I. Preventive Effect of Sodium Butyrate on Neuroinflammation

In this section, building on existing literature support, we further verified the preventive effect of sodium butyrate on inflammation in microglia and neurons through a series of cell biology experiments. All cell lines used in the experiments were immortalized cell lines, and their use complied with relevant ethical guidelines. Our experimental results are presented below.

Biosafety Verification of Sodium Butyrate

To investigate whether butyrate itself exhibits additional toxicity to cells, we treated mouse microglial BV-2 cells and hippocampal neuronal HT22 cells with different concentrations of sodium butyrate. After treatment, cell viability was measured using flow cytometry following incubation with CCK-8 reagent. The concentration gradient of sodium butyrate was established based on previous studies on butyrate alleviating neuroinflammation (Wang et al., 2022). Detailed protocols for this experiment can be found in the Protocol section.

Figure 1. Cell viability of BV-2 and HT22 cells treated with varying concentrations of NaB. (BV-2: left; HT22: right; control: without NaB treatment)

The results showed that, compared to the control group (untreated with sodium butyrate), sodium butyrate at concentrations ranging from 0.1 mM to 1 mM did not affect the viability of BV-2 or HT22 cells. At a concentration of 2 mM, sodium butyrate caused a slight decrease in the viability of both cell types, although the effect was minimal. These findings indicate that butyrate at concentrations between 0.1 mM and 2 mM does not exert toxic effects on neural cells and does not interfere with their normal physiological activities.

Protective Effect of NaB Pretreatment Against Neuroinflammation via Modulation of BV2 Polarization

Previous studies have shown that systemic inflammation may allow inflammatory mediators to cross the blood-brain barrier and affect microglia, further exacerbating neuroinflammation (Zhang et al., 2024). This process, along with other mechanisms, can lead to neuronal degeneration and damage, ultimately triggering Alzheimer's disease (AD) . Therefore, the state of microglia is critical to the pathogenesis of AD. Meanwhile, butyrate has been reported to exert therapeutic effects on microglia under inflammatory conditions. Therefore, in this series of experiments, we aimed to investigate whether butyrate can prevent microglial inflammation and, by protecting microglia, mitigate further damage to neurons.

As for the experiment, we pretreated microglial cells with different concentrations of sodium butyrate. After 24 hours of incubation with sodium butyrate, lipopolysaccharide (LPS) was applied to induce inflammation in microglia, simulating the inflammatory mediators that cross the blood-brain barrier and stimulate microglial activation in AD. The treated cells were incubated with CCK-8 reagent and divided into two groups for simultaneous assessment of cell viability and intracellular reactive oxygen species (ROS) levels using flow cytometry. Detailed protocols for this experiment can be found in the Protocol section.

Figure 2. Effect of NaB pretreatment on LPS-induced loss of cell viability in BV-2 cells. (Control: untreated; LPS: 100 ng/mL LPS only)

Cell viability assays revealed that compared to the blank group (no treatment), the LPS-treated control group (0 mM sodium butyrate) exhibited a significant reduction in cell viability, indicating that LPS successfully mimicked AD-related inflammatory damage to microglia. Pretreatment with 0.1-1 mM sodium butyrate restored cell viability to levels comparable to the blank group, with the most pronounced effect observed in the 1 mM treatment group. However, pretreatment with 2 mM and 5 mM sodium butyrate resulted in certain cytotoxicity, as cell viability decreased with increasing concentrations, though it remained higher than that of the LPS-treated control group. This suggests that sodium butyrate at concentrations above 2 mM may have adverse effects on microglia but still retains a strong preventive effect against inflammation.

Figure 3. Effect of NaB pretreatment on LPS-induced intracellular ROS levels in BV-2 cells, assessed by DCFH-DA fluorescence. (Control: untreated; LPS: 100 ng/mL LPS only)

The results of ROS level detection were consistent with those of cell viability assays. Compared to the LPS-treated control group, NaB at concentrations of 0.1–2 mM alleviated LPS-induced inflammation in BV2 cells to varying degrees. Notably, after treatment with 1 mM NaB, the proportion of ROS-positive cells decreased by 60.3% (from 97.5% to 37.2%), returning to levels similar to those of the blank group (30.5%). However, after pretreatment with 5 mM NaB, the proportion of ROS-positive cells increased by 24% compared to the 1 mM NaB treatment group (from 37.2% to 61.2%), indicating that higher concentrations of NaB may exhibit certain cytotoxicity, consistent with the cell viability results. These findings suggest that the inhibitory effect of sodium butyrate on oxidative stress may be highly concentration-dependent, with an optimal concentration of 1 mM, below or above which the protective effect diminishes.

These results demonstrate that low concentrations of sodium butyrate exert strong anti-inflammatory and protective effects. Sodium butyrate at concentrations of 0.1–1 mM effectively prevents microglial inflammation, and microglia pretreated with these concentrations exhibit enhanced resistance to externally induced inflammation and oxidative stress. Among these, 1 mM sodium butyrate provided the most optimal preventive effect.

After demonstrating that sodium butyrate can protect microglia, we further investigated whether this protective effect could mitigate inflammation-induced damage to neurons, thereby more fundamentally preventing the onset of AD.

In the pathogenesis of AD, neuroinflammation propagates sequentially from microglia to neuronal cells. To simulate this process, we stimulated mouse microglial BV-2 cells by using LPS to establish an inflammatory model and co-cultured them with mouse hippocampal neuronal HT22 cells. The levels of reactive oxygen species (ROS) and the pro-inflammatory cytokine TNF-α in the culture supernatant were measured using flow cytometry to evaluate the exacerbating effect of microglial polarization on neuronal inflammation—a process often regarded as a critical marker of AD progression. During the pretreatment phase, we selected NaB concentration gradients of 1 mM and 2 mM, based on the significant protective effects observed in previous microglial experiments.

Figure 4. Effect of NaB pretreatment on the levels of ROS and TNF-α in a LPS-stimulated BV-2/HT22 co-culture system. (Control: untreated; LPS: 100 ng/mL LPS only)

The results showed that, compared to microglia treated only with LPS, microglia pretreated with 1 mM and 2 mM NaB induced lower levels of ROS and TNF-α secretion in neuronal cells after LPS induction. The 1 mM NaB pretreatment exhibited the strongest inhibitory effect on LPS-induced responses, reducing ROS and TNF-α secretion levels by 27.4% (from 55.6% to 28.2%) and 11.1% (from 25.9% to 14.8%), respectively. This indicates that NaB pretreatment not only protects microglia, but also enables the protected microglia to further reduce damage to neurons when exposed to external inflammatory stimuli. These findings suggest that butyrate can protect other neural cells in the brain by modulating microglial polarization (inflammation), thereby preventing the onset of AD.

II. Construction and Functional Verification of the "AND Gate" System

Plasmid Construction and Transformation Verification

We constructed several primary plasmids, spanning four design-build-test-learn (DBTL) cycles. Detailed information on these plasmids can be found in the "The Design-Build-Test-Learn Cycles" section of our Engineering Framework. During experimentation, we successfully transformed four of these plasmids into the INVSc1 yeast strain. The designs of these constructs will be presented subsequently. To confirm the plasmid was successfully integrated into yeast, we first screened transformants using appropriate dropout media, followed by verification of target plasmid presence through colony PCR. The PCR amplification products were analyzed via agarose gel electrophoresis, and correct insert sequences were further confirmed by sequencing.

Our plasmid design maps and experimental results are presented below. We explicitly state that we designed these plasmid maps using SnapGene software. We extend our special gratitude to SnapGene for providing our team with complimentary sponsorship, enabling efficient design and documentation of our constructs.

Figure 5. a. Map of plasmid pGBKT7-pFBP1-EGFP ; b. Map of plasmid pGBKT7-pMsn2-EGFP; c. Map of plasmid pGBKT7-pTRR1-EGFP; d. Map of plasmid pRS416-TAAR5-GPA1-EGFP.

Figure 6. The PCR result (left) and schematic diagram (right) of pFBP1-EGFP fragment. The band was identical to the expected length of 1719 bp.

Figure 7. The PCR result (left) and schematic diagram (right) of pMsn2-EGFP fragment and pTRR1-EGFP fragment. The band was identical to the expected length of 1834 bp and 1273 bp.

Figure 8. The PCR result and schematic diagram of TAAR5-GPA1-EGFP fragment. The full fragment was amplified in two segments: TAAR5 (1186 bp) and EGFP (721 bp), both of which showed bands identical to the expected length.

Figure 9. Positive plasmid transformations confirmed by auxotrophic selection. (11: pGBKT7-pFBP1-EGFP; 12: pGBKT7-pMsn2-EGFP; 13: pGBKT7-pTRR1-EGFP; 16: pRS416-TAAR5-GPA1-EGFP)

ROS Response of pGBKT7-pTRR1-EGFP Transformants

After successfully transforming the pGBKT7-pTRR1-EGFP plasmid into yeast, we analyzed the transformants' response to ROS signals by monitoring the expression of green fluorescent protein (EGFP).

Figure 10. Fluorescence microscopic analysis of pGBKT7-pTRR1-EGFP transformants' response to ROS stress. (Control: untreated)

We firstly used fluorescence microscopy to detect EGFP emission, verifying whether the transformed plasmid expressed functionally. The observed minimal background fluorescence in untreated controls suggests a basal level of promoter activity. Results demonstrated that after treatment with 5 mM hydrogen peroxide (H₂O₂), the proportion of yeast cells exhibiting green fluorescence was significantly higher in treated samples compared to untreated controls. This indicates that yeast cells harboring the pGBKT7-pTRR1-EGFP plasmid can perceive external oxidative stimuli, subsequently activating the expression of the downstream gene product (EGFP in this case).

To quantitatively assess the ROS-responsive capability of the pGBKT7-pTRR1-EGFP plasmid, we applied varying concentrations of H₂O₂ to simulate reactive oxygen species (ROS) signals detectable by the PTRR1 promoter element. We then measured the corresponding EGFP fluorescence intensity using flow cytometry. To ensure accurate evaluation of the plasmid's responsiveness, we designed a wide range of H₂O₂ concentrations to avoid missing potential response dynamics due to insufficient dosage or minimal variation. Additionally, we selected two incubation periods (1 hour and 2 hours) to determine the optimal response timing for plasmid functionality in yeast. Detailed experimental protocols can be found in the Protocol section.

Figure 11. Flow cytometric analysis of EGFP-positive yeast after incubation with H₂O₂. (1 hour: left; 2 hours: right)

We implemented two control groups to normalize the detected EGFP-positive rates. First, we used wild-type INVSc1 yeast (WT) without plasmid transformation to account for autofluorescence from the medium and cells themselves, thereby establishing a baseline for calculating absolute positive rates. Second, we included an untreated control group (0 mM H₂O₂) to determine the basal expression level of the gene, which served as the relative zero point for calculating ROS-induced changes in relative positive rates.

Experimental results demonstrated that after 1-hour and 2-hour H₂O₂ treatments, EGFP expression in pGBKT7-pTRR1-EGFP transformants exhibited a dose-dependent positive correlation with ROS signal intensity (i.e., H₂O₂ concentration) within a specific range. The 2-hour incubation group showed a more consistent correlation. Specifically, the EGFP-positive rate of yeast cells increased with rising H₂O₂ concentration, peaking at 20 mM with values of 69.09% (1h) and 61.9% (2h). These peak values showed significant differences compared to the untreated control, representing increases of 26.92% (1h) and 30.59% (2h), respectively. This confirms that gene expression in pGBKT7-pTRR1-EGFP transformants can be not only significantly induced by ROS stimulation but also upregulated with increasing signal strength, validating its functional capacity as a ROS-responsive genetic element.

TMA Response of pRS416-TAAR5-GPA1-EGFP Transformants

After successfully transforming the pRS416-TAAR5-GPA1-EGFP plasmid into yeast, we analyzed the transformants' response to trimethylamine (TMA) concentration signals by monitoring the expression of the green fluorescent protein (EGFP).

Figure 12. Fluorescence microscopic analysis of pRS416-TAAR5-GPA1-EGFP transformants' response to TMA signal. (Control: untreated)

We initially used fluorescence microscopy to detect EGFP emission, verifying functional expression of the transformed plasmid. Results demonstrated that after treatment with 20 mM TMA, the majority of transformants exhibited green fluorescence, whereas untreated controls showed virtually no fluorescence. This indicates that yeast cells harboring the pRS416-TAAR5-GPA1-EGFP plasmid can detect increased environmental TMA concentrations, subsequently activating the expression of the downstream gene product (EGFP in this case).

To quantitatively assess the responsiveness of the pRS416-TAAR5-GPA1-EGFP plasmid to the Alzheimer's disease biomarker TMA, we applied varying TMA concentrations to simulate signals detectable by the TAAR5 receptor-mediated pathway. We then measured corresponding EGFP fluorescence intensity using flow cytometry. It is important to emphasize that in preliminary experiments, we observed that shorter TMA treatments (within 2 hours) did not significantly affect fluorescence expression. After multiple optimization rounds, we extended the incubation period to 6 hours to ensure complete signal transduction and sufficient downstream gene expression, thereby obtaining more accurate measurements. Detailed experimental protocols can be found in the Protocol section.

Figure 13. Flow cytometric analysis of EGFP-positive yeast after 6-hour incubation with TMA.

Following the same control groups used for pGBKT7-pTRR1-EGFP transformants, we employed wild-type INVSc1 yeast and untreated transformants to normalize the detected EGFP-positive rates. Experimental results demonstrated that after 6 hours of TMA treatment, even 1 mM TMA significantly increased the positive rate of yeasts from 0.09% to 7.33%, representing a 7.24% change relative to the control group. While higher TMA concentrations resulted in additional increases, the overall enhancement was limited - the positive rate increased by only 1.15% across the 1-50 mM TMA concentration range. Nevertheless, these results still confirm that the pRS416-TAAR5-GPA1-EGFP plasmid can respond sensitively to low environmental TMA concentrations (1 mM), and activate downstream TAAR5 receptor expression. This enables rapid response to TMA stimulation during early AD stages (when intestinal TMA concentrations remain low) and initiation of downstream gene expression (butyrate synthesis).

Growth Characteristics of S. boulardii in Simulated Colonic Environment

During the plasmid construction and functional verification phase, we selected the commonly used S. cerevisiae INVSc1 strain as the host to enhance transformation efficiency, shorten experimental cycles, and rapidly verify plasmid expression. However, since INVSc1 cannot survive at 37°C (human body temperature), it is unsuitable for final therapeutic applications. S. boulardii, the only yeast strain approved by the FDA as a probiotic, has been confirmed as a member of the species S. cerevisiae through genotypic and proteomic analyses (Herbrecht & Nivoix, 2005). Therefore, we selected S. boulardii as the final delivery vehicle for our therapeutic objectives. To validate its viability as a live biotherapeutic vehicle, we cultured S. boulardii in media simulating the colonic environment, monitored OD600 to plot growth curves, and performed data fitting using a Logistic model to assess its survival and proliferation capacity under gut-like conditions. Detailed experimental protocols and medium composition can be found in the Protocol section.

Figure 14. Growth curve of S. boulardii under simulated colonic environment.

Results showed that the growth curve of S. boulardii in intestinal simulation medium followed a sigmoidal pattern, with the inflection point (maximum growth rate) occurring at approximately 14.9 hours. The model demonstrated excellent goodness-of-fit (R² = 0.9931), indicating high agreement between the model and experimental data. This model provides crucial temporal reference for subsequent experiments, enabling precise determination of cultivation duration and stimulation timepoints.

Figure 15. Assessment of the growth curve model fit. a. Residuals vs. fitted values plot to check for homoscedasticity; b. Q-Q plot for normality of residuals; c. Residuals vs. time plot to detect autocorrelation; d. Observed data vs. model fit.

To evaluate the model's generalizability and representativeness, we performed comprehensive model diagnostics. Results confirmed that residuals met fundamental linear model assumptions: they showed unbiased distribution, homoscedasticity, normal distribution, and no detectable time-dependency. Furthermore, the fitted curve showed excellent agreement with measured data. In conclusion, this model reliably reflects the growth dynamics of S. boulardii in gut-like environments, enabling dry lab simulations of its intestinal growth and butyrate production.

III. Construction and Functional Verification of the Engineered Yeast Colonization System

Adhesion of Engineered Yeast (Syn) to Mouse Colon Epithelial Cells (MCEC)

To establish an intestinal yeast colonization model, we employed a Saccharomyces cerevisiae strain engineered to express the LcGC adhesion protein (Syn) to simulate the colonization system (Feng et al., 2022). Meanwhile, We used mouse colon epithelial cells (MCEC) to mimic the intestinal surface. Based on the two aforementioned materials, we evaluated the colonization capacity of the engineered yeast in the colon through quantitative analysis using confocal microscopy and flow cytometry. All cell lines used in the experiments were immortalized, and their application complied with relevant ethical guidelines.

We first stained the cells and yeast with two distinct dyes: DAPI for nuclear staining (blue fluorescence) and AF594 conjugated to concanavalin A for yeast staining (red fluorescence). This approach allowed us to characterize the spatial distribution of yeast and cells.

Figure 16. Confocal microscopy analysis of Syn yeast adhesion to MCECs. (From left to right: bright field, DAPI (nuclei, blue), Alexa 594-ConA (yeast, red), and a merged image)

Confocal microscopy after co-incubation revealed that, compared to wild-type (WT) yeast, Syn yeast (red fluorescence) extensively surrounded MCEC cells (blue fluorescence), indicating effective adhesion to the cell surface. These results demonstrate that the engineered yeast Syn exhibits significantly enhanced intestinal epithelial adhesion and colonization capabilities compared to the wild-type strain.

To further quantify the colonization capacity of Syn yeast, we co-cultured yeast with cells at different multiplicities of infection (MOI) and measured the green fluorescence emitted by each cell using flow cytometry. In this experiment, we performed fluorescence measurements at the individual cell and individual yeast levels, thus obtaining the green fluorescence intensity from single cells and single yeast cells. By dividing the former by the latter, we calculated the number of yeast cells adhered to each cell surface. This method enabled us to quantitatively analyze the adhesion ability of Syn yeast to colon epithelial cells.

Figure 17. Flow cytometric quantification of Syn yeast adhesion to MCECs.

Table: Green Fluorescence Intensity at MOI=10

Sample / Replicate	Replicate 1	Replicate 2	Replicate 3
MCEC Cells
Mean Fluorescence Intensity	2428	2014	2156
Syn Yeast
Blank Control	26.7	-	-
Mean Fluorescence Intensity	176 (Invalid)	229	238
Mean Fluorescence Intensity (Background Subtracted)	149.3 (Invalid)	202.3	211.3

The results demonstrated that, at an MOI of 10, the EGFP positive rate of cells colonized by Syn yeast remarkably reached its peak (62.2%) , which was significantly higher than that observed in groups with either lower or higher MOI values (58.9% higher than the MOI=100 group and 41% higher than the MOI=1 group). Regarding the observed decrease in positive rate at higher MOI, we inferred that during the incubation process, intraspecific competition among yeast cells led to nutrient depletion in the culture medium, resulting in partial yeast cell death or detachment from the cell surface, which consequently contributed to the reduced rate detected in the final measurement. Meanwhile, at an MOI of 10, the green fluorescence intensity of individual MCEC cells was measured at approximately 2199, while that of single yeast cells was about 206.8. Based on these measurements, we calculated that under MOI=10 conditions, each cell was colonized by an average of 10.6 Syn yeast cells. These findings indicate that Syn yeast possesses a remarkable capacity to adhere to cell surfaces.

Reference

Details

Feng, L. A., Liang, B., Zeng, X., Shi, C., Yin, H., Feng, Y., Chen, Y., & Yu, Q. (2022). Engineered bacterium-binding protein promotes root recruitment of functional bacteria for enhanced cadmium removal from wastewater by phytoremediation. Water Research, 221, 118746.

Herbrecht, R., & Nivoix, Y. (2005). Saccharomyces cerevisiae fungemia: An adverse effect of Saccharomyces boulardii probiotic administration. Clinical Infectious Diseases, 40(11), 1635–1637.

Wang, X., Duan, C., Li, Y., Lu, H., Guo, K., Ge, X., Chen, T., Shang, Y., Liu, H., & Zhang, D. (2022). Sodium butyrate reduces overnutrition-induced microglial activation and hypothalamic inflammation. International Immunopharmacology, 111, 109083.

Zhang, J., Zhang, Y., Wang, J., Wang, J., & Zhang, J. (2024). Recent advances in Alzheimer’s disease: Mechanisms, clinical trials and new drug development strategies. Signal Transduction and Targeted Therapy, 9(1), 211.