close
Skip to main page content
U.S. flag

An official website of the United States government

Dot gov

The .gov means it’s official.
Federal government websites often end in .gov or .mil. Before sharing sensitive information, make sure you’re on a federal government site.

Https

The site is secure.
The https:// ensures that you are connecting to the official website and that any information you provide is encrypted and transmitted securely.

Access keys NCBI Homepage MyNCBI Homepage Main Content Main Navigation
. 2025 Jun;642(8069):990-998.
doi: 10.1038/s41586-025-08955-7. Epub 2025 May 7.

Deep origin of eukaryotes outside Heimdallarchaeia within Asgardarchaeota

Jiawei Zhang  1 , Xiaoyuan Feng  2 , Meng Li  2 , Yang Liu  2 , Min Liu  3 , Li-Jun Hou  4 , Hong-Po Dong  5
Affiliations

Deep origin of eukaryotes outside Heimdallarchaeia within Asgardarchaeota

Jiawei Zhang et al. Nature. 2025 Jun.

Abstract

Research on the morphology, physiology and genomics of Asgard archaea has provided valuable insights into the evolutionary history of eukaryotes1-3. A previous study suggested that eukaryotes are nested within Heimdallarchaeia4, but their exact phylogenetic placement within Asgard archaea remains controversial4,5. This debate complicates understanding of the metabolic features and timescales of early eukaryotic ancestors. Here we generated 223 metagenome-assembled nearly complete genomes of Asgard archaea that have not previously been documented. We identify 16 new lineages at the genus level or higher, which substantially expands the known phylogenetic diversity of Asgard archaea. Through sophisticated phylogenomic analysis of this expanded genomic dataset involving several marker sets we infer that eukaryotes evolved before the diversification of all sampled Heimdallarchaeia, rather than branching with Hodarchaeales within the Heimdallarchaeia. This difference in the placement of eukaryotes is probably caused by the previously underappreciated chimeric nature of Njordarchaeales genomes, which we find are composed of sequences of both Asgard and TACK archaea (Asgard's sister phylum). Using ancestral reconstruction and molecular dating, we infer that the last Asgard archaea and eukaryote common ancestor emerged before the Great Oxidation Event and was probably an anaerobic H2-dependent acetogen. Our findings support the hydrogen hypothesis of eukaryogenesis, which posits that eukaryotes arose from the fusion of a H2-consuming archaeal host and a H2-producing protomitochondrion.

PubMed Disclaimer

Conflict of interest statement

Competing interests: The authors declare no competing interests.

Figures

Fig. 1
Fig. 1. Phylogenomic analysis of 53 concatenated archaeal markers (GTDB.ar53) in GTDB r207.
ML tree inferred using IQ-TREE under the LG + C60 + F + G + PMSF model, based on a set of 579 archaeal taxa (411 Asgard archaea, 51 DPANN archaea, 47 Euryarchaea and 70 representatives of TACK archaea). Only Asgard archaeal lineages are shown in the tree. Two cultured Asgard archaeal strains are highlighted with a red star. Newly identified genomes in this study are displayed by coloured bars in the outermost ring. Bootstrap support values of at least 95% are represented by black dots. The scale bar denotes the average expected number of substitutions per site.
Fig. 2
Fig. 2. Taxonomic profiles and clustering of contigs/scaffolds in Njordarchaeales genomes.
a,b, Percentage of contigs/scaffolds assigned to Thermoproteota (formerly known as the TACK superphylum) or Asgardarchaeota based on count for each of the ten Njordarchaeales representatives; classification of contigs/scaffolds was determined using CAT (a) or MMseqs2 (b) tools. c,d, Hierarchical clustering of contigs/scaffolds in Njordarchaeales B7_G17 and B20_G9 (c) and B62_G16 and S143_49 (d) genomes based on their sequence composition and differential mean coverage across different metagenomes.
Fig. 3
Fig. 3. Phylogenomic analyses of several sets of concatenated marker proteins, showing the placement of eukaryotes relative to genomically sampled Asgard archaea.
a, ML phylogenetic analysis of 67 concatenated marker proteins based on 461 archaeal taxa and 14 eukaryotes (inferred using IQ- IQ-TREE under LG + C60 + F + G + PMSF model), using 50 TACK archaea as the outgroup (tES67 alignment; 13,348 sites). The number below Lokiarchaeales represents 150 MAGs and the genomes of two cultured Asgard strains. Meanwhile, the alignment was SR4-recoded and its Bayesian inference was performed using the CAT + GTR model (Supplementary Fig. 6a; two chains; 50,000 generations). PP support for the node of eukaryotes and the closest Asgard relatives is shown (0.69). b, ML phylogenetic analysis of 97 concatenated marker proteins based on 411 Asgard archaeal taxa and 14 eukaryotes (inferred using IQ- IQ-TREE under the LG + C60 + F + G + PMSF model) (S97 alignment; 20,067 sites). This tree is unrooted. In addition, Bayesian inferences of the SR4-recoded S97, S150 and NM57 supermatrices were performed using the CAT + GTR model (Supplementary Figs. 7 and 8b). PP support for the node of eukaryotes and the closest Asgard relatives is shown (PP = 1 for S97, PP = 0.8 for S150 and PP = 1 for NM57). c, Evolution of bootstrap support for the grouping of eukaryotes with Heimdallarchaeia (Heimdallarchaeia-sister, EHeim) or Hodarchaeales (Hodarchaeales-sister, EHod) in phylogenetic trees inferred from S97 and NM57 datasets, as the fastest-evolving sites were progressively removed.
Fig. 4
Fig. 4. Metabolic reconstruction of key Asgard archaeal ancestors and distribution of WLP in lineages of Heimdallarchaeia.
a, Transition from the Asgard common ancestor to Lokiarchaeia and Heimdallarchaeia ancestors. Based on ALE results, it is inferred that the LAsCA was a H2-dependent chemolithoautotroph. The archaeal WLP was inferred to be present in all four ancestors; thus, it could also have been present in the LAECA. Four genes related to acetogenesis (pta, ack, acs and acd) were predicted to be present in the Heimdallarchaeia ancestor, suggesting that the LAECA may have been an anaerobic H2-dependent acetogen. b, Distribution of key enzymes of the WLP in Heimdallarchaeia. As a basal branch within Heimdallarchaeia, Hodarchaeales possessed a complete WLP, supporting presence of the pathway in the Heimdallarchaeia ancestor. During the transition from Hodarchaeales to Heimdallarchaeaceae and Kariarchaeaceae, these key enzymes of the WLP appeared to be progressively lost. Fully filled circles indicate that the gene was detected in at least half of the genomes of the clade. Half-filled circles indicate that the gene was detected in fewer than half of the genomes of the clade. EMP, Embden–Meyerhof–Parnas; NOPPP, nonoxidative pentose phosphate pathway; RuMP, ribulose monophosphate pathway; sp/np-ED, semi/non-phosphorylative Entner–Doudoroff pathway; TCA cycle, citrate cycle; AA, amino acid metabolism; β, β-oxidation; H4F, tetrahydrofolate methyl branch; H4MPT, H4MPT methyl branch tetrahydromethanopterin; Fdh, formate dehydrogenase; Mvh/Hdr, F420-non-reducing hydrogenase and heterodisulfide reductase complex; Pyr, pyruvate; Ech, energy-converting hydrogenase; F420, coenzyme F420-reducing hydrogenase; Mcr, methyl-CoM reductase; Nucl, nucleotide; H, Heimdallarchaeaceae; K, Kariarchaeaceae; G, Gerdarchaeales; Hod, Hodarchaeales; W, Wukongarchaeia; fwd, formylmethanofuran dehydrogenase; ftr, formylmethanofuran-tetrahydromethanopterin N-formyltransferase; mch, methenyltetrahydromethanopterin cyclohydrolase; mtd, methylenetetrahydromethanopterin dehydrogenase; mer, 5,10-methylenetetrahydromethanopterin reductase; cdh, acetyl-CoA decarbonylase/synthase, CODH/ACS complex.
Extended Data Fig. 1
Extended Data Fig. 1. The maximum likelihood phylogenomic analysis of the S67 dataset.
The tree was inferred using IQ-TREE under the LG + C60 + F + G + PMSF model, based on 579 archaeal taxa (11,860 sites, 67 concatenated proteins, 579 taxa). This tree was rooted to DPANN and Euryarchaeota. Bootstrap support values ≥ 95% are represented by black dots. The scale bar denotes the average expected number of substitutions per site.
Extended Data Fig. 2
Extended Data Fig. 2. Evolution of ultrafast bootstrap support for either the monophyly of Njordarchaeales and Heimdallarchaeia (NHeim, red line) or the monophyly of Njordarchaeales and Korarchaeia (NKor, blue line).
The bootstrap values were obtained from phylogenomic trees inferred from the S67 (a) and NM57 (b) datasets, based on 579 archaeal taxa, with the fastest-evolving sites progressively removed. The trees were inferred using IQ-TREE under the LG + C60 + F + G + PMSF model.
Extended Data Fig. 3
Extended Data Fig. 3. The maximum likelihood phylogenomic analysis of NM57 marker set, showing the position of contigs in two larger clusters of Njordarchaeales B7_G17_GCA_029856635.
Contigs of Njordarchaeales B7_G17_GCA_029856635 were grouped into three clusters (Fig. 2c). a, Phylogenetic position of contigs in cluster 1. b, Phylogenetic position of contigs in cluster 2. The trees were inferred using IQ-TREE under the LG + C60 + F + G + PMSF model and rooted to DPANN and Euryarchaeota. Bootstrap support values ≥ 95% are represented by black dots. The scale bar denotes the average expected number of substitutions per site.
Extended Data Fig. 4
Extended Data Fig. 4. Evolution of ultrafast bootstrap support for the monophyly of either eukaryotes and Heimdallarchaeia (EHeim, red line) or the monophyly of eukaryotes and Hodarchaeales (EHod, blue line).
The bootstrap values were obtained from the phylogenies inferred from the tES67 dataset (a) and AsES67 dataset (b), as the fastest-evolving sites were progressively removed. The trees were inferred using IQ-TREE under the LG + C60 + F + G + PMSF model.
Extended Data Fig. 5
Extended Data Fig. 5. Phylogenetic analyses and site-exclusion treatment of the S150 dataset for 411 Asgard archaeal taxa and 14 eukaryotic taxa.
a, The maximum likelihood phylogenomic analysis based on the S150 dataset (32,277 sites, 150 concatenated proteins, 425 taxa). The trees were inferred using IQ-TREE under the LG + C60 + F + G + PMSF model. Bootstrap support values ≥ 95% are represented by black dots. The scale bar denotes the average expected number of substitutions per site. b, Evolution of ultrafast bootstrap support for the monophyly of either eukaryotes and Heimdallarchaeia (EHeim, red line) or the monophyly of eukaryotes and Hodarchaeales (EHod, blue line), in the phylogenies inferred from the S150 dataset. The trees were inferred using IQ-TREE under the LG + C60 + F + G + PMSF model.
See this image and copyright information in PMC

References

    1. Imachi, H. et al. Isolation of an archaeon at the prokaryote-eukaryote interface. Nature577, 519–525 (2020). - DOI - PMC - PubMed
    1. Rodrigues-Oliveira, T. et al. Actin cytoskeleton and complex cell architecture in an Asgard archaeon. Nature613, 332–339 (2023). - DOI - PMC - PubMed
    1. Betts, H. C. et al. Integrated genomic and fossil evidence illuminates life’s early evolution and eukaryote origin. Nat. Ecol. Evol.2, 1556–1562 (2018). - DOI - PMC - PubMed
    1. Eme, L. et al. Inference and reconstruction of the heimdallarchaeial ancestry of eukaryotes. Nature618, 992–999 (2023). - DOI - PMC - PubMed
    1. Liu, Y. et al. Expanded diversity of Asgard archaea and their relationships with eukaryotes. Nature593, 553–557 (2021). - DOI - PMC - PubMed

LinkOut - more resources