Within the E. coli genome, there is the naturally occurring toxin-antitoxin system whose production is altered in response to various types of stress. In layman’s terms, a toxin-antitoxin system consists of two genes: one coding for the toxin, or “poison”, and one coding for the antitoxin, or “antidote”.
There are three different types of toxin/antitoxin systems, all with different products effectively committing apoptosis. A general overview of all types are listed below.
Type 1: Inhibition takes place when the antitoxin RNA binds to the complementary toxin mRNA. If there is not enough antitoxin RNA being transcribed, toxin proteins will be produced, inducing toxicity through cell membrane damage. Toxin RNA has a half life of ~20 minutes, while antitoxin RNA has a half life of ~30 seconds.
Type 2: both genes code for separate proteins, which bind to each other in a normal, unstressed state. While undergoing stressful conditions, the production of antitoxins will drastically decrease, allowing the toxin protein to act as a pseudo-RNAase, cleaving mRNA.
Type 3: The most recently discovered, inhibition of this toxin requires the interplay between a toxin protein and an antitoxin RNA gene. There is only one example of this system so far - the ToxIN system from the bacterial plant pathogen Erwinia carotovora.
For our purposes (tuning population ratios of symbiotic strains), Type 2 systems were determined to be ideal, since they have the greatest chance of longevity/sustainability as proteins, rather than RNA strands. There are also at least thirty-three toxin-antitoxin pairs in E. coli alone, giving us a greater selection to work with in our project.
In a Type 2 system (diagrammed above), the antitoxin gene is usually upstream of the toxin gene, and the antitoxin is usually the more unstable of the two, degrading much more rapidly than the toxin. As this is the case, antitoxin proteins are produced in a much larger quantity in order to counteract the toxin. Antitoxin and toxin pairs are coded into proteins and bind to each other to prevent an accumulation of toxin. In stressful situations – when there is DNA damage, drastic change in temperature, or lack of nutrients – stress-induced proteases cleave antitoxins and leave the toxins to cleave the mRNA strands.
In order for us to work on our project, we had to pick a few toxin-antitoxin pairs to work with. We chose MazE/MazF, RelB/RelE, and YefM/YoeB based on the size of the antitoxin (~95-115 aa). Each pair is classified under the Type 2 category.
How could Toxin/Antitoxin pairs be used to tune or regulate strain ratios in E. coli?
Although toxin-antitoxin pairs usually function within one cell, we propose to use the system as a way of controlling ratios of two different cells (almost like a kill switch, but perhaps the antitoxin could be a "life switch"). So, instead of using amino acids to feed cells and allow them to survive, we could feed cells varying amounts of a toxin or antitoxin to determine their presence in a co-culture.
We designed primers and PCR'd each individual gene off of the E. coli genome and used Seamless Cloning (Life Technologies) to insert the gene into the plasmid we were using: pCDFDuet (Novagen). We then transformed each plasmid into E. coli, grew them up, and performed experiments to see if:
the expression of the individual toxin or antitoxin had an effect on growth
the toxin or antitoxin could be leaked into the media
a strain producing antitoxin could affect the growth of a strain producing toxin (or vice versa)