Phylogenomic Analyses Reveal that Panguiarchaeum Is a Clade of Genome-Reduced Asgard Archaea Within the Njordarchaeia

Abstract

The Asgard archaea are a diverse archaeal phylum important for our understanding of cellular evolution because they include the lineage that gave rise to eukaryotes. Recent phylogenomic work has focused on characterizing the diversity of Asgard archaea in an effort to identify the closest extant relatives of eukaryotes. However, resolving archaeal phylogeny is challenging, and the positions of 2 recently described lineages-Njordarchaeales and Panguiarchaeales-are uncertain, in ways that directly bear on hypotheses of early evolution. In initial phylogenetic analyses, these lineages branched either with Asgards or with the distantly related Korarchaeota, and it has been suggested that their genomes may be affected by metagenomic contamination. Resolving this debate is important because these clades include genome-reduced lineages that may help inform our understanding of the evolution of symbiosis within Asgard archaea. Here, we performed phylogenetic analyses revealing that the Njordarchaeales and Panguiarchaeales constitute the new class Njordarchaeia within Asgard archaea. We found no evidence of metagenomic contamination affecting phylogenetic analyses. Njordarchaeia exhibit hallmarks of adaptations to (hyper-)thermophilic lifestyles, including biased sequence compositions that can induce phylogenetic artifacts unless adequately modeled. Panguiarchaeum is metabolically distinct from its relatives, with reduced metabolic potential and various auxotrophies. Phylogenetic reconciliation recovers a complex common ancestor of Asgard archaea that encoded the Wood-Ljungdahl pathway. The subsequent loss of this pathway during the reductive evolution of Panguiarchaeum may have been associated with the switch to a symbiotic lifestyle, potentially based on H2-syntrophy. Thus, Panguiarchaeum may contain the first obligate symbionts within Asgard archaea besides the lineage leading to eukaryotes.

Keywords: Panguiarchaeum; Asgard archaea; evolution; phylogenetics; symbiosis.

Fig. 1.

The phylogenetic placement of Njordarchaeia. a) Schematic of the phylogenetic hypotheses tested in this study. Korarchaeota topology: Njordarchaeia branch sister to Korarchaeota (note that Korarchaeota was recovered as a sister group to a clade comprising both Thermoproteota and Asgard archaea in this study). Asgard archaea topology: Njordarchaeia branch within Asgard archaea. b) Maximum-likelihood phylogenetic analysis of the concatenated 50% top-ranked marker proteins (n = 43) and streamlined set2 (303 taxa) using a custom site-heterogeneous substitution model (LG + EDM0256LCLR + G + F). Scale bar: average substitutions per site. The number of taxa in each collapsed clade is shown by the number in parentheses next to the clade name. GCA_029856705.1* is the closest genome to the first proposed type material of Njordarchaeum guaymasis (see supplementary discussion, Supplementary Material online). The GCA_024720975.1^TS is the proposed type material for Panguiarchaeum  symbiosum. The tree is rooted by “Euryarchaea” (including Halobacteriota, Thermoplasmatota, Hydrothermarchaeota, Methanobacteriota, Methanobacteriota_A, and Methanobacteriota_B). c) Maximum-likelihood phylogenetic analysis of the concatenated 50% top-ranked marker proteins (n = 43) and streamlined set3 (71 taxa) using the LG + CAT-PMSF model. Scale bar: average substitutions per site. The number of taxa in each collapsed clade is shown by the number in parentheses next to the clade name. d) Evaluation of phylogenetic signals for Asgard archaea and Korarchaeota topology under LG + G + F, LG + C60 + G + F, Poisson + CAT-PMSF, and LG + CAT-PMSF models using the streamlined set 3 (71 taxa), listed in an order of increasing model fit. Sites underlying the signals were binned by substitution rates (based on LG + C60 + G + F) and effective amino acids per site (based on LG + CAT-PMSF). e) Distribution of amino acid diversity scores for the simulated alignments under LG + G + F, LG + C60 + G + F, one representative Poisson + CAT-PMSF, and LG + CAT-PMSF simulation using the streamlined set 3 (71 taxa). The dashed line indicates the amino acid diversity score from the real dataset. Z-scores quantifying the difference between the compositional heterogeneity of the real dataset and that of the simulated datasets from each model are listed in brackets for each model. Note that Z-scores of LG + C60 + G + F and Poisson + CAT-PMSF are similar, but the right tail of Poisson + CAT-PMSF is closer to the real dataset, indicating a better fit. f) Impact of progressive removal of the most compositionally biased sites ranked by chi-square score on the statistical support of Asgard archaea and Korarchaeota topology. The support values, UFBOOT and SH-aLRT, were estimated on the 50% top-ranked markers of streamlined set2 (303 taxa) based on LG + C60 + G + F and LG + EDM0256LCLR + G + F models.

Fig. 2.

Hierarchical clustering of contigs in 4 MAGs from Panguiarchaeaceae (a, b) and Njordarchaeaceae (c, d) based on sequence composition and depth of coverage across different metagenomes. Note that the tip of the hierarchical clustering tree represents splits of long contigs at a size of 20,000 bp. This analysis demonstrates that metagenomic contamination is not a source of conflicting phylogenetic signals. The ring indicating coverage across the source metagenomes is highlighted with a shaded area. The locations of marker gene sequences used in this study (ring labeled top50) and marker gene homologs from a previous study (ring labeled S150) are shown as bar plots, with bar height corresponding to the number of genes. Notably, our marker gene sequences are located on clusters of contigs that share similar coverage profiles across metagenomes, as are the majority of the S150 homologs. Support for Asgard archaea and Korarchaeota topologies is shown in a stacked bar plot, with genes on the same contig that show conflicting topological support highlighted in thick outlined boxes. e) Barplot showing the number of genes that confidently reject Asgard archaea, Korarchaeota, or neither of the two topologies under different models.

Fig. 3.

Presence of Asgard cluster of orthologues (AsCOGs) across major archaeal lineages. Panguiarchaeaceae and Njordarchaeaceae shared a number of AsCOGs homologs with other Asgard archaea, in contrast to Korarchaeota. AsCOG presence was determined across 966 archaeal genomes, and the occurrence of AsCOGs within any taxonomic clade of interest (based on the genome taxonomy database) was recorded as a percentage. Clusters with fewer than 2 genomes were not included. The number of genomes for each cluster is shown in parentheses.

Fig. 4.

Metabolic characteristics of Njordarchaeia. Overall, Njordarchaeia are thermophilic fermentative heterotrophs, with core pathways of varying completeness, suggesting some core metabolites must be obtained from other organisms in the environment. Full circles indicate that a gene is present in all or more than 50% of the MAGs, while half circles indicate that a gene is found in at least one, but less than half, of the MAGs. Open circles denote genes absent from all MAGs of a group. A detailed list of genes encoded by Njordarchaeia can be found in supplementary data S7, S8, and S9, Supplementary Material online.

Fig. 5.

Gene-tree-aware ancestral genome reconstruction of streamlined set 2 (303 taxa) focusing on Njordarchaeia including Panguiarchaeum lineages. Njordarchaeia, in particular Panguiarchaeum, might have experienced genome reduction from their common ancestor with a larger ancestor Wukongarchaeia and Heimdallarchaeia (0.53 to 0.84 Mpb on the stem: nodes 571 to 563). a) Inferred species of asgard archaea, thermoproteotal, and korarchaeotal lineages from Fig. 1b. The colors on the branches represent the number of gene losses, and the bar plots on the branches of interest show the number of originations, duplications, and transfers. The size of circles at nodes represents the inferred ancestral genome size. The number of taxa in each collapsed clade is shown by the number in parentheses next to the clade name. b) The presence probability of genes of interest at the ancestral nodes leading to Panguiarchaeum. Node numbers of ancestral nodes are shown and referred to in Panel a). Full circles indicate a presence probability (PP) ≥ 0.75, half circles show a PP < 0.75 but ≥ 0.5, while PP < 0.5 is shown in open circles (see supplementary data S13, Supplementary Material online).

Fig. 6.

Abundance-based co-occurrence of Njordarchaeales and Sulfolobales. For plotting, the counts of both orders Njordarchaeales (including Panguiarchaeaceae) and Sulfolobales, were summed in each sample and plotted. For visualization purposes, the counts were square root transformed.