Although our engineered bacteria won`t and don`t need to release into the environment, we have a responsibility to apply / create a geneguard device just in case. The genes that we chosen are from toxin-antitoxin system, which originally exist in microbial plasmid, and be used as “Plasmid Addiction”. So our goals are not only to prevent horizontal gene transfer ( HGT ) between the engineered bacteria and other wild species , but also to get a better stability between different generations through “Plasmid Addiction”. Furthermore, the use of these toxin genes, such as ccdB, can reduce the using of antibiotics. This last feature is already applied by some biotechnology companies.
——————————————————————————————————————————
Figure 1. Schematic diagram framework ( details will be showed below)
Introduction
Biological safety is one of the core issues in synthetic biology. And most of the safety devices contain toxin - antitoxin genes. Then we came up with a idea: can we build a multi-functional safety device, not only solve the major problem of HGT, but also have more significant functions?
One of these functions is improving stability of the plasmid which contains exogenous genes in engineered bacteria --- that is, the ability to be stable in engineered bacteria between different generations.
In order to do that, this safety device contains two separate toxin - antitoxin systems ( four functional genes, in all) . One of them ensure the plasmids that contain exogenous genes are not easy to lose, which means these plasmids are addicted to our specific engineered bacteria. Another pair can guarantee that these plasmids will die if HGT happens.
But it`s a little late when we began to create and construct this module. It`s not easy to get these relevant genes, too. So we do not have enough measurements and data. In the future, we`ll focus on this module and do enough experiments about the toxicity and stability.
Here are schematic diagrams about this device. For more information on our Human Practices——Safety, click here.
——————————————————————————————————————————
Super problems
1. Horizontal gene transfer (HGT)
No matter in conventional genetic engineering or synthetic biology, we transfect the chassis bacteria, such as Escherichia coli, with the vector where the target genes(or parts, devices) are inserted and then get the engineered bacteria. But here comes problems, especially when we need to apply or release the engineered organisms to the environment or into other living creatures.
HGP is the most important concerns as soon as we talk about bio-safety and GM crops. Engineered bacteria contain non-natural genes, which means these genes don`t belong to them inartificially and maybe easier to lose. This could cause uncertain danger or even biohazard when those exogenous gens transfer to other species.
From all the materials and past bio-safety devices, we found the toxin-antitoxin system(TA system) in bacterial plasmids which used as “Plasmid Addiction” are great. If HGP happens, the “toxin gene a”(which can produce a stable protein, toxin A) in plasmid can kill non-engineered bacteria. The “antitoxin gene a”(which can produce an unstable protein, antitoxin B) in genome can protect our engineered bacteria.
2. Stability in different generations---Plasmid Addiction
In almost all microbic systems, low copy number of plasmids (or other vectors) may cause loss of the target genes and genetic instability of the engineered bacteria. But we need the exogenous genes(and parts, devices, systems which insert in engineered bacteria ) can be stable, so they can work for a long time. It is better if the plasmid or vector can addict to our own chassis bacteria.
Toxin-antitoxin systems contain a pair of stable toxin and an unstable antidote[1] . When the plasmids are missing, this unstable antidote is decomposed by the enzyme quickly, and the toxin causes plasmid-missing bacteria to die. That`s why we called this mechanism “Plasmid Addiction” because bacteria cannot be alive without plasmid.
In order to achieve both HGP blocking and a higher stability, it means we need at least two pairs of toxin-antitoxin genes. These two pair of genes should not affect each other, by the way.
3. Antibiotic substitutes
It is common to use antibiotics for selecting target genes during the microbiological experiments, especially in selection process. But antibiotics can cause bacteria tolerance and improve their resistance, which is not good for our people and environment.
No matter use the method of insertional inactivation, or the toxicity itself(such as, ccdB ), we do not need more metabolites for selection. This feature is already applied in commercial run.
4. Universality and other problem
If we want this geneguard can be used in more projects, we should consider the carrying capacity of plasmid or vector. Fortunately, genes from toxin-antitoxin system always in 200—500bp. Their size are suitable for a geneguard device.
On the other hand, genes from toxin-antitoxin system are naturally exist in microbes` plasmids. Traditionally, they are used for “Plasmid Addiction” and there`s no harm to our environment , cause this kind of system already run for a very long time out of laboratory.
Super Power
There are so many potential genes can be chosen in toxin-antitoxin systems. Some of them interact with each other in the form of proteins, which already got fully research. The others can interact through RNA—RNA interaction and post-transcriptional control: only the “antidote RNA” degrade, the “toxin RNA” can translate into protein[2]
We select genes form the first form because it`s not that complicated in mechanism. ccdB/A are the most wildly researched genes. We name them “toxin a”(ccdB) and “antidote a”(ccdA). Another pair genes are from E.coli, too. We name them “toxin b”(parE) and “antidote b”(parD).
Figure 2. Engineered plasmid and engineered genome
a) Plasmid Vector: Insert “toxin a” and “antidote b” in the Plasmid vector.
b) Bacteria Genome: Insert “toxin b” and “antidote a” in the bacteria genome through transposon.
Figure 3. Engineered bacteria alive(Plasmids contain)
c) Our engineered bacteria contain both these kind of plasmid vector and bacteria genome. The toxin and antidote proteins can interact with each other, so the bacteria alive.
Figure 4. Non-engineered bacteria dead(HGT happens, plasmids transfect)
d) When HGT happens, the plasmid can transfect to other microbes. Due to the lack of antidote b and its production in the new host (non-engineered bacteria), toxin A will cause new host to die. Then HGP won`t happen because the plasmid is lethiferous.
Figure 5. Engineered bacteria dead(Plasmids lost)
e) When the plasmid is missing, the lost of the antidote b will cause our engineered bacteria to die. Because in the toxin-antitoxin system, toxin proteins are always stable, antidote proteins are unstable and have a short half-life period. So there are not enough Antidote B to block Toxin B. And the Plasmid Addition is established.
f) This “Addiction” feature can be used as selection, too. Because the bacteria will die if they don`t contain this necessary plasmid. Insertional inactivation is another method because we can take the advantage of the toxicity. “Antibiotic Free” is not a dream anymore.
Note: “toxin a” , “antidote a” stands for gene; “Toxin A” , “Antidote A” stands for protein.
———————————————————————————————————————————
Modelling
This system includes 4 proteins:
1. Toxin1 can sentence the bacteria to death in certain concentration.
2. Toxin2 is another kind of toxin which has the same effect as toxin 1.
3. Antidote1 is the antidote to toxin1.
4. Antidote2 is the antidote to toxin2.
In our system, toxin1 and antidote2’ genes are settled on the genome, antidote1 and toxin2 are on the plasmid, thus two antidote genes can prevent bacteria from death caused by two toxin genes. Like Siamese twins, whenever plasmids and genome are separated (wild type cells get the plasmid or artificial cells lost their plasmid), toxin genes on both side would kill them.
See more details on our modeling —“geneguard”, please click here.
———————————————————————————————————————————
Discussion & Future Work
It is sorry to say that we do not have enough time to measure this geneguard device. But we will and we guarantee to do this experiment as soon as we can. On the other hand, there`s more tasks should be done in the future, such as more selection of different genes from toxin-antitoxin system, different construction methods.
1. Toxicity
Does the toxin gene in plasmid have enough toxicity? Can these toxin proteins kill all the other microbes (non-engineered bacteria) as soon as HGP happens? Toxicity measurement is the first we should consider. It`s important to “Addiction”, too. We need them kill our engineered bacteria when plasmid lost.
2. Stability
Toxin and antidote can be stable in natural state, and we need them to be stable in our engineered bacteria. We have to consider both strength and rate of transcription, translation. Modeling is a best way to solve this problem.
3. Selection
Toxin-antitoxin system is a huge family. No one can say which is the most suitable for different engineered bacteria until they do enough experiment. Because these genes have different mechanisms and characteristics. We will choose as many pairs as we can in the future, and use mathematical modeling to model the system and optimize the efficiency of different combinations.
4. Construction
One pair of genes: toxin a in plasmid and antidote a in genome, which can block the HGP;
Another pair of genes: toxin b in genome and antidote b in plasmid, which can establish addiction mechanism and be more stability.
Can we build this in another way?
One pair of genes: toxin a in plasmid and antidote a in genome, which can block the HGP. This part is the same;
Another pair of genes: both of toxin b and antidote b are in plasmid, which can establish plasmid addiction, too. Because toxin is more stable than antidote. When plasmid lose, Toxin B can exist longer than Antidote B, engineered bacteria will die. In fact, this is the natural establishment of “Plasmid Addiction”[2] .
———————————————————————————————————————————
References
[1] Hanna Engelberg-Kulka, Gad Glaser, Addiction modules and programmed cell death and antideath in bacterial cultures. Annu. Rev. Microbiol. 1999. 53:43-70
[2] Ming-xi Hu, Xiao Zhang, Er-li Li, Yong-jun Feng. Recent advancements in toxin and antitoxin systems involved in bacteria programmed cell death. Microbiology, vol. 2010, Article ID 781430
[3] A Jaffé, T Ogura, S Hiraga. Effects of the ccd function of the F plasmid on bacteria growth. J. Bacteriol. 1985, 163(3):841. P. 841-849
"
