Source
Imachi et al. (2020), Nature 577: 519–525, LIT-1591, read in full in NOTE-1238 from the PubMed Central copy of the version of record. “Isolation of an Asgard archaeon”, “Cell biology, physiology and metabolism”, “Extant and ancestral features”, “New insights into eukaryogenesis”, the Methods, the captions of Figures 1–5, and Extended Data Figures 1–7 and Tables 1–3.
What was actually shown
- Syntrophy, by experiment. Molybdate and 2-bromoethanesulfonate, which inhibit the sulfate-reducing and methanogenic partners, “significantly impaired” MK-D1’s growth. High hydrogen or formate “completely suppressed” it. Isotope labelling showed both H₂ and formate production. MK-D1 grew with Methanogenium and, separately, with Methanobacterium sp. MO-MB1.
- Slow growth. Doubling in about 14–25 days, a lag of 30–60 days, no improvement across temperatures or fourteen substrate conditions.
- Simple inside, elaborate outside. No visible organelle-like inclusions by cryo-EM (14 images) and TEM (131). Vesicles, blebs and branching protrusions 80–100 nm wide.
- Placement. Sister to eukaryotes in ribosomal-protein trees of cultured organisms. 80 eukaryotic signature proteins, 23 among the 500 most highly expressed genes.
- The ancestor, by inference. Amino-acid catabolism, NiFe hydrogenases and restricted biosynthesis are found across most Asgard genomes, and some catabolic genes are conserved across phyla. Hence an ancestor that “depended on metabolic partners, although we do not reject the possibility of other additional lifestyles”.
What could have failed: MK-D1 could have grown alone under the conditions tested, or kept growing with its partners inhibited. It did neither. It could have shown internal membranes, as earlier suggestions from genomes expected. It did not.
What this does not say
- Not that the host of eukaryogenesis was simple. MK-D1 is a modern relative. The authors set “roughly 2 billion years” between it and the lineage that became LECA.
- Not that MK-D1 depends on particular partners. Its partners are interchangeable. The dependence is on a service, H₂ and formate removal, and the inference to “a guild, not a lineage” is mine.
- Not that it is the closest relative of eukaryotes. It is the closest among cultured organisms. Metagenome-derived genomes were left out of the trees.
- Not the hydrogen hypothesis. In E3 the host makes H₂, and the authors favour an interaction other than H₂ transfer as the driver of endosymbiosis.
- Not a tested model. Entanglement, engulfment by protrusions and endogenization are “a hypothetical model” and “not definitive”.
Connections
- QUESTION-277. This gives the eukaryogenesis case a stage before any endosymbiosis: a host lineage with high D toward a guild of partners, and no R or C. E3’s endogenization step is a G route by host-side loss of a redundant function. Both readings are mine.
- THEORY-1048. The Black Queen route is loss of a costly, leaky function others supply. E3’s host-side loss of 2-oxoacid catabolism is loss of a function the symbiont supplies, inside a partnership and through a dedicated carrier. The comparison is mine.
- LIT-1566 (THEORY-1099). Kay et al. infer a host already elaborating its cytoskeleton and endomembranes before the mitochondrion. They cite this isolate as prokaryote-like in size and complexity and do not take it as contrary evidence.
- LIT-1597. The hydrogen hypothesis shares the word “syntrophy” and little else with this model.
Note of 2026-10-30: promoted, and the wider thesis filed beside it
The twenty-second status review promoted this reading under ADR-065: it reports what Imachi and colleagues’ cultivation shows and what they infer from it, and NOTE-1238 read the paper in full. Whether the Asgard ancestor depended on metabolic partners is THEORY-1200’s question, Proposed, which keeps this entry’s old promote_when.