Large attachment organelle mediates interaction between Nanobdellota archaeon YN1 and its host

Abstract DPANN archaea are an enigmatic superphylum that are difficult to isolate and culture in the laboratory due to their specific culture conditions and apparent ectosymbiotic lifestyle. Here, we successfully isolated and cultivated a coculture system of a novel Nanobdellota archaeon YN1 and its host Sulfurisphaera ohwakuensis YN1HA. We characterized the coculture system by complementary methods, including metagenomics and metabolic pathway analysis, fluorescence microscopy, and high-resolution electron cryo-tomography (cryoET). We show that YN1 is deficient in essential metabolic processes and requires host resources to proliferate. CryoET imaging revealed an enormous attachment organelle present in the YN1 envelope that forms a direct interaction with the host cytoplasm, bridging the two cells. Together, our results unravel the molecular and structural basis of ectosymbiotic relationship between YN1 and YN1HA. This research broadens our understanding of DPANN biology and the versatile nature of their ectosymbiotic relationships.

DPANN organisms are distinguished by their diminutive size, reduced genomes, and limited metabolic abilities.Therefore, the majority of the DPANN thrive in mutualistic, commensal, or parasitic associations with a variety of archaeal and bacterial hosts [2,[4][5][6][7][8][9].Despite their widespread occurrence, and important role in microbial ecology and environment, very little is known about their cell biology, metabolic potential, and molecular/structural basis of their symbiosis.A considerable hurdle in the study of DPANN lies in the difficulty of isolating and establishing new DPANN coculture systems.Consequently, very few DPANN coculture systems have been established and investigated to date [4,5,[10][11][12][13].
We have established a new coculture system composed of a novel Nanobdellota archaeon YN1 and its host Sulfurisphaera ohwakuensis YN1HA.We used microbiology, metagenomics and metabolic pathway analysis, f luorescence microscopy, and highresolution electron cryo-tomography to characterize the YN1-YN1HA interaction.Our analyses revealed the growth characteristics and genomic features of YN1 and uncovered an attachment organelle in the YN1 envelope that bridges the host and DPANN cytoplasm, facilitating this ectosymbiotic relationship.

Sampling
Approximately 50 ml of hot water samples containing a small quantity of mud were collected in polypropylene tubes at an acidic hot spring Yunoike in Kirishima geothermal area located in Kagoshima prefecture, Japan.Temperature and pH at the sampling site were 78 • C and pH 2.7, respectively [14].The sample in a plastic tube was transported to the laboratory under ambient temperature and used for enrichment culture.

Enrichment culture
Enrichment culture was conducted using modified Brock's basal salt medium [15] supplemented with 0.1% (w/v) yeast extract (MBSY medium).The pH of the medium was adjusted to 2.9.One hundred microliters of each sample was inoculated into 10 ml MBSY medium in a 20 ml screw-capped test tube (16.5 × 105 mm, ST-16.5L,Nichiden-rika glass, Kobe, Japan) and incubated on static condition at 80 • C. Microbial growth of the enrichment cultures was regularly monitored by measuring the optical density at 600 nm.When growth entered stationary phase, 2 ml of each enrichment culture was centrifuged (20 000 rcf, 25 • C, 30 min) and the resulting pellet was used for DNA extraction for 16S ribosomal RNA (rRNA) gene amplicon sequencing.The rest of the enrichment culture was stored in a freezer at −80 • C with 10% (v/v) glycerol as a cryoprotectant reagent.The cryostock was later used for the establishment of pure coculture composed of a Nanobdellota archaeon (strain YN1) and its host S. ohwakuensis (strain YN1HA).

16S rRNA gene amplicon sequencing
Microbial DNA from the enrichment culture was extracted using Extrap Soil DNA Plus Kit version 2 (Nippon Steel and Sumikin Eco-Tech).The extracted DNA was used for the 16S rRNA gene amplicon sequencing.The V3-V4 region of 16S rRNA was amplified using the DNA polymerase KOD Fx Neo (TOYOBO) with primers A340F [16] and 806rB [17] (with barcode for MiSeq System).The following thermal cycler protocol was used: an initial denaturation of 2 min at 94 • C, 25 cycles of 10 s denaturation at 98 • C, 30 s annealing at 50 • C, followed by 68 • C for 15 s, and a final extension of 68 • C for 5 min.The PCR product was sent to a sequencing company (Fasmac, Kanagawa, Japan), and the 16S rRNA gene amplicon sequencing was carried out using a MiSeq sequencer (300 bp × 2; Illumina).The microbial community structure of the enrichment culture was investigated by the QIIME2 pipeline [18], followed by BLASTN search against nt database in NCBI.

Primer design
The specific primers targeting the 16S rRNA gene for a Nanobdellota archaeon (NanoF: GGCGAAATGCAGTAATCCCG, NanoR2: TAACGGCTTCCCTATCCCAC) detected in the enrichment culture were designed by using Primer BLAST [19], while those for other detected archaea (Sulfolobales species) were designed as previously described [20].

Establishment of pure coculture and isolation of pure host
To establish a pure coculture only composed of a Nanobdellota archaeon and its host, dilution to extinction method was carried out three times under the same condition as enrichment culture.The presence of a Nanobdellota archaeon was confirmed by quantitative PCR (qPCR) with specific primers, and the most diluted culture containing the Nanobdellota archaeon was always chosen for the next purification step (dilution to extinction).The same PCR cycling conditions as described previously [20] were used.A pure culture of host was isolated by single colony isolation from the pure coculture using an MBSY plate medium solidified with 0.7% (w/v) gellan gum (Wako), 10 mM MgSO 4, and 2.5 mM CaCl 2 .

Genome sequencing and assembly
Microbial cells were collected by centrifugation (15 880 rcf, 10 min) from the pure coculture.Genomic DNA was extracted from the cells using Genomic-Tip 100/G (QIAGEN).To obtain long reads, a DNA library was prepared following the protocol as described by Oxford Nanopore Technologies (NBE_9065_v109_revZ_ 14Aug2019), followed by sequencing with MinION sequencer using R9 f low cell (Oxford Nanopore Technologies).To obtain short reads, the genomic DNA was sent to Novogene Bioinformatics Technology (China) to be sequenced on a NovaSeq 6000 platform (2 × 150 bp; Illumina).The short and long reads were quality-filtered with using the same protocol as described previously [21] and coassembled using Unicycler (ver.0.4.8)[22].To ensure the purity of the coculture, the quality-filtered short reads were mapped against genome assemblies using CoverM v.0.7.0 [23].

Growth monitoring
The growth of a Nanobdellota archaeon YN1 and its host S. ohwakuensis YN1HA in pure coculture or pure culture was monitored by qPCR with specific primers for each species.The same protocols as described previously [20] was used.In brief, microbial DNAs regularly extracted from cultures were applied to qPCR with specific primers targeting 16S rRNA gene, which is a single copy gene for each strain.Cultivation was conducted in triplicate to compare YN1HA growth between coculture and pure culture and in duplicate to examine the growth temperature and pH ranges for YN1 and YN1HA.

Light microscopy and scanning electron microscopy
General cell morphology was assessed by phase contrast microscopy (Optiphot-2; Nikon) and differential interference contrast microscopy (BX51; Olympus) at room temperature.The cells stained with SYBR Green I by the same protocol as described previously [20] were observed by f luorescent microscopy (BX51; Olympus).Scanning electron microscopy (JSM-7500F; JEOL) was carried out using the same protocols as described previously [20].Samples for these observations were collected on Day 3 or Day 4.

Fluorescent in situ hybridization
Fluorescent in situ hybridization (FISH) was conducted using the method as previously described [35] with some modifications.Samples for FISH analysis was collected on Day 3. The FISH probes used for YN1 (Nano-R4_TEX: GTATTCCCGTGGCGACTGC) and YN1HA (SFB-R_FAM: CGGTTACTAGGGATTCCTCG), both targeting the 16S rRNA gene were designed using Primer BLAST [19].The probes were labelled with either Texas-red or 6-carboxyf luorescein at their 5 end.Cells were fixed with 2% paraformaldehyde for overnight, washed in phosphate-buffered saline (PBS), and resuspended in PBS before spotting 15 μl onto a silane-coated glass slide (S8111; MATSUNAMI).The spotted sample was dried; incubated in 100 μl of 0.25 N HCl solution for 30 min at room temperature; rinsed in distilled water; and dehydrated by serial treatment with 50%, 80%, 90%, and 100% ethanol solutions (3 min each, two times).Hybridization was carried out with the probes (final concentration: 1 pmol/μl each) in buffer (0.9 M NaCl, 20 mM Tris-HCl, 0.1% SDS, pH 7.5) at 48 • C overnight.After hybridization, the slide was washed in buffer (0.9 M NaCl, 20 mM Tris-HCl, 0.1% SDS, 5 mM EDTA, pH 7.5) at 48 • C for 20 min, lightly rinsed in distilled water, and enclosed with mounting medium (SlowFade Gold Antifade reagent; Life Technologies).The f luorescence signals were detected with an epif luorescence microscope (BX63; Olympus) fitted with filter sets specific for Texas red and f luorescein.

Sample vitrification and CryoET data acquisition
Samples for cryoET were collected on Day 3. Before vitrification, archaeal cocultures were mixed in a 2:1 ratio with 10 nm colloidal gold beads precoated with 1% BSA (Sigma-Aldrich, Australia).The mixture was added onto glow-discharged copper R2/2 Quantifoil holey carbon grids (Quantifoil Micro Tools GmbH, Jena, Germany).Grids were blotted for 6-8 s (under 100% humidity conditions) and plunged into liquid ethane using a Vitrobot Mark IV, FEI (Thermo Fisher Scientific).Micrographs of YN1 cocultures were acquired using an FEI Titan Krios G4, 300 keV FEG transmission electron microscope (Thermo Fisher Scientific), equipped with a BioQuantum K3 Imaging Filter (slit width 20 eV), and a K3 direct electron detector (Gatan).Tilt series were collected automatically using FEI Tomography 5.10 from −60 • to +60 • at 2 • intervals with a defocus of −6 μm and a total electron f luence of 120 e − /Å 2 and a pixel size of 3.39 Å.

Tomogram reconstruction
Before processing of tilt series, tilt images of poor quality were identified and removed through visual inspection.The cleaned tilt series were aligned using fiducial tracking in IMOD (Version 4.11.5)[36,37] and subsequently binned four times (13.56 Å/pix).Tomo3D (version 2.2) [38] was used to generate Simultaneous Iterative Reconstruction Technique (SIRT)-reconstructed tomograms from aligned tilt series, which were then filtered to enhance contrast using the deconvolution filter in IsoNet [39].

Tomogram segmentation and visualization/3D segmentation
Visualization and segmentation of tomograms were conducted using the Dragonf ly software (https://www.theobjects.com/dragonf ly/index.html).Tomograms were preprocessed using built-in filters, including histogram equalization, gaussian, and unsharp filters.Neural network 5-class U-Net (with 2.5D input of 5 slices), were trained on tomogram slices to recognize YN1 cytoplasmic membrane, YN1 outer layer, YN1HA cytoplasmic membrane, YN1HA outer layer, cones = pink, cytoplasmic filaments, intercellular filaments, intercellular sheath filament, and ribosomes and further fine-tuned to achieve high-quality segmentation for automated 3D segmentation.The software's built-in tools for exporting 2D images and creating 3D movies were utilized.

Nanobdellota DPANN-host coculture system
From an initial enrichment culture, three archaeal taxa were detected: S. ohwakuensis [15], Scl.solfataricus [40], and a novel archaeon belonging to the Nanobdellota [41] (Fig. 1A).Using the enrichment culture as a starting material, we attempted to establish a pure coculture by dilution to extinction method [42].During each purification attempt, we confirmed the presence of the Nanobdellota archaeon using qPCR with specific primers, and the most diluted culture containing the Nanobdellota archaeon was always chosen for the next purification step.After three rounds of purification attempts, a pure coculture composed of a Nanobdellota archaeon (strain YN1) and S. ohwakuensis (strain YN1HA) was successfully established (Fig. 1B).The purity was checked by shotgun genome sequencing of genomic DNA extracted from the coculture.As a result of hybrid assembly of HiSeq short and Nanopore long reads, two complete circular genomes having 773 326 bp (for YN1) and 2 761 125 bp (for YN1HA), respectively, were obtained.No other contigs encoding any prokaryotic marker gene and none of the other sequence fragments were obtained from the hybrid assembly.Besides, 99.7% of the short reads (6 737 386 out of 6 755 440 quality-filtered reads) were mapped to the two genomes, indicating that this coculture can be reliably considered a pure coculture.
The genome of strain YN1 has 889 protein coding sequences (CDSs), 3 rRNAs, and 43 transfer RNA (tRNAs), with a guaninecytosine (GC) content of 23.4%, whereas that of S. ohwakuensis YN1HA has 3130 CDSs, 2 rRNAs, and 47 tRNAs, with a GC content of 32.6% (Supplementary Tables 1 and 2).Phylogenomic analysis using amino acid sequences of 53 archaeal marker genes derived from the database [30] placed YN1 within the phylum Nanobdellota (Fig. 1C).The closest relative of YN1 was Nanobdella aerobiophila MJ1 T , whose host was reported to be Metallosphaera sedula MJ1HA.This was in good agreement with the fact that YN1 and N. aerobiophila, but not other cultivated representatives in the phylum Nanobdellota, can grow under toxic conditions.The topology of phylogenetic tree based on the 16S rRNA gene sequence was almost the same as that of the based on 53 archaeal marker genes (Fig. 1C and D).The 16S rRNA gene sequence similarity between strain YN1 and N. aerobiophila MJ1 T was only 86.5%, significantly below the highly reliable threshold value of 94.5% [43] for defining prokaryotic genera.The growth temperature of YN1 (55 • C-95 • C) is higher than that of N. aerobiophila (60 • C-75 • C).In addition, the host species of YN1 is different from that of N. aerobiophila (i.e.M. sedula).These differences indicate that strain YN1 represents a novel genus within the phylum Nanobdellota.Therefore, we propose a novel genus and species "Candidatus Nanofervidus parviconus" gen.nov.sp.nov. to accommodate strain YN1 (Na.no.fer'vi.dus.Gr. masc.n. nânos, a dwarf; L. masc.adj.fervidus, hot; N.L. masc.n.Nanofervidus, a thermophilic small-sized microorganism.Parviconus par'vi.co'nus.L. masc.adj.parvus, little; L. masc.n. conus, cone; N.L. fem.n. parviconus).
Growth conditions of the coculture system were investigated to establish an optimum.The growth temperature and pH ranges for YN1 were 55 • C-95 • C and pH 2.0-3.0,respectively, while those for S. ohwakuensis YN1HA were 55 • C-95 • C and pH 1.5-5.5, respectively (Supplementary Fig. 1A and B).The optimal conditions for both strains were established as 80 • C and pH 3.0.Using these conditions, the growth of YN1 did not occur on the first day (Days 0-1) and the ratio of YN1/YN1HA was decreased from 90.5 to 7.3.After that, YN1 started to grow from Day 1 when its host YN1HA entered the early exponential phase (Fig. 1B).During the middle exponential phase of YN1HA (from Day 1 to Day 2), YN1 exhibited a growth rate almost identical to that of YN1HA, while the ratio of YN1/YN1HA was 3.9 to 7.3.From Day 2 to Day 3, while YN1HA was in the late exponential to early stationary phase, YN1 entered exponential phase and the ratio of YN1/YN1HA increased to 70.At this point, the growth of YN1HA was inhibited compared to pure culture.On Day 4, when YN1HA was in the late stationary phase, the YN1/YN1HA ratio became maximum of 233.After Day 5, YN1 entered the death phase, and the growth of YN1HA was slightly recovered reaching a maximum cell density at Day 7. At this point, the ratio of YN1/YN1HA decreased to 1.5, which was the lowest value observed during this growth experiment (Fig. 1B).FISH analysis confirmed that nanosized cells with ∼300 nm in diameter were "Ca.Nanofervidus parviconus" YN1, while relatively larger cells with 1-2 μm in diameter were S. ohwakuensis YN1HA (Fig. 1E).While most YN1HA cells were surrounded by nanosized YN1 cells in the exponential phase (Days 3-4), many free YN1 cells away from the YN1HA cells (Supplementary Fig. 2A and B).Although some of the nanosized cells could be derived from cell debris or vesicles, we believe that most of the nanosized cells observed were YN1, since almost no small f luorescent objects were observed in the YN1HA pure culture (Supplementary Fig. 2C and D).
We observed cell division of YN1 on the surface of the host (Supplementary Fig. 3A-D) and away from the host (Supplementary Fig. 3A-D) in our scanning electron micrographs.During the exponential phase (Days 3-4), the majority of YN1HA cells were surrounded by 1-17 YN1 cells under the phase/f luorescence microscope, although many detached YN1 cells were also observed (Supplementary Figs 2 and 3E).When YN1 cells were physically separated from the YN1HA by filtration (pore size: 0.45 μm), YN1 growth was completely inhibited.
A cocultivation screen showed that YN1 only grows on S. ohwakuensis strains as a host but not on other genera and species (i.e.Metallosphaera, Sulfuracidifex, Sulfodiicoccus, Saccharolobus, and Sulfurishaera javensis).Growth of coculture was confirmed by qPCR targeting the 16S rRNA gene of each species after 1-2 weeks of cultivation.Conversely, YN1 was able to grow on different strains of S. ohwakuensis (YN1HA and TA-1 T ).The motility of individual YN1 cells, on the surface of the host, was not clearly observed under the phase contrast microscope at room temperature.YN1 cells attaching to a host cell seemed to adhere firmly to the point of attachment (SI movie 1).

CryoET uncovers an attachment organelle used by YN1 to interact with its host
Electron cryo-tomography (cryoET) is unparalleled in capturing a broad range of biological phenomena from microns to subnanometer scale in their near-native conditions [44][45][46][47].CryoET does not require molecular biology and therefore can be used as an exploratory method to identify and characterize systems that are genetically intractable [48,49].Here, we utilized cryoET to investigate the structural basis of ectosymbiosis between YN1 and YN1HA at molecular resolution.In our tomograms, the two cell types are clearly distinguishable by their size; YN1 cells measure between ∼300 and 600 nm in diameter compared to the host cells, which measure ∼1-2 microns in diameter (Fig. 2A-C, SI movie 2).Our data revealed that multiple YN1 cells interact with their host YN1HA; we observed as many as 5 YN1 per YN1HA (2.4 average YN1 per YN1HA, n = 29) (Fig. 2A, Supplementary Table 3).Our cryoET observations were further confirmed by scanning electron microscopy imaging (Fig. 2C-E).Intriguingly, at the interface between YN1 and YN1HA, we observed a cone-shaped large macromolecular structure imbedded in the envelope of YN1 cells (Fig. 2A, SI movie 2).Density for the YN1 inner membrane can be observed on the cytosolic side of the cone.However, the outer S-layer appears to be completely replaced by the cone-shaped assembly.Analysis of these structures in segmented volumes showed that they form open-ended cone-like structures with a wide opening at the base (cytosol facing) and narrow opening at the top (extracellular facing) (Fig. 2B, SI movie 3).We refer to these large structures as "cones" and the narrow opening at the interface of YN1 and YN1HA as "portals."From our 15 tilt-series, we found a total of 31 YN1 cells and 29 cone structures.All but two YN1 cells contained a cone with a maximum of 1 cone per YN1 cell (Supplementary Table 3).Our measurements of these cone structures showed that they had an average lateral length of 165.9 nm (SD = 51.4nm, n = 14), average base diameter of 288.4 nm (SD = 82.2nm, n = 14), and average height of 97.6 nm (SD = 27.6 nm, n = 14) (Supplementary Table 3).We estimated the average lateral surface area of the cones at 8.5 × 10 4 nm 2 (± 4.7 × 10 4 nm 2 ), which demonstrates their enormous size and overall distribution within our dataset.We found that the cross-sectional thickness of the cones and the size of the portal were much more consistent between different cones; the average cross-sectional thickness and portal diameter was 20.8 nm (SD = 1.8 nm, n = 14) and 20.4 nm (SD = 2.6, n = 14), respectively.In 15 cone structures, we also observed an intercellular filament that passed from the cytoplasm of the YN1 cell to the cytoplasm of the host cell.The filaments passed through the portal of the cone and were attached to the cone via a cytoplasmic gate-like structure (Fig. 2A  inset and B).Analysis of the segmented volume revealed that the cytoplasmic membrane of the host cell is contorted through its outer layer (S-layer) to the cone portal (Fig. 2A inset and B inset, SI movie 2).Finally, in 11 of the 31 YN1 cells, we observed filament bundles in the cytoplasm.Using automated segmentation, we trained a u-net to segment these filaments, revealing two different types of intercellular filament ( Fig. 2B, SI movie 3, purple filaments and light-yellow filaments).It is unclear if these structures are related to the cones and involved in intercellular bridging, but their measured diameters (6.6 nm, n = 11) were close to those of the intercellular filaments present in the cones.Seemingly, the cone, portal, and intercellular filament together form an "attachment organelle" that facilitates the interaction between YN1 and YN1HA, and this interaction creates a cytoplasmic bridge for resource exchange.

CryoET reveals different stages of interaction between YN1 and YN1HA
Close inspection of our tomograms revealed different stages of interaction between YN1 and YN1HA.Based on the distance between the YN1 and YN1HA, we found three distinct relationships-(i) YN1 cells completely detached from their host, (ii) distally attached with their host, and (iii) forming a tight junction with their host.Interestingly, the detached YN1 cells still harboured the attachment organelle (cone, portal, and filament) (Fig. 3A and B, SI movie 4).This suggests that attachment organelle could exist prior to interaction with the host and/or could remain intact even after YN1 is detached from their host.The distally attached YN1 cells interacted with the host by the intercellular filament structure that protrudes out of the portal from a distance of ∼40 nm ( Fig. 3C and D, SI movies 5 and 6).Based on previously published cryoET data on sheathed filaments [50,51], we suggest that the intercellular filaments are likely coated in a proteinaceous sheath (Fig. 3C and D, SI movie 6), which was not observed in the fully developed interaction (proximal attachment) depicted in Fig. 2A and B. Finally, we observed YN1 cells proximally attached with the host, bridging the two cytoplasms (Fig. 2A and B, SI movie 3).In a few examples of proximal attachment, the attachment organelle seemingly did not form any cytoplasmic bridge between the two cells (Fig. 3E  and F).The attachment organelle appeared in direct contact with the host outer layer, and the intercellular sheathed filament was visibly inside the host and extended more than 400 nm into the host (Fig. 3F, SI movie 6).The biogenesis of this enormous attachment organelle must require considerable resources from YN1, suggesting that this structure is critical for host-DPANN interaction and resource exchange.

Metabolic pathway analysis of YN1 reveals key dependencies on the host cell
The metabolic pathways of YN1 were predicted on the basis of the genomic annotation (Fig. 4).As in the case of other Nanobdellota archaea, the YN1 genome lacks most of the genes in the central carbon metabolism pathways, including tricarboxylic acid (TCA) cycle; pentose phosphate pathway; respiration; adenosine triphosphate (ATP) synthesis' carbon fixation; and in the biosynthesis of amino acids, nucleotides, cofactors, vitamins, and lipids.Several genes related to glycolysis and gluconeogenesis were encoded in the YN1 genome; however, in glycolysis, no genes from glucose to glucose-6-phosphate (G6P), from fructose-6-phosphate (F6P) to fructose 1,6-bisphosphate (FBP), and from phosphoenolpyruvate (PEP) to pyruvate were found (Fig. 4).Similarly, in gluconeogenesis, no genes from glycerate 3-phosphate (3PG) to glyceraldehyde-3-phosphate (GAP) were found.In contrast, most of the genes involved in genetic information processing (e.g.transcription, translation, and replication) were encoded in the YN1 genome.Kato et al. recently revealed the cell surface structure of N. aerobiophila MJ1HA T and suggested it has putative S-layer proteins (MJ1_0160 and MJ1_0707) [10].The YN1 genome also encoded these proteins (YN1_5000 and YN1_2560) with high amino acid sequence similarity to MJ1_0160 (69.5%) and MJ1_0707 (61.7%), respectively.Genes involved in archaellum formation were encoded in the YN1 genome (FlaBDEHIJF).Consistent with the gene detection, the archaellum-like filaments were observed under the scanning electron microscopy (Fig. 2E).Genes involved in defence systems such as types I and IV of restriction and modification systems were identified.The YN1 genome did not contain the CRISPR-Cas system as there was no CRISPR repeats, and only the Cas4 gene was present.1).Grey indicates the lack of a gene.Coloured boxes indicate different cellular pathways.A grey box with red cross indicates pathways that are not present in YN1.The total number of genes is indicated in parentheses near the gene name.

Discussion
The nature of DPANN biology and life cycle is enigmatic.Based on the information of their reduced genome, DPANN must require external resources to proliferate, but how they obtain resources and utilize them remains underexplored due to the limited number of coculture systems established to date [4,8,10,20,52,53].Previous work has shown that some DPANN may opt for a parasitic lifestyle, instead of an ectosymbiotic relationship [52].Furthermore, it has often been suggested that proliferation of the DPANN would occur while attached to the host cell [2,54].Our results also show that the YN1 genome lacks most of the genes in the central carbon metabolism, suggesting that it ought to be an obligate symbiont of other microorganisms in its natural habitat.This was also confirmed by our cocultivation experiment, where YN1 did not grow without its host, indicating that YN1 is an obligate symbiont of S. ohwakuensis YN1HA.The growth of S. ohwakuensis YN1HA was significantly inhibited by the presence of YN1 (Fig. 1B), suggesting YN1HA-YN1 interaction might be parasitic.We have observed the narrow range of growth pH in YN1 (pH 2.0-3.0)compared with its host YN1HA (pH 2.0-5.5).A similar observation was reported for N. aerobiophila [41], suggesting that this characteristic is a common feature in Nanobdellota.The presence of most genes involved in genetic information processing and the observation of dividing YN1 cells while detached from their host (Fig. 2D) might imply that YN1 has the capability to undergo cell division away from the host cell, although we cannot exclude the possibility that dividing cells were detached from the host during sample processing.Since YN1 cells adhere firmly to the host (SI movie 1), YN1 cell division is likely to occur while they are attached onto the host.
Homologous genes for the recently suggested putative S-layer proteins of Nanobdellota [10] were found in the YN1 genome (YN1_5000 and YN1_2560), suggesting that these genes may be involved in S-layer formation.The presence of the putative S-layer proteins, having no sequence homology with that of YN1HA, could possibly explain why the cell wall structure of YN1 is different from that of YN1HA (Fig. 4).Additionally, the phenomenon of selective lipid component uptake from host species, as reported in three cultivated representatives of DPANN archaea [41,55,56], may also occur in YN1.This selective lipid acquisition could serve as an alternative explanation for the distinct cell wall structures observed between YN1 and its host YN1HA, potentially arising from differences in their respective lipid compositions.
The presence of genes involved in archaellum formation [57] (FlaBDEHIJF) and the observation of archaellum-like structures in our micrographs suggest the potential for motility in YN1 (Fig. 2E).Recently, using live cell imaging at the optimum growth temperature of 65 • C, Kato et al. showed that sole N. aerobiophila cells (without host attachment) were not motile despite the presence of archaellum-like filaments [10].On the contrary, Gaisin et al. reported that free cells of N. aerobiophila are motile [58].Under our experimental conditions, however, we failed to observe motility of YN1 under light microscopy.It is possible that specific required conditions were not met for motility during our experiment.The diameter of the archaellum-like filaments observed under the SEM was ∼18-21 nm (Fig. 2E), which is relatively larger than that of a typical archaellum of 10-14 nm [59].The relatively larger size of the archaellum filaments was probably caused by a plasma osmium coating (5 s), which could result in a coating thickness of a few nanometres on the surface of the samples.Given the coating thickness, the actual diameter of the archaellum-like filaments should be slightly smaller, which would be consistent with the diameter for a typical archaellum of 10-14 nm.Upon attaching to a host cell, the YN1 cells appeared to adhere firmly to the point of attachment (SI movie 1), suggesting a robust connection between cells of YN1 and YN1HA.The intracellular filament visualized through cryoET may play a role in tethering YN1 firmly to the host cell's surface, impeding its detachment while facilitating the acquisition of essential nutrients for its proliferation.
The presence of the solo Cas4 gene (with no CRISPR repeat) in the YN1 genome indicates that the CRISPR system does not work in YN1 as a defence system but may have a role for other purposes.It has been reported that most members of the DPANN superphylum encode solo-Cas4, and Cas4 could perform uncharacterized defence (possibly anti-defence) or repair functions in these microorganisms [60].
Our cocultivation experiments confirmed that YN1 grows on different strains of S. ohwakuensis (YN1HA and TA1 T ) but not on the other genera and species examined so far.(M.sedula DSM5348 T , M. hakonensis DSM7519 T , M. cuprina JCM15769 T , Sulfodiicoccus acidiphilus HS-1 T , Sulfuracidifex tepidarius JCM16833 T , Saf. metallicus DSM6482 T , Scl. solfataricus JCM8930 T , Scl. shibatae DSM5389 T , Scl. caldissimus HS-3 T , S. acidocaldarius DSM639 T ) [40,[61][62][63][64][65][66].This is consistent with the published literature, which has shown that all cultivated Nanobdellota archaea can only use a specific host [7,41,[67][68][69].For instance, a recently described N. aerobiophila MJ1 T showed that it only grows with the original host, M. sedula MJ1HA, but not with the type strain of M. sedula TH2 T as well as other several different genera and species [41].A similar observation was also reported for Nanoarchaeum equitans [69], the first cultivated species in the phylum Nanobdellota [7].In contrast, YN1 was able to grow with different strains of S. ohwakuensis, a unique observation in phylum Nanobdellota.This suggests that this taxon could have a narrow range of host specificity.The fact that the YN1 cells passed through the membrane filter grew with different strains of S. ohwakuensis possibly indicates that the free YN1 cells observed under microscope are living cells and capable of proliferation (Fig. 2D).The archaellum, typically known for its functions in chemotaxis and surface attachment [57], could in principle support free YN1 cells in locating and reattaching to host cells.
DPANN archaea are dependent on other organisms to provide resources for proliferation, previous work has shown that cell-cell connections can form between DPANN archaea and their host; including connecting filaments and cytoplasmic bridges [10,11,70].More recently, cryoET and sub-tomogram averaging were used to determine the in situ structure of proteinaceous tubes that bridge host and DPANN cells [71].We identified an attachment organelle in the YN1 cell envelope that is used to attach to the YN1HA host cell to form a cytoplasmic bridge.This organelle was comprised of a cone structure with a portal opening, and a filament extension that passes through the portal.When YN1 made intimate contact with the host (proximal attachment), the intercellular filament traversed inside the host and appeared to have a sheath.Similar cone-shaped protein complexes have been previously observed in other microorganisms, e.g.virusassociated pyramids (VAPs) were described in Sulfolobus islandicus [72].The VAP structure did not have a portal and, although the inner layer of the VAP formed a continuous density with the inner membrane, the VAP was not embedded within the cell envelope as reported in this study.Archaeal cones have also been described in Thermococcus kadokaraensis [49] and Pyrococcus furiosus [73] and were described as an organization center for archaellum assembly.However, these cone structures were cytoplasmic and without a portal.The attachment organelle of YN1 is embedded within the envelope, has filaments passing through the cone's portal, and is therefore structurally unique.Although it is possible that these previously described structures are evolutionarily linked to the cone structures observed here, the structures are considerably different and likely have completely different functions.In one instance, attachment organelles from different YN1 cells appear to be facing and possibly interacting with each other (Supplementary Fig. 4A and B, SI movie 7).However, the portals of the attachment organelles are not in plane and there is no density connecting them.Recently, the same attachment organelles were observed in an unrelated DPANN-host coculture system comprised of Nanobdella aerobiophila and its host M. sedula [58], which suggests that these structures may be ubiquitous across the Nanobdellota phylum and may interact with a range of hosts.
Our results show that YN1 Nanobdellota assembles an attachment organelle to form direct cytoplasmic bridge with the host cell.We have also described attachment organelles in YN1 cells that are detached from the host, either primed to form an interaction or a relic from a previous interaction.Finally, we described two further attachment stages where the filament, and filament sheath, of the attachment organelle appear to be inserted into the host cytoplasm, while the cone structure remained distally connected.We predict that YN1 has a biphasic lifestyle, an obligate ectosymbiont that uses its host to acquire the essential nutrients for growth, but once fuelled up, YN1 can detach and proliferate in isolation.We propose that the attachment organelle is used to latch onto the host cells and form an intimate interaction that results in the formation of a cytoplasmic bridge between the host cell and YN1, allowing for resource exchange.Many steps in this process remain unanswered, the protein(s) that comprise the attachment organelle are unknown, as is the mechanism by which they assemble this massive macromolecular complex.We identified different types of filaments (and filament bundles) present in the DPANN cytoplasm, but further investigation is required to confirm their identity and functions (Supplementary Fig. 4A-E).
Our work here established a new Nanobdellota coculture system and characterized the structural basis of host-DPANN interaction through a complex attachment organelle.The proteins that comprise this complex remain unknown and are the subject of future research, as is the method by which this enormous organelle assembles into the YN1 membrane, permeates the host's outer layer, and fuses with the cytoplasm.This research broadens our understanding of DPANN biology and reveals the versatile nature of their ectosymbiotic relationships.

Figure 1 .
Figure 1.YN1 and YN1HA coculture system, metagenomics, growth, and association.(A) Pie chart depicting the archaeal community structure of initial enrichment culture, based on the 16S rRNA amplicon analysis.(B) Growth curve analysis of pure YN1HA and in coculture with YN1 using 16S-rRNA quantification.The error bars indicate the standard deviation based on three culture replicates.(C) The maximum-likelihood tree was reconstructed using iqtree2 with the LG + F + R6 model.The tree was reconstructed based on 53 archaeal maker protein sequences obtained from GTDB-tk (release214).The scale bar represents 0.1 amino acid substitutions per sequence position.Bootstrap values are indicated at nodes.The numbers next to the nodes at the collapsed clades are the numbers of sequences used.(D) The maximum-likelihood tree was reconstructed using iqtree2 with the GTR + F + R3 model.The tree was reconstructed based on 16S rRNA gene sequences obtained from Genbank.The scale bar represents 0.1 nucleotide substitutions per sequence position.Bootstrap values are indicated at nodes.(E) FISH of YN1-YN1HA coculture.Red (small cells): Nanobdellota archaeon YN1, green (large cells): S. ohwakuensis YN1HA.

Figure 2 .
Figure 2. (A) 2D slice through a 3D tomogram of a representative YN1-YN1HA interaction showing YN1 and YN1HA cells.Pink arrows indicate example cone structures, and purple arrows indicate cytoplasmic filaments; zoomed-in YN1-YN1HA interaction is shown in the inset.(B) Segmentation analysis of the volume shown in (A), YN1 cytoplasmic membrane = light blue, YN1 outer layer = dark blue, YN1HA cytoplasmic membrane = light green, YN1HA outer layer = dark green, cones = pink, cytoplasmic filaments 1 and 2 are shown in purple and light yellow, respectively, intercellular filaments = red, and ribosomes = grey.(C) SEM analysis of an example coculture used in this work showing small YN1 cells attached to larger YN1HA cells.(D) SEM analysis showing example cell division events of YN1 cells separate from the host cell.(E) SEM analysis showing the presence of archaellum.Red, white, and green arrows indicate YN1HA cells, YN1 cells, and YN1 cell division sites, respectively.Yellow arrows indicate the location of example archaellum filaments.

Figure 4 .
Figure 4. Overview of the potential metabolic capability of YN-1.Metabolic pathways were constructed based on the annotation of predicted genes (Supplementary Table1).Grey indicates the lack of a gene.Coloured boxes indicate different cellular pathways.A grey box with red cross indicates pathways that are not present in YN1.The total number of genes is indicated in parentheses near the gene name.