Design

Purpose

Due to the inherent characteristics of polypeptides, maintaining a high concentration near target cells and enabling their entry into cells to exert their effects becomes a challenge in treatment and experimentation. The Type VI secretion system (T6SS) of Gram-negative bacteria can deliver effector proteins to the bacterial surrounding environment or into cells. Therefore, we hope to explore the capability of T6SS to act as an exogenous polypeptide delivery system, thus developing new drug delivery methods.

Deletion of Toxic Genes

In this experiment, we use Vibrio cholerae and Pseudomonas aeruginosa as chassis cells. As bacteria of the BSL-2 group, how to minimize their toxicity to ensure the safety of experimenters is our primary concern. The toxicity of these two strains mainly comes from their metabolic products (toxins), Type III secreted proteins, and Type VI secreted proteins. Fortunately, our PI has researched the Type VI secretion system, so we already have bacteria with reduced toxicity, deleting T6SS effectors. Based on this, the toxicity of Vibrio cholerae has been greatly reduced. Therefore, we mainly focused on modifying Pseudomonas aeruginosa.

In this experiment, we mainly want to reduce the toxicity brought by the T3SS of P. aeruginosa (see safety section for details). Firstly, using the suicide plasmid pEXG2.0 as a template (where gen^R is a resistance gene, used to screen whether the suicide plasmid is successfully imported into the bacteria, and sacB expresses levansucrase protein, which can decompose sucrose, and at the same time makes fructose polymerize into levan, mediating bacterial death, used for the second screening of whether the bacteria successfully lose the plasmid), the up and downstream of 800-1000bp of the target knockout gene was cloned and the complete knockout plasmid was constructed through the Gibson assembly method. This plasmid was introduced into PAO1 by mating. After multiple screenings and PCR checks, a monoclonal bacterium that successfully knocked out the target fragment was finally obtained. After knocking out toxicity, we used fluorescent microscopy to observe whether the Type VI of Pseudomonas aeruginosa, with some toxicity genes knocked out, can be normally assembled.

Figure 1. The basic structure of T6SS ^[1]

Figure 2. Principle of homologous recombination

Construction of Expression Vectors

The Type VI components used in this experiment are Hcp1, Tse6 (Pseudomonas aeruginosa), and VgrG3 (Vibrio cholerae). The Type VI secretion system is mainly composed of the membrane complex, baseplate, and tail sheath complex, and the two proteins we selected are associated with the spike at the baseplate place. Tse6 contains a PAAR domain, which can be assembled with Type VI component proteins, and also has an effector domain, which performs functions like eliminating enemies in the natural environment. Simply put, in this experiment, what we did can be understood as replacing the effector domain with the target polypeptide, so that the exogenous polypeptide can be assembled with Type VI and secreted. The same is true for VgrG3, but it is assembled through the VgrG domain. For Tse6 and VgrG3, the Type VI secretion system secretes only one protein at a time, and if a large amount of protein needs to be delivered, Hcp1 will be used, as a single activation of T6SS can deliver tens of Hcp1. Usually, Hcp does not contain effector proteins, so we will connect Hcp1 and the target polypeptide through the AAAGGG linker, which is provided by a previous study[2].

Figure 3. Schematic of three types of fusion protein we designed

For Vibrio cholerae, we choose pBAD33 as the expression vector, which contains the arabinose-inducible operon and can express a large amount of protein under the induction of arabinose; the chloramphenicol resistance gene is a selection marker. For Pseudomonas aeruginosa, we choose pPSV37 as the expression vector, which contains the lactose-inducible operon and can express a large amount of protein under the induction of lactose; the gentamicin resistance gene is a selection marker. First, we use AlphaFold 2 to predict the structure of the fusion protein to observe whether fusion with the target protein will affect the structure of the Type VI domain or the structure of the Drug. Then through the Gibson assembly method, the cloned target fragment and vector are ligated and introduced into the corresponding bacteria.

Figure 4. Plasmids we utilized in our project, Cm: Chloramphenicol, Gen: Gentamicin

Experimental Verification

The Type VI Secretion System of normal Pseudomonas aeruginosa is membrane perturbation mediated. The Pseudomonas aeruginosa strains used in this experiment have knocked out the retS gene so that Type VI can be constitutively activated. Before performing secretion experiments, we first compare whether retS-mediated Type VI activation is normal and also want to explore whether extracellular DNA can enhance the activation of Type VI. Firstly, this is explored through the killing assay. After figuring out better secretion conditions, we will first enrich the bacteria, induce secretion under corresponding conditions, and finally check through Western blot, to observe whether secreted proteins appear in the supernatant.

Engineering

Knock out cycle 0:

The first round involved the knockout of Pseudomonas aeruginosa type III regulatory gene vfr and type III secretion proteins exoS and exoT, as well as the second step synthesis enzyme phzS for pyocyanin. The first three genes were successfully knocked out. However, no bacteria survived the second selection step when knocking out phzS, possibly due to toxicity from accumulated intermediates. Considering that the first-step synthesis enzyme may have other functions, we decided to abandon the knockout of this gene.

Expression plasmids construction cycle 0:

In the first cycle, all target peptides were constructed after codon optimization. Some were fused directly by building overlap on the primers, while others were synthesized by companies, cloned via PCR to get the fragment, and then assembled through Gibson assembly. Apart from segments like Lunasin, Adlyxin, Exenatide, Teriparatide, and Teduglutide, all other plasmids were successfully constructed and sequenced normally. The failure may be due to the difficulty in recovering short gene fragments during product recovery and the inefficiency of Gibson assembly for short fragments. In the next cycle, we plan to amplify the PCR cloning system and extend the adsorption time during product purification.

Expression plasmids construction cycle 1:

On the one hand, we repeated the Etelcalcetide experiment and confirmed that the peptide is toxic to bacteria. On the other hand, we re-conducted experiments for unsuccessful fragments, according to cycle 0's optimization results. Besides the plasmid successfully constructed in the former round, Only the Tse6-Teduglutide fusion protein was successfully built, proving that short fragments are indeed difficult to construct. We considered using direct overlap building on primers but chose to validate only the successfully constructed proteins due to time constraints and the long synthesis cycle.

Test

Secretion test cycle 0:

At this point, we only tested strains Tse6-KLA and Hcp-minihepcidin. However, Western blot results showed that only Hcp-minihepcidin was secreted intracellularly, and was unable to be detected extracellularly. We hypothesize that prolonged induction may reduce T6SS activity, and intracellular protein might be degraded. Therefore, the induction time needs to be controlled. Also, since Tse6 is scarce and hard to detect even if secreted, we need to enrich the supernatant protein with TCA precipitation. It also suggests that T6SS might not be functional, so a control group with Hcp1 is needed to observe whether the T6SS of this strain is functional. We also tested the secretion of VgrG3 fusion proteins. Firstly, the result was strange because there were some bands indicating it existed smaller proteins with 3V5 tags. It might suggest that VrgG3 was easily degraded in cells. We also discovered that the level of tag in the supernatant of the negative control group was higher than we expected. We conjecture that this is because of cell lysis, but the signal level of RpoB is undetectable in the supernatant, and the level of RpoB is low in cells. Therefore, we will change our RpoB antigen and test for another round.

Secretion test cycle 1:

Secretion test cycle 2:

In this round, we test the expression and secretion of Hcp1-Minihepcidin, VgrG3-Dulaglutide, VgrG3-Parkin, VgrG3-KLA, and VgrG3-Tenecteplase. For P. aeruginosa, its expression remains normal in the cell, but we cannot detect any signal in the supernatant. For V. cholerae, this is the second round of our test, and we are surprised to discover that our plasmid can be expressed stably. We discovered that the level of RpoB was too low in test cycle 0, so we replaced our RpoB antigen with another product with a stronger level of signal. As a result, we witnessed a high level of RpoB in our supernatant in this cycle, proving that the high level of the signal in the negative control group was due to cell lysis. In this group, as we change some induce conditions, we see that no RpoB was detected in some groups, so we may draw a result that our induction condition is optimized.