Genome-resolved metagenomics reveals unexpected diversity and host range of Candidatus Lariskella (Rickettsiales: Midichloriaceae)

Abstract

The intracellular endosymbiont Candidatus Lariskella (Alphaproteobacteria, Candidatus Midichloriaceae) has been found across a wide diversity of terrestrial arthropods, including ticks, true bugs, beetles, fleas, wasps and moths. Despite its prevalence, little is known about the biology of Ca. Lariskella, nor do we grasp the full extent of its host range. Here, we report the first known occurrence of Ca. Lariskella infecting a population of free-living marine nematodes (Enoplida, Thoracostomopsideae). This novel nematode-infecting Ca. Lariskella was found to be closely related to insect-infecting strains, despite the drastic shift in both host taxonomy and habitat. TEM and FISH microscopy showed Ca. Lariskella is localized within both the nematode somatic cells and developing oocytes, confirming its status as a nematode endosymbiont and strongly suggesting maternal transmission. This finding led us to reassess the host range of Ca. Lariskella. We screened the SRA database for Ca. Lariskella sequences and performed genome-resolved metagenomics on SRA entries positive for Ca. Lariskella. We recovered 16 novel Ca. Lariskella metagenome-assembled genomes from SRA entries, including from novel hosts such as ants and treehoppers. However, we did not encounter further instances of Ca. Lariskella within nematodes or marine invertebrates, which we attribute to the relatively poor sampling of these groups. Overall, our findings illustrate the ability of Ca. Lariskella to infect both arthropods and nematodes as well as hosts from both terrestrial and marine environments.

Supplementary Information

The online version contains supplementary material available at 10.1186/s12864-026-12881-x.

Introduction

The Rickettsiales are a large, well-studied order of endosymbiotic bacteria within the Alphaproteobacteria. Due to their endosymbiotic nature, Rickettsiales typically possess small genomes (1-1.5 Mb) with streamlined gene contents [1]. Rickettsiales have been widely studied for decades, initially primarily for their role in diseases such as typhus (Rickettsia prowazekii), Rocky Mountain spotted fever (Rickettsia rickettsii), and anaplasmosis (Anaplasma phagocytophilum) [2]. However, diverse Rickettsiales have also garnered increasing evolutionary interest, as their obligate symbiotic lifestyle has resulted in a wide array of unique adaptations, such as supplementing host nutrition [3], aiding host defense [4], manipulating host apoptosis [5], or manipulating host sex determination [6].

Candidatus Midichloriaceae (hereafter Midichloriaceae) comprises a recently discovered family within the Rickettsiales order, which was formally described in 2013 [7]. Currently, the Midichloriaceae represent, by far, the least studied among the three families of Rickettsiales, in part due to its novelty, and in part because many Midichloriaceae infect poorly sampled host taxa such as protists [8]. Despite being less studied, Midichloriaceae have been discovered across a vast phylogenetic range of hosts such as in multiple protist supergroups [9–12], placozoans [13, 14], corals [15], multiple arthropods [16–21], as well as in secondary infections of mammals [22–24]. Additionally, Midichloriaceae have been found to exhibit several unique evolutionary traits, such as being the only group in Rickettsiales known to maintain flagellar genes [8, 25, 26], and including the only known bacterial species that can reside within the host mitochondria [21].

Midichloriaceae appear to primarily infect aquatic hosts. However, the family contains two ‘terrestrial’ genera: Candidatus Lariskella (hereafter Lariskella) and Candidatus Midichloria (hereafter Midichloria). Both Lariskella and Midichloria have primarily been studied in ticks, where they appear to be widespread [17, 21, 27, 28] and are hypothesized to play a nutritional role [29]. Notably, both Lariskella and Midichloria are transmitted to mammals during the tick’s blood meal, though it has not been confirmed whether either bacterium can replicate within mammals or cause disease [22–24].

While Midichloria is only known to infect ticks, Lariskella is known to infect a far broader range of arthropod hosts. It has been reported in ticks [17, 27, 28, 30], stinkbugs [20], leaf-footed bugs [31], weevils [18], psyllids [19], beetles [16], and fleas [32]. In addition to these published reports, circularized Lariskella genomes have been sequenced from wasps (NCBI Refseq GCF_964020185.1) and moths (NCBI Refseq GCF_964019805.1).

Here, we expand the Lariskella host range, describing Lariskella as an endosymbiont of a marine nematode (Enoplida, Thoracostomopsideae) using a combination of genomics and microscopy. In addition, we screened the SRA database for Lariskella, identifying several further novel hosts, including ants and treehoppers, and reconstructed metagenome assembled genomes (MAGs) for multiple Lariskella strains. The screening of the SRA did not reveal any further nematode or marine hosts of Lariskella, which we hypothesize is due to the limited sampling for these groups. The phylogeny of the novel Lariskella MAGs also demonstrated that the nematode-infecting Lariskella is closely related to insect-infecting lineages, despite the shift in host and habitat. We hypothesize that this pattern may indicate a relatively recent acquisition by the nematode, supporting the ability of Lariskella for host switches.

Materials and methods

Nematode sampling

From March 2023 until June 2024, sediment samples were taken from the Okinawan coast (26.4723631548844, 127.83168712213792). Nematodes were extracted from the sediment through a magnesium chloride decanter method: 100 mL of 7.5% MgCl2 was added to approximately 200 mL of sediment. Sediment was stirred with a spoon and left to sit for 10 min, after which it was stirred again, and the solution was sieved through a 32 μm sieve. The contents of the sieve were washed into a Petri dish using sand-filtered seawater. Nematodes were then picked from the Petri dish using an Irwin loop under a dissecting microscope (Olympus SZX16). Nematodes picked for microscopy were washed once in filter-sterilized (0.22 μm) seawater to remove debris and then transferred directly to the appropriate fixative. All other nematodes were cleared of surface contaminants by a triple serial transfer in filter-sterilized seawater using sterile pipette tips. Cleaned nematodes were then frozen at − 70°C awaiting processing.

Illumina sequencing

For 15 nematodes, shotgun metagenomes were obtained using Illumina sequencing. DNA was extracted from the nematodes using the MasterPure Complete DNA and RNA purification kit (Thermo Fisher Scientific). The kit was used following the manufacturer’s instructions, with the exception of a prolonged (2 h) proteinase K incubation to facilitate full lysis of the nematode. Illumina libraries were prepared using the NEBNext Ultra II PCR plus kit (New England Biolabs) and sequenced on the SP flow cell of the NovaSeq 6000 platform with 150 bp paired-end reads. The generated reads were thereafter quality checked using Fastqc v0.11.9 (https:/www.bioinformatics.babraham.ac.uk/projects/fastqc/).

Lariskella identity and diagnostic PCR

The initial identification of Lariskella was performed by 16S rRNA gene reconstruction using phyloFlash. To experimentally confirm the presence of Lariskella, a diagnostic PCR with Lariskella-specific primers was performed in 10 nematodes. We amplified the Lariskella rrl-rrf intergenic spacer using the Lariskella-specific Mtz-5S-F and Mtz-23S-R primers [33] and Q5 high-fidelity DNA polymerase (New England Biolabs), with the following thermocycling conditions: initial denaturation at 94 °C for 4 min; then 30 cycles of 94 °C for 1 min, 62 °C for 1 min, 72 °C for 80 s; followed by a final elongation at 72 °C for 4 min. To confirm that the bands were the intended products, the obtained bands were Sanger sequenced as described above. Resulting Sanger sequences were compared against the Lariskella rRNA operon obtained during Illumina sequencing using EMBOSS Needle [34], to ensure the PCR product originated from the Lariskella symbiont.

Lariskella genome binning

Based on the 16S and 18S rRNA gene sequences, both the host (N = 15) and Lariskella (N = 4) were identical across the shotgun-sequenced nematode samples. The reads for these samples were thereafter trimmed using Fastp v0.23.2 [35] and co-assembled using Megahit v1.2.9 [36]. The coverage for this assembly was then calculated by Bowtie2 v2.4.2 [37], and sorted, compressed, and indexed using SAMtools v1.12 [38]. The assembly was then binned using Metabat2 v2.12.1 [39], CONCOCT v1.1.0 [40], and SemiBin2 v2.0.2 [41], and the resulting bins were refined with DASTool v1.1.6 [42]. It was then assessed which of the metagenome assembled genomes (MAGs) corresponded to Lariskella through a threefold method: first, a MEGAN v6.23.3 [43] lowest common ancestor assignment analysis was run to confirm the approximate composition of the MAG. Second, the 16S rRNA genes in the MAGs were predicted using barrnap v0.9 (https://github.com/tseemann/barrnap), and aligned by BLASTn v2.10.0 [44] against the 16S rRNA gene reconstructed by phyloFlash. Finally, the MAG was aligned with the Mummer nucmer algorithm v3.23 [45] against the complete Lariskella reference genomes obtained from NCBI (GCF964019805.1 and GCF964020185.1). For all MAGs of Lariskella origin, the best MAG was chosen based on completeness and duplication scores inferred by CheckM2 v1.0.2 [46]. The best quality MAG was found to originate from CONCOCT. CONCOCT filters out contigs smaller than 1000 bp by default, and these are thus not present within the Lariskella MAG.

SRA screening and Lariskella phylogeny

The NCBI SRA database was screened for Lariskella by querying the STAT assignments of SRA entries for Lariskella using bigQUERY on Google Cloud. Entries that had at least 1000 reads assigned as Lariskella were downloaded and checked for Lariskella 16S rRNA gene sequences using PhyloFlash. Entries for which a Lariskella 16S rRNA gene was recovered were binned using the method described above. The MAGs corresponding to Lariskella were then identified using GTDB-Tk v2.4.0 [47], and their quality determined using CheckM2.

The phylogenetic positions of the nematode MAG, the SRA Lariskella MAGs, and the two publicly available Lariskella genomes (GCF_964020185.1 and GCF_964019805.1) were inferred from a multigene phylogenetic tree. Prokka v1.14.6 [48] was used to predict protein-coding sequences, which were piped into Orthofinder v2.5.4 [49] to find single-copy orthologs shared across the 25 genomes. The protein sequences of the 114 shared single-copy orthologs were aligned using MAFFT v7.505 [50], set to -localpair and with 1,000 iterations. The alignment was then trimmed using trimAl v1.4.rev22 [51] set to automated1, and concatenated using a Python script. A maximum likelihood phylogenetic tree was reconstructed from the concatenated alignment using IQ-TREE2 v2.0.7 [52], with 1,000 bootstraps. IQ-TREE2 model finder chose JTTDCMut+F+R3 as the best model based on the Bayesian Information Criterion.

Host 18S rRNA gene sequencing and phylogeny

The 18S rRNA gene sequences of the nematode host were assembled from the metagenomes and approximately identified using phyloFlash v3.4 [53]. To determine the host’s exact phylogenetic position, an 18S rRNA gene tree was reconstructed. 18S rRNA gene sequences of nematodes from Enoplida were downloaded from SILVA [54], with Triplonchida 18S rRNA gene sequences used as an outgroup. The downloaded sequences were aligned with the metagenome-reconstructed 18S rRNA gene sequence using MAFFT set to a maximum of 1,000 iterations, and using the -localpair argument. A phylogenetic tree was reconstructed from the alignment using IQ-TREE2 set to 1,000 bootstraps using the GTR + F + R4 model selected by the IQ-TREE2 modelfinder based on the Bayesian Information Criterion. As a comparison, trees were also reconstructed using the GTR + G + I model frequently used in nematode 18S rRNA gene phylogeny literature [55, 56] and both trees had identical topologies with only minor differences in branch lengths (Supplemental figure S1). We confirmed the similarity of the 18S rRNA sequences using BLASTn megablast on the NCBI website (3-6-2025) [44].

Double labelling of oligonucleotide probe for fluorescence in situ hybridization (DOPE-FISH) microscopy

Multiple nematodes (5–20 individuals per slide) were picked and placed in a 1 mL drop of filter-sterilized (0.22 µm) sea water. They were then transferred to a 1.5 mL tube containing 100 µl of 1:1 glacial acetic acid: ethanol solution, where they were fixed for 10 min. Fixed nematodes were washed in 100% ethanol two times for 5 min, and incubated in 1:1 methanol: PBS-T (1X PBS buffer containing 0.1% Tween 20) for 10 min. Afterwards, they were incubated in 1% paraformaldehyde (Nacalai) PBS-T, which was washed away with two PBS-T washes of 2 min each. Nematodes were prepared for hybridization with three washes in hybridization buffer (0.9 M NaCl, 0.02 M Tris-HCl, 0.01% sodium dodecyl sulphate (SDS, Sigma-Aldrich) + 30% formamide; Wako) of 5 min each. The nematodes were then hybridized in 100 µL of hybridization buffer containing DAPI DNA stain (0.01 µg/µL; Roche), and 0.1 µM of Lariskella-specific probe (JBioS) overnight at 42°C. The Lariskella-specific probe (CTGTGATAGTCCGGCCGAAC) was purified through high performance liquid chromatography purification. This probe was then labelled at both the 5’ and 3’ ends with Alexa Fluor 647. The DNA sequence of the probe was determined to be fully Lariskella-specific using SILVA’s TestProbe v3.0 function [57]. After hybridization, excess DAPI and probe were washed away with three PBS-T washes of 10 min. Nematodes were suspended in MilliQ water and pipetted onto a glass slide. After removing excess water, the nematodes were mounted using ProLong Diamond Antifade Mountant (Thermo Fisher Scientific) and imaged using a confocal fluorescence microscope (Microscope: Nikon A1, detector: x2 MA PMT, x2 GaAsp, x1 T-PMT, lasers: 404.6 nm and 637.8 nm).

Transmission electron microscopy (TEM)

Five adult female nematodes were picked from sediment samples. These nematodes were fixed by briefly (2 min) submerging in the fixative (2.5% glutaraldehyde; Fujifilm Wako and 2% paraformaldehyde; Fujifilm Wako in 0.1 M HEPES; Nacalai). The anterior of the nematode (from the mouth to slightly behind the end of the pharynx) was dissected from each nematode. The posterior of the nematodes was incubated in the fixing solution overnight at 4 °C, awaiting TEM post-fixation steps. The processed nematodes were assessed for Lariskella infection by performing the Lariskella diagnostic PCR described above on the dissected pharynges. One nematode was found to be Lariskella-positive, and its posterior was prepared for TEM.

The posterior of the Lariskella-positive nematode was washed with 0.1 M HEPES for 5 min three times, and post-fixed with 2% osmium tetroxide for 1 h at room temperature. After post-fixation, the nematode was en bloc stained with 2% uranyl acetate overnight at 4 °C. The nematode was dehydrated by serial transfer through increasing concentrations of ethanol (30%, 50%, 70%, 90%, 100%), incubating for 10 min in each, and two times 10 min in 100% ethanol. This was followed by a final dehydration using 100% acetone for two times 10 min. Low viscosity resin (Agar Scientific, catalog no. R1078) was then infiltrated by replacing with 2:1 (vol: vol) of acetone: resin for 2 h at room temperature while rotating in an TAITEC Rotator RT-50. After 100% of the resin was replaced, the resin was infiltrated under a 1-hour vacuum and incubated overnight at room temperature. The resin was then polymerized for 48 h at 60°C. The resin block was trimmed using glass knives, and ultrathin sections  (50–60 nm) were cut using a Leica Ultra-microtome EM (UC7). Ultrathin sections were collected onto the 0.5% formvar (SPI-CHEM) coated copper one-slot grids (Nisshin-EM). The air-dried grids were then examined using a JEOL 1400 Flash transmission electron microscope.

Sex-specific infection rate experiment

Lariskella diagnostic PCR was performed on adult nematodes of known sex to determine per-sex infection rates. Nematodes were collected, and their sex was determined based on morphology using an inverted microscope (Nikon Eclipse Ti2-E). Due to the difficulty of determining sex in juveniles, only adults were considered. Adults with a sclerotized gubernaculum were determined as males. The gubernacula had clear autofluorescence under 488 nm light, which was used to more easily identify adult males. Adults lacking a gubernaculum, but containing ovaries with visible eggs, were considered as fully developed adult females. No hermaphrodites were observed. The presence of Lariskella was then confirmed in the sexed nematodes using the Lariskella diagnostic PCR described above. In total, 35 sexed nematodes were tested with PCR (23 females, 12 males). Furthermore, where possible, adult nematodes from FISH slides were sexed as well, and assessed visually for Lariskella signal. In total, 23 sexed nematodes were assessed using FISH (16 females, 7 males). To assess whether male and female infection status followed the same distribution, chi-square goodness of fit tests were performed on the PCR and microscopy results individually.

Analysis of reproductive manipulation gene presence in Lariskella

To identify potential reproductive manipulation genes in the nematode-infecting Lariskella, we compiled a curated reference gene set (Supplemental sequences) including putative male-killing genes (wmk, oscar, Spaid), cytoplasmic incompatibility genes (cifA and cifB), and parthogenesis induction genes (pifB and piff). We generated custom BLAST databases containing these reference genes. Protein-coding genes from Lariskella were predicted using Prokka and queried against this database using BLASTp (E-value ≤ 1e^− 5, minimum sequence identity ≥ 30%). Only wmk-like genes resulted in BLAST hits above 50% identity.

The protein domains present in the inferred wmk-like orthologs were identified using InterProScan v5.60-92.0 [58]. InterProScan reported only 1 domain shared among all wmk orthologs, which was annotated as a Cro/C1-type DNA-binding domain. Alphafold2 v2.1.1 [59] was used to infer the protein structures, which were visualized in iCn3D [60]. Structural alignments were performed using TM-align. These structural alignments showed a helix-turn-helix structure shared among all identified orthologs, the position of which corresponded to the Cro/C1-type DNA-binding domains reported by InterProScan.

Functional annotation of Lariskella genomes

The Lariskella MAGs and the two reference Lariskella genomes were annotated using eggNOG-mapper v2.1.12 [61] in genome mode, using the bundled MMseqs2 BLASTx as the gene prediction method, and the eggNOG bacterial database. The eggNOG-mapper COG assignments were extracted and counted using a Python script.

ANI was assessed using ANIclustermap (https://github.com/moshi4/ANIclustermap) using the default settings, including ANI calculation by fastANI [62].

To infer the potential nutritional contribution by Lariskella, the completion of metabolic pathways synthesizing nutrients essential for nematodes was assessed using the genome data. Nematode essential nutrients were based on Caenorhabditis elegans as a guide [63]. The KEGG KO numbers were extracted from the eggNOG-mapper annotation using a Python script, and used to assess the completion of the KEGG modules corresponding to the synthesis of essential nutrients in KEGG mapper reconstruct [64, 65]. If there were multiple redundant KEGG modules, only the most complete bacterial pathway was considered.

Results

Lariskella identity, genome and phylogeny

The Lariskella infection within this nematode population was first discovered by the presence of a Lariskella arthropodarum-related (99.6% identity) 16S rRNA gene in the nematode metagenome. Diagnostic PCRs with Lariskella-specific intergenic primers were performed to experimentally confirm the presence of Lariskella. In total (including the sex-specific infection rate experiment), diagnostic PCRs were performed on 50 individuals, of which ten gave visible bands indicating Lariskella infection.

Using binning, a MAG of the suspected Lariskella was obtained from 15 co-assembled nematode metagenomes. This MAG was compared against two publicly available Lariskella genomes. The MAG is fragmented (321 contigs) but is highly similar to the references in assembly size (1.5 Mb), CheckM2 completeness score (96.9%), and GC content (35%) (Supplemental Table S1). The MAG was aligned against the references and showed a high degree of nucleotide-level similarity, further confirming its identity as Lariskella (Supplemental Figure S2).

To assess the phylogenetic position of this novel Lariskella lineage, we reconstructed a 114-gene multigene phylogenetic tree using 2 NCBI Lariskella references and 16 MAGs reconstructed from SRA data (Fig. 1A). In this tree, the nematode Lariskella is nested within insect-infecting lineages, clustering most closely to a Lariskella reference genome infecting Lepidoptera.

Fig. 1.

A) 114-gene multigene phylogenetic tree of the genus Lariskella, showing the position of the assembled Lariskella MAGs, named based on their host species. The tree was inferred by maximum-likelihood based on the JTTDCMut + F+R3 model, and rooted with Cytobacter and Midichloria species as outgroups. B) Cropped phylogenetic tree based on nematode 18 S rRNA gene sequences, demonstrating that the nematode host belongs to the Thoracostomposidae. The tree was inferred by maximum-likelihood based on the GTR + F+R4 model. The full tree is shown in Supplemental Figure S1

We also reconstructed a phylogenetic tree based on the 18S rRNA gene of the nematode host (Fig. 1B). While the host 18S rRNA gene did not match exactly with any known reference, the phylogenetic tree places it within the family Thoracostomopsidae of the order Enoplida. The nematode clusters closest to a Mesacanthoides species with 97% sequence identity.

DOPE-FISH microscopy and TEM

The localization of the novel Lariskella strain was visualized using DOPE-FISH microscopy (Fig. 2A-C). The localization of the Lariskella-specific probe shows Lariskella has a patchy distribution of intracellular infections throughout the somatic tissue of the nematode host. The ovaries of the nematode clearly deviate from this patchy distribution, and instead display highly abundant infections of Lariskella in both the maternal ovarian tissue and the developing oocytes.

Fig. 2.

A to C) DOPE-FISH microscopy images of a whole-mounted gravid female nematode. Pink indicates the Lariskella-specific probe. Blue indicates DAPI, which stains all DNA and visualizes host nuclei. A) Whole nematode. B) Zoomed image of one of the paired ovaries, including infected oocytes C) Zoomed image of somatic tissue, including gut epithelium and cuticle. D) and E) TEM micrographs of a nematode oocyte. D) Electron micrograph showing the inside of an infected oocyte. E) Electron micrograph within an oocyte, zoomed in on a Lariskella cell

To further confirm the intracellular nature of the nematode Lariskella, TEM micrographs of various nematode tissues were inspected. The micrographs show the presence of bacteria-like structures within the oocytes (Fig. 2D-E), as well as within the somatic cells (Supplemental Figure S3).

Sex-specific infection rates

Diagnostic PCR showed an infection rate of 26% for female nematodes (6/23 infected) (Fig. 3). Adult males were found to be free of infection, with none of the tested males (N = 12) showing bands during the PCR. These results matched the DOPE-FISH results, where no Lariskella signal was observed in any of the processed males (N = 7). In contrast, 7 out of 16 adult females showed a clear Lariskella signal.

Fig. 3.

Bar plot displaying the infection status of tested nematodes. Sexed nematodes were examined for Lariskella infection using diagnostic PCR (N = 35), and DOPE-FISH microscopy (N = 23). For both the PCR and microscopy experiments a chi-square goodness of fitness test confirmed a significant difference in the female and male infection status (PCR: p = 0.0175, microscopy: p = 0.0196)

SRA screening

To determine the prevalence of Lariskella infection in nematode or marine hosts, we screened the complete NCBI SRA database for Lariskella sequences (Supplemental Table S2). We found 305 entries that contained at least 1000 reads that were identified as potential Lariskella. For 33 of these entries, a Lariskella 16S rRNA gene was recovered from the metagenome. From 16 entries, high-quality (> 80% completion) Lariskella MAGs were recovered (MAG statistics in Supplemental Table S1). We assessed the COG counts, completion, and genome sizes of these MAGs (Fig. 4). None of the SRA entries that yielded Lariskella MAGs or 16S rRNA gene sequences were of aquatic or nematode origin. Instead, the Lariskella MAGs originated from a variety of terrestrial arthropods: treehoppers (Horiola picta, Vanduzea arquata), ants (Temnothorax longispinosus), ticks (Ixodes persulcatus), and beetles (Doliops duodecimpunctatus, Sambus kanssuensis). Of these, treehoppers and ants were not previously known to harbor Lariskella.

Fig. 4.

Heatmap showing the COG counts (logarithmically scaled) found per Lariskella genome or MAG, named according to their host species. The bar plot shows the genome size and Checkm2 completion per assembly. Genomes were clustered based on their phylogeny described in Fig. 1

Notably, the diversity of Lariskella MAGs appears to constitute several distinct species of the Lariskella genus. According to an ANI analysis (Supplemental Figure S4), the Lariskella strains within ticks comprise a distinct species from the other hosts. Arguably, the Lariskella within Temnothorax longispinosus might also constitute its own species. Lariskella from treehoppers, moths and the marine nematode appear to form a single species. The MAGs obtained from beetles were identified as Lariskella by GTDB-Tk¸but have less than 80% identity with any other Lariskella genome. Thus, these beetle-derived lineages appear to instead comprise a sister genus, phylogenetically close but distinct from Lariskella.

Lariskella metabolic capacity

The Lariskella genomes encode complete pathways required to synthesize the essential vitamins pyridoxine (B6) and biotin (B7) (Fig. 5). However, none of the pathways for essential amino acids were fully encoded by any of the Lariskella genomes. The pathways for lysine, heme, and tetrahydrofolate are nearly complete, and could potentially be completed by uncharacterized enzymes or through substrate exchange with the host. The nematode-infecting Lariskella appears not to deviate from the metabolic profiles of other Lariskella. The Lariskella-allied lineages infecting beetles do slightly deviate, as these appear to encode more complete pathways for nicotinamide and arginine than found for the Lariskella lineages.

Fig. 5.

Heatmap displaying the completion of biosynthesis pathways in the two publicly available Lariskella genomes and Lariskella MAGs generated in this study. Genomes were clustered based on their phylogeny found in Fig. 1. Pathway completion was found by calculating KEGG module completion based on K-terms generated by eggnog-mapper annotation. The shown pathways correspond to nutrients essential to nematodes, based on findings in Caenorhabditis elegans

Discussion

In this study, a wild population of marine nematodes was confirmed to be infected by Lariskella through multiple experiments. The presence and identity of the Lariskella symbiont were confirmed through metagenomics, diagnostic PCR, and DOPE-FISH microscopy. In addition, the intracellular localization of the nematode-infecting Lariskella was confirmed using DOPE-FISH microscopy and TEM. All results point towards the Lariskella being a vertically transmitted intracellular symbiont of the nematode host. This novel host represents the first known nematode host of the genus Lariskella, as well as the first known nematode host for bacteria of the family Midichloriaceae. Additionally, this marine nematode represents the first known aquatic host found for Lariskella, which was previously considered exclusively terrestrial [8].

The phylogeny suggests a recent infection of nematodes

It was unexpected for the nematode-infecting Lariskella to be so closely related to the insect-infecting Lariskella. Nematodes occur in both marine and terrestrial environments, and throughout their evolution have frequently transitioned between these two habitats [66]. Therefore, we initially suspected that marine nematodes were ancient hosts of Lariskella, potentially representing a link from the ancestrally marine hosts of Midichloriaceae to the primarily terrestrial hosts of Lariskella. However, the phylogenetic tree contradicts this notion, with the nematode-infecting Lariskella instead nested among the insect-infecting Lariskella lineages. This nested placement of Lariskella, combined with its high nucleotide-level similarity to insect-infecting Lariskella strains, suggests the nematode population was likely infected by Lariskella relatively recently. Since the Thoracostomopsidae are primarily predators [67], and as predation can facilitate the horizontal transmission of endosymbionts [68], we suspect the nematode population might have been infected through the ingestion of infected arthropods.

Microscopy confirms an intracellular localization

Microscopy confirmed the identity of the symbiont as Lariskella, with clear binding of the Lariskella-specific probe. The bacterial morphology in the TEM micrographs also resembled the Lariskella observed in insect cells [20], though the nematode-infecting Lariskella are notably more compact and less filamentous. Additionally, the microscopy demonstrated the intracellular localization of the Lariskella symbionts, confirming Lariskella represents a true intracellular infection of the nematodes and did not originate from external contamination. Furthermore, both the TEM and DOPE-FISH microscopy confirmed Lariskella colonized the developing oocytes within the gravid females, strongly suggesting maternal transmission of Lariskella. Additionally, all oocytes within infected females displayed abundant Lariskella signal, suggesting a high rate of transmission. However, despite this suspected transmission efficiency, the overall Lariskella infection rate in nematodes was surprisingly low. Most notably, male nematodes seemed to be entirely free of infection during DOPE-FISH microscopy (Supplemental Figure S5). While some Lariskella probe signal could be observed within males (Supplemental Figure S6), these signals did not form clear cell-shaped structures or co-localized with the DAPI stain, and thus most likely originated from incomplete washing of unbound probe. This was confirmed with diagnostic PCR which also failed to detect Lariskella within male nematodes.

The observed Lariskella infection rate might suggest reproductive manipulation

The overall infection rate of the nematode was found to be approximately 20%. This is far lower than infection rates typically reported for Lariskella, with studies reporting Lariskella frequencies around 47%-52% in ticks [33, 69, 70], 77%-100% in Nysius stinkbugs [20], 66%-100% in leaf-footed bugs [31], and 100% in psyllids [19]. Additionally, our results suggest that all male nematodes were free of Lariskella infection. A reduced prevalence of Lariskella in males is a known phenomenon. Female ticks have abundant Lariskella infections, while males only harbor small to negligible numbers of Lariskella [17, 28, 71]. Becker et al. (2023) suggested that Lariskella might be maternally transferred to all offspring, but that the Lariskella titer diminishes in males over the course of their development [28], similar to what has been observed for the related bacteria Midichloria mitochondrii [72]. However, while low in abundance, male ticks still harbor Lariskella, and Lariskella sequences have been detected in males by both qPCR [33] and amplicon sequencing [28, 71]. In contrast, we failed to detect any Lariskella within males using diagnostic PCR or DOPE-FISH microscopy. Notably, the DOPE-FISH microscopy was sensitive enough to visualize individual Lariskella cells and still failed to detect Lariskella infection in males.

Our observed lack of male infection led us to cautiously hypothesize that Lariskella might be a reproductive parasite in nematodes. Lariskella is a reproductive manipulator in insects, where it has recently been demonstrated to be capable of inducing cytoplasmic incompatibility [31]. However, cytoplasmic incompatibility relies on infected males, and our results suggest an absence of infected males in the nematode population. Instead, the infection pattern observed in nematodes appears indicative of a male-killing or feminizing effect, as this would explain both the lack of male infection and the lower male prevalence. Additionally, a feminizing or male-killing effect would be more in line with the observed infection rate. As mentioned above, Lariskella tends to occur at higher frequencies in arthropods, which is considered typical of symbionts inducing cytoplasmic incompatibility [73]. In contrast, feminizing or male-killing symbionts are typically maintained at lower frequencies [6], to allow for sufficient male births to sustain the population. Our observed female infection rate (26%) appears in line with those observed for feminizing symbionts (15%-42%) [74–76] or male-killing symbionts (< 40%) [6]. We also found potential orthologs of putative male-killing genes within the analyzed Lariskella genomes (Supplemental_sequences). The inferred orthology was found to be based on a shared Cro/C1-type helix-turn-helix DNA-binding domain, which is commonly encountered in putative male-killing genes (Supplemental Figure S7) [77–79]. However, attempts to culture the nematode have failed, and we thus lack the ability to experimentally test for male-killing or feminization.

Expanded host range of Lariskella

The novel finding of a marine nematode host contrasts with the known host range of Lariskella, which is limited to terrestrial arthropods. To assess whether this novel host represents a wider pattern, we performed an exhaustive screening of the literature and the NCBI SRA database, compiling a comprehensive list of currently known Lariskella hosts (Supplemental Table S3). No other nematode or marine hosts were found. This might indicate that Lariskella infection of such hosts is rare. Alternatively, the failure to find Lariskella could be attributed to an undersampling of marine invertebrates, and particularly nematodes. Nematodes numbered an impressive 95,641 entries in the SRA database (on October 10th, 2025). However, the sampling represents only a fraction of nematode diversity, with only 10 species accounting for over 80% of all entries, and Caenorhabditis elegans representing over two thirds of all nematode entries.

Regardless of its prevalence, the capability to infect nematodes represents an interesting parallel between Lariskella and Wolbachia. Wolbachia is a group of highly studied Alphaproteobacteria famous for their success as endosymbionts, and it is estimated to infect 50% of all arthropods [80]. Wolbachia are known to infect arachnids and insects [80], as well as nematodes [81, 82], all of which are now known to also harbor Lariskella (Supplemental Figure S8). The nested placement of nematode-infecting Lariskella among insect-infecting strains also represents a parallel with Wolbachia, as several studies indicate that nematode-infecting Wolbachia supergroups are also nested among insect-infecting supergroups [83, 84]. It warrants mention that Wolbachia are also known to infect Isopoda [85], for which no Lariskella have been discovered to date. However, like nematodes, isopod diversity is undersampled (4735 SRA entries), and we suspect further sampling of isopods might reveal Lariskella infection, particularly as our findings illustrate the ability of Lariskella to infect marine hosts. If this is indeed true, it would raise interesting questions regarding the evolution of Rickettsiales and host choice, as Wolbachia and Lariskella are only distantly related, yet seem to have independently converged on a similar host range.

Lariskella is not an obligate nutritional symbiont in nematodes

Lariskella has been hypothesized to play a nutritional role in ticks, acting as a potential source of B vitamins [29]. We confirmed the ability of Lariskella to synthesize several essential B vitamins, which could potentially support host nutrition. However, as Lariskella is absent in the majority of the nematode population, we can conclude that the nematode host does not rely on Lariskella to source essential B vitamins. The supplementation of B vitamins might instead be a method by which Lariskella could offset the fitness cost of infection, allowing it to spread in the host population more effectively, as has been shown in Wolbachia [86]. Though, notably Lariskella presence or absence has been shown to not affect the fitness of an insect host [31], suggesting Lariskella might not significantly contribute to nutrition.

Conclusions

Here, we confirmed that Lariskella is an endosymbiont of a marine nematode species. This illustrates that Lariskella can infect nematodes and marine hosts, implying a vastly larger host range than currently described. However, screening of the SRA database did not reveal any further nematode or marine hosts. We attribute this to an undersampling of nematodes and marine invertebrates, although we cannot rule out the possibility that the infection in nematodes could be a rare exception.

Authors’ contributions

AH led the sampling, data analysis, and manuscript writing. YM aided in TEM microscopy, while PP assisted in the reproductive manipulation gene analyses. The principal investigator FH oversaw the project, and assisted in data interpretation and manuscript writing.

Funding

The work of FH’s group is partly supported by a Research Grant from HFSP (RGEC29/2024;DOI:10.52044/HFSP.RGEC292024.pc.gr.194160).

Data availability

The nematode data used in this study are publicly available on NCBI as BioProject PRJNA1272136, SRA-derived *Lariskella* MAGs are available from FigShare: 10.6084/m9.figshare.30943895. Commands and scripts used for the analyses are available from [https://github.com/ECBSU/ECBSU\_manuscripts\_code/tree/main/Hagenbeek\_nematode\_lariskella].

Declarations

Ethics approval and consent to participate

Not applicable.

Consent for publication

Not applicable.

Competing interests

The authors declare no competing interests.