Design
Based on the application of ferroptosis in the killing module, we were naturally curious if the engineered bacteria could be converted to magnetotactic bacteria that can be guided by an external magnetic field. After reviewing the literature, we found that a research team has successfully linked magnetic Fe3O4 nanoparticles (MNP) covalently to E. coli, acting as a catalyst for a Fenton-like reaction1. Taking inspiration from this study, we designed an amino-modified MNP that can link to N-acetyl muramic acid in the bacteria’s wall for the purpose of anchoring the MNP to the engineered attenuated Salmonella (Figure 1). By altering the strength of the magnetic field in vitro, we aimed to enable the engineered bacteria to specifically target tumor tissue.
Figure 1. Schematic representation of conjugated Fe3O4 nanoparticles on the surface of engineered bacteria.
Build
We treated the bacteria solution with magnetic Fe3O4 nanospheres with a diameter of 100 ~ 200 nm for 30 minutes on an ultrasonic vortex apparatus, and then shaked 3 h at 37℃, 220 rpm. After 3 h stir, the modified bacteria were separated by differential centrifugation (3000 rpm, 5 min) and washed for three times.Test
We placed the magnet on the side of the engineered bacteria coupled with Fe3O4 nanoparticles, and observed that the engineered bacteria moved to the side with the magnet (Figure 2). In other words, engineered bacteria can directly move under an external magnetic field.
Figure 2. Engineered bacteria coupled to Fe3O4 nanoparticles were attracted by magnets. a. Distribution of bacteria when the magnet was first placed. b,c. Bacterial distribution after 24 h of magnet placement. Red frames outlined in the figure are engineered bacteria attracted by magnets.
Learn
Although we successfully conjugated Fe3O4 nanoparticles on the surface of engineered bacteria, we realized that magnetic targeting lacks specificity and has an imprecise targeting range. Furthermore, in deep tumor tissue, the effectiveness of engineered bacteria targeting may be compromised. Additionally, through a literature review, we discovered that attaching Fe3O4 nanoparticles to the surface of engineered bacteria could potentially impact their invasiveness into cells 2.Design
When our team communicated with the SCAU-China, they suggested us take advantage of the antigen-antibody to improve specificity and targeting ability. In order to achieve more precise targeting in vivo and ensure the smooth entry of engineered bacteria into tumor cells, we changed our thought and focused on the antigen-antibody conjugation, a more precise targeting method. We reviewed the literature, found the single chain fragment variable (scFv) of carcinoembryonic antigen (CEA). Inspired by the design of the 2018 HZAU-China team and the literature, we decided to conjugate scFv to the lipoprotein (Lpp) and the outer membrane protein (OmpA) of E. coli3,4 (Figure 3). So that the scFv can be fixed on the surface of VNP20009 and target CEA on the surface of tumor cells.
Figure 3. Schematic representation of the engineered bacteria expressing CEA single chain fragment variable on the outer membrane.
Build
To ensure that scFv could be successfully anchored to the cell surface by the Lpp-OmpA expression system, we conducted molecular modeling of the scFv used in this project. And we employed molecular docking and corresponding calculations to predict the structure of the antibody and its affinity for CEACAM5 (Figure 4-5).
Figure 4. Prediction of the transmembrane regions of Lpp-OmpA-scFv.
Notes: The angle of inclination is defined as the angle between the first principal axis of the protein and the normal of the membrane. The angle of rotation is the angle between the second principal axis of the atomic set and the normal of the plane defined by the first principal axis and the normal of the membrane.
Figure 5. The docking of antibody molecules with antigen molecules.
Notes:(a) Results of molecular docking. Lpp-OmpA-scFv is shown in green and CEACAM5 is shown in orange. (b) Schematic representation of the molecular docking surface. Among them, amino acids that can form hydrogen bonds are shown.
Test
Western blot was performed to verify whether the scFv could be expressed in the engineered bacteria. Our data suggested that the scFv successfully expressed in VNP20009 (Figure 6).
Figure 6. WB analysis of the expression of specific single chain antibody fragments (scFv).
Notes: Lpp-OmpA-scFv-GFP molecular weight is about 66 kDa and GAPDH is a reference protein in cells with a molecular weight of 36 kDa.
Figure 7. Diagrammatic sketch of engineered bacteria with weak green fluorescence.
Learn
By reviewing the literature, we found that the VH and VL of scFv are connected by a linker, and the transient dissociation of VH and VL leads to the instability of scFv and the downstream GFP protein5. This implies that we need to seek more methods to make green fluorescence stably expressed.Design
In order to ensure that the engineered bacteria could express stable green fluorescence, we selected VNP20009 competent cells with stable GFP conversion as plasmids receptors.Build
We prepared VNP20009 competent cells stably transfected with GFP and electroporated the constructed plasmids into it.Test
We infected NUGC-3 and BGC-823 with the constructed engineered bacteria and negative control simultaneously. For NUGC-3 with high-CEA-expression, the infection efficiency of engineered bacteria (green fluorescence) was higher than that of negative control (red fluorescence), but there was no significant difference in the infection efficiency of BGC-823 which was with low-CEA-expression. (Figure 8).
Figure 8. Fluorescent microscopy results of BGC-823 and NUGC-3 cells co-infected by engineered bacteria, and VNP20009 for 90 minutes. Both the engineered bacteria and negative control infections were at the MOI of 1:50. Red frames outlined in the figure were engineered bacteria that infected cells.
Learn
To determine whether scFv really worked, we compared the efficiency of different bacteria in infecting NUGC-3 and BGC-823. However, we could not exclude the possibility that there was an interaction between the engineered bacteria and the negative control that affected their infection efficiency, which meant that further improvement of our experiments was necessary.In addition, NUGC-3 cells were in a poor state and died in large numbers after a short time of infection, which resulted in a majority of bacteria being washed off with PBS along with the dead cells during the infection process.
Design
In this attempt, we selected the human colorectal cancer cell line LS174T, which also has a high expression of CEA, as the experimental group.Build
In this attempt, we added control experiments in which the engineered bacteria and the negative control infected the cells separately.Test
By analyzing the results of bacterial infection of LS174T cells and BGC-823 cells, we got the same conclusions as above. Therefore, the scFv promoted the infection of bacteria no matter when the engineered bacteria and the negative control infected the cells separately or simultaneously (Figure 9-11).
Figure 9. Microphotographs of LS174T cells with scFv delivery by bacterial infection for 2 hours. Both the engineered bacteria and negative control infections were at the MOI of 1:50.
Figure 10. Microphotographs of BGC-823 cells with scFv delivery by bacterial infection for 2 hours. Both the engineered bacteria and negative control infections were at the MOI of 1:50.
Figure 11. Microphotographs of BGC-823 and LS174T cells co-infected by engineered bacteria and VNP20009 for 2 hours. Both the engineered bacteria and negative control infections were at the MOI of 1:50.
Learn
The experimental results demonstrated that the scFv could be expressed on the surface of VNP20009 and effectively enhanced the targeting ability of engineered bacteria towards tumor cells. However, since the bacterial infection of cells occurs within a confined space, it becomes imperative to further enhance the targeting efficacy of engineered bacteria to meet the need of efficient tumor cell recognition in a dynamic bloodstream environment.Design
To enable the engineered bacteria to have stronger tumor cell targeting ability, we tried to find out the amino acid sites that are involved in antigen binding through modeling. Based on the results of modeling we hoped to improve the affinity of the scFv by mutating the simulated sites.Build
We performed interaction-based virtual amino acid mutagenesis of antibody-antigen complexes by Discovery Studio software. After the structure file from the molecular docking was uploaded, the complex protein was subjected to a CHARMm force field, and the amino acid residues within 3 Å range around the ligand were selected. By mutating all of the above amino acids to ALA and ranking the mutational energy results of the individual mutations, we finally obtained five amino acids that are critical for antibody-antigen interaction (LEU196, ARG177, ILE193, TYR190, ASN194) (Figure 12).
Figure 12. Schematic representation of amino acids critical for antibody affinity.
Notes: The molecule labeled in yellow represents anti-CEA scFv; The molecule labeled in blue represents CEACAM5. Amino acids in the antibody that resulted in decreased antibody affinity after mutation to ALA are highlighted in red, whereas amino acids with increased antibody affinity after mutation are highlighted in magenta.
Test
In order to gain more detailed information on how mutations enhance antibody affinity, we performed saturation mutagenesis on the aforementioned five amino acids. From our findings, we identified ASN195 as a crucial site for improving the affinity of anti-CEA scFv. Substituting ASN195 with TRP, GLN, PHE, GLU or HIS may significantly enhance the affinity of anti-CEA scFv.We also calculated mutation energy to evaluate the effect of mutations on protein stability (Figure 13). By calculation, we found that, except for ASN195 mutated to GLU and GLN, several other mutations with increased affinity did not affect antibody stability.
Figure 13. Mutation energy (Binding) analysis in single mutation of key amino acids in scFv.
Learn
Based on the above modeling, we can perform site-directed mutagenesis of the scFv and test the affinity of the modified antigen-antibody complex.2 Suh, S. et al. Nanoscale Bacteria-Enabled Autonomous Drug Delivery System (NanoBEADS) Enhances Intratumoral Transport of Nanomedicine.Adv Sci (Weinh) 6, 1801309 (2019). https://doi.org:10.1002/advs.201801309
3 Francisco, J. A., Stathopoulos, C., Warren, R. A., Kilburn, D. G. & Georgiou, G. Specific adhesion and hydrolysis of cellulose by intact Escherichia coli expressing surface anchored cellulase or cellulose binding domains. Biotechnology (N Y) 11, 491-495 (1993). https://doi.org:10.1038/nbt0493-491
4 Bereta, M. et al. Improving tumor targeting and therapeutic potential of Salmonella VNP20009 by displaying cell surface CEA-specific antibodies. Vaccine 25, 4183-4192 (2007). https://doi.org:10.1016/j.vaccine.2007.03.008
5 Wang, J. L. et al. Construction and activity analysis of a recombinant immunotoxin composed of PE38 and a disulfide stable single-chain antibody. Xi Bao Yu Fen Zi Mian Yi Xue Za Zhi 22, 74-77 (2006).
Design
Through reviewing articles, we noticed that the glucose oxidase (GOx) from Aspergillus niger has the ability to efficiently catalyze O2 to generate H2O21. H2O2 can be catalyzed by Fe2+ to produce radical HO• through Fenton reaction. It should be noted that HO•, also known as the hydroxyl radical, is categorized as a reactive oxygen species (ROS). This particular ROS has the capability to intensify the lipid peroxidation process of cell membranes. Ultimately, such intensification is responsible for inducing ferroptosis in tumor cells. Engineered plasmids or proteins carried by recombinant S. typhimurium can be delivered into tumor cells. The unique characteristics of S. typhimurium enable targeted protein to be delivered into tumor cells. We designed to engineer VNP20009 to constitutively express glucose oxidase and deliver the protein to the cytoplasm after invasion of tumor cells. In the catalysis of GOx, glucose will be turned into gluconic acid, which can also reduce the energy source of tumor cells, further strengthening the ferroptosis in tumor cells.Build
We have constructed a prokaryotic expression plasmid (pFPV25.1-GOx) that can be expressed in S. typhimurium (Figure 1-2). We intend to transform the plasmid expressing GOx into VNP20009.
Figure 1. The expression of GOx
Figure 2. The map of pFPV25.1-GOx
Learn
S. typhimurium VNP20009 has the ability to survive within host cells. However, due to its encapsulation in intracellular vesicles, it exhibits limited or minimal impact on the host cell cytoplasm. By reviewing the literature, we found that the efficiency of protein delivery after entry into cells by S. typhimurium was not appreciable. We need to find a new mode of delivering functional proteins.Design
Like many bacterial pathogens of animals, S. typhimurium have evolved a specialized protein-secretion system to deliver bacterial proteins into host cells. S. typhimurium has a special secretion system: type III secretion system (T3SS), which is closely related to bacterial colonization, infection and survival in host cells2. We designed to use the secretion and translocation signals of type III secreted protein Salmonella outer protein E (SopE) to fuse with GOx. So, engineered VNP20009 can express GOx and inject GOx into the tumor cells.Build
We integrated the translocation signal of Salmonella exoprotein E (SopE), a type III secretory protein, atthe N-terminus of Gox (Figure 3-4). The gene fragment was synthesized by a commissioned company. Subsequently, we assembled the fragments into the expression plasmid (pFPV25.1) using gibson assembly. We then transferred the ligation products into DH5α competent cells and extracted plasmids for identification. The successfully constructed plasmids (pFPV25.1-SopE-FLAG-GOx) were transferred into VNP20009 competent cells. As a result, the engineered VNP20009 (VNP-SopE-GOx) is supposed to utilize T3SS to facilitate the expression of GOx and deliver it into tumor cells upon contact.
Figure 3. The expression of SopE-FLAG-Gox
Figure 4. The map of pFPV25.1-SopE-FLAG-GOx
Test
Recombinant S. typhimurium (VNP-SopE-GOx) was grown at appropriate
conditions. The proteins from the culture supernatant and bacterial lysates were prepared for an
immunoblotting assay. The results of Western blot showed that VNP20009 electroporated with plasmid
pFPV25.1-SopE-FLAG-GOx expressed SopE-FLAG-GOx protein successfully (Figure 5).
Figure 5. Western blot analysis shows the expression of gene of interest (SopE-FLAG-GOx) in VNP-SopE-GOx. The theoretical molecular weight of SopE-FLAG-GOx is about 76 kDa, and the theoretical molecular weight of GAPDH is about 36 kDa.
Salmonella VNP-SopE-GOx was cultured at 37℃. Engineered bacterial suspensions were prepared using DMEM medium at a multiplicity of infection (MOI) of 150 bacterial cells per mammalian cell. Additionally, a control group was established using the unmodified VNP20009 strain. Following infection, cells were harvested and prepared for SDS-PAGE and Western blot experiments.
However, the results of the Western blot showed no any specific band. The immunoblotting results failed to meet our expectation.
Learn
At the first time, the temperature to culture the VNP-SopE-GOx is 37℃. From the western blot result of VNP-SopE-GOx infecting BGC-823, we did not get any specific band. After consulting Dr. Yihao Li, the postdoctoral of center of biology science and technology, Beijing Normal University at Zhuhai, we found that at 37℃, inclusion bodies will form in VNP20009. Inclusion bodies are high-density (1.3 mg/ml) insoluble protein particles encapsulated by membranes when foreign genes are expressed in prokaryotic cells. According to the advice from Dr. Li, we adjusted the bacteria culture temperature to 30℃.Test
Salmonella VNP-SopE-GOx was cultured at 30℃. Engineered bacterial suspensions were prepared using DMEM medium at a multiplicity of infection (MOI) of 150 bacterial cells per mammalian cell. Additionally, a control group was established using the unmodified VNP20009 strain. Following infection, cells were harvested and prepared for SDS-PAGE and Western blot experiments. The results showed that the recombinant type III secretion system could effectively present SopE-FLAG-GOx fusion protein to BGC-823 cells, and that SopE-FLAG-GOx fusion protein could be processed by BGC823 cells, and its secretory related protein SopE could be further excised3 (Figure 6).
Figure 6. Western blot analysis shows the injection of FLAG-GOx into BGC-823 cells by VNP-SopE-GOx via T3SS. The theoretical molecular weight of FLAG-GOx is about 64 kDa, and the theoretical molecular weight of GAPDH is about 36 kDa.
After infection, we used the hydrogen peroxide detection kit (Beyotime) to detect the H2O2 level of the cell culture supernatant. The results demonstrated that the H2O2 level in the culture supernatant of tumor cells infected with VNP-SopE-GOx is significantly higher compared to the control group and tumor cells infected with VNP20009 (Figure 7).
Figure 7. H2O2 levels in the cell culture supernatant
We also used the CCK-8 cell counting kit (Vazyme) to detect the viability of the cells after infection. The results of the experiment indicated that tumor cells infected with VNP-SopE-GOx exhibited lower viability compared to non-infected tumor cells and those infected with VNP20009. Additionally, the cell viability was significantly increased when tumor cells are treated with Fer-1 and infected with VNP-SopE-Gox. There was no significant difference in cell viability between tumor cells infected with VNP20009 without Fer-1 treatment and those treated with Fer-1 and infected with VNP20009. Therefore, this study demonstrated that VNP-SopE-GOx infection can effectively induce ferroptosis in tumor cells (Figure 8).
Figure 8. Cell viability of cancer cells treated with bacterial infection and co-treated with bacterial infection and Fer-1.
Learn
H2O2 is one of the substrates of aquaporins (AQPs)4. So, the concentration of H2O2 in the culture medium can reflect the level of H2O2 in the tumor cells. Fer-1 is a selective inhibitor of ferroptosis. The conclusion we learned was that GOx can catalyze glucose to generate hydrogen peroxide, which can further promote ferroptosis in tumor cells. To further verify the GOx can induce lipid peroxidation of the tumor cell membranes, we intended to detect the lipid peroxidation level of the tumor cells infected by the VNP-SopE-GOx.To enhance the safety of our project, we need to select the condition that specifically exists in the tumor tissue to activate the expression of GOx. Tumor microenvironment (TME) often exhibits overall oxygen deficiency. After communicating with USTC and SCAU-China, we decided to use the specific anaerobic promoter pnirB. The pnirB is induced in anaerobic conditions and is up regulated in response to nitrite and nitrate. Based on the aforementioned considerations, we have selected the pnirB to initiate the expression of GOx.
We intended to use the anaerobic promoter pnirB to replace the constitutive promoter prpsM. Therefore, the engineered VNP20009 can drive the expression of GOx in anaerobic conditions.
Figure 9. The anaerobic expression of Gox
Figure 10. The map of pFPV25.1-pnirB::SopE-FLAG-GOx
2 Galán, J. E. & Collmer, A. Type III Secretion Machines: Bacterial Devices for Protein Delivery into Host Cells. Science 284, 1322-1328 (1999). https://doi.org:10.1126/science.284.5418.1322
3 Kubori, T. & Galán, J. E. Temporal regulation of salmonella virulence effector function by proteasome-dependent protein degradation. Cell 115, 333-342 (2003). https://doi.org:10.1016/s0092-8674(03)00849-3
4 Zhang, M. et al. Aquaporin OsPIP2;2 links the H2O2 signal and a membrane-anchored transcription factor to promote plant defense. Plant Physiol 188, 2325-2341 (2022). https://doi.org:10.1093/plphys/kiab604
Design
Facultative anaerobic strains of S. typhimurium can preferentially accumulate in tumor tissues and produce a potential anti-tumor effect. The progress of RNAi technology for cancer therapy faces the challenge of how to efficiently deliver RNAi to the cytoplasm of cancer cells. Therefore, we engineered bacteria to deliver shRNA expressing plasmids into the cytoplasm of tumor cells in order to improve treatment efficacy and safety. The cystine transporter solute carrier family 7 member 11 (SLC7A11) is highly expressed in many tumor cells to protect cancer cells from oxidative stress and ferroptosis1. To inhibit the expression of SLC7A11, we engineered S. typhimurium to deliver RNAi.Build
We commissioned companies to synthesize short hairpin RNAs (shRNA) sequences targeting SLC7A11 and ligated into the eukaryotic expression plasmids (pSilencer 2.1-U6-Neo). The plasmids were transformed into attenuated S. typhimurium.
Figure 1. Construction of pSilencer 2.1-U6-shRNA-SLC7A11-Neo
Test
We co-cultured the engineered bacteria with mammalian cells for 2 hours, and then performed qRT-PCR experiments to analyze the levels of SLC7A11 mRNA. However, no significant silencing of target genes was observed.Learn
The S. typhimurium VNP20009 can survive and proliferate within host cells, but as it is encapsulated by intracellular vesicles, it is unable or releases only minimal perturbations into the cytoplasm of host cells. Therefore, VNP20009 is not suitable as a vector for delivering shRNA expressing plasmids to silence genes in mammalian cells2. We need to further engineer Salmonella to improve efficiency.Design
In addition to utilizing bacteria as the shRNA-expressing plasmid carrier, a trans-kingdom RNAi (tkRNAi) technology was developed based on E. coli to deliver RNAi for gene silencing in mammalian cells. This method involves producing high amounts of shRNA and releasing them into the cytoplasm of host cells, both in vitro and in vivo3. We referred to the tkRNAi system in E. coli and aimed to apply it to Salmonella. Plasmids were constructed to retain the HlyA gene encoding listeriolysin O, which we believed would facilitate product entry into the cytoplasm. Additionly, the invasin gene Inv was deleted from the E. coli tkRNAi system, as it had no impact on our system due to our inherently invasive chassis. We chose T7 promoter and enhancer which were commonly used in engineered bacteria that mediate trans-kingdom RNAi4.Build
We constructed plasmids named pSilence-SLC7A11, which included T7 promoter-driven expression cassette and anaerobically induced the expression of HlyA. Oligonucleotides encoding shRNA targeting SLC7A11 were inserted into the plasmid. Plasmids were transformed into VNP20009 competent cells.
Figure 2. Construction of pSilence-shRNA-SLC7A11
Test: After 2 hours of co-culturing the engineered bacteria with mammalian cells, we conducted qRT-PCR experiments to confirm the changes in mRNA levels. However, there was no obvious target gene silencing.
Learn
After analysis, we found that the T7 promoter could not function directly in S. typhimurium, which is due to the inability of S. typhimurium VNP20009 to express T7 RNA polymerase.Design
T7 RNA polymerase is a highly specific RNA polymerase that recognizes the T7 promoter sequence5. It can specifically recognize the T7 promoter sequence and activate the transcription of T7 promoter. Consequently, we incorporated the T7 RNA polymerase expression cassette into the plasmid, enabling the proper functioning of the T7 promoter. In order to visually verify the feasibility of this system, we also designed shRNA sequences targeting GFP. By observing the intensity of fluorescence, we could determine whether the target gene was silenced.Build
The T7 RNA polymerase expression cassette was integrated into our plasmid by homologous recombination. By virtue of Gibson Assembly, the shRNA sequence targeting GFP was inserted into the plasmid. The plasmids were successfully constructed as confirmed by sequencing. The plasmids were transformed into VNP20009 competent cells.
Figure 3. Design of the plasmid
Test
We first used engineered bacteria targeting GFP to infect human gastric adenocarcinoma cells BGC-823 that transiently transfected with GFP eukaryotic vector. After the engineered bacteria and BGC-823-GFP cells were co-cultured for two hours in serum-free medium, the cells were washed and a fresh complete medium was used. After 24 hours we observed it under a fluorescence microscope.
Figure 4. Results of shGFP delivery by bacterial infection 24 hours later. Both the engineered bacteria and VNP20009 infections were at the MOI of 500. Control was BGC-823 cells that not transfected with the GFP plasmid.
Learn
The above results proved the feasibility of this system. The engineered bacteria successfully silenced GFP by cross-kingdom RNAi delivery of shRNA. We then conducted a follow-up experiment using the system to silence SLC7A11 and quantitatively analyze its effects.Design
As the engineered bacteria targeting GFP successfully silenced the gene and affected its function, we decided to silence SLC7A11 by this system. The shRNA sequence in the plasmid was replaced from targeting GFP to targeting SLC7A11. We still chose BGC-823 cells and infected them with bacteria delivered shRNA targeting SLC7A11.Build
By virtue of enzyme-cut and link up, the shRNA sequence targeting SLC7A11 was inserted into the plasmid. The plasmids were successfully constructed as confirmed by sequencing. Then they were transformed into VNP20009 competent cells.
Figure 5. Construction of pSilence-SLC7A11
Test
After being exposed to bacteria, cells were washed and replenished with fresh complete medium containing antibiotics. In order to assess the expression level of the SLC7A11 gene, we collected the samples and conducted quantitative polymerase chain reaction (qPCR). The qPCR results indicated a decrease in the expression of SLC7A11 in BGC-823 cells post-infection with genetically modified bacteria. To further confirm the impact on cellular function after down-regulating SLC7A11, we analyzed the glutathione (GSH) content. The results demonstrated a decrease in GSH content in cells where SLC7A11 was down-regulated, suggesting that the down-regulation of SLC7A11 affected the metabolic pathway of the cells. The assessment of cell viability using CCK-8 assay revealed a significant decrease in tumor cell viability when SLC7A11 expression was down-regulated. These findings collectively indicate that our therapy has effectively produced positive outcomes in BGC-823 cells.
Figure 6. qPCR results after bacterial infection and siRNA-SLC7A11 transfection. a. The SLC7A11 expression level of BGC-823 after siRNA transfection, control was BGC-823 without any treatment, and siRNA-control was siRNA transfected with other genes. Results were significant between siRNA-SLC7A11 and control groups (*** p<0.001). b. The SLC7A11 expression level of BGC-823 after engineered bacterial infection with different MOI. VNP20009 without functional plasmids at an MOI of 1:5000 was used as control group. The results of VNP-shSLC7A11 (1:3000), VNP-shSLC7A11 (1:5000) and VNP-control groups have significant differences (* p<0.05, ** p<0.01)
Figure 7. Results of the GSH detection. a. Standard curve of the GSH content varying with the absorbance value. b. The intracellular GSH content in BGC-823 cells after VNP20009 and engineered bacteria infection. Results between VNP-shSLC7A11 (1:500) and VNP20009 (1:500) infection have significant differences (**p< 0.01).
Figure 8. Results of cell viability test after infection. The siRNA-control is the siRNA transfected with the other genes. The results between siRNA-SLC7A11 and siRNA-control groups have significant differences; and results between engineered bacteria with different MOI and VNP20009 (1:3000) infection have significant differences (****p < 0.0001).
Learn
We achieved specific gene silencing of SLC7A11 by shRNA-expressing Salmonella VNP20009-mediated RNAi. The inhibition of SLC7A11 interfered with lipid homeostasis of tumor cells and significantly inhibited cell growth activity. Next, we hope to further confirm the function of the SLC7A11 to ensure the applicability of our therapy. We learned from Professor Xiaotian Zhang that different SLC7A11 expression level dictates differential responses to oxidative stress in cancer cells, which could be complicated. For certain tumor cell lines with high expression of SLC7A11 and sensitivity to disulfidptosis 6, inhibiting the expression of SLC7A11 does not align with the purpose of our project. Therefore, we will explore other target genes for Xc- system/GSH/GPX4 inhibition in the future.2 Cabezudo, I., Lobertti, C. A., Véscovi, E. G. & Furlan, R. L. E. Effect-Directed Synthesis of PhoP/PhoQ Inhibitors to Regulate Salmonella Virulence. J Agric Food Chem 70, 6755-6763 (2022). https://doi.org:10.1021/acs.jafc.2c01087
3 Xiang, S., Fruehauf, J. & Li, C. J. Short hairpin RNA-expressing bacteria elicit RNA interference in mammals. Nat Biotechnol 24, 697-702 (2006). https://doi.org:10.1038/nbt1211
4 Guo, H. et al. Targeting tumor gene by shRNA-expressing Salmonella-mediated RNAi. Gene Ther 18, 95-105 (2011). https://doi.org:10.1038/gt.2010.112
5 Borkotoky, S. & Murali, A. The highly efficient T7 RNA polymerase: A wonder macromolecule in biological realm. Int J Biol Macromol 118, 49-56 (2018). https://doi.org:10.1016/j.ijbiomac.2018.05.198
6 Liu, X. et al. Actin cytoskeleton vulnerability to disulfide stress mediates disulfidptosis. Nat Cell Biol 25, 404-414 (2023). https://doi.org:10.1038/s41556-023-01091-2
Design
Initially, we planned to utilize the conventional toxin-antitoxin system to sustain two operative plasmids in our project. This involves incorporating the toxin gene into plasmid pFenton, while introducing the antitoxin gene into plasmid pSilence. Consequently, if pSilence is no longer present, the expression of toxin in plasmid pFenton will result in the eradication of the modified bacteria. Nevertheless, if plasmid pFenton is lost, the TA system will become nonfunctional. As a result, we devised the first generation of "logical suicide circuit". Our intention was to engineer specific components of the circuit within each of the two plasmids, so that in the event of one plasmid being lost, the modified strain would initiate suicide.Build
The "logical suicide circuit" was introduced in two plasmids (Figure 1). The corresponding gene fragments were amplified by PCR and introduced into the plasmid by homologous recombination to construct plasmids that can be used to verify the logic suicide circuit. Plasmid pSilence constitutively expressed LacI and toxin Hok. The expression of tetR was regulated by the repressible promoter pLacO, which is repressed by the LacI protein. Similarly, the expression of Sok was regulated by the repressible promoter pLtetO, which is repressed by the tetR protein. If pSilence is lost, LacI protein levels will gradually degrade, resulting in cessation of its production. Subsequently, pLacO will promote the expression of tetR, leading to repression of Sok transcription. As Sok degrades, Hok mRNA will translate into the toxin protein and eliminate the engineered bacteria. If pFenton is lost, the Hok-Sok system will induce bacterial death. (see in design).
Figure 1. Schematic representation of the first generation of "logical suicide circuit"
Test
The "logical suicide circuit" was introduced to experts and other teams for their advice and gain a lot from them, especially in the communication in 10th Conference of China iGEMer Community (CCiC) . It was suggested that we should pay more attention to the induction of the suicide of engineered bacteria after the end of the mission. Otherwise, the engineered bacteria might harm the normal cells.Learn
We realized that it was necessary for artificially manipulating the initiation of engineered bacterial suicide to strengthen the safety of our project. Especially in the later stage of treatment, if the volume of the tumor tissue decreases and the tumor regress, the engineered bacteria are likely to invade the normal tissue cells. Then it came to the second generation of "logical suicide circuit".Design
The second generation of the "logical suicide circuit" was improved based on the initial generation (refer to Figure 2). In order to enable the engineered bacteria to be induced after treatment with doxycycline (Dox), we altered the arrangement of the LacI and tetR expression cassettes in the circuit. Doxycycline is a tetR acycline analogue that can bind to tetR, consequently preventing tetR from suppressing the pLtetO promoter. As a result, the pLtetO can express the LacI protein normally, which then inhibits the transcription of the antitoxin Sok, ultimately leading to the death of the engineered bacteria caused by the toxin Hok.
Figure 2. Schematic representation of the second generation of "logical suicide circuit"
Build
To validate the second generation of "logical suicide circuit", pET-T7::Hok-Sok-LacI and pGEN-prpsm::tetR plasmids were re-constructed by restriction enzyme ligation and homologous recombination to validate the circuit.
Figure 3. Construction of plasmid pET-T7::Hok-Sok-LacI and pGEN-prpsM::tetR to validate the third generation of "logical suicide circuit"
Test
Firstly, the system was modeled to predict its functionality. The modeling results predicted that the expression of toxin Hok would surpass that of antitoxin Sok, which is logically impossible (Figure 4). Additionally, the pET-T7::Hok-Sok-LacI and pGEN-prpsM::tetR were introduced into BL21 (DE3) competent cells simultaneously, and they were plated on LB-agar plates containing ampicillin, kanamycin, and IPTG. The presence of IPTG enables the binding with LacI, thus preventing LacI from repressing the expression of Sok and ensuring the survival of the engineered bacteria. However, it was observed that the engineered bacteria did not proliferate on the LB-agar plates supplemented with IPTG (Figure 5).
Figure 4. Schematic representation of the concentration of each substance over time. The black curve is the concentration of Hok, which is significantly higher than Sok
Figure 5. The plasmid pET-T7::Hok-Sok-LacI and pGEN-prpsM::tetR were transformed into BL21 (DE3) competent cells, which was plated on the LB-agar plates with ampicillin, kanamycin and IPTG. No colonies grow up after more than 36 hours culture at 37℃
Learn
The model was highly significant in predicting the performance of our system. We conducted model validation and experiments, which revealed that the strength of toxin protein expression driven by the T7 promoter was excessively high compared to the expression of antitoxin. Consequently, we determined the need to employ a weaker promoter to drive toxin protein expression.Design
Therefore, it comes to the third generation of "logical suicide circuit". The T7 promoter that drives toxin protein expression becomes a weaker trp promoter than the T7 promoter.
Figure 6. Schematic representation of the third generation of "logical suicide circuit"
Build
The plasmid pET-trp::Hok-Sok-LacI was constructed with pET-T7::Hok-Sok-LacI. We removed T7 promoter from pET-T7::Hok-Sok-LacI by reverse PCR. The trp promoter gene fragments were amplified by PCR and introduced into the plasmid by homologous recombination. However, later we found that BL21 (DE3) competent cells will endogenous express LacI protein. As a result, we failed to validate "logical suicide circuit" in BL21 (DE3).
Figure 7. Construction of plasmid pET-trp::Hok-Sok-LacI to validate the third generation of "logical suicide circuit"
Test
The model initially tested the third generation of the "logical suicide circuit" and predicted that it would function as intended (see in model). However, it is a pity that we don’t have enough time to validate complete "logical suicide circuit" in our chassis strain. We only tested the functionality of pLacO and pLtetO, and evaluated pET-Hok-Sok-lacI separately in BL21 (DE3). The results confirmed the functionality of pLacO and pLtetO as expected, as well as the successful performance of the Hok-Sok system (see in results).Learn
By studying the background of the E. coli BL21 (DE3) strain, we found that the strain carried the DE3λ prophage on the chromosome, which in turn carried the lacI and sam7I genes and the T7 RNA polymerase gene controlled by the lacUV5 promoter. Therefore, E. coli BL21 (DE3) will endogenous express LacI protein, we need to validate complete "logical suicide circuit" in a new bacteria strain.We designed the "logical suicide circuit" according to our needs and built step by step. It was tested with different and effective methods. During these processes, we learned from problems that we met and then improve the system. We finally designed the "logical suicide circuit" that meets the requirements through “DBTL”.
