Bacteria-Phage Population Model
The OhioState 2021 team created a steady state population model and that can be found here. This year, we attempted to remove the steady state assumption from our population model. To do that we contacted Dr. Lenski and he suggested the Runge-Kutta method to solve the system of ODEs.
Two iterations were completed with the population model, one where the concentration of free phage (P’) and uninfected bacteria (S’) were set equal to the concentration of free phage (P) and uninfected bacteria (S) at time t. However, this iteration produced results that were incorrect, so another iteration was run to fix the issue. For the next iteration, P’ and S’ were set equal to P and S at t-τ, where τ is the latent period of time. More information about the population model can be found here.
Cellular Automata Model
In talking with Dr. Strathdee, author of The Perfect Predator and Co-Director at the Center for Innovative Phage Applications and Therapeutics, we saw how important phage cocktails are. If one only administers one type of phage or phage that all target the same bacterial receptor, there is a high risk the bacteria mutate or evolve to be resistant to that phage. However, using many different phages can keep selective pressure on the bacteria and help stop it from mutating away from the phage. We created a python model to visually explain this complex topic, by using cellular automata to show how a bacterial colony develops in the presence of either one or two types of phages. More information about the population model can be found here.
This model was not created as is right away. Instead it was created over many iterations. The first draft of this model tested one phage type and one bacteria type. Once this model was shown to run successfully, a second type of bacteria and phage were coded in. With two phage and two bacteria in the model we could finally represent a basic phage cocktail. Each bacteria will be assumed to be different, whether one type evolved from the other or if they are simply different. The two phages will act upon each bacteria of the same type only. When this model was constructed and ran, Figure 1 resulted.
A weird pattern in the data is seen. If the two phage heat maps are overlapped, they essentially fill out the whole grid. So no type 1 phage are getting into the areas where type 2 phage are. Looking back at our initial assumptions provides a clue. One says that only one type of phage can surround a bacteria at a time. This is not realistic, phages can surround any type of bacteria, their types only impact what bacteria they can infect. In order to fix this, the phages were allowed to occupy the same space. That is, both phage type 1 and type 2 could be surrounding the same type of bacteria; however, only the phage that could infect the bacteria actually went on to infect. In this way, the phages could spread throughout the grid space, infecting and killing as much bacteria as they could. Figure 2 shows the updated results and the lack of cohesiveness between the two phage maps.
Wet Lab
As we were starting our protocols for cloning our phage DNA, we did not want to spend valuable time correcting easily avoidable mistakes made in previous iterations of our project. We contacted Dr. Brian Ahmer to talk about our plans for our constructs. Based on our talk, we resolved to use both the T5 and pR promoters to highly upregulate protein production of our construction. Here, we took the lac operon out of the T5 promoter to remove a limiter in its operation. In addition, we also made a rrnB-pR fusion promoter after our own research to find something that E. coli cannot suppress.
At first, we planned to use luciferase, however after the OhioState 2021 team had a few troubles regarding its expression, we decided to use a much stronger fluorophore in mCherry. Using this, we hoped to have the strongest possible expression in which to compare our promoters.
One of the decisions that must be made when deciding to make the phage reporter is promoter design. So we decided to experimentally test 4 reporters and test them in another reporter assay for the promoters.
We wanted to have the strongest possible promoter sequence for our phage reporter, so our diagnostic time would be as fast as possible. The team researched some common strong promoters, and specifically liked the idea of a constitutive promoter, which contains fewer or no inhibiting sequences. We then designed 4 different parts. We ordered the required sequences to spec, and began the cloning process.
First, we isolated the backbone from pET28a to give a basis so we could test our diagnostic construct. Then, we began ligating the sequences into the backbone, and hit our first issues. After 2 attempted and failed ligations, we didn’t have enough backbone to continue.
We consulted the graduate students in Dr. Wood’s lab, as they are very experienced with ligations and know ways to improve on further iterations. They recommended that we reisolate the backbone for all following attempted ligations. However, even once we implemented this change there was still trouble: the transformations weren’t taking. No colonies containing our specific construct had survived the transformation process. Despite this, we reconfirmed the size of our four inserts, resulting in the gel below showing that they were the approximate sizes that were expected.
