Metabolic Diversity and Evolutionary History of the Archaeal Phylum “Candidatus Micrarchaeota” Uncovered from a Freshwater Lake Metagenome

The recently described superphylum DPANN includes several phyla of uncultivated archaea with small cell sizes, reduced genomes, and limited metabolic capabilities. One of these phyla, “Ca. Micrarchaeota,” comprises an enigmatic group of archaea found in acid mine drainage environments, the archaeal Richmond Mine acidophilic nanoorganisms (ARMAN) group. Analysis of their reduced genomes revealed the absence of key metabolic pathways consistent with their partner-associated lifestyle, and physical associations of ARMAN cells with their hosts were documented. However, “Ca. Micrarchaeota” include several lineages besides the ARMAN group found in nonacidic environments, and none of them have been characterized. Here, we report a complete genome of “Ca. Micrarchaeota” from a non-ARMAN lineage. Analysis of this genome revealed the presence of metabolic capacities lost in ARMAN genomes that could enable a free-living lifestyle. These results expand our understanding of genetic diversity, lifestyle, and evolution of “Ca. Micrarchaeota.”

ABSTRACT

Acidophilic archaea of the archaeal Richmond Mine acidophilic nanoorganisms (ARMAN) group from the uncultured candidate phylum “Candidatus Micrarchaeota” have small genomes and cell sizes and are known to be metabolically dependent and physically associated with their Thermoplasmatales hosts. However, phylogenetically diverse “Ca. Micrarchaeota” are widely distributed in various nonacidic environments, and it remains uncertain because of the lack of complete genomes whether they are also devoted to a partner-dependent lifestyle. Here, we obtained nine metagenome-assembled genomes of “Ca. Micrarchaeota” from the sediments of a meromictic freshwater lake, including a complete, closed 1.2 Mbp genome of “Ca. Micrarchaeota” Sv326, an archaeon phylogenetically distant from the ARMAN lineage. Genome analysis revealed that, contrary to ARMAN “Ca. Micrarchaeota,” the Sv326 archaeon has complete glycolytic pathways and ATP generation mechanisms in substrate phosphorylation reactions, the capacities to utilize some sugars and amino acids as substrates, and pathways for de novo nucleotide biosynthesis but lacked an aerobic respiratory chain. We suppose that Sv326 is a free-living scavenger rather than an obligate parasite/symbiont. Comparative analysis of “Ca. Micrarchaeota” genomes representing different order-level divisions indicated that evolution of the “Ca. Micrarchaeota” from a free-living “Candidatus Diapherotrites”-like ancestor involved losses of important metabolic pathways in different lineages and gains of specific functions in the course of adaptation to a partner-dependent lifestyle and specific environmental conditions. The ARMAN group represents the most pronounced case of genome reduction and gene loss, while the Sv326 lineage appeared to be rather close to the ancestral state of the “Ca. Micrarchaeota” in terms of metabolic potential.

IMPORTANCE The recently described superphylum DPANN includes several phyla of uncultivated archaea with small cell sizes, reduced genomes, and limited metabolic capabilities. One of these phyla, “Ca. Micrarchaeota,” comprises an enigmatic group of archaea found in acid mine drainage environments, the archaeal Richmond Mine acidophilic nanoorganisms (ARMAN) group. Analysis of their reduced genomes revealed the absence of key metabolic pathways consistent with their partner-associated lifestyle, and physical associations of ARMAN cells with their hosts were documented. However, “Ca. Micrarchaeota” include several lineages besides the ARMAN group found in nonacidic environments, and none of them have been characterized. Here, we report a complete genome of “Ca. Micrarchaeota” from a non-ARMAN lineage. Analysis of this genome revealed the presence of metabolic capacities lost in ARMAN genomes that could enable a free-living lifestyle. These results expand our understanding of genetic diversity, lifestyle, and evolution of “Ca. Micrarchaeota.”

INTRODUCTION

Over the past decade, our knowledge of archaeal diversity has significantly expanded through the application of metagenomics and single-cell genomics (1–4). The recently described superphylum DPANN (1) (named after candidate divisions “Candidatus Diapherotrites,” “Candidatus Parvarchaeota,” “Candidatus Aenigmarchaeota,” “Candidatus Nanohaloarchaeota,” and phylum Nanoarchaeota) includes several phyla of archaea with small cell sizes, reduced genomes, and limited metabolic capabilities (1, 5–7). The first reported member of this group was Nanoarchaeum equitans, representing the phylum Nanoarchaeota. This archaeon has one of the smallest known archaeal genomes (0.49 Mbp) with only 552 genes and grows only in association with the host, Ignicoccus hospitalis, which supplies essential nutrients to N. equitans (8, 9).

Another enigmatic group of archaea, archaeal Richmond Mine acidophilic nanoorganisms (ARMAN), was initially found in biofilms in acid mine drainage (AMD) from Iron Mountain (CA, USA) (10). ARMAN archaea have around 1-Mb genomes and are among the smallest microorganisms described to date, with cell volumes of 0.009 μm³ to 0.04 μm³ (10–12). Phylogenetically, ARMAN archaea belonged to two candidate phyla, “Candidatus Micrarchaeota” and “Candidatus Parvarchaeota” (1, 3, 13). Metagenomic and 16S rRNA bioprospecting studies showed that “Ca. Parvarchaeota” are limited to AMD and hot springs, while “Ca. Micrarchaeota” have a broader habitat distribution (13). Physical interactions of ARMAN cells with their hosts from the archaeal order Thermoplasmatales via pili-like structures were observed using three-dimensional (3D) cryo-transmission electron microscopy (11, 14). Although connection with Thermoplasmatales was predicted for 16 “Ca. Micrarchaeota” species (13), only two cultured consortia have been reported. The first is a coculture of “Candidatus Mancarchaeum acidiphilum” Mia14 with Cuniculiplasma divulgatum (15). The second culture included ARMAN-1 related “Ca. Micrarchaeota” species, Cuniculiplasma divulgatum, another Thermoplasmatales archaeon, and a fungus (16). Recently, the co-occurrence with Thermoplasmatales in enrichment cultures was demonstrated for ARMAN-related “Ca. Micrarchaeota” species from several geothermal sites (17).

Analysis of 39 ARMAN-related genomes obtained from metagenomic studies (metagenome-assembled genomes [MAGs]) revealed the presence of pathways for carbon, nitrogen, and iron cycling and the absence of biosynthesis pathways for amino acids and nucleotides, consistent with their partner-associated lifestyle (13). However, although the first high-quality draft genomes of ARMAN archaea were reported 10 years ago (11), only the following two complete, ungapped genomes of “Ca. Micrarchaeota” members, enabling accurate metabolic reconstruction, are known: “Candidatus Micrarchaeum acidiphilum” ARMAN-2 (13) and “Ca. Mancarchaeum acidiphilum” Mia14 (15). The two organisms were found in AMD-related environments and belonged to the candidate family “Micrarchaeaceae,” defined in the genome-based taxonomy system (18). ARMAN-2 and Mia14 genomes are about 1 Mbp in size, and they lacked central glycolytic pathways and capacities for the biosynthesis of nucleotides and most amino acids, consistent with a parasitic/symbiotic lifestyle (13, 15). However, unlike the majority of DPANN members, ARMAN archaea seem to be capable of respiration as indicated by the presence of a nearly complete tricarboxylic acid (TCA) cycle and an aerobic respiratory chain with a terminal cytochrome bd-II oxidase (13). Genetic potential for iron oxidation, a typical energy-generating reaction in acid mine drainage environments (19), was reported as well (13). The second genome, Mia14, encoded only two components of a respiratory chain, cytochrome bd oxidase and V-type ATP synthase (15).

However, the genetic diversity of “Ca. Micrarchaeota” is not limited to the ARMAN lineage. At present, the Genome Taxonomy Database recognizes two candidate orders (UBA10214 and UBA8480) in addition to the “Micrarchaeales,” comprising ARMAN-related members of this phylum. Besides AMD sites, “Ca. Micrarchaeota” were detected in geothermal sites, soils, hypersaline environments, and freshwater (13, 17). Previously, we investigated the microbial communities of water and bottom sediments of the subarctic meromictic Lake Svetloe, located in the North European part of Russia (20–22). Lake Svetloe represents a rare type of meromictic freshwater lake with a high concentration of ferrous iron and dissolved methane and low concentration of sulfate in the anoxic zone (22). Archaea accounted for about half of the microbial communities in the sediments and were mostly represented by methanogenic Euryarchaeota, “Candidatus Bathyarchaeota,” and DPANN lineages (21).

To improve our knowledge of diversity and genetic potential of “Ca. Micrarchaeota” from nonacidic environments, we obtained nine new MAGs from the metagenome of the sediments of Lake Svetloe and assembled a complete, closed genome of a member of the new order-level lineage of “Ca. Micrarchaeota.” The metabolic potential of this archaeon, designated Sv326, was predicted based on a genome analysis to reveal its evolutionary history, lifestyle, and potential roles in the environment. Genomic information revealed striking differences between the Sv326 archaeon and ARMAN-related “Ca. Micrarchaeota” and indicated that Sv326 is probably a free-living organism with fermentative metabolism, while ARMAN lineages have evolved toward a parasitic lifestyle.

RESULTS

Assembly of the complete genome of a new member of “Ca. Micrarchaeota.”

To obtain MAGs of microbial community members, we sequenced the metagenome of sediments of the meromictic Lake Svetloe using Illumina and Oxford Nanopore techniques. Analysis of the taxonomic affiliation of the obtained MAGs showed that nine of them belong to the candidate phylum “Ca. Micrarchaeota” (Table 1). Most “Ca. Micrarchaeota” MAGs consisted of several dozen contigs and had an estimated completeness of less than 90%. However, the use of long MinION reads enabled us to obtain the most abundant “Ca. Micrarchaeota” MAG, designated Sv326, as two circular contigs. These contigs represented a chromosome (1,178,830 bp) and a plasmid (25,211 bp) of the Sv326 archaeon. The relative abundance of this genotype in the microbial community, determined by the fraction of the Sv326 MAG in the entire metagenome, was 6.39%. This is the third complete, closed genome of the member of “Ca. Micrarchaeota” after ARMAN-2 and Mia14.

TABLE 1.

Main characteristics of “Ca. Micrarchaeota” MAGs obtained in this work

MAG ID	MAG size (bp)	Completeness (%)	Contamination (%)	GC (%)	No. of contigs	Coverage (contigs)	No. of tRNA genes
Sv326a	1,178,830	95.33	1.87	47.8	1	138.71	46
Sv368	945,707	89.72	7.48	47.1	58	16.63	43
Sv174	1,004,840	87.85	2.8	47.7	55	11.44	44
Sv348	1,097,012	84.03	0	54.2	89	48.3	24
Sv319	986,370	76.87	2.18	54.1	87	9.64	40
Sv217	965,041	88.55	5.14	51.4	131	7.29	38
Sv175	624,665	63.86	0.47	51.4	166	5.53	30
Sv246	863,371	81.31	1.4	53.7	105	18.05	43
Sv301	411,752	65.68	0.93	49.4	109	5.32	27

^aChromosome.

The size of the Sv326 genome is significantly larger than that of the acidophilic ARMAN-related “Ca. Micrarchaeota” Mia14 (0.95 Mbp) and ARMAN-2 (1.01 Mbp) and is comparable to the genome sizes of some free-living Crenarchaeota (e.g., 1.30 Mbp for Ignicoccus hospitalis). As a result of the analysis of the Sv326 genome, a 16S-23S rRNA operon, a separately located 5S rRNA gene, and 46 tRNA genes encoding all 20 amino acids were detected. The 16S rRNA gene contains a 510-nucleotide (nt)-long intron encoding putative homing endonuclease. According to the annotation results, 1,362 potential protein-coding genes were called, and functions of only 862 (63%) were predicted as a result of comparison with NCBI databases. Interestingly, among them, there are 249 overlapping genes, and the average intergenic space is only 68 nt, which indicates genome streamlining. Compared with other “Ca. Micrarchaeota” genomes (13), Sv326 has a higher frequency of overlapping genes (18.3% versus 10%) and approximately the same average gene length (854 bp versus 843 bp).

Phylogenetic placement of Sv326 archaeon.

A search for Sv326-related microorganisms based on genome similarity showed that its closest relative is “Candidatus Micrarchaeota archaeon” SpSt-758 (23), which is found in hydrothermal sediments of black smokers in the Atlantic Ocean with an average amino acid sequence identity (AAI) of 71.04% (see Table S1 in the supplemental material). AAI values between Sv326 and other previously known members of “Ca. Micrarchaeota” did not exceed 49.5%. Following the criteria proposed for delineation of the phylogenetic position of uncultivated microorganisms (24), Sv326 and “Ca. Micrarchaeota archaeon” SpSt-758 are different species of the same genus. The Sv368 and Sv174 MAGs belong to the same genus, sharing about 71% AAI with Sv326.

To analyze the phylogeny of “Ca. Micrarchaeota,” we constructed a phylogenetic tree based on concatenated sequences of 122 conserved marker genes, including all “Ca. Micrarchaeota” MAGs obtained in this work and 23 other “Ca. Micrarchaeota” genomes. The results obtained (Fig. 1) show that Sv326, Sv368, Sv174, and the other three MAGs form a separate phylogenetic lineage of the order level, along with the candidate orders “Microarchaeales” (comprising Mia14 and ARMAN-2 archaea), UBA10214, and UBA8480, recognized in the Genome Taxonomy Database (GTDB) (18).

FIG 1.

Genome-based phylogenetic analysis of “Ca. Micrarchaeota.” MAGs obtained in this study are marked in red. Taxonomy is shown according to the GTDB (o_, order). GenBank assembly accession numbers are shown after the genome names. The levels of support for internal branches assessed using the Bayesian test in PhyML are indicated at the nodes.

The three “Ca. Micrarchaeota” MAGs obtained in this study belong to UBA10214, and two MAGs were assigned to UBA8480 (Fig. 1). MAG Sv301 formed a distinct branch close to the root of the “Ca. Micrarchaeota” and, according to GTDB classification, could represent a novel class. Unfortunately, the low quality of this assembly (65.7% completeness, total length 411,752 bp) is insufficient for a comprehensive analysis of Sv301.

A nucleotide BLAST search against the NCBI NR and IMG databases for 16S rRNA gene sequences related to Sv326 revealed 4 environmental clones that could represent the same genus considering a 95% sequence identity threshold (25) (see Table S2 in the supplemental material). These sequences have been detected in the sediments of Lake Yellowstone (USA) and Lake Kivu (Rwanda). Several dozen more distantly related clones (85 to 92% identity) have been found in lake sediments, microbial mats in hot springs, groundwater, wetlands, and peat soil (Table S2). On the phylogenetic tree based on 16S rRNA gene sequences, Sv326 and related clones form a monophyletic branch clearly separated from clones representing “Microarchaeales” and other candidate orders of “Ca. Micrarchaeota” (see Fig. S1 in the supplemental material).

Predicted central metabolic pathways of Sv326 archaeon.

The Sv326 genome contains a nearly complete set of genes encoding enzymes of the archaeal version of an Embden-Meyerhof glycolytic pathway and gluconeogenesis, including ADP-dependent phosphofructokinase/glucokinase, glucose-6-phosphate isomerase, bifunctional fructose-1,6-bisphosphate aldolase/phosphatase, triosephosphate isomerase, NAD(P)-dependent glyceraldehyde 3-phosphate dehydrogenase, phosphoglycerate kinase, phosphoglycerate mutase, and enolase (Fig. 2; see also Table S3 in the supplemental material). Although fructose-1,6-bisphosphate aldolase/phosphatase has been described as a gluconeogenic enzyme due to irreversibility of phosphatase reaction (26), we suppose that together with a phosphofructokinase it could convert fructose-6-phosphate into trioses in glycolysis. The pyruvate kinase was missing, but its function in the conversion of phosphoglycerate to pyruvate is probably performed by phosphoenolpyruvate synthase as shown in the euryarchaeon Thermococcus kodakarensis (27). Enzymes of the Entner-Doudoroff pathway, found in “Ca. Mancarchaeum acidiphilum” Mia14 (15), were absent in the Sv326 genome.

FIG 2.

Overview of the metabolism of the Sv326 archaeon. Enzyme abbreviations are as follows: GK, phosphofructokinase/glucokinase; GPI, glucose-6-phosphate isomerase; FBAP, bifunctional fructose-1,6-bisphosphate aldolase/phosphatase; TIM, triosephosphate isomerase; GPDH, glyceraldehyde 3-phosphate dehydrogenase; PGK, phosphoglycerate kinase; PGM, phosphoglycerate mutase; ENO, enolase; PEPS, phosphoenolpyruvate synthase; POR, pyruvate:ferredoxin oxidoreductase; ACS, acetyl-CoA synthetase; ATR, transaminase; RPE, ribulose-phosphate 3-epimerase; RPI, ribose 5-phosphate isomerase; RK, ribokinase; TKT, transketolase; TAL, transaldolase; HPI, 6-phospho-3-hexuloisomerase; HPS, 3-hexulose-6-phosphate synthase; ADH, alcohol dehydrogenase; RPP, ribose-phosphate pyrophosphokinase; AMPP, AMP phosphorylase; RBPI, ribose-1,5-bisphosphate isomerase; RuBisCO, ribulose-1,5-bisphosphate carboxylase; SOD, superoxide dismutase; HD, hydrogenase; PPase, pyrophosphatase; ABC, ABC-type transporter; P-ATP, P-type ATPase. Compounds abbreviations are as follows: Sed-7P, sedoheptulose-7-phosphate; GAP, glyceraldehyde-3-phosphate; DHAP, dihydroxyacetone phosphate; 1,3-bis-PG, 1,3-biphosphoglycerate; 3-PG, 3-phosphoglycerate; 2-PG, 2-phsophoglycerate; PEP, phosphoenolpyruvate; PRPP, phosphoribosyl pyrophosphate; R15bP, ribose 1,5-bisphosphate; Ru15bP, ribulose 1,5-bisphosphate; Fd(o) and Fd(r), oxidized and reduced forms of ferredoxin.

Pyruvate could be decarboxylated to acetyl coenzyme A (CoA) by pyruvate:ferredoxin oxidoreductase, producing reduced ferredoxin. Then acetyl-CoA synthetase (ADP-forming) could form acetate from acetyl-CoA with concomitant ATP production. The Sv326 genome also encodes an acetyl-CoA synthetase (AMP-forming) capable of performing the reverse reaction of acetyl-CoA production from acetate, CoA, and ATP.

The genome encoded all genes of the nonoxidative branch of the pentose phosphate pathway. Ribulose-5-phosphate (Ru5P) could also be generated from fructose-6-phosphate by 6-phospho-3-hexuloisomerase and 3-hexulose-6-phosphate synthase with concomitant formaldehyde production. The same enzymes could enable the reverse reaction of formaldehyde fixation. Formaldehyde could be made from methanol by an alcohol dehydrogenase encoded in the genome, although the specificity of this enzyme could not be predicted.

Ru5P could be converted by ribose-5-phosphate isomerase into ribose-5-phosphate (R5P) followed by its phosphorylation by ribose-phosphate pyrophosphokinase to make phosphoribosyl pyrophosphate (PRPP), a key intermediate for nucleotide biosynthesis. Consistently, the Sv326 genome encodes a complete set of genes for de novo biosynthesis of purine and pyrimidine nucleotides.

Interestingly, the Sv326 genome harbors a gene encoding ribulose-1,5-bisphosphate carboxylase type IIIb (RuBisCO), a key enzyme of the Calvin cycle of CO₂ fixation. However, the gene encoding the upstream enzyme of this cycle, phosphoribulokinase, was absent as in other archaeal genomes (28). Archaeal type III RuBisCO proteins are known to be involved in AMP metabolism rather than in the Calvin cycle (29). In the first stage of this pathway, AMP phosphorylase replaces the adenine base by a phosphate group to generate ribose 1,5-bisphosphate (R15bP). The latter becomes converted to ribulose 1,5-bisphosphate (Ru15bP) by ribose-1,5-bisphosphate isomerase, and finally, RuBisCO catalyzes the conversion of Ru15bP and CO₂ to two molecules of 3-phosphoglycerate, which enters the Embden-Meyerhof glycolytic pathway (30). All of these genes are present in the Sv326 genome, and the corresponding proteins have 42 to 45% amino acid sequence identity with functionally characterized enzymes from Thermococcus kodakarensis (30), suggesting the ability of this archaeon to utilize AMP. Sufficient amounts of AMP could be formed, for example, in glycolysis by ADP-dependent phosphofructokinase or by ribose-phosphate pyrophosphokinase in the course of PRPP synthesis.

Analysis of the Sv326 genome revealed the absence of all genes of the TCA cycle. Components of the aerobic respiratory chain (NADH dehydrogenase, succinate dehydrogenase, cytochrome bc₁ complex, and cytochrome oxidases) and terminal reductases that could perform anaerobic respiration were also not found.

Reduced ferredoxin, generated by pyruvate:ferredoxin oxidoreductase, could be reoxidized by group 4e [NiFe] hydrogenase. Such hydrogenases form respiratory complexes that couple oxidation of ferredoxin to reduction of protons to H₂ with proton or sodium translocation across the membrane (31). The presence of a potential subunit with transmembrane domains in the hydrogenase operon indicates that this could be the case in the Sv326 archaeon. The transmembrane ion gradient could also be generated by a pyrophosphate-energized sodium pump, while membrane-bound ATP synthases were missing. Therefore, the Sv326 archaeon lacked respiratory capacities like most members of the DPANN superphylum (32) and is probably confined to fermentative metabolism.

The search for hydrolytic enzymes did not reveal proteases and glycosyl hydrolases containing N-terminal secretion signals that enable the extracellular hydrolysis of complex polymers. The presence of amino acid transporters, transaminases, NADP-specific glutamate dehydrogenase, and enzymes for further processing of 2-keto acids suggests that the Sv326 archaeon could import and ferment some amino acids. The presence of several enzymes involved in the metabolism of carbohydrates (e.g., GH57 family alpha-amylase/alpha-mannosidase and sugar kinases) indicates the possibility of utilizing sugars as growth substrates. Particularly, ribose could be phosphorylated by ribokinase, and the resulting R5P could enter the pentose-phosphate pathway or other reactions leading to the production of fructose-6-phosphate and/or 3-phosphoglycerate as shown in Fig. 2. The Sv326 archaeon could also generate ATP through fermentation of pyruvate that could be taken up by passive diffusion (33). Genes encoding the fatty acids beta-oxidation pathway were not found, and neither were ones for the utilization of glycerol.

Pili and secretion.

Four operons encoding type IV pili systems were found in the genome, each of which includes genes for pilus assembly proteins TadB and TadC and pilus assembly ATPase of the CpaF family. No genes related to the formation of archaellum were found, which is in agreement with the absence of motility in other “Ca. Micrarchaeota” (13).

Genes for the Sec protein secretion system were identified, including genes for preprotein translocase subunits SecY, SecF, and SecD, signal-peptide peptidase, signal recognition particle protein Srp54, and receptor FtsY (Table S3). The presence of N-terminal Sec signal peptide was predicted for 105 potential proteins, while signal peptides for the Sec-independent Tat secretion system were not identified. Ninety-nine of 105 potentially secreted proteins were annotated as hypothetical with unknown functions. Thirty-three signal-peptide carrying proteins also contained from 2 to 11 transmembrane helices and could be associated with the cell surface.

Mobile elements and plasmid.

An interesting feature of the Sv326 genome is the presence of a region containing three tandem repeats of a 26,004-nt-long sequence. Three copies of the repeat unit are completely identical and are separated from each other and from flanking genome regions by 14-bp-long repeats. Each of the repeat units was predicted to contain 39 genes, mostly annotated as encoding hypothetical proteins. Five genes were predicted to encode a site-specific tyrosine recombinase, a S-adenosylmethionine (SAM)-dependent methyltransferase, a concanavalin A-like lectin/glucanase superfamily protein, a very short patch repair protein, and a DNA breaking-rejoining enzyme. The presence of putative integrase-like genes and short terminal repeats suggests that this repeat is probably a novel mobile element integrated in the Sv326 genome.

The Sv326 MAG comprised a circular 25,211-bp-long contig, presumably representing a plasmid designated as pSv326-1. It was sequenced with 2,335-fold average coverage, a value 16.8 times higher than the Sv326 chromosome. Two regions of this plasmid (1,398 and 1,034 bp long) were identical to the Sv326 chromosome, further supporting the hypothesis that the plasmid belongs to the Sv326 archaeon. The plasmid harbored 48 genes, of which only three were tentatively annotated. Genes of SAM-dependent methyltransferase and concanavalin A-like lectin/glucanase superfamily protein were present in the repeat element mentioned above (∼80% amino acid identity). The third annotated gene was predicted to encode ParB-like plasmid partitioning protein. To the best of our knowledge, this is the only report of a plasmid in “Ca. Micrarchaeota” and DPANN archaea in general.

MAGs of the members of the candidate orders UBA10214 and UBA8480.

UBA10214 and UBA8480 are order-level candidate divisions of the “Ca. Micrarchaeota,” recognized in the GTDB taxonomy. Although all “Ca. Micrarchaeota” MAGs obtained in this study and assigned to these orders were only medium quality drafts (Table 1), we used the best assemblies, Sv348 for UBA10214 and Sv246 for UBA8480, to get insights into the metabolic potential and lifestyle of these archaea. Both MAGs contained a nearly complete set of genes for the Embden-Meyerhof glycolytic pathway and the subsequent conversion of pyruvate to acetate via acetyl-CoA (Table S3). The nonoxidative pentose-phosphate pathway was present in Sv348 but missing in Sv246 MAG. Both MAGs lacked any of the components of the aerobic respiratory chain, terminal reductases for anaerobic respiration, and membrane-bound ATP synthases. The TCA cycle is incomplete in both MAGs, although Sv348 harbors most of the relevant genes (Table S3). Both MAGs contained genes for type III RuBisCO involved in AMP salvage metabolism. The most striking difference between Sv348 and Sv246 genomes is the presence of a nearly complete set of genes for de novo biosynthesis of nucleotides in Sv348 and their absence in Sv246 (Table S3). Although the Sv246 MAG is shorter in size than Sv348 and has lower estimated completeness, genes of nucleotide biosynthesis pathways are dispersed in the genomes (as occurred in Sv326 and Sv348), and their simultaneous absence in the assembly by chance seems to be unlikely.

DISCUSSION

At present, in addition to the Sv326 genome, the complete genomes of two “Ca. Micrarchaeota” members are known, “Candidatus Micrarchaeum acidiphilum” ARMAN-2 (13) and “Candidatus Mancarchaeum acidiphilum” Mia14 (15). Both of these genomes represented the ARMAN lineage found in acid mine drainage environments (10, 11, 15) and were assigned to the candidate family “Micrarchaeaceae” in the GTDB taxonomy. Comparison of the predicted proteomes revealed that 820 of the 1,125 (73%) Sv326 protein-coding genes are missing in either ARMAN-2 or Mia14, 263 genes are present in three genomes, while 16 and 26 genes are shared with only Mia14 and ARMAN-2, respectively (Fig. 3). The ARMAN-2 and Mia14 genomes contain fewer genes and are more closely related to each other, sharing more than a half of predicted proteomes.

FIG 3.

Homologous protein-coding genes of the Sv326 archaeon, “Ca. Micrarchaeum acidiphilum” ARMAN-2, and “Ca. Mancarchaeum acidiphilum” Mia14. Genes from all three genomes were clustered into homologous groups using the BlastClust program, and two or more genes were merged into one cluster if the region of similarity covered >90% of the shorter protein with a minimum of 30% identity. The numbers of shared and specific gene clusters are shown.

The notable features of the ARMAN-2 and Mia14 genomes are a complete absence of glycolytic pathways of central metabolism, present in Sv326, including the Embden-Meyerhof pathway, pentose-phosphate pathway, RuBisCO, and other enzymes involved in ATP metabolism (see Table S3 in the supplemental material). Both genomes also lacked pathways for de novo biosynthesis of nucleotides, most amino acids, and cofactors. All of these critical metabolic deficiencies indicate that ARMAN-related archaea are devoted to a partner-dependent parasitic/symbiotic lifestyle, and their associations with Thermoplasmatales hosts have been documented (15, 16). Contrary to Sv326, the ARMAN-2 genome had most genes of the TCA cycle (but succinyl-CoA synthetase was missing) and a respiratory chain that comprised of NADH dehydrogenase, succinate dehydrogenase, cytochrome bc₁ complex, and V-type H⁺-transporting ATP synthase (13). The presence of a terminal oxidase in ARMAN-2 was not reported, but a cytochrome bd-II oxidase was found in the closely related ARMAN-1 genome (13). The ARMAN-2 genome also encodes carbon monoxide dehydrogenase that could supply electrons in the respiratory chain. The Mia14 genome has more limited respiratory capacities. It lacks the TCA cycle, and its respiratory chain consisted only of cytochrome bd oxidase and V-type ATP synthase.

Thus, ARMAN-2 and Mia14 seem to be a group evolving toward the parasitic/symbiotic lifestyle, in which they receive key nutrients and cellular building blocks from their hosts. Adaptation to such a lifestyle allowed “Micrarchaeaceae” members to lose key pathways involved in central metabolism. More complex is the question of the origin of energy-related pathways in these archaea. They could be present in the lowest common ancestor (LCA) of all “Ca. Micrarchaeota” and be lost in the Sv326 lineage and partially in Mia14. Alternatively, the TCA cycle and respiratory chain were absent in the “Ca. Micrarchaeota” LCA and was gained in the course of the evolution of the “Micrarchaeaceae.” Phylogenetic analysis of genes of energy-related pathways argues for the second hypothesis and indicated that these genes were mostly acquired from the Thermoplasmatales (13).

Phylogenetic analysis of the DPANN superphylum indicated that the candidate phylum “Ca. Diapherotrites” is a sister lineage to “Ca. Micrarchaeota,” and these two phyla have a common ancestor (13, 15). A draft genome of a member of this phylum, “Candidatus Iainarchaeum andersonii,” characterized in detail by Youssef et al. (7), had a size of about 1.24 Mbp, short average gene length, high coding density, and a high number of overlapping genes. These traits are also observed in “Ca. Micrarchaeota” genomes. Nevertheless, metabolic reconstruction suggested that “Ca. Iainarchaeum andersonii” could be a free-living organism capable of fermenting a narrow range of substrates, including ribose, polyhydroxybutyrate, and some amino acids (7). The genome contained most but not all of the genes of the Embden-Meyerhof pathway, some TCA cycle genes, and genes for the conversion of pyruvate to acetate with ATP production. Like the Sv326 archaeon, “Ca. Iainarchaeum andersonii,” has all genes for ribose utilization via the AMP salvage pathway, including type III RuBisCO, and nucleotide biosynthesis. Genes for NADH dehydrogenase, complexes II and III, and terminal oxidases were not found, but the V-type ATP synthase was present. These data support the hypothesis that the LCA of “Ca. Micrarchaeota” had complete fermentative pathways and lacked an aerobic respiratory chain.

Evolution of the “Ca. Micrarchaeota” probably involved losses of important metabolic pathways in some lineages and gains of specific functions in the course of adaptation to a partner-dependent lifestyle and particular environmental conditions. The Sv326 lineage appeared to be rather close to the ancestral state of the “Ca. Micrarchaeota” in terms of metabolic potential. Limited gene loss occurred in Sv348 (order UBA10214), while a more pronounced loss of functions (e.g., pentose-phosphate pathway and nucleotide biosynthesis) was detected in Sv246 (order UBA8480). The ARMAN group (“Micrarchaeaceae”) probably represents the most pronounced case of genome reduction among “Ca. Micrarchaeota” and gene loss accompanied by the gain of specific functions (e.g., aerobic respiratory chain).

Analysis of the Sv326 genome showed that this archaeon has key metabolic pathways necessary for heterotrophic autonomous growth, including glycolysis and ATP generation mechanisms in substrate phosphorylation reactions, AMP metabolism, pathways for de novo nucleotide biosynthesis, and isoprenoid biosynthesis, among others. This predicted lifestyle matches the environmental conditions in which the Sv326 archaeon was identified. Contrary to ARMAN-2 and Mia14 found in AMD biofilms (10, 15), which are metal-rich aerobic or microaerophilic ecological niches providing optimal conditions for respiratory metabolism, Sv326 was found in the sediments from a freshwater lake. The bottom sediment of Lake Svetloe is a strictly anoxic organic-rich ecological niche in which anaerobic electron acceptors are scarce, and the main microbial processes are fermentation of organic compounds and methanogenesis (21). Such conditions fit the metabolic capabilities of the Sv326 archaeon that were predicted from the genomic analysis. We suppose that Sv326 is a free-living scavenger, and its ecological role in lake sediments consists of the fermentation of low-molecular-weight organic compounds with the formation of acetate and hydrogen, which become consumed by methanogenic archaea, performing the final step of organic matter decomposition.

Description of “Candidatus Fermentimicrarchaeum limneticum.”

The genome of Sv326 is the first complete genome of a member of a new order-level lineage of the “Ca. Micrarchaeota.” Taking into account that it meets the criteria for a finished MAG (34) and considering the recently published consensus statement regarding naming uncultivated Archaea and Bacteria (35) we propose the following taxonomic names for the novel order, family, and species of Sv326.

Description of “Candidatus Fermentimicrarchaeum” gen. nov. Fermentimicrarchaeum (Fer.men.ti.micr.ar.chae′um. L. neut. gen. n. fermenti, of a fermentation process; Gr. adj. micros, small; N.L. neut. n. archaeum (from Gr. adj. archaios -ê -on, ancient), ancient one, archaeon; N.L. neut. n. Fermentimicrarchaeum, a small archaeon performing fermentation processes).

Description of “Candidatus Fermentimicrarchaeum limneticum” sp. nov. Fermentimicroarchaeum limneticum (lim.ne′ti.cum. Gr. n. limnê, pool of standing water, lake; L. neut. suff. -ticum suffix denoting belonging to; N.L. neut. adj. limneticum, belonging to a lake).

Not cultivated. Inferred to be an anaerobic, obligate organotroph that obtains energy through fermentation of low-molecular-weight organic substrates. Represented by the complete genome obtained from the metagenome of sediments of subarctic freshwater meromictic Lake Svetloe, Russia.

Based on this, we propose the following names for the order and family:

“Candidatus Fermentimicrarchaeales” ord. nov.

“Candidatus Fermentimicrarchaeaceae” fam. nov.

MATERIALS AND METHODS

Site description, sampling, and metagenomic DNA isolation.

The freshwater Lake Svetloe (65.0498 N, 41.0626 E) is a permanently stratified meromictic lake located in the Arkhangelsk region of the Russian Federation. The lake has a maximum depth of about 39 m, and the chemocline zone is located at a depth of 20 to 24 m. Lake Svetloe represents a rare type of freshwater meromictic lake, as the concentrations of ferrous iron and dissolved methane in the anoxic layer are high while sulfate is scarce (22). The water has a neutral pH (7.33 to 7.46) and a constant temperature of about 4°C at depths below 8 m (22). The physicochemical parameters of the lake water and sediments have been reported in previous studies (20, 22). Briefly, the concentration of sulfate in the bottom water was below 2 μM, nitrate was 8.3 to 20.2 μM, ammonium was 143 to 214 μM, Fe³⁺ was 240 μM, Fe²⁺ was below 7.3 μM, and methane was 920 μM (20, 22).

A sample of bottom sediments was taken with a stratometer-type tube corer on 19 May 2019. For the analysis, a slice of the sediment’s core corresponding to depths of 5 to 15 cm was chosen. The total DNA was extracted from 4 g of sediment using a PowerSoil DNA isolation kit (MO BIO Laboratories, Inc., Carlsbad, CA, USA). A total of about 1.5 μg DNA was obtained.

Sequencing of metagenomic DNA and assembly of a complete genome of Sv326 archaeon.

A metagenomic DNA sample was sequenced using the Illumina (Illumina, USA) platform. The TruSeq DNA library was sequenced on the Illumina HiSeq 2500 instrument in a paired reads mode (2 × 150 nt). Adapter sequences and low-quality reads were removed using Cutadapt (36) and Sickle (https://github.com/najoshi/sickle), respectively. A total of 408.38 million high-quality reads (66.5 Gbp) was obtained. Contigs were assembled using the Megahit v.1.2.9 (37) and binned into MAGs using MetaBAT v.2.12.1 (38).

The taxonomic assignment of the obtained MAGs was performed using the Genome Taxonomy Database Toolkit (GTDB-Tk) v.1.1.1 and Genome Taxonomy Database (GTDB) (18). Completeness of the MAGs and their redundancy (possible contamination) were estimated using the CheckM v.1.05 tool (39).

Metagenomic DNA was additionally sequenced on MinION (Oxford Nanopore, UK) using a Ligation Sequencing kit 1D protocol and R9.4 flow cell. Sequencing resulted in 8,190,486 reads with a total length of ∼10.49 Gbp. MinION reads were de novo assembled using Flye v.2.6 (40). Contig sequences were corrected using Pilon v.1.2.2 (41) with two iterations of Illumina reads mapping. Obtained polished “Flye” contigs were binned into MAGs using MetaBAT v.2.12.1 (38).

One MAG, designated Sv326, and assigned to “Ca. Micrarchaeota,” was represented by one unclosed contig. Paired-end Illumina reads were mapped to this contig using Bowtie2 v.2.3.4.1 (42). The mapped Illumina reads were de novo assembled using SPAdes v.3.13.0 (43); when only one read of a pair mapped to the contig, both reads were included in the assembly. Mapping of Illumina reads and de novo assembly were repeated several times using the last SPAdes assembly as a reference for read mapping and recruitment.

The final SPAdes assembly was scaffolded into two circular contigs by mapping MinION reads to the SPAdes contigs with the BWA v.0.7.15 tool (44) and joining the contigs by npScarf (45) using the overlap information provided by MinION reads. Gaps were filled in using Illumina consensus sequences of the SPAdes assembly graph (–spadesDir parameter of npScarf).

The sequence of the genome region comprising three tandem copies of 26-kb-long repeats was assembled manually using SPAdes assembly graph visualized using the Bandage tool (46). Connections between all contigs that could be connected to repeats were also checked using long MinION reads. Finally, two circular contigs, 1,178,830 bp and 25,211 bp, were obtained for the Sv326 MAG.

Genome annotation and analysis.

Gene search and annotation was performed using the RAST server 2.0 (47). The annotation was then checked and manually corrected by the comparison of predicted protein sequences with the National Center for Biotechnology Information (NCBI) databases. The N-terminal signal peptides were predicted by Signal P v.5.0 for archaea (http://www.cbs.dtu.dk/services/SignalP/), and the presence of transmembrane helices was predicted by TMHMM v.2.0 (http://www.cbs.dtu.dk/services/TMHMM/).

The average amino acid identity (AAI) between the genomes was determined using the aai.rb script from the Enveomics collection (48).

Phylogenetic analysis.

The data set used for genome-based phylogenetic analysis included Sv326, 23 other members of “Ca. Micrarchaeota” from all putative orders recognized in GTDB, and three “Ca. Diapherotrites” genomes. For these genomes, a multiple alignment of concatenated 122 archaeal single-copy marker genes was carried out using the GTDB toolkit v.1.1.1 (49). This multiple alignment, 5,124 amino acids in length, was used to construct the maximum likelihood phylogenetic tree using PhyML v.3.3 (50) with the default parameters. The level of support for internal branches was assessed using the Bayesian test in PhyML.

Data availability.

The raw metagenomic sequences obtained in this study have been deposited in the NCBI Sequence Read Archive under accession numbers SRR12594517 (Illumina reads) and SRR12594518 (MinION reads). The annotated sequences of “Ca. Micrarchaeota” MAGs have been deposited in the GenBank database under BioProject number PRJNA644262. Genome data has been deposited in GenBank under CP058998 for the chromosome and CP058999 for the plasmid.