It was recently in the news that researchers have genetically engineered tobacco with 40% more efficient photosynthesis. I had first seen this kind of research at a seminar during my graduate years at MIT. The presenter noted that the genetically engineered plants grew faster and showed memorable side-by-side pictures of scrawny-looking normal plants next to their larger and lusher engineered brothers. I tried to find out more after the fact, but had forgotten the name of the presenter and lab. When I searched online for the research, I found lots of people proposing doing this kind of thing, but not the lush success story I had just seen. I wanted to learn more not because it was super cool and hugely important (which it was), but because it was the first (and to date only) example of what I thought was a dangerous genetically modified organism.
Unambiguously safe engineering
Back when I was working in the world of genetically engineered organisms, I would often talk to non-scientists about actual and potential uses of synthetic biology. Invariably, one question always came out in those conversations: But what if one accidentally escaped, wouldn’t that be a disaster? Anyone who works in genetic engineering and synthetic biology has heard this question many times. It prompts a sigh, then a firm no, and if we are feeling patient, an explanation that that is not how ecology works.
The PEPCK supermouse we made in Richard Hanson’s lab ran faster, ran longer, lived longer, and was fertile for longer than a normal mouse. When people learned about this mouse, they would sometime worry what would happen if such a mouse escaped. Would the city be overrun by fast, immortal, virile rodents? No, that is simply impossible. These mice were created by over-expressing a single protein in their muscles. If over-expression of this protein would create super fit mice (evolutionarily not athletically), then it would have already happened by random chance. Our data actually revealed the fatal flaw: the mice required a lot more food. An escaped supermouse would have quickly starved to death.
Fundamentally, if an organism escapes in the environment, it must compete with similar organisms in that environment in order to survive. Any simple change we make to an organism is just not going to improve the fitness of that organism relative to the wild type. Any increase or decrease of existing genes cannot make the organism more fit or evolution would have already done it. This includes gene knockouts, the most popular type of genetic engineering. If removing one or more genes made the organism more competitive, it would have already happened.
The issue is the same for changes that introduce new genes, like adding vitamin A to rice, programming bacteria to make gasoline, or making fish glow-in-the-dark. There is no reason to think that these changes would do anything but lower the fitness of the organisms in the wild. In a survival-of-the-fittest world, survive they would not.
Improving photosynthesis
The core enzyme of photosynthesis is rubisco. This is the enzyme that takes carbon dioxide out of the air and incorporates it into sugar. As anyone familiar with chemistry knows, this is a ridiculously difficult reaction to preform and it is a miracle that life does it at all. It is made even more difficult because oxygen likes to undergo the same reaction. Oxygen is smaller in size (two atoms versus carbon dioxide’s three) and is in higher concentration (21% versus carbon dioxide’s 0.04%) so it gets into the active site of the enzyme and attaches to the sugar instead. This poisons the sugar and the plant is forced to spend a lot of energy removing the oxygen through a process known as photorespiration. For years, protein engineers tried to improve photosynthesis by tweaking rubisco to either attach carbon dioxide better and oxygen worse in order to improve photosynthesis in desirable plants. This failed despite tremendous effort, and in retrospect, it is obvious that it should have. Any possible improvement to this molecule would have already been made in its multibillion-year history.
So if evolution is so efficient, how did synthetic biologists manage to improve photosynthesis in the case that made the news? Well, I used the terms “simple” and “tweak” and “existing” to describe the changes that evolution can make. If an improvement can only be obtained by supplying several new genes all at once, evolution is ineffective because the odds are too low that all of the required genes would arrive at once. The recent tobacco paper made several improvements to photorespiration, including one that was made to arabidopsis previously. This improvement took the glycolate oxidation pathway out of E. coli and put it into the chloroplast of the plant. This alone boosted growth by 10%. Plants have been around billions of years, yet not once did a plant evolve this game-changing attribute. This is because it took transferring five genes at once, a feat that is simply beyond the capabilities of natural evolution.
My concern
So why am I worried about this plant? Because the safety of synthetic biology has long rested on the assumption that the ecology is at a stable equilibrium, the native organisms are already optimized for their environment. This work to improve photorespiration appears to violate that assumption. When we make mice that can run faster than all the other mice out there or bacteria that can make oil better than all the other bacteria out there, there is no risk to the biosphere. But when we make a plant that is better at growing and utilizing energy and carbon dioxide than all the other plants out there, I am unable to convince myself that this is still safe. If one of these plants escapes, why would it not slowly spread as it outcompeted all the plants of similar size? Each species is specially adapted for its niche, which makes it hard for other species to enter, but 40% better energy utilization is an unrivaled superpower.
The research so far been on the arabidopsis and tobacco. Neither of these plants would I like to see everywhere. If the researchers accomplish their next goal of engineering food crops, maybe it would not be that bad of an outcome to accidentally blanket the planet with food. But I don’t think that is even close the worst-case scenario. Genes can hop between plants through a process called horizontal gene transfer. We may make a prolific wheat plant with a genetic construct, but it is doubtful that wheat is the most prolific plant that could possibly hold that construct. Those genes will eventually be stolen by other photosynthetic lifeforms. When the five genes were spread around the E coli genome, the odds were vanishingly small that they would all transfer at once to a welcoming plant host. But to build the construct the researchers connected all the genes together in a single short segment of DNA and modified all the proteins so that they would end up in the chloroplast. I don’t know the overall rate of horizontal gene transfer from plants, but a single segment of DNA is certainly more likely to jump species. If we start growing crops with this construct, eventually pieces of those plants will end up in the ocean, and eventually some algae will absorb the DNA and receive an competitive advantage the world has not seen in eons. This is the outcome that worries me the most.
Conclusion
It may sound like I think this research should be stopped. But that is not at all what I am trying to get at. I am simply not knowledgeable enough about the details of this engineering or ecology to make recommendations. I want to raise this issue because I don’t see anyone else raising it. The best outcome of this post is that someone points out why I am wrong about the dangers of improving photosynthesis. Maybe photorespiration is important for the health of offspring. Maybe photorespiration is important for robustness to changes in the environment. Things like these may doom a genetically engineered plant that escaped but would still leave the plant useful to us. There is a huge upside to this research if that 40% increase in biomass can be translated to food mass in crops. People who are starving will be fed. People everywhere will see the cost of their food go down.
I think that the chance of a worst-case scenario (a single organism overrunning a huge chunk of the biosphere) is low but not zero. One mitigation would be to split the genes into different DNA segments before putting them into plants will be grown out in the field. Splitting the genes up will eliminate the risk of horizontal gene transfer. A plant already bred for food is far less likely to spread uncontrollably than an algae that picked up the construct.
I want to see this research continue. The potential benefit to humanity is too great to simply stop. But I would love to see experts address or mitigate the dangers before unleashing this on the world’s fields.