The Cell Within the Cell

[SOFIA] Okay, so picture the most extreme version of "it takes a village." You are a cell — a single eukaryotic cell — and living inside you, past your membrane, is another whole organism. Not a virus hijacking you. A guest. Sometimes a paying guest, sometimes a squatter, and sometimes... you're not sure which.

[DANIEL] That's the thread today, isn't it? The cell within the cell. Endosymbiosis.

[SOFIA] It's the story of how life got complicated in the first place! Mitochondria, chloroplasts — those were free-living bacteria that got engulfed a billion-plus years ago and never left. That's not folklore, that's the origin of the eukaryotic cell. So when we find organisms doing this today, in real time, we're basically watching reruns of the biggest merger in the history of life.

[DANIEL] Hm. And I'd flag for anyone listening from outside the field — endosymbiosis is a spectrum, not a category. On one end you've got obligate, fully domesticated partners like mitochondria that can't live alone anymore. On the other end you've got facultative lodgers that come and go. Most of the interesting biology is in the messy middle.

[SOFIA] Which is exactly where these papers live. And the reason this matters beyond "cool trivia" — if you're an engineer like me, an intracellular partner is a delivery system. It's a chassis that already knows how to get inside a host cell and survive. If you understand the rules, you can borrow them.

[DANIEL] So let's define the toolkit before we get to results, because a couple of these methods do the heavy lifting. Dual RNA-seq — you sequence the messenger RNA from both partners at once, host and symbiont, from the same sample. You get two transcriptomes, and crucially, you see what each one turns on or off in the other's presence.

[SOFIA] And then transposon sequencing — Tn-seq. You blast a bacterial genome with a jumping element that inserts randomly and knocks out whatever gene it lands in. Make a library of thousands of mutants, throw them all at a condition, and sequence to see who drops out. The genes you can't afford to lose are the ones that matter.

[DANIEL] The barcoded version, RB-TnSeq, tags each insertion with a little DNA barcode so you can just count barcodes instead of re-sequencing insertion sites every time. Cheaper, so you can run many conditions. Keep that in mind — it becomes the workhorse.

[SOFIA] Alright. Roots. 2017, Burns and colleagues, and honestly this is the paper that hooked me. There's a spotted salamander whose embryos have green algae — Oophila — living literally inside the cells. It's the only known mutualistic endosymbiosis in a vertebrate. A vertebrate! We're supposed to have immune systems that don't tolerate this.

[DANIEL] And they did dual RNA-seq on it. Two findings that surprised me. First, the intracellular algae stop photosynthesizing — they shift toward fermentation. The alga that's supposed to be paying rent in oxygen and sugar goes quiet and starts fermenting instead.

[SOFIA] Which is wild, right? The whole reason you'd want an alga inside you is the photosynthesis. And once it's inside, it abandons the day job.

[DANIEL] Second piece — the host side. The salamander cells attenuate NF-κB signaling. That's the master inflammatory pathway. They flagged OLFM4, MUC1, TNIP1 as the dampeners. So the host is actively quieting its own immune alarm to keep the guest.

[SOFIA] And that's the turning point conceptually. It reframes endosymbiosis as a negotiation. The host suppresses immunity, the symbiont changes its metabolism. Both sides remodel. That's the template every later paper is really testing in a different system.

[DANIEL] Now, one caution — this is a transcriptome. It's correlation, expression changes. It tells you what's associated with the symbiotic state, not what's causal. To get causal you need to break genes. Which is where the field goes next.

[SOFIA] 2020, Malassezia. This is the yeast all over your skin — the one behind dandruff and some eczema. And the question was: where'd it get the trick to survive there? Turns out it acquired a flavohemoglobin gene, YHB1, by horizontal gene transfer — twice, independently, from two different skin bacteria, Brevibacterium and Kocuria.

[DANIEL] Flavohemoglobin detoxifies nitric oxide under aerobic conditions. NO is one of the weapons your skin uses against microbes. So the fungus stole a bacterial defense gene to neutralize it.

[SOFIA] And I love this because it's the same negotiation, different currency. Instead of remodeling metabolism, the microbe grabs a gene off the shelf — off another microbe's shelf — to survive the host environment.

[DANIEL] But credit to Ianiri and colleagues for restraint here. They tested whether YHB1 drives skin pathogenesis and it doesn't. It suppresses an allergen, it detoxifies NO, but it's not the disease factor. That's the kind of negative result I trust the whole paper more for.

[SOFIA] Then 2021 gets really intimate. Midichloria — a bacterium in ticks that lives inside the mitochondria. Not inside the cell. Inside the organelle inside the cell.

[DANIEL] The endosymbiont's endosymbiont. And there'd been a model that it behaves like Bdellovibrio — a predatory bacterium that invades, replicates, and lyses its prey. Bursts the mitochondrion.

[SOFIA] And this paper says: no. They did quantitative TEM — electron microscopy, counting — plus network simulation, and the colonization pattern in tick oocytes follows a power law that the lyse-and-burst model just can't produce.

[DANIEL] Right, this is a direct contradiction, and it's the good kind. They pit two mechanistic models against a distribution and one of them fails to fit. Their answer is that Midichloria travels through the mitochondrial network — mitochondria fuse and connect — and spreads organelle to organelle without killing them.

[SOFIA] So it's a commuter, not a predator. Uses the mitochondrial highway. And you only see that because someone insisted the mechanism has to reproduce the actual numbers.

[DANIEL] That's the through-line tightening — we've gone from "what genes change" to "which mechanistic model survives the quantitative test."

[SOFIA] And then 2023 brings the causal firepower. Akkermansia muciniphila — a gut bacterium everyone's excited about for metabolic health. They ran a transposon screen and pinned down the MUL system: pili plus a periplasmic complex that actively imports mucin — your gut mucus glycoprotein — into intracellular compartments they call mucinosomes.

[DANIEL] And the screen shows MUL is required to compete against other gut microbes for that mucin, and it's linked to repressing the host's own colonic sterol biosynthesis genes. So now we have the causal handle Burns couldn't get in 2017 — knock out the gene, watch the interaction break.

[SOFIA] The companion review that same year, Voogdt and colleagues, is basically the field saying: RB-TnSeq is now the standard way we interrogate gut commensals — chemical genomics, host interaction, in vivo fitness. The method matured into infrastructure.

[DANIEL] And the newest one zooms out spatially. SmT — a multimodal Visium array reading 16S, 18S, ITS, and host polyA at the same time, at 55-micron resolution, in the Arabidopsis leaf surface.

[SOFIA] Bacteria, fungi, and plant, mapped together in tissue. And the punchline is quantitative — how strongly two kingdoms interact is predicted by how much they share those microbial hotspots. Physical co-location predicts interaction strength.

[DANIEL] So the arc: transcriptomes told us the partners talk, HGT and Tn-seq told us which genes do the talking, the Midichloria work told us the mechanism has to fit the numbers, and now we're putting all of it back in space.

[SOFIA] From "there's a cell inside a cell" to "here are the genes, the mechanism, and the map." And for someone who wants to engineer a delivery chassis — that map is where I'd start looking next. That's where we'll leave it.