Cell Within Cell Delivery Systems

[SOFIA] Okay, so today we're doing a whole story, not a single paper — and the story is about the cell inside the cell. Life living inside other life. Bacteria inside bacteria, bacteria inside mitochondria, symbionts you literally cannot pull out of their host no matter how hard you try.

[DANIEL] Which is a bigger category than people think. Most of the biology on this planet isn't free-living. It's tucked inside something else.

[SOFIA] Right! And here's why I love this as an engineer — every one of these relationships is a delivery problem that evolution already solved. Something is moving DNA, or protein, or a whole organism, across a membrane into another living thing. That's the thing I spent my postdoc trying to do to non-model bacteria, badly, and nature's out here doing it at a hundred percent efficiency.

[DANIEL] Let's define the field for someone coming in cold, though. When we say endosymbiont, we mean an organism that lives inside the cells of another organism. Sometimes it's a loose partnership, sometimes it's so tight the symbiont can't survive on its own anymore — it's shed genes, it's dependent, it's basically becoming an organelle.

[SOFIA] Which is the origin story of the mitochondria and the chloroplast, by the way. Those were free-living bacteria a couple billion years ago. So this isn't a weird edge case — this is how complex cells got built in the first place.

[DANIEL] And the technical problem that runs through all of it: how do you study something you can't isolate? Classical microbiology is "grow it in a dish, purify it, sequence it." A strict endosymbiont breaks every step of that. It won't grow alone. You can't separate it from the host. So the field had to invent new ways to even see what's going on.

[SOFIA] That's the through-line. Every paper in this arc is somebody building a tool to look at a relationship they couldn't look at before. And the story starts, weirdly, with a fight between bacteria.

[DANIEL] 2017. Bartonella. This is the toxin-antitoxin paper.

[SOFIA] So Bartonella is a genus of bacteria — some cause disease in mammals — and this group was chasing where its host-targeting weapons came from. They built a sensor called CRAfT. And the design is genuinely clever, Daniel, you'll like this. You fuse the protein you think is getting exported to Cre recombinase. That's the donor cell.

[DANIEL] And Cre is an enzyme that flips or excises DNA at specific loxP sites. So it's a reporter that only does something if the protein physically arrives in the other cell.

[SOFIA] Exactly. The recipient bacterium has a loxP-flanked resistance switch. If your Cre-fusion protein actually gets translocated in, Cre flips the switch, and the recipient survives selection. So you're reading protein delivery as antibiotic resistance.

[DANIEL] And the numbers here are what sold me. A Cre-relaxase fusion transferred at roughly a hundred percent per donor cell. Versus transfer of the plasmid itself at 0.1 to 1 percent. That's a two-to-three log gap.

[SOFIA] Which tells you the protein and the DNA are moving through the same machine — a conjugative type IV secretion system, the T4SS — but the protein's the main event, the DNA's kind of coming along for the ride.

[DANIEL] And the evolutionary punchline: the toxin they tracked, VbhT, is a FicT-family toxin that started life as half of a toxin-antitoxin module — a little bacterial suicide-pact system — and it got repurposed into an effector that one bacterium shoots into another. An intermediate, caught in the act, between an internal module and a host-targeting weapon.

[SOFIA] So the roots of the whole "cell within a cell" story start with, how does one cell reach into another at all. And here's a machine that does it.

[DANIEL] Turning point number two is really a tooling turning point. 2021, the co-sequencing paper.

[SOFIA] This is the one that made me go, oh, that's how you beat the isolation problem. So the subjects are feminizing microsporidia — microsporidia are these obligate intracellular parasites, weird fungi-adjacent things — and these particular ones live inside a crustacean, Gammarus, and they're strictly vertically transmitted. Passed mother to offspring, only.

[DANIEL] Which means, critically, they never make a purifiable environmental spore. There's no life stage you can catch and clean up. You genuinely cannot get symbiont DNA away from host DNA.

[SOFIA] So they just... didn't. They sequenced everything together — host plus symbiont, one messy pile of DNA — and then built a computational strategy to pull the assemblies apart afterward.

[DANIEL] And I want to give them credit, because "deconvolve two genomes from a co-sequenced sample" is the kind of claim I'd normally poke at hard. But they resolved three separate microsporidia species — Nosema granulosis and two Dictyocoela — plus a draft of the Gammarus host. That's not one lucky separation, that's four assemblies out of one soup.

[SOFIA] And then the payoff — feminization. These symbionts convert genetic males into functional females, because a symbiont that's only passed through the mother wants more daughters.

[DANIEL] And they compared to Wolbachia, a bacterium that does the same feminizing trick, and showed it evolved independently — once in a eukaryote, once in a bacterium. Convergent manipulation. Different candidate genes, DNA-binding and membrane-fusion functions.

[SOFIA] Same year, there's this gorgeous mechanistic paper I have to bring up — Midichloria.

[DANIEL] Midichloria mitochondrii. Best-named organism in the arc. It lives inside the mitochondria of ticks. Inside the organelle.

[SOFIA] A bacterium inside the powerhouse of the cell! And the open question was, how does it get around? One model — the Bdellovibrio-like model — says it invades a mitochondrion, chews it up from the inside, bursts out, moves to the next. Predatory.

[DANIEL] And they rejected that. Quantitative transmission electron microscopy plus network simulation — they showed Midichloria travels through the mitochondrial network, mito-to-mito, without lysing the organelles. And that movement pattern reproduces the power-law colonization they see in tick oocytes.

[SOFIA] So it's a commuter, not a wrecking crew. And notice what's happening across these — 2017 it's a genetic reporter, 2021 it's computational deconvolution and quantitative imaging. The tools keep getting better at seeing the unseeable.

[DANIEL] And then the arc moves into the gut and out onto plants, where the tooling really matures.

[SOFIA] Akkermansia muciniphila. Gut bug everybody's excited about for metabolic health. And a transposon screen — you knock out genes across the whole genome, sequence which mutants survive — found this MUL system, a pili-periplasmic complex that actively imports mucin. It pulls host mucin into intracellular compartments they call mucinosomes.

[DANIEL] And the fitness angle: knock out MUL and Akkermansia can't compete against other gut microbes for mucin, and it loses its ability to repress the host's colonic sterol biosynthesis genes. So the uptake machinery is doing double duty — competition and host signaling.

[SOFIA] And that transposon-screen approach is exactly what the RB-TnSeq review in the same batch is about — barcoded transposon libraries, letting you run fitness screens on gut commensals in a live animal.

[DANIEL] Randomly barcoded transposon sequencing. Each insertion carries a unique barcode, so you can track thousands of mutants at once, cheaply, across conditions and through a host gut.

[SOFIA] And then the tool I think is the most fun — SmT, spatial metatranscriptomics. In the Arabidopsis phyllosphere, the leaf surface. They put down Visium-style arrays with probes for bacterial 16S, fungal 18S and ITS, and host polyA — all at once, at 55-micron resolution.

[DANIEL] So you're not just asking who's there, you're asking who's next to whom. And the result that survives scrutiny: interkingdom interaction strength — bacteria with fungi — is quantitatively predicted by how much they share these microbial hotspots.

[SOFIA] Which closes the loop! We started with one cell reaching into another cell, and we've arrived at mapping entire host-bacteria-fungi neighborhoods in space.

[DANIEL] The through-line is the isolation problem giving way, tool by tool. Reporter, then deconvolution, then quantitative imaging, then barcoded screens, then spatial multi-kingdom maps.

[SOFIA] And where it's heading — I want these tools pointed back at the hard endosymbionts. Spatial transcriptomics inside a tick oocyte, barcoded screens on things we can't culture. That's the next chapter. We'll leave it there for today.