Our biodetection system is composed of a sea sponge and of a genetically modified bacteria strain that lives inside this sponge. During our research, we have often wonder about the best way to safely contain our modified bacteria if Sponge Patrol was ever to be used by industrials or water treatment plants, and thought about many problems and several potential solutions.
In order to safely contain the bacteria out of the environment, we could put them in a hermetic aquarium, with a filtration system that doesn't let any bacteria leave the aquarium. We would then need to put samples of the water we want to test in the aquarium.
But we also thought of other ways to use our sponge-bacteria system, such as releasing them in the oceans in regions that are likely to contain important concentrations of pollutants. The release of Genetically Modified Micro-Organisms (GMMO) in the environment is very controversial though, as it raises several safety issues. As of yet, there has been no legal release of GMMO in the oceans. But then we thought of a new possible way to contain the GMMO: what if we designed an epibiosis strain of Pseudovibrio denitrificans, such that it would not be dependent from its host and unable to live away from it? Would such an epibiosis be a safe containment system?

In order to better understand the legal and safety issues related to the release of GMMO in sea water, and to define whether epibiosis could be an efficient containment system of GMMO, we digged into the litterature and discussed those points together.

• What are the risks involved in the release of GMMO in the oceans?

• Can epibiosis be an efficient containment system?

As far as we know, releasing GMO or GMMO in the oceans has presently not been legalized anywhere in the world. There is a Canadian campany that is breeding genetically modified salmons, AquaBoutny Technologies, but the fish are physically contained in off-shore water tanks. These fish are not yet commercialized, though the company has asked for an authorization to sell their product.

Since we lacked actual examples of GMMO released in the sea from which we could have retrieved data, we decided to first study some experiments that have been done in soils in order to evaluate the impact of released GMMO in the environment.
In 1993-1994, Ian Thompson and his team conducted a research at the University of Oxford Wytham Field Station, where they released free-living GM bacteria in fields. For their experiment, they mostly used an engineered Pseudomonas fluorescens whose presence they could easily detect in samples taken from the fields. Thompson reports that the GM bacteria survived a whole season in the field, before disappearing because it couldn't compete with the existing, natural bacterias present in the soil and in the plants. They also noticed that the bacteria survived better in some environments than others: they survived longer in the glasshouse where there was less competition, and they became as numerous as natural bacteria in sugar beets for a while. The overall result, though, was that the GM bacteria did not survive in the long-term, because it could not compete with the existing microorganisms.
Other similar experiments have been made in different soils, and the result so far has always been the same: the GM bacteria was not able to compete with natural species, and rapidly disappeared.

Thompson also co-founded a company, Microbial Solutions, where some GMMO are used as biosensors in water. Since the release of GMMO in water outside labs has not been authorized, they are only used in physically contained environments. But as reported in the report of the 2013 Workshop on "Synthetic biology: containment and release of engineered micro-organisms", Thompson believes than even if they were released in the environment, these bacterias would also be unlikely to compete with natural species, because realizing the additional function - in this case, sensing toxic compound in industrial waste streams and signal their presence - would be costly in terms of energy for the cells. It would be a loss of energy for the bacteria, for which they would not gain any advantage. Those engineered bacterias would thus be less competitive than bacterias that do not realize this function.

This lead us to believe that for the same reason, our engineered Pseudovibrio would have a very low chance of survival if released in the environment. In fact, it probably wouldn't even live very long inside the sponge, if we put our bacteria in presence of all the other microorganisms that live inside Spongia officinalis. Because of its additional biosensing function, our strain would probably either get rid of the function, or disappear given a certain amount of time.
The fact that our bacteria may not survive for a long period of time is necessarily a problem if we want to use the sponge as a biosensor, though. We would simply need to inject new GM Pseudovibrio in the sponge every time we want to test the water. The time it would take for the bacteria to produce fluorescent proteins if the water was polluted should indeed be shorter than the time it would take for other bacterias to overcome our Pseudovibrio.


However, the low-probability of survival of our GM bacteria in the long term does not suffice to say that releasing it in the environment would be harmless. First, because the possibility that the engineered Pseudovibrio would survive in the sponge despite our expectations is not null. Second, because even if the bacteria strain doesn't survive for a very long time, it may still have negative consequences on its environment. And the environment we have to take into account is not only the sponge and its microbiome, but also other marine species since the bacteria is not, as of yet, contained inside the sponge.

To evaluate the consequences of a GM micro-organism on an ecosystem as complex as the ocean, though, is near impossible because we don't have enough data to make a relevant model. The behavior and impact of the micro-organism could moreover be highly variable given the specific conditions of each location. The only data we can try to rely on, is our knowledge of the DNA of the micro-organism: knowing the different genes in the bacteria's genome, we may be able to predict whether it could be harmful. For example in the Sponge Patrol project, the inducible promotors and the genes coding for fluorescent proteins we want to put in Pseudovibrio are unlikely to have a negative effect on the ecosystem.
But we cannot be certain that no unexpected effect or no mutation will occur, that could be detrimental in some way. Very often in synthetic biology, scientists don't get the expected result when they design and engineer a biological system with new functions. One example is that gene expression can change a lot depending on the context. Another is that two parts put together in a construction can have an emergent, unpredicted function, instead of just the additional functions of both the parts. Since the sea is a very rich and hardly known ecosystem, we truly can't know if the release of our bacteria could have detrimental effects.

When we first started to work on Pseudovibrio denitrificans, we were very interested by the fact that it was living inside a sponge, or to be precise at the surface of a sponge. This epibiosis is a form of commensalism where an organism, called epibiont, lives at the surface of another, called basibiont: the epibiont is harmless to its host, but benefits from it (often from nutrients produced by the host). We then had this thought: if the bacteria depended on nutrients produced by the sponge to survive, then it may not be able to live away from it; and hence this epibiotic relationship would provide an efficient biological containment of the GMMO. We read the literature about the two organisms, and soon discovered that Pseudovibrio denitrificans was not dependent on the sponge: even though the microorganism can be found in the microbial florae of the sponge, some similar strains of denitrificans can also be found in seawater and in other organisms.

We still decided to work with Pseudovibrio, even though it was probably able to leave its host, because of its other interesting characteristics :
- We wanted to introduce a marine bacteria in iGEM, in order to tackle the huge problem of water pollution.
- This "denitrificans" bacteria appeared to already have a system that allowed it to degrade and use nitrite as a source of energy.
- Since it was already part of the microbial florae of the sponge, we thought that our GMMO had a better chance to survive inside the sponge than other marine bacterias, which meant that we could use the huge filtration power of the sponge to our advantage.

And though we knew that our bacteria could live away from the sponge, the idea of a strict dependence with its basibiont stuck. Throughout the summer, we kept wondering: what is it WAS dependent on the nutrients produced by the sponge? We discovered that many bacterias, if not Pseudovibrio, had been studied and were thought to be symbiotic with sponges. In 2002, Hentschel et al. mentioned different symbiotic relations that have been found between sponges and micro-organisms, like nutrient acquisition and processing of metabolic waste. And some of those bacterias were believed to have been living inside the sponges for millions of years. It doesn't prove that these bacterias are dependent from the sponge, but at least it shows that life inside the sponge was favorable enough for those microorganisms that they would prefer to stay inside this host than leave away from it. We then thought that it could be very interesting to either engineer Pseudovibrio to make it symbiotic with its host, or to use insert our Biobricks in one of those symbiotic bacteria: maybe then the engineered bacteria would safely stay inside inside the sponge, and not spread into the seawater, even we were to put Spongia officinalis back into the oceans.

We discovered that there were actually very few papers that mentioned symbiotic or epizoic GMMO, nor biological dependence to an host as a possible biological containment. So we decided to research and discuss ourselves whether a strict epibiosis could be an efficient biological containment of GMMO - by strict epibiosis, we mean a relation such that the epibiont would not be able to survive without its host.

1) Could we engineer the bacteria so that they would become strictly dependant of the sponge?

Our first question was whether it was possible to transform Pseudovibrio denitrificans to make it strictly dependent of the sponge. We currently do not know whether Pseudovibrio gains anything from its epibiosis with the sponge. The most probable, since it can be found both in the sponge, in other organisms, and in sea water, is that there may be a small advantage to living at the surface of the sponge, but not one important enough that the bacteria would necessarily prefer to stay there. If we wanted to engineer a bacteria that would be strictly dependent of the sponge, it should be a priori possible with a better knowledge of the sponges and of the micro-organism. We currently don't know Spongia officinalis nor Pseudovibrio denitrificans well enough, since they were rarely studied. But we could for example search for a nutrient used by Pseudovibrio that would be secreted by the sponge ; then we would try to identify the genes in the bacteria's DNA that are involved in the biosynthesis pathway of this nutrient. We would then simply need to knock-out the genes coding the synthesis of this nutrient in the bacteria, and obtain an auxotrophic strain. This bacteria would not be able to grow without an external source of this nutrient: it would be able to grow in the sponge which produces this nutrient, but not in seawater where it wouldn't be able to find it. In such a system, the host sponge would provide to the bacteria a necessary element for its growth and survival, and this would be the basis of a strict epibiotic relationship between the two organisms. But would it really be sufficient to contain the GM bacteria inside its host, if we were to put the sponge in the ocean?

2) Would a strict epibiosis be an efficient containment system?

Our idea was that a strict epibiosis — or a strict symbiosis — could be an improved biological containment system. Usually, biological containment imply that the genetically engineered organism need to be given a nutrient that is only found in the lab in order to survive. But if the nutrient was produced by a host organism rather than by scientists in a lab, then we wouldn't need to have scientists feeding the micro-organism everyday, and we wouldn't need to keep the organism inside or near a lab. Theoretically, the host organism with the modified bacteria could be released in the wild without any risk of bacteria leaving the host, since they need the nutrient produced by the host. However throughout our researches and reflexion, we came across several issues that could be problematic for the efficiency of this biological system. The first is that the nutrient for which Pseudovibrio would be auxotrophic could potentially be found elsewhere than in the sponge ; for example, it could be produced by an organism living close to the sponges. Thus the bacteria could colonize this other organism, instead of staying in the sponge. The second is that Horizontal Gene Transfert (HGT) may occur, and that the synthetic genes could be passed to other micro-organisms, which would not be contained inside the sponge. Hence the synthetic genes would escape our containment system and spread in the ocean. The third is that if we engineered the bacteria to make it auxotrophic, then it may be even less competitive compared to the other micro-organisms living in the sponge. The sensing system coupled with a production of fluorescent proteins is already an important physiological burden for the bacteria ; adding an auxotrophy would probably add another selective pressure on our strain, and make it even more likely to either eject the sensing system, find a way to circumvent the auxotrophy, or not be selected and disappear. The same would be true for any "kill switch" we would try to add on our bacteria: as pointed out by Wright et al., “the higher the complexity of a safety device, the more prone it may be to disturbance and failure. It is therefore important to understand the expression ‘cost’ of each component, as several in tandem will place an undesirable physiological burden on the host and in turn act as a selective pressure to eject the system" (Wright, Stan, Ellis 2013). Any engineered bacteria with additional functions that give them no benefit, have a very low chance of survival if we released it in the wild. Though they can grow in controlled environments, in the presence of competitive organisms they are very likely to get rid of the biological containment system.

To have bacteria that lives at the surface of the sponge is very interesting in order to profit from its filtration power. However, it is probably not possible at our current knowledge to create an entirely fail-proof containment system based on a strict dependence of the bacteria to nutrients produced by the sponge. This strict dependence would decrease the probability of our engineered bacteria to spread into the oceans, though. And since we believe that our bacteria would not be likely to survive long in the wild because of its lack of competitiveness, and that the modifications brought to its genome are unlikely to be dangerous to the environment, placing a sponge with our bacteria in the oceans would probably not be very dangerous for the environment. But does a low chance at escaping, a low chance at survival, and a low risk of harm to the ecosystem, suffice to say that it would be safe? The risk, however thin, can never entirely disappear.