culture conditions, both strains maintained high cell densities, with their proportions largely consistent with the initial inoculation ratios. This quantitative data confirms that EcN and Lp can stably coexist and grow well within the established co-culture system.
A. 4 h post-incubation; B. 6 h post-incubation; C. 8 h post-incubation; D. 10 h post-incubation.
