Plasmids That Go Anywhere

[THEO] Okay, picture this. You've got a piece of DNA — a plasmid, a little loop of genetic instructions — and it works beautifully in E. coli. You want to put it in some other microbe. A fungus, a predatory bacterium, whatever. And it just... dies. Nothing happens. The DNA goes in and the cell shrugs.

[DR. MARA] Which is the story of most of microbiology, honestly. We have a handful of organisms we can engineer easily — E. coli, baker's yeast, a couple of others — and then thousands of species that do genuinely interesting things and are almost completely off-limits, because we can't get DNA to stick.

[THEO] And "stick" is the key word, right? It's not just getting the DNA inside.

[DR. MARA] Right. Two separate problems. One, physically get the DNA across the membrane — electroporation, conjugation, whatever your method. Two, once it's inside, keep it. That's where the plasmid needs an origin of replication that the host machinery actually recognizes, so it gets copied every time the cell divides. If it doesn't replicate, it's diluted away in a few generations and it's gone.

[THEO] So the dream is a plasmid that goes anywhere. Broad host range. You hand it any weird microbe and it just sets up shop.

[DR. MARA] That's the dream. And it's mostly a dream. What we actually have is decades of people fighting this fight organism by organism. But there's a real through-line if you follow it, and it starts somewhere unglamorous — an accident, in 1991.

[THEO] I love that it's an accident.

[DR. MARA] Gems and colleagues, working in Aspergillus nidulans — a filamentous fungus. Fungi were a nightmare for this. When you transform most filamentous fungi, the DNA integrates into the chromosome at random, at low frequency, and you're stuck with whatever spot it landed in. There was no good free-floating episomal plasmid — nothing that just stayed as a plasmid and replicated on its own.

[THEO] Episomal meaning it's a separate little unit, not baked into the genome.

[DR. MARA] Exactly. So they're reisolating a gene bank and they pull out this clone with a 6.1-kilobase insert. They call it AMA1. And this insert boosts transformation efficiency 250-fold.

[THEO] Two hundred and fifty! That's not a tweak, that's a different regime.

[DR. MARA] And the reason is the payoff — AMA1 lets the plasmid replicate episomally, sitting at ten to thirty copies per cell as a free-form plasmid. It wasn't integrating. It was behaving like a real plasmid in an organism that didn't have one. And it worked across the genus — in A. niger and A. oryzae too.

[THEO] So that's the first proof that "goes anywhere," at least within a family, is even possible. Find the right replicator and the door opens.

[DR. MARA] That's turning point one. The replicator is the key. But notice what the AMA1 story is about — getting the plasmid to persist. Fast forward to 2010, C. cinerea — that's an ink-cap mushroom, a model for fungal development — and Collins and colleagues run into the next wall.

[THEO] The DNA's in and staying, but...

[DR. MARA] But now: does the gene you carried actually get expressed? They build nice modular vectors — a multiple cloning site so you can drop genes in easily, and an optional intron at the 5' end, because fungi often need an intron for good expression.

[THEO] An intron being a chunk that gets spliced out of the RNA before it's translated. And for some organisms, if it's not there, the gene barely turns on.

[DR. MARA] Right. So they get DsRed and mRFP — two red fluorescent proteins — expressing for the first time. Success. But then GUS and LUC, two standard reporter enzymes, fail. Even with the intron.

[THEO] Huh. So it's not a splicing problem.

[DR. MARA] That's the important negative result. Whatever's blocking GUS and LUC is downstream of splicing — some post-splicing barrier they couldn't pin down. And the lesson stacks on top of AMA1: getting the plasmid to persist is necessary but not sufficient. Expression is its own separate fight, and it's organism- and even gene-specific. There's no universal "it'll just work."

[THEO] So by 2010 the field knows it's fighting on at least three fronts — get it in, keep it, express it. Where does the story go from there?

[DR. MARA] It broadens out of fungi and gets systematic. And 2024 is when a whole cluster of papers land, each attacking a different chassis with the same philosophy: build a real toolkit, not a one-off. Take Haemophilus influenzae — historically huge, it's where the first bacterial genome was sequenced, and yet chronically underexplored as an engineering chassis.

[THEO] Why so hard?

[DR. MARA] Restriction-modification. The cell's immune system against foreign DNA. Restriction endonucleases chop up incoming DNA that isn't marked as "self." So your plasmid gets shredded before it can do anything.

[THEO] It's the bouncer at the door checking IDs, and your DNA doesn't have the right stamp.

[DR. MARA] Good analogy. So the move — and this is the 2024 Haemophilus work — is delete the bouncers. They knock out the two native type II restriction enzymes, HindIIR and HindIIIR, and electroporation efficiency measurably improves. Then they build pSU20-derived multi-host plasmids with selection markers and fluorescent markers. First real synthetic biology toolkit for that organism.

[DR. MARA] And notice how that connects back — AMA1 was about replication, this is about survival against the host's defenses. Different front, same war.

[THEO] And there were others that year?

[DR. MARA] Behrendt and colleagues in Zymomonas mobilis — an ethanol-producing bacterium, big for biofuels. They build seven shuttle vectors, each using a different native replicon, and they span a hundred-fold expression range. That's the toolkit mindset — give people a dial, not a single setting.

[THEO] A hundred-fold from swapping the origin? That's a lot of range just from where it replicates.

[DR. MARA] The replicon influences copy number, which influences dosage. And one of theirs, pZP605, was the only plasmid stable without antibiotic selection across fifty-plus generations — which matters enormously if you want something to stay put in a bioreactor without constantly dosing drugs.

[THEO] Then there's the predator one. Bdellovibrio.

[DR. MARA] Bdellovibrio bacteriovorus — a bacterium that hunts and eats other bacteria. Salgado and colleagues. The problem there is bizarre: it doesn't divide by simple binary fission, so plasmids partition unevenly into daughter cells. Heterogeneous. Some cells get lots, some get none.

[THEO] So the plasmid approach itself is shaky.

[DR. MARA] So they do the first Tn7-mediated chromosomal integration in that organism — Tn7 is a transposon system that drops your cargo into a defined spot in the genome. Put it in the chromosome and the partitioning problem disappears. And they found P_J/EliR, an inducible promoter with a 350-fold dynamic range, the only tightly controlled one of five they tested.

[THEO] It's funny — the arc started with getting off the chromosome, going episomal, and here they go back onto it deliberately.

[DR. MARA] Because the goal was never episomal for its own sake. It was control. Whatever gives you reliable, tunable expression in that organism wins.

[THEO] And the last one flips the whole thing sideways, doesn't it? The barcoding paper.

[DR. MARA] Coaux-Seq. Instead of engineering one organism, they take three-kilobase genomic fragments from eleven different bacteria, barcode them, express them in E. coli, and screen across twenty auxotroph knockouts. Cheap BarSeq readout. They experimentally validated 53 protein functions — including a TauE-family protein working as a sulfate importer, when that family was only known to export.

[THEO] So E. coli becomes the universal test bench for everyone else's genes.

[DR. MARA] That's where it's heading. The 1991 fight was one plasmid, one organism. Now it's parallel, barcoded, across whole panels. Same dream — go anywhere — just finally at scale.

[THEO] From an accident in a mushroom cousin to reading thousands of genes at once. That's a good place to hand off. More after this.