Interesting that old school serial dilution ALE performs the best. How were you changing the selection pressure - a linear increase in salt concentration each day or something else?
We didn't change selection pressure over the course of the ALE, mostly to keep it consistent with the BioBloom and Additive Engineering selections, which were much shorter. Everything was passaged in LB(miller)+4% salt (w/v), which is about the highest concentration that can sustain passages at this dilution volume.
Did you see any meaningful difference with HGT when you added DNA on a transposon vs a plasmid? Did DNA from the plasmid get integrated into the chromosome or remain plasmid-carried? I'm not surprised that AE worked better, I'd imagine that there are a ton of mutations which confer salt tolerance, and allowing it to run over multiple generations just increases the chances for more of those mutations to pile up in cells over time. Super cool work!
The biggest difference between plasmid and transposon HGT approaches was that most of the winning transposon strains acted like a transposon-knockout screen, where the insertion location mattered more than the new sequence. It seems easier to get insertional effects than to acquire new genes with horizontal gene transfer.
The plasmid DNA wasn't integrated into the genome as far as we saw; it just remained stable on the plasmid (with antibiotic selection).
Another strange thing was that when we pulled winners from the selections and tested them clonally in our IC50 experiment, all of the Transposon-winners had higher salt IC50s than the parent, but not all of the plasmid-winners did. We might have selected for general growth instead of salt tolerance in a few cases. There's a couple of possible explanations, it's possible that the higher copy number of the plasmid lets different kinds of hits win.
+1 that I'm not surprised to see the endogenous E coli genome able to re-regulate itself through SNPs to tolerate higher salt. I'm interested to try HGT again when the selective pressure is more unusual. I'd hypothesize that there will be stressors that you just can't tolerate with SNPs alone, and you need whole new genes/enzymes for. Antibiotic resistance is the classic example, but there's plenty of Mars-relevant ones I expect to fall in this category too, like perchlorate tolerance.
Thank you both for your quick responses! The work you guys do is so neat and I love that you write about it. Yes, plasmid carriage is such a nuanced topic, I'd be curious how fast they lose the plasmid when they aren't under selection. I agree that there are probably gonna be a ton of stressors out there- perchlorate I hadn't thought about! Cold and radiation are the two I would think most. Are you guys also working to develop better genetic tools in some of the weird extremophiles as well? I know that's a HUGE question (I work in a system with a horrendous genetic model myself), mostly curious as to your approach.
Part of why we want to work with HGT is because we'd love to be able to get the advantages of extremophiles without needing to develop genetics in every single one. It's so difficult to domesticate non-model microbes, and there are SO MANY that we might want to work with. Hence the desire to explore ways to transfer DNA from extremophiles into microbes where we already have genetic tools.
It will only get you so far though! For the moment we're working on engineering microbes that can use nutrients directly in Martian regolith, but are protected from cold/radiation. As we venture toward microbes that work in minimal greenhouses or outdoors it may not cut it anymore to start with a non-extremophile chassis and add DNA. So I suspect that more direct extremophile work is in our future :D
Oh absolutely!! To both of those things. E coli is a fantastic model for many things yeah.... genetic engineering in non models. Definitely a whammy. I've had many a professor ask me "Well can't you just do the knockouts?" in my organism and I have to be like... well... no. XD
The idea of greenhouse work in this system is fascinating! Are you actually working with plants or just to actually test these guys in as close to Martian soil as possible? How are you going to simulate the harsh light/cold conditions?
Yes, one does not simply do genetics in any old non-model microbe 🤣
For greenhouses, what we're doing at the moment is working with some simple mathematical models to figure out how to properly determine the day/night and seasonal temperature, pressure, light, and radiation profiles as a function of what exactly you mean by 'greenhouse': everything from fully enclosed + heated buildings to literally a plastic bag outside have been proposed and could work in some cases. So mostly still trying to get to a clean problem statement. Much of 2026 will probably be spent actually determining which organisms could be grown in these conditions.
Interesting that old school serial dilution ALE performs the best. How were you changing the selection pressure - a linear increase in salt concentration each day or something else?
We didn't change selection pressure over the course of the ALE, mostly to keep it consistent with the BioBloom and Additive Engineering selections, which were much shorter. Everything was passaged in LB(miller)+4% salt (w/v), which is about the highest concentration that can sustain passages at this dilution volume.
Nobody dares to read this! Are you one of them?
https://thegonersclub.substack.com/p/consciousness-is-a-trick-of-meat
Did you see any meaningful difference with HGT when you added DNA on a transposon vs a plasmid? Did DNA from the plasmid get integrated into the chromosome or remain plasmid-carried? I'm not surprised that AE worked better, I'd imagine that there are a ton of mutations which confer salt tolerance, and allowing it to run over multiple generations just increases the chances for more of those mutations to pile up in cells over time. Super cool work!
The biggest difference between plasmid and transposon HGT approaches was that most of the winning transposon strains acted like a transposon-knockout screen, where the insertion location mattered more than the new sequence. It seems easier to get insertional effects than to acquire new genes with horizontal gene transfer.
The plasmid DNA wasn't integrated into the genome as far as we saw; it just remained stable on the plasmid (with antibiotic selection).
Another strange thing was that when we pulled winners from the selections and tested them clonally in our IC50 experiment, all of the Transposon-winners had higher salt IC50s than the parent, but not all of the plasmid-winners did. We might have selected for general growth instead of salt tolerance in a few cases. There's a couple of possible explanations, it's possible that the higher copy number of the plasmid lets different kinds of hits win.
+1 that I'm not surprised to see the endogenous E coli genome able to re-regulate itself through SNPs to tolerate higher salt. I'm interested to try HGT again when the selective pressure is more unusual. I'd hypothesize that there will be stressors that you just can't tolerate with SNPs alone, and you need whole new genes/enzymes for. Antibiotic resistance is the classic example, but there's plenty of Mars-relevant ones I expect to fall in this category too, like perchlorate tolerance.
Thank you both for your quick responses! The work you guys do is so neat and I love that you write about it. Yes, plasmid carriage is such a nuanced topic, I'd be curious how fast they lose the plasmid when they aren't under selection. I agree that there are probably gonna be a ton of stressors out there- perchlorate I hadn't thought about! Cold and radiation are the two I would think most. Are you guys also working to develop better genetic tools in some of the weird extremophiles as well? I know that's a HUGE question (I work in a system with a horrendous genetic model myself), mostly curious as to your approach.
Part of why we want to work with HGT is because we'd love to be able to get the advantages of extremophiles without needing to develop genetics in every single one. It's so difficult to domesticate non-model microbes, and there are SO MANY that we might want to work with. Hence the desire to explore ways to transfer DNA from extremophiles into microbes where we already have genetic tools.
It will only get you so far though! For the moment we're working on engineering microbes that can use nutrients directly in Martian regolith, but are protected from cold/radiation. As we venture toward microbes that work in minimal greenhouses or outdoors it may not cut it anymore to start with a non-extremophile chassis and add DNA. So I suspect that more direct extremophile work is in our future :D
Oh absolutely!! To both of those things. E coli is a fantastic model for many things yeah.... genetic engineering in non models. Definitely a whammy. I've had many a professor ask me "Well can't you just do the knockouts?" in my organism and I have to be like... well... no. XD
The idea of greenhouse work in this system is fascinating! Are you actually working with plants or just to actually test these guys in as close to Martian soil as possible? How are you going to simulate the harsh light/cold conditions?
Yes, one does not simply do genetics in any old non-model microbe 🤣
For greenhouses, what we're doing at the moment is working with some simple mathematical models to figure out how to properly determine the day/night and seasonal temperature, pressure, light, and radiation profiles as a function of what exactly you mean by 'greenhouse': everything from fully enclosed + heated buildings to literally a plastic bag outside have been proposed and could work in some cases. So mostly still trying to get to a clean problem statement. Much of 2026 will probably be spent actually determining which organisms could be grown in these conditions.