The goal of this wiki page is to present an engineering success of Phagent and all the steps that we went through.
Our engineering success is split between two sides of the project: both wetlab and drylab for which we met difficulties and found solutions to reach our engineering success.
Phagent aims at engineering phages with oncolytic genes, so they can infect bacteria of cancer microbiomes, that will in turn produce anti-cancer molecules in situ. For this year, we wanted to address a simple question: can we introduce the GFP gene in a phage, infect bacteria and have them produce GFP? And if this works, what is the impact of phage infection and GFP production on the bacteria population?
The scheme above details the 4 main steps of our project, and how they were implemented at the dryLab and WetLab levels.
1. Native phage vector
We focused on phage M13 that can infect the bacteria Escherichia coli.
We were given a purified M13 single-stranded DNA. We had to transform a population of XL1-Blue cells with the purified single-stranded DNA. We then grew that population overnight and extract ed double stranded phage vector. The main issue was to work on a phage without any selection pressure. Generally we never work with this kind of element in microbiology, so we focused on adding an antibiotic resistance gene to the phage vector in order to facilitate the next experiments and only keep infected bacteria.
2. Engineered phages
We decided to extract a cassette containing a gene encoding for the sfGFP and a resistance gene to kanamycine from the pFab217 vector that we had in the lab, and clone it into the phage vector (named p7560) which contains the M13 wild type genes. To maximize our chances of successful cloning, we used three different strategies in parallel. This workflow allowed us to avoid wasting time on troubleshooting and got us more efficient in the end. Our first approach was digestion / ligation, which eventually failed. We had big issues using this strategy because even after dephosphorylation of pFab217, the vector religated over itself, leading to false positive colonies. The second approach was Golden Gate Assembly which consists of attaching short cohesive sequences at the end of the amplified DNA sequence that will turn into single stranded cohesive ends once they are digested by the enzyme Bsa I. Last we tried the Gibson Assembly which consists of creating cohesive ends thanks to primers, the DNA fragments then anneal together thanks to the cohesive ends without requiring the use of restriction enzymes. This last approach was successful: we obtained recombinant phage vectors containing both sfGFP and Kanamycine resistance genes, to use later for the rest of the experiments.
3. Phage production
We then transformed XL1-Blue with our recombinant phage vector and purified the phages released in the culture supernatant by PEG-NaCl precipitation. We approximated their quantity using nanodrop and prepared frozen stocks of recombinant phages. Everything is now ready to get the proof of concept, but unfortunately we ran out of time to go any further in the project. To conclude, we managed to get to step 3 of our project, which is already a success as we could produce clean material to work with. However the difficulties we went through slowed the experiments advance. The first problem we met was in the preliminary experiments verifying with simple transformation that the cells could receive the phage vector we were working on (p7560). Time has been lost with miniprep results that were not satisfying, we had to redo the experiment to be sure that the E. coli could receive the phage vector. Then, while engineering the phages, some bacteria culture on petri dishes were unusable because of antibiotics dosage problem. The antibiotic wasn’t selective enough and colonies grew even without the resistance gene.
4. Bacteria infection
a. GFP production
Here we wanted to infect bacteria with the engineered phage and to assess GFP production by fluorescence measurements in a plate reader. Several conditions can be tested in parallel (phage / bacteria ratio, time after infection, growth conditions….)
b. Evolution of the bacterial population
Here we wanted to use FACS to study the balance between infected bacteria (GFP+) and susceptible bacteria (GFP-). This is very important to understand how infection and exogenous protein production impact the bacterial population. So we explored this aspect by an in silico approach: We decided to mention this side of the project as an engineering success as it is in the heart of Phagent and we had to face difficulties, find solutions to the problems, and demonstrate ingenuity. In order to reach the final model we went through different phases:
- Writing of a first model, first sketches, and equations (beginning of July)
- Reading literature to find parameters values (mid July)
- Coding a first stochastic model on Python following the Gillespie algorithm (end of July)
→ First problem: our algorithm wasn’t adapted to the time range we wanted to simulate. The time steps were too short and the number of iterations of the algorithm was too large. Moreover, the orders of magnitude of individuals (bacteria and phages) were too high for the calculation capacity of the computers we had.
- Finding another way to simulate the populations of phages and bacteria. We decided to abandon the stochastic side of the simulation even if it is biology and it is not determined and focused on the numerical resolution of differential equations. (beginning of August)
- First simulation results (mid August)
→ Second problem: the metabolic cost couldn’t be well studied because of the type of equation we gave to it: it was linear and the bacteria growth rate became negative with the growing phage population.
- Adaptation of the equation involving the metabolic cost (mi August).
- Final model and results to study the equilibrium states depending on the infection strategy and the metabolic cost (end of August).
Beyond this primary proof of concept, the idea would be to repeat these experiments with therapeutic proteins and secretion signals for proteins that need some. We would use the same infection conditions found for the first experiment. This step would imply using the composite parts we created for the project. They would all be useful but we can in particular cite BBa_K3700004, BBa_K3700009 and BBa_K3700011 each concerning one of the three therapeutics proteins that have been selected. The ultimate step would be to add bacteria to spheroids and infect them with phages modified with therapeutic proteins. Spheroids are cancer cells grown in 3D. These experimental models mimick tumors on which we can study the impact of the infected bacteria.