Engineering Uncooperative Cells

[SOFIA] Okay, so here's a question that sounds trivial and absolutely is not: how do you get a piece of DNA into a cell that doesn't want it? And I mean really doesn't want it. For the handful of organisms we love — E. coli, yeast, a few others — this is a solved problem, right? You order competent cells, you heat-shock, you're done by lunch.

[DANIEL] For those organisms. Which is maybe a few dozen out of millions.

[SOFIA] Right! And that's the whole story today. Everything cool in biology right now — the bacteria that eat lignocellulose, the ones doing photosynthesis in weird ways, fungi that make the enzymes in your laundry detergent — most of those are what the field politely calls "recalcitrant." You cannot easily transform them. So the tools we have for engineering biology only work on the organisms that happen to be easy, not the ones that are interesting.

[DANIEL] And I want to be careful with a word there, because people blur it constantly. Transformation — for bacteria and archaea and fungi, that's getting naked DNA in. Not transfection. Transfection is for eukaryotic cells, mammalian cell culture. If you're talking about a bacterium, it's transformation, conjugation, or electroporation.

[SOFIA] Thank you, yes. And those three are the toolkit. Conjugation is cell-to-cell transfer through a bridge the bacteria build themselves. Electroporation is the brute-force one — you hit cells with a voltage pulse, punch temporary holes in the membrane, DNA slips through. And natural transformation is when the organism actually has machinery to pull DNA in from its environment on purpose.

[DANIEL] Which some do and most don't. And when they do, there are barriers waiting inside. Restriction-modification systems — the cell's immune memory that chops up foreign DNA it doesn't recognize. You can get the plasmid through the membrane and still have it shredded a second later.

[SOFIA] So that's the problem space. Let's trace how people chipped at it, because the arc here is really a shift in how the field even thinks about the question. Start in 1994 — this A. niger paper. Aspergillus niger, the black mold, industrial workhorse for enzymes and citric acid.

[DANIEL] And the standard move for filamentous fungi back then was protoplasting. You strip the cell wall off with lytic enzymes, because the wall is the barrier, and then you've got a fragile naked cell you can coax DNA into. It's finicky and it's slow.

[SOFIA] And this paper just said — what if we don't? They took intact germinating conidia, the spores as they're waking up, and electroporated them directly. No protoplasting. Got 100 colonies per microgram of plasmid.

[DANIEL] Which is modest. A hundred per microgram is not a lot.

[SOFIA] It's not a lot! But the point wasn't the number, it was that you could skip the worst step. And a little lytic enzyme pretreatment — not full protoplasting, just a touch — doubled it. So the idea planted here is: meet the organism where it is. Catch it at a vulnerable moment, the germinating spore, instead of tearing it apart.

[DANIEL] A developmental window as a delivery window. That's a real idea.

[SOFIA] Jump to 2005, Choi and colleagues, Pseudomonas aeruginosa. And this one I love because it's so practical. The whole bottleneck for a lot of species was making the cells electrocompetent — the prep, washing out all the salts so your voltage pulse doesn't just arc.

[DANIEL] Salts conduct. If the buffer's conductive the pulse goes through the solution instead of the cells, and sometimes it literally arcs and you lose the sample.

[SOFIA] So they built a ten-minute microcentrifuge prep. Fast, few washes. And it gave a ten-thousand-fold improvement in efficiency over what they'd had.

[DANIEL] Hm. Now that's the number that earns attention. Ten-thousand-fold. And the consequence they claim is bigger than convenience — it meant they could do chromosome-level manipulation by electroporation and retire the older, clumsier routes, transduction and conjugation, for that organism.

[SOFIA] Which is the turning point in miniature: a better prep didn't just save time, it changed which experiments were possible.

[DANIEL] For one species. And here's where the field gets honest with itself. 2009, Aune and Aachmann, a comparative review of artificial transformation methods across recalcitrant bacteria. And the headline finding is almost deflating: no single method wins.

[SOFIA] Right, there's no silver bullet.

[DANIEL] You have to match the mechanism of the method against the specific barrier of your organism. Thick cell wall? One approach. Aggressive restriction-modification system? You need to bypass it, maybe methylate your DNA in advance to disguise it. The review's real contribution is reframing the question — stop looking for the universal protocol, start diagnosing the barrier.

[SOFIA] And then 2012 throws a genuinely strange wrench in. The mitochondrial transformation paper. This one's a little outside the bacterial story but it belongs here, because it asks: can an organelle import DNA?

[DANIEL] And isolated mitochondria — both plant and mammalian — actively take up double-stranded DNA through VDAC, the voltage-dependent anion channel. There's a door in the outer membrane and DNA goes through it.

[SOFIA] A channel that normally handles metabolites, moving DNA. And the plant and mammalian cases differ — plant mitochondria need more, an additional transporter, ANT, and these terminal inverted repeat sequences on the DNA. Mammalian don't.

[DANIEL] Which tells you the import isn't one conserved thing, it's at least two different mechanisms. And philosophically it connects to the next piece — natural uptake machinery.

[SOFIA] Which is 2014, the big natural transformation review. About eighty bacterial species documented as naturally competent.

[DANIEL] And the structure of the finding is the interesting part. The uptake proteins — the machinery that physically pulls DNA across — are conserved across those eighty species. But the regulatory logic, when the cell decides to turn competence on, is all over the map.

[SOFIA] Conserved hardware, totally different software.

[DANIEL] And the review weighs in on the old "why does this exist" debate. Is natural transformation about eating DNA for nutrients, or about getting new genes and repairing your chromosome? And the evidence they marshal favors genetic diversity and DNA repair over nutrition.

[SOFIA] Which matters for engineers, because if an organism already has import machinery, you might coax it rather than fight it. Okay — and this is the good stuff — the 2018 piece pulls the whole arc together. The non-model prokaryote toolbox review out of NREL and CU Boulder.

[DANIEL] Phototrophs, autotrophs, cellulose-degraders. The organisms the bioeconomy actually wants — things that fix carbon or break down plant waste — and they're the ones we can't engineer.

[SOFIA] And the framing has completely flipped from 1994. Back then it was one organism, one trick. By 2018 the argument is: we need design rules. A systematic way to build a genetic toolbox in a new organism — transformation method, then CRISPR, then dealing with the regulatory barriers inside.

[DANIEL] And they're explicit that those rules are poorly synthesized. Everyone's solving it one bug at a time. The paper's whole move is to say the chassis itself should be a design variable — you choose and build your organism's toolkit deliberately, not inherit E. coli because it's convenient.

[SOFIA] So the through-line, start to finish: catch a spore at the right moment, then speed up a prep, then admit there's no universal method, then realize organisms have their own doors you can use — and finally, stop treating each organism as a one-off and build the rules.

[DANIEL] And where it's heading is combinatorial and high-throughput. Screening delivery conditions across dozens of species at once instead of one PhD, one bug, five years.

[SOFIA] Which is exactly the kind of thing we'll get into another day. For now — the unengineerable are slowly becoming engineerable, and it's because people stopped looking for one answer. Daniel, thank you.

[DANIEL] Always.