Grown in Japan, complex life’s closest relatives have no nucleus and shun oxygen

The closest living relatives of every animal, plant and fungus on Earth are now growing in a laboratory in Japan, and they look nothing like the ancestor many biologists had pictured. A team at JAMSTEC, the Japan Agency for Marine-Earth Science and Technology, has grown the first pure cultures of Heimdallarchaeia, a group of single-celled archaea from which complex cells are thought to have emerged. Inside, the cells are plain: no nucleus, no internal compartments, nothing resembling the organelles that define our own cells.

The surprise is on the outside. Each cell sprouts long, branching arms that grab hold of partner microbes it cannot live without, and the cells sicken when exposed to oxygen even though they carry genes for handling it. Together, those traits sketch a portrait of our deepest single-celled ancestor as a slow, oxygen-shy microbe that lived by partnership, not a miniature version of the complex cell it would eventually become.

The finding lands on one of biology’s oldest questions: how a simple cell without a nucleus turned into the elaborate eukaryotic cell that makes up every human body. The leading answer holds that an archaeal cell formed a close partnership with a bacterium, and that the bacterium was eventually swallowed and became the mitochondrion, the power plant inside our cells. Which archaeal lineage played host, and what it was like, has been argued over for years mostly from fragments of DNA pulled out of mud and seawater.

Heimdallarchaeia sit at the centre of that argument. They belong to the phylum Promethearchaeota, better known by its former name, Asgard archaea, and genetic family trees place them as the archaeal branch nearest to eukaryotes. Until now nobody had seen a living one up close, so proposals that they already had complex internal structures, or could breathe oxygen like we do, rested on genome readings alone.

How they did it

The microbes came from an unusual place: a biofilm coating the inside of a natural gas well on the Boso Peninsula in Chiba Prefecture, east of Tokyo. Salty, methane-rich groundwater wells up there and periodically erupts like a small geyser, so the film is regularly exposed to air. The team picked the site on purpose, hoping to catch relatives of eukaryotes that had some history with oxygen.

Growing them took three years. The researchers fed the cultures tiny amounts of protein-rich food, including casamino acids, peptone and powdered milk, and used rounds of antibiotics to kill off competing bacteria, transferring the survivors to fresh medium again and again. The result was two strains, called HC1 and SC1. HC1 now lives in a stable pairing with a single methane-producing partner microbe; SC1 lives in a trio with two such partners.

These are not fast growers. HC1 takes roughly a week or more to double, and SC1 two to three weeks. A gut bacterium such as E. coli can double in about 20 minutes. The team then examined the cells with scanning and transmission electron microscopes, cryo-electron tomography that reconstructs a frozen cell in 3D, and fluorescent labelling, and sequenced both genomes, which came out at 7.91 and 6.75 million DNA letters.

Simple inside, busy outside

Across dozens of electron micrographs, neither strain showed any permanent internal membrane, the hallmark of eukaryotic cells. A few cells contained small bubble-like structures that came and went, but no nucleus, no endoplasmic reticulum and nothing a cell biologist would call an organelle. The cells divide in two like ordinary microbes.

Their surfaces told a different story. After a growth phase, most cells extended multiple arms about 370 nanometres thick that branch and bulge, and those arms often held more volume than the cell body itself, about 1.4 times as much on average. The arms were laced with filaments resembling actin, the protein that gives our own cells their shape, and they physically latched onto the partner microbes. Cells in the middle of dividing never had them.

That partnership is not optional. Both strains break down protein fragments and release hydrogen as waste, a reaction that stalls unless a partner keeps mopping the hydrogen up. When the researchers blocked the methane-producing partners with an inhibitor, the archaea stopped growing entirely.

One more trait caught the team’s attention. When HC1 got more food, it did not divide faster. Instead its cells grew bigger and sprouted more arms. Wide-area imaging of roughly 15,000 cells showed larger and more numerous arms per cell, and protein per cell rose from 66 to 226 femtograms. Typical bacteria and archaea pour extra energy into reproduction. Trading size against reproduction is, the authors note, something eukaryotes do.

The oxygen puzzle

HC1 carries genes for terminal oxidases, enzymes that in animals sit at the end of the chain that burns oxygen for energy. Some researchers had read similar genes in Heimdallarchaeia DNA as a sign that the ancestor of eukaryotes already breathed oxygen. The cultures say otherwise. Adding even trace oxygen, or a 1% oxygen atmosphere, cut HC1’s cell numbers roughly a hundredfold.

HC1 could survive oxygen only with help. When oxygen-consuming bacteria isolated from the original enrichment were added to the culture, HC1 grew despite low levels of oxygen, though it produced no more cells than before. Its genome also lacks the machinery needed to route electrons from food to oxygen. The team concludes the oxidases work as a defence, scrubbing away oxygen, much as they do in other strict anaerobes. Ferritin, a protein that limits oxygen damage involving iron, was among the three most highly expressed proteins in HC1.

A genome-wide test pointed the same way. The researchers trained a classifier on enzymes that distinguish organisms that need oxygen from those that avoid it, and it sorted known microbes correctly about 95% of the time. Applied across the whole Asgard phylum, it pointed to the group as broadly anaerobic, with or without oxidase genes.

That fits an idea known as the Ox-Tox model: the bacterium that became the mitochondrion first earned its place by soaking up oxygen that was toxic to its archaeal host, long before it became the cell’s energy supplier.

What it doesn’t settle

These microbes are cousins of the ancestor, not the ancestor itself. They have had as long to evolve as we have, some two billion years by most estimates, and any trait seen today may have been gained or lost along the way. The team’s own analysis suggests oxygen-handling genes moved sideways between related groups, so whether the ancestor carried them at all remains open.

The sample is also small. Only three strains from this branch of the tree have been examined under the microscope: HC1, SC1 and an enrichment culture reported earlier by another group. All three lack organelles, but a lineage this diverse could include members with more elaborate interiors. The occasional internal bubbles in some cells have no explanation yet, and the authors call them neither primitive organelles nor artefacts.

Lab conditions matter too. Cell shapes changed when cells were chemically fixed for imaging, and the arms tended to break off, a warning the authors themselves flag about reading anatomy from preserved samples. The cultures are pure pairings and trios with partners, not single-species cultures, because the archaea cannot grow alone.

A lineage now open to experiments

The JAMSTEC group grew the first member of the wider Asgard group, a strain called MK-D1 from deep-sea sediment, and reported it in 2020. In 2023, a team in Vienna described a second, Lokiarchaeum ossiferum, with similar arms. The new strains belong to the branch closest to eukaryotes, grow faster than MK-D1 and have larger cells, which makes them far easier to study.

The team has proposed names for both: Candidatus Margulisarchaeum peptidiphilum for HC1, a genus name carrying the surname of Lynn Margulis, the biologist who championed the theory that mitochondria began as captured bacteria, and Candidatus Flexarchaeum multiprotrusionum for SC1.

Common questions about the archaea behind complex life

What are Asgard archaea?

Asgard archaea, now formally called Promethearchaeota, are a group of single-celled microbes first identified from environmental DNA, named after Norse mythology. Their genomes carry many genes once thought unique to eukaryotes, which is why they are seen as the lineage complex cells arose from.

Did humans evolve from archaea?

In a deep sense, yes. The leading model holds that the first eukaryotic cell formed when an archaeon related to today’s Heimdallarchaeia partnered with a bacterium that became the mitochondrion. Every animal, plant and fungus descends from that merged cell.

Do Heimdallarchaeia have a nucleus?

No. The cultured strains HC1 and SC1 have no nucleus and no permanent internal membranes. Their complexity lies on the outside, in long branching arms that connect to partner microbes.

Can these archaea live with oxygen?

Barely. They are anaerobic and their numbers fall about a hundredfold with trace oxygen, but one strain survived low oxygen when oxygen-eating bacteria shared its culture.

The study was published in Nature on October 28, 2026, and was first posted as a preprint in February 2025. JAMSTEC says the two strains will now serve as experimental models, with the team turning to their cell structures and the way they bind to partner microbes to trace how an archaeal cell with arms on the outside ended up building a nucleus on the inside.

Reference: Imachi, Nobu et al., “Eukaryotes’ closest relatives are internally simple syntrophic peptidotrophic archaea,” Nature, 2026. DOI: 10.1038/s41586-026-11170-7

Tags: archaea, Asgard archaea, cell biology, eukaryotes, evolution, Heimdallarchaeia