We assembled the selected components (see the main text of the Wiki for details). Referring to the design concept of the "split-TEV" system in the literature, we split the TEV protease into N-terminal and C-terminal fragments, which were then linked to the N-terminus and C-terminus of the LIdR protein (A protein for sensing lactate) respectively, completing the first-generation design of our sensor, designated as LS1.0. However, it failed to achieve the desired performance (Figure 1).

Figure 1. This figure presents the experimental results of the LS1.0 sensor verification using the dual-luciferase reporter gene assay. The ordinate (RLU, relative light unit) reflects luciferase enzyme activity; the results show that even without lactate stimulation (0mM), the TEV enzyme activity of the LS1.0 group is still high, indicating background activity issues in the first-generation sensor.RLU represents the ratio of firefly luciferase fluorescence intensity to Renilla luciferase fluorescence intensity. Each group included 3 biological replicates, and the data are presented as mean ± Standard Deviation (SD). Data were analyzed using one-way analysis of variance (ANOVA). "ns" indicates no significant difference compared with the control group, while "*" indicates a significant difference with p < 0.05 relative to the control group.

After team discussions, we attributed this issue to the excessively short physical distance between the split TEV protease fragments on the LIdR protein, which caused the TEV protease to become active and function even before lactate binding. Therefore, we made an improvement: we split the LIdR protein into two parts (N-terminal and C-terminal), which were then linked to the N-terminal and C-terminal of the TEV protease respectively, resulting in our LS2.0. We conducted experiments with it, but unfortunately, the results were still unsatisfactory (Figure 2).

Figure2. This figure presents the experimental results of the LS2.0 sensor verification using the dual-luciferase reporter gene assay. The ordinate (RLU) represents TEV enzyme activity; similar to LS1.0, the results indicate that the LS2.0 group still exhibits high TEV enzyme activity in the absence of lactate (0mM), meaning the background activity problem was not resolved.RLU represents the ratio of firefly luciferase fluorescence intensity to Renilla luciferase fluorescence intensity. Each group included 3 biological replicates, and the data are presented as mean ± Standard Deviation (SD). Data were analyzed using one-way analysis of variance (ANOVA). "ns" indicates no significant difference compared with the control group.

We observed that the LS2.0 still exhibited high TEV protease activity in the absence of lactate. To address this, we decided to further optimize the sensor based on the aforementioned design. In this iteration, we abandoned the fixed linkage pattern of connecting the N-terminus of LIdR to the N-terminus of TEV and the C-terminus of LIdR to the C-terminus of TEV. Instead, we altered the N/C-terminal connection modes between the two proteins and constructed 8 different combinations (see the main text of the Wiki for details).（Figure 3）

Figure3. This figure provides a visual diagram of the structural composition of each sensor in the LS3.1-LS3.8 series. It clearly marks the connection relationship between Lac-N/Lac-C and TEV-N/TEV-C in each sensor, serving as a design basis for subsequent background screening.

Subsequently, we proceeded to conduct further exploration on these 8 sensor combinations and successfully identified the optimal combination for the lactate sensor.

Based on the summary of failure causes from the two previous screenings, we first conducted a background screening in this iteration (Figure 4). Subsequently, we performed multiple rounds of repeated verification on the three combinations with low background, and finally confirmed that LS3.5 was the sensor with high sensitivity and stability we required (Figure 5).

Figure4. This figure presents the background screening results of the LS3.1-LS3.8 sensors. The ordinate (RLU) reflects the TEV enzyme activity of each sensor without lactate stimulation; the results help screen out sensors with low background activity (e.g., LS3.5, LS3.6, LS3.8).RLU represents the ratio of firefly luciferase fluorescence intensity to Renilla luciferase fluorescence intensity. Each group included 3 biological replicates, and the data are presented as mean ±Standard Deviation (SD).

Figure5. This figure shows the sensitivity and stability verification results of the screened low-background sensors (LS3.5, LS3.6, LS3.8) under different lactate concentrations (0mM, 1mM, 5mM). The ordinate (RLU) indicates TEV enzyme activity; the results confirm that LS3.5 has the highest sensitivity and stability, as it shows a more obvious response to lactate concentration changes while maintaining low background.RLU represents the ratio of firefly luciferase fluorescence intensity to Renilla luciferase fluorescence intensity. Each group included 3 biological replicates, and the data are presented as mean ± Standard Deviation (SD). Data were analyzed using one-way analysis of variance (ANOVA). "ns" indicates no significant difference compared with the control group, while "*" indicates a significant difference with p < 0.05 relative to the control group.

We selected lactate oxidase (LOx) as our tool enzyme. However, given that LOx is a cytoplasmic enzyme, we modified it to enable secretion outside the cell. Specifically, we added a signal sequence of rat follicle-stimulating hormone beta subunit (rat FSHB) to the N-terminus of the LOx sequence (Figure 6) to facilitate the secretion of the LOx enzyme out of the cell.

Figure6. This figure shows the structural design of the engineered secreted lactate oxidase (sLOx). It illustrates that a signal sequence derived from the rat follicle-stimulating hormone β-subunit (rat FSHB) was fused to the N-terminus of the lactate oxidase (LOx) sequence, and a FLAG tag was also included in the structure (with the total length marked as approximately 1250 base pairs). The rat FSHB signal sequence is responsible for directing the newly synthesized sLOx protein into the cell secretory pathway, ensuring that sLOx is secreted extracellularly to exert its lactate-degrading function, rather than being retained in the cytoplasm.

This is the secreted lactate oxidase (sLOx) we designed.

We first conducted a secretion assay for the designed sLOx enzyme, and ultimately confirmed via Western blot that it could be secreted into the Culture Medium (Figure 7).

Figure7. This figure displays the results of the Western blot assay used to verify sLOx protein secretion. The blot includes two sample types: culture medium (extracellular fluid) and cell lysate (intracellular fluid), with both NC (negative control) and LS3.5 (sensor-sLOx system) groups tested. The presence of the sLOx protein band in the culture medium of the LS3.5 group (while no obvious band is detected in the NC group's culture medium) confirms that the engineered sLOx protein was successfully secreted from 293T cells into the extracellular space.

Subsequently, we conducted experiments on the extracellular lactate degradation by sLOx protein. We combined the optimized lactate sensor (screened out in previous steps) with the designed sLOx, which constitutes our LS3.5 lactate-responsive system. The lactate-degrading function of this system was verified by detecting the lactate content in the Culture Medium.(Figure 8)

Figure8. This figure shows the results of the extracellular lactate degradation assay. The ordinate represents the relative concentration of extracellular lactate, and the abscissa shows the NC group and the LS3.5 group (sensor-sLOx system group). The results clearly indicate that compared with the NC group, the extracellular lactate concentration in the LS3.5 group is significantly lower, directly demonstrating that the secreted sLOx protein has a stable and efficient lactate-degrading function, which can reduce the local extracellular lactate concentration as expected.