Cellular Co-habitation And Ancient Origins

[SOFIA] Okay, so here's a question that sounds like a joke but isn't: what happens when one cell decides to live inside another one? Not eat it, not infect-and-kill it — move in. Take up residence. Because that single trick — a cell within a cell — is arguably the most important thing that ever happened in the history of life on Earth.

[DANIEL] Right. Every eukaryote listening to this — every plant, animal, fungus — is carrying the descendants of a bacterium that moved in and stayed. We call them mitochondria now.

[SOFIA] The power plant in your cells was once free-living. And the chloroplast in a leaf, same story — a cyanobacterium that got engulfed and never left. Endosymbiosis. That's the word. One organism living stably inside the cells of another.

[DANIEL] And it's worth being precise about why that's hard, because it's not obvious a cell should tolerate a tenant. Your cells have machinery whose whole job is to detect and destroy things inside them. So for an endosymbiont to persist, something has to give on both sides — the host stops killing it, the symbiont stops being lethal, and over evolutionary time you get this negotiated truce.

[SOFIA] And the fun part for me, as someone who thinks about moving DNA and proteins between cells for a living — endosymbiosis is the ultimate version of that problem. How do you get material across the boundary between two organisms? How does a symbiont talk to its host? That's a delivery problem, and it's a communication problem.

[DANIEL] So the arc we want to trace is how a handful of groups, coming at this from totally different angles, have been filling in the mechanics. How do things get in, how do they persist, and how far can you push it.

[SOFIA] Let's start at the root — 2017, the Bartonella work. And this one's lovely because it starts with something that looks like nothing to do with symbiosis. Toxin-antitoxin modules.

[DANIEL] Which are everywhere in bacteria. A little genetic cassette — a toxin that would kill the cell, and an antitoxin that keeps it in check. Classic selfish-element stuff.

[SOFIA] But the Bartonella people found that one of these, a Fic-domain toxin called VbhT, isn't just poisoning its own cell. It's getting shipped into other bacteria through a type IV secretion system — a T4SS, the molecular syringe bacteria use for conjugation.

[DANIEL] And the clever bit is how they proved the protein actually crossed over. They built a sensor — Cre fused to the relaxase in the donor cell, and a loxP switch in the recipient. If the fusion protein makes it across, Cre flips the switch, and the recipient lights up with resistance. The reported detection for the Cre-relaxase fusion was essentially every donor cell — around a hundred percent — versus a fraction of a percent for the plasmid itself transferring.

[SOFIA] So you can distinguish "the protein went across" from "the whole plasmid went across." That sensitivity is the whole game.

[DANIEL] And conceptually it reframes the toxin. VbhT is an evolutionary intermediate — it started as an inter-bacterial weapon and it's on its way to becoming a host-targeted effector. You can watch a weapon becoming a diplomatic tool.

[SOFIA] Which is exactly the muscle an endosymbiont needs. Okay — jump to 2021, and two papers that could not be more different in style. First, the microsporidia one, and this is a genomics headache turned into a method.

[DANIEL] This is the problem of the unpurifiable symbiont. These feminizing microsporidia — Nosema, two Dictyocoela species — live inside their crustacean host, Gammarus, and they're strictly vertically transmitted. Mother to offspring, never out into the environment. They don't make a spore you can go collect.

[SOFIA] So you can't separate them from the host to sequence them. You're stuck.

[DANIEL] Their answer was to stop fighting it. Sequence everything together — host plus symbiont DNA in one pool — and deconvolve the assemblies computationally afterward. They pulled out three microsporidia genomes plus a draft of the Gammarus host.

[SOFIA] And then the payoff — they compared feminization in these microsporidia to feminization in Wolbachia, which is a bacterium.

[DANIEL] And concluded it evolved independently in the eukaryote and the bacterium. Convergent manipulation of host sex. The candidate genes they flagged had DNA-binding and membrane-fusion functions — which, membrane fusion, there's your crossing-the-boundary theme again.

[SOFIA] The same year, Midichloria — and this one I just love the name. Midichloria mitochondrii. It lives inside the mitochondria of ticks. A bacterium inside the organelle that was itself once a bacterium.

[DANIEL] A cell within a cell within a cell, effectively. And the question was mechanistic: how does it move through the mitochondrial population? There was a model borrowed from Bdellovibrio — a predatory bacterium that invades prey, replicates, and lyses it on the way out.

[SOFIA] Burst-and-kill. And if Midichloria did that, you'd expect a certain signature in how it spreads through the mitochondria in a tick's eggs.

[DANIEL] This is the paper I respect, because they didn't just tell a story — they did quantitative TEM, counted the colonization, saw a power-law distribution, and then built a network simulation to ask which model actually produces that pattern. And the Bdellovibrio-like model got rejected. Midichloria traverses the mitochondrial network, moving organelle to organelle, without lysing them.

[SOFIA] So it's a commuter, not a wrecking crew. And that directly contradicts the predatory model people had floated.

[DANIEL] It's a clean example of using simulation to falsify a hypothesis instead of just gesturing at plausibility. The data pick a winner.

[SOFIA] Now 2023 shifts gears — this is where the tools get industrial. Akkermansia muciniphila, the gut bug everybody's excited about metabolically, and a transposon screen.

[DANIEL] Transposon insertion sequencing — you knock out genes across the whole genome by inserting transposons randomly, then sequence to see which mutants survive a given condition. It's how you go from "this organism has these genes" to "these genes actually matter for this job."

[SOFIA] And the job here is eating mucin — the gunk lining your gut. They found this MUL system, a pili-and-periplasmic complex, actively importing mucin into intracellular compartments they call mucinosomes.

[DANIEL] And MUL isn't just for feeding — the mutants lose out competing against other gut microbes, and the system reaches out and represses the host's own colonic sterol biosynthesis genes. So the bug is reshaping host metabolism.

[SOFIA] Which is the endosymbiosis theme scaled down to the gut lumen — a microbe talking back to host gene expression. And the review that comes alongside it, the barcoded transposon work in gut commensals — that's the methods backbone.

[DANIEL] RB-TnSeq. You tag each transposon mutant with a DNA barcode, so instead of resequencing insertion sites every time, you just count barcodes. It makes chemical genomics and in-vivo fitness screens cheap enough to run at scale across non-model gut bugs. That's the enabling layer under all of this.

[SOFIA] And then 2024 — okay, this is the good stuff. Somebody just built an endosymbiosis. From scratch, in a dish.

[DANIEL] The Cyanidioschyzon chloroplast paper. They took chloroplasts from a thermophilic red alga — a compact 243-gene chloroplast genome — and put them inside CHO-K1 cells. Hamster ovary cells. Mammalian.

[SOFIA] And the chloroplasts kept doing photosystem-II electron transport for at least 48 hours inside an animal cell. Actual photosynthetic activity, running, inside a cell that has never in its evolutionary history photosynthesized.

[DANIEL] Forty-eight hours, and it's electron transport they measured — I'd want to see how long before it degrades, and whether it's doing anything useful for the host cell. But as a proof of principle it's remarkable. The first confirmed photosynthesis in an animal cell.

[SOFIA] So look at the whole arc. 2017, how proteins cross between cells. '21, how to even read a symbiont you can't isolate, and how one moves without killing. '23, the tools to dissect what symbiont genes actually do. And '24, we start engineering the thing nature spent two billion years on.

[DANIEL] From watching endosymbiosis to building a first draft of it. That's the turning point.

[SOFIA] A cell within a cell, on purpose. We'll be watching that one. More after the break.