Cultivating the Ancestors (4|4)

by Christoph
 

Having one or two 'Asgard' archaea under the microscope – after having cultivated them with great effort and even more patience – and looking them in the face is exciting, but a bit unsatisfying if they are cousins. Are they mavericks or rather typical for "Lokis"? Here are the portraits of two more distant re­la­tives, both also cousins: Margulisarchaeum peptidophila HC1 and Flexarchae­um multipro­tru­sionis SC1. They became known through a very recent preprint by Imachi, Nobu et al. (2025).
 

Note  Since the text would become overly long if always their full names were given, I will limit myself in the following to their "license plates", the strain identi­fiers. I will refer to 'Ca. Promethearchaeum syntrophi­cum' strain MK‑D1 as MK‑D1 (part 1), to 'Ca. Lokiarchaeum ossiferum' strain Loki‑B35 as Loki‑B35 (part 3), to 'Ca. Margulisarchaeum peptidophila' strain HC1 as HC1, and to 'Ca. Flexarchaeum multiprotrusionis' strain SC1 as SC1. See here a cladogram of the 'Asgard' ar­chae­a including the mentioned strains.
 

Cultivating Margulis­archaeum peptidophila HC1 and Flexarchaeum multiprotrusionis SC1

(click to enlarge)
Figure 9. Microscopy characterization of HC1. a–n, HC1 cultures. a–e, SEM images showing single (a), di­viding (b), protrusion-producing (c), and methanogen attached (d) cells and a magnified view of d (e). f, QFDE image. g, h, Ultrathin section. i, Fluorescence image of cells stained with DAPI (blue) and hybridized with nuc­leotide probes targeting HC1 (green) and Methano­calculus sp. (red). j, Immunofluorescence image of cells stain­ed with SYBR Green I (yellow/green) and labeled with an HC1 actin-specific antibody (red) (Extended Data Fig. 7). k–n, Slice images from cryo-tomograms (k, m) and cor­responding 3D volume segmentations (l, n). Vacuole- or intracellular vesicle-like structures were infrequently observed (m, n). White arrows indicate Me­thano­cal­culus sp.. Images c, f, g–n were obtained from cultures in late-exponential phase and others from mid expo­nential phase. SEM, QFDE, ultrathin section, FISH, and immunofluorescence images are representative of n=98, 48, 60, 21, and 36 recorded images for HC1. For HC1, cryo-tomography images were obtained from 8 re­corded tomograms. Scale bars, 1 μm (a–j) and 500 nm (k,m). Source

Strains HC1 and SC1 were enriched by Imachi, Nobu et al. (2025) from a sample taken at a gas pro­duct­ion well in the Boso Peninsula, Chiba Pre­fecture, Japan. At this deliberately chosen sampling site methane-bearing ancient brine upwells into the well and supports a biofilm on the walls con­tain­ing populations identical/similar to those found in the anoxic, methane-bear­ing marine sedi­ments in which MK‑D1 was found (see part 1). Enriched in this biofilm, which was intermittently exposed to air (oxy­gen) due to periodic eruptions of brine that caused the water level to tempo­ra­ri­ly drop, were small populations of 'Asgard' archaea, which they tar­geted for enrichment/isola­tion.
 

In subsequent sub-cultures, the researchers re‑introduced reducing agents (sodium sulfide and cysteine) and added peptone as an additional energy source. The 'Hodarchaeales' popu­lation detected by specific primers among the 'Asgard' ar­chaea (see here the cladogram for your orien­ta­tion) in the cultures reached up to 30% in re­lative abundance and maximum cell densities up to 10⁶ 16S rRNA gene copy numbers/mL, with me­thanogenic Methanocalculus sp. (Euryarchaeota, order Methanomicrobia­les), Bac­te­roidota, and Gammaproteobacteria as the remaining mi­crobes in the co-culture. The latter two bacterial groups were eliminated by anti­biotic supplement during subculturing, which was done by opti­mizing cultivation temperature and energy source types and concentrations in parallel. Three years after the original sam­pling, they isolated two 'Hodarchaea­les', designated HC1 (72.6%) and MC2, in a co-culture with a single partnering Methanocalculus sp. population (27.4%), in an an­aer­obic medium supple­mented with casamino acids, peptone and yeast ex­tract at 30°C. HC1 required 25–35 days to reach full growth with a dou­bling time of ap­prox­i­mately 7–12 days, reach­ing a maximum cell density of 10⁶ 16S rRNA gene copies/mL. HC1 cells were morphologi­cal­ly si­milar to MK-D1 and Loki-B35, with coccoid cells, albeit with larger dia­meters of on average 1.3 μm and multiple protrusions extending from the cell surface (Figure 9, a,c,d).
 

Using a similar cultivation strategy, they isolated another 'Hodarchaeales' strain, designated as SC1, in a purified tri-culture with Methanocalculus sp. strain MC3 and Methanolobus sp. strain MLB, with relative abundances of approximately 40%, 60%, and <0.3%, respectively. The tri-culture is stably maintained in anaerobic medium supplemented with casamino acids, pow­dered milk, polypeptone peptone and yeast extract and required approximately 60 days to reach maxi­mum growth with a doubling time of approximately 14–20 days, reaching a maxi­mum cell density of 10⁷ 16S rRNA gene copies. The 16S rRNA gene sequence identity between SC1 and HC1 was 77.75% based on BLAST analysis. SC1 cells were morpholo­gi­cally similar to HC1 but with a smaller diameter, on average 0.9 μm (Figure 10, o,q,r).
 

Metabolism and cell structure of M. peptidophila HC1 and F. multiprotrusionis SC1

(click to enlarge)
Figure 10. Microscopy characterization of SC1. o–ab, SC1 cultures. o–s SEM images showing single (o), di­viding (p), protrusion-producing (q), and methanogen-attached (r) cells and a magnified view of r (s). t, QFDE image. u, v, Ultrathin section. w, Fluorescence image of cells stained with DAPI (blue) and hybridized with nucleotide probes targeting SC1 (green) and Methano­calculus sp. (red). x, Immunofluorescence image of cells stained with SYBR Green I (yellow/green) and labeled with an SC1 actin-specific antibody (red) (Extended Data Fig. 7). y, Cryo-EM snapshot. z, aa, Slice images from cryo-tomograms. ab, Ultrathin section of a cell con­taining a vacuole- or intracellular vesicle-like structure. White arrows indicate Methanocalculus sp.. Images t–x, z–ab were obtained from cultures in late-exponential phase and others from mid exponential phase. SEM, QFDE, ultrathin section, FISH, and immunofluorescence images are representative of n=46, 44, 14, 9, and 11 re­corded images for SC1. For SC1, cryo-tomography ima­ges were obtained from 5 recor­ded tomograms. Scale bars, 1 μm (p–r, w, x) and 500 nm (o, s–v, y–ab). Source

Imachi, Nobu et al. (2025) found by BLAST ana­lysis that the 16S rRNA gene sequence identity between HC1 and SC1 is 77.75%. This allowed to place Margulisarchaeum peptidophila HC1 and Flexarchaeum multiprotrusionis SC1 as "cousins" in separate (taxonomic) families within the order Hodarchaeales of the class Heimdallarchaeia, one of the presently known 10 classes of the phylum 'Asgard' archaea (see here the cladogram). Except that the genomes of HC1 (7.91 Mbp) and SC1 (6.75 Mbp) are larger than those of MKD1 (4.46 Mbp) and Loki-B35 (6.04 Mbp), little can be learned from these numerical values (an aside: most free‑living bac­teria have genome sizes in the range of 3–6 Mbp). More informative is a comparison of their cell physiology and cell struc­ture when it comes to the question of which lineage of the 'Asgard' archaea comes closest to the ancestor of the eukaryotes.
 

All four "Lokis" have an anaerobic (anoxic), peptide-degrading, mesophilic and syntrophic life­style. 'Syntrophic' meaning that they live in obligate symbiosis with methanogenic archaea, and lack many pathways for biosynthesis of amino acids, vitamins, and nucleotides. Their small coccoid cells have a simple internal structure, that is, no endomembrane systems, only infrequently observed vesicle-/vacuole-like structures, and no membrane-based compartmen­talization of the chromosome. To put it casually in a nutshell: to become eukaryotic cells, they would have to learn how to do/make all of this first. Yet, they have and shed membrane vesi­cles (blebs), and typical archaeal S-layer-like structures.
 

In the electron microscope, HC1 and SC1 cells show multiple membrane-based cytosol-con­nected protrusions of uniform diameter (~340 nm), va­rying lengths, branching structure, and bulbous expansions (Figure 9;b; Figure 10, q,r). These protrusions could always be seen in growing cells, not – or much less frequently – in cells in the process of division (Figure 9, b; Figure 10, p). Dividing cells always have a conspicuous ring, a 🛟 life­buoy , around their "waist". The molecular structure of the ring is unknown, but on magnification it looks a little like a string of uniform blebs cov­ered with the S-layer. This had already been observed in MK-D1 cells, and like these, dividing cells have (to-be) daughter cells of equal size, indi­cating binary division and not budding. Imachi, Nobu et al. (2025) found in feeding experiments that cells with increased nutrient supply grow a bit larger with more protrusions but maintained their slow-growth characteristic and doubling rates. They conclude form this that unlike most prokaryotes the "Lokis" separate growth and cell division processes. It needs to be said that, given the slow growth and the low division rate, it is a real feat to actually find any cells in a sample right in the process of division.
 

The researchers found the protrusions of HC1 and SC1 decorated with filaments of varying lengths and numbers (Figure 9, d,e; Figure 10, q–s; Figure 12). Such filaments are not known from the protrusions of MK-D1 and Loki B-35. It is still unknown whether these are different types of filaments, what their molecular structure is and what function they have.
 

(click to enlarge)
Figure 11. Immunofluorescence staining of HC1 and SC1 cells with Lokiactin-specific antibodies imaged by a combination of confocal laser scanning micro­scopy (CLSM) and stimulated emission depletion (STED). a, HC1 culture. b, SC1 culture (the same cell as in Figure 10, x). The detection of the Loki­actin is shown in red (abberior STAR580-labeled sec­ond­ary antibody) and DNA shown in green (SYBR Gre­en-I). The overlay of Locactin and DNA stains appears yellow ("merged"). White arrows indicate Methano­cal­culus sp. cells. Scale bars, 3 μm Source

As in Loki-B35 and (not yet proven, but very likely) in MK-D1, virtually all protrusions of HC1 and SC1 cells show well-defined Lokiactin filaments (Figure 11; click here or on the image to see it enlarged). Lokiactin is also present in the cell bodies, but there it appears as a blob in the "stain" and not resolved as filaments. The researchers prepared the cells so carefully and gently that they came to rest on the microscope slide like octopuses with outstretched arms and without rupturing of the fragile protrusions. From such images, they could measure that the cell volume, the cytoplasm, was distributed roughly equally between the protrusions and the cell bodies, while only the latter contained genomic DNA (SYBR-Green staining).
 

I said above that all four cultured "Lokis" have an anaerobic (anoxic), peptide-degrading, meso­philic and syntrophic lifestyle. But there’s a twist. Imachi, Nobu et al. (2025) had detected Loki­ar­chae­ales (the order of MK-D1 and Loki-B35), Heimdall­archaeales, and Hodarchaeales (the order of HC1 and SC1) in the original enrichment cultures, but in further sub­cultures in which reducing agents were omitted to leave trace amounts of oxygen in the medium, only one of the Hod­archaeales was enriched and the others lost. Apparently, HC1 but not SC1 has a tolerance for (very) low levels of oxygen. A genome analysis for HC1 revealed the presence of genes that encode enzymes indicative of adapta­tions to a more oxidized habitat: cytochrome bd oxidase (Cbd)-like protein, cytochrome c oxidase (Cox), catalase, and superoxide dismutase. In HC1 cultures, the researchers verified the expression of these genes and also the presence of heme B, which is a cofactor for three of the above enzymes. However, they have no indications for the presence of complete respiratory chains, and they consider oxygen reduction in HC1 by, for example, the termi­nal oxidase Cbd a supplementary metabolism.
 

A personal wrap-up

(click to enlarge)
Figure 12. SEM image of "blebby" MK-D1 cell with pro­trusions. The tips of the protrusions appear slightly en­larged. Scale bar, 500 nm. Source

I have been following this story with increasing joy since 2014, when I first heard about the "Lokis" (intro part 1). It was fascinating to learn of their unusual cell shape with these pro­tru­sions – which I like to call "tiny arms" with an actin cytoskeleton – from the images of P. syn­tro­phicum MK‑D1 in the 2019 preprint (and Hiro Imachi's first ever tweet on 𝕏) and the images of L. ossi­fe­rum Loki-B35 in the paper by Rodrigues-Oliveira, Woll­we­ber, Ponce-Toledo et al. (20­22). It is encouraging to now be reasonably certain that this was not an artifact caused by sample preparation – the curse of electron microscopy, as practitioners know – but is a com­mon morphological fea­ture of 'Asgard' archaea. Or, more cautiously: ...of at least four species of two 'Asgard' lineages (here again the cladogram).
 

The engulfment of an alphaproteobacterium (proto-mitochondrion) by an archaeon during eukaryogenesis can now be imagined in a completely different way than through one of the known eukaryotic pathways for endocytosis. Differently, that is, not as a singular, unprepared and thus ac­cidental event but as a sequence of potentially preparatory steps by syntrophic partners from certain 'Asgard' archaeal and bacterial and/or other archaeal lineages (I had in­troduced the E3 model in part 2). It is easy to imagine that the evolution of the "tiny arms" by the 'Asgard' archeon al­lowed the syntrophic consortium to become more locally cohesive by embracing, which was conducive to metabolite exchange (Figure 12). An even more intimate embrace would be possible if the "tiny arms" had "fingers", as suggested by the filaments at the protrusions seen in HC1 and SC1 (see above, Figure 13). Thus, entanglement likely pre­ceded en­gulfment.
 

(click to enlarge)
Figure 13. SEM images of HC1 methanogen-attached (e) cells and short filaments on the surface of protrusions (f). Source

The fortunate circumstance that after 2 billion years of evolution, 'Asgard' descendants are still found who, if sluggishly, thrive comfortably with and in entanglement suggests that this was not a temporary and necessary intermediate step toward the engulfment of a syntrophic part­ner, but an evolutionary success in itself. Another – but not the last – intermediate step to­wards LECA, the still mysterious Last Eukaryotic Common Ances­tor, was probably the de­velop­ment of oxygen tolerance in the ancestral 'Asgard' archaeon with its entirely anaerobic (an­oxic) metabolism, at the latest during or after the Great Oxidation Event (GOE) 2.460–2.060 bil­lion years ago. In the genome of HC1, which was enriched from a low‑oxy­gen habitat, Imachi, Nobu et al. (2025) found several genes encoding proteins that would allow the cells to tolerate O2 but not to grow by O2 respiration (that is, by "aerotolerant anaerobiosis"; what a tongue twister!). One of these genes, en­coding catalase, was probably acquired via horizontal gene transfer (HGT) from an ancestral alphaproteo­bacterium. Since other genes that can me­­­diate O2 tolerance were "passed around" via HGT between different 'Asgard' families over hundreds of millions of years, it cannot be said at pre­sent that we fully understand the trajectory of the development of O2 tolerance in the 'Asgard' archaea towards LECA. Except that it had oc­curred, and that ancestors of the Hodarchaeales HC1 and SC1 are presently the prime candidates. But we can be curious to see what other 'Asgard' archaea have "up their sleeve" when they are once cultivated (pun intended: "tiny arms" they have, in all likeyhood).
 

Do you want to comment on this post? We would be delighted to hear from you! Please send us an email, or comment on Mastodon or Bluesky. We'll be posting comment threads that arise below (with your permission). Feel free to subscribe if you'd like to get regular notifications about new blog posts.