The Chromosome-level Genome of the Ctenophore Mnemiopsis leidyi A. Agassiz, 1865 Reveals a Unique Immune Gene Repertoire

Abstract

Ctenophora are basal marine metazoans, the sister group of all other animals. Mnemiopsis leidyi is one of the most successful invasive species worldwide with intense ecological and evolutionary research interest. Here, we generated a chromosome-level genome assembly of M. leidyi with a focus on its immune gene repertoire. The genome was 247.97 Mb, with N50 16.84 Mb, and 84.7% completeness. Its karyotype was 13 chromosomes. In this genome and that of two other ctenophores, Bolinopsis microptera and Hormiphora californensis, we detected a high number of protein domains related to potential immune receptors. Among those, proteins containing Toll/interleukin-1 (TIR2) domain, NACHT domain, Scavenger Receptor Cystein-Rich (SRCR) domain, or C-type Lectin domain (CTLD) were abundant and presented unique domain architectures in M. leidyi. M. leidyi seems to lack bona fide Toll-like Receptors, but it does possess a repertoire of 15 TIR2 domain-containing genes. Besides, we detected a bona fide NOD-like receptor and 38 NACHT domain-containing genes. In order to verify the function of those domain-containing genes, we exposed M. leidyi to the pathogen Vibrio coralliilyticus. Among the differentially expressed genes, we identified potential immune receptors, including four TIR2 domain-containing genes, all of which were upregulated in response to pathogen exposure. To conclude, many common immune receptor domains, highly conserved across metazoans, are already present in Ctenophora. These domains have large expansions and unique architectures in M. leidyi, findings consistent with the basal evolutionary position of this group, but still might have conserved functions in immunity and host–microbe interaction.

Significance.

Our knowledge of immunity in basal metazoans is still in its infancy. Ctenophores may provide unique insights, being the sister group to all other animals. Here, we generated a chromosome-level genome and explored the immune repertoire of the ecologically important species Mnemiopsis leidyi along with two other recently-characterized ctenophore genomes. We find that ctenophores feature many protein domains known to be involved in invertebrate immunity, albeit with large expansions and unique architectures. Even unique, some of these genes had also a functional role during bacterial pathogen exposure. These results provide first insights into how the immune repertoire in M. leidyi and other ctenophores is organized, advancing our knowledge on the evolution of the immune system in Metazoa.

Introduction

Ctenophora, or comb jellies, are a unique phylum of exclusively marine invertebrates. They feature a mostly pelagic life-style as zooplankton feeders (only very few benthic species exist), and are distributed across all oceans, temperature gradients and depths (Tamm 2014; Dunn et al. 2015). Ctenophores are the earliest diverging phyla in the evolution of metazoans, containing around 200 described species and likely many more undescribed to date (Tamm 2014; Dunn et al. 2015). The lively debate on whether or not sponges or comb jellies appeared first in the tree of life (Ryan et al. 2013; Moroz et al. 2014; Halanych 2015; Pisani et al. 2015; Whelan et al. 2015; Feuda et al. 2017; Kapli and Telford 2020) was recently settled. Schultz et al. (2023) demonstrated that ctenophores are the sister group to all other metazoan phyla based on chromosomal synteny. This is in line with many other comparative studies of the last decade that revealed that this phylum is unique in terms of many traits such as neural system and locomotion as well as cell type organisation (Martindale et al. 2002; Jager et al. 2010; Moroz et al. 2014). Up to date, studies in ctenophores have shed light on the evolutionary origins of bioluminescence, the evolution of the neural system and muscular organization, Wnt signaling, axis patterning and locomotion, as also wound healing and adult regeneration (Martindale et al. 2002; Schnitzler et al. 2012; Jager et al. 2013; Moroz 2015; Jager and Manuel 2016; Babonis et al. 2018; Ramon-Mateu et al. 2019; Burkhardt 2022). In contrast, only few studies have focused on the immunological repertoire of ctenophores (Bolte et al. 2013; Moroz et al. 2014; Traylor-Knowles et al. 2019), while the innate immune system in other nonbilaterian phyla, such as sponges and cnidaria, showed an impressively high molecular complexity (Wiens et al. 2007; Augustin and Bosch 2010; Srivastava et al. 2010; Detournay et al. 2012; Hentschel et al. 2012; Bosch 2013, 2014; Krishnan et al. 2014; Riesgo et al. 2014; Degnan 2015; Pita et al. 2018; Parisi et al. 2020; Emery et al. 2021; Schmittmann et al. 2021). This fact leads to the necessity to study further the immune repertoire in ctenophores, expecting to discover similar complexity and unique features in this group.

In the animal world, hosts live in an environment full of microbes in which they have evolved in symbioses ranging from reciprocal benefits to antagonistic relationships. Under this scenario, the animal innate immune system acts not only as defense mechanism against microbial infections but it is also key for maintaining and specifying interkingdom communication between host and their microbiota (Koropatnick et al. 2004; Nyholm and Graf 2012; Bosch 2013; Chu and Mazmanian 2013; Gerardo et al. 2020). Like other aquatic organisms, ctenophores have established mutualistic relationships with microbial communities (Daniels and Breitbart 2012; Breitbart et al. 2015; Jaspers et al. 2019), evolving together as holobionts (Bosch and McFall-Ngai 2011). As such, it is expected that ctenophores already possess developed immune mechanisms to control their microbiota and distinguish between mutualists and pathogens. A crucial first step in animal–microbe interaction is the recognition of Microbial-Associated Molecular Patterns (MAMPs), such as lipopolysaccharides (LPSs), peptidoglycan (PGN), and flagellin, by host Pattern-Recognition Receptors (PPRs), which further trigger downstream pathways including several transcription factors and modulators (Zheng et al. 2005). There is a large variety of PRRs among animals, of which the most common ones are the Toll-Like Receptors (TLRs), the Nucleotide Oligomerization Domain (NOD)-Like Receptors (NLRs), the C-Type Lectins (CTL) and the scavenger receptors (SRs) (Rosenstiel et al. 2009). Apart from the typical PPRs, other receptors which have been related to microbial recognition are the G-Protein-Coupled Receptors (GPCRs), and cytokine receptors (Reboul and Ewbank 2016; Liongue et al. 2016). Most of these receptors are characterized by specific protein domains and motifs (Dierking and Pita 2020). In marine invertebrates, these receptors are highly divergent, and they can be either present fully or with parts of their original architecture, containing some of the homologous domains' architecture (e.g. Riesgo et al. 2014). Previously, it has been reported that the ctenophore Pleurobrachia bachei lacks the common PRRs and further mediators of the innate immune system (e.g. TLRs and TIR domain, Myd88, NLRs, RLRs, etc.) found in all other bilaterians and sponges (Moroz et al. 2014), whereas the ctenophores Mnemiopsis leidyi and Hormiphora californensis possess several of these conservative genes or their related domains (Traylor-Knowles et al. 2019). Consequently, it becomes crucial to identify the immune repertoire in ctenophores in order to extend deeper our understanding on their biology as well as the origin and evolution of the innate immune system in metazoans.

Chromosome-level genomes have become indispensable to characterize the gene repertoire of nonmodel (marine) species (e.g. Kenny et al. 2020; Cazet et al. 2023). So far, the majority of chromosome-scale genome assemblies come from bilaterians, and only a low number of chromosome-level genome assemblies is currently available from nonbilaterian animals, mainly from cnidarians (Zimmermann et al. 2023; Li et al. 2020; Kon-Nanjo et al. 2023; Cazet et al. 2023) and the freshwater sponge Ephydatia muelleri (Kenny et al. 2020). Regarding the Ctenophora phylum, two chromosome-level genomes have been recently published from the species H. californensis and Bolinopsis microptera (Schultz et al. 2021, 2023).

Here, we generated a chromosome-level assembly of the widespread ctenophore species M. leidyi Agassiz, 1865 with Pacbio and Hi-C technologies, in order to further explore the immune repertoire in this animal. M. leidyi is native to the western Atlantic Ocean, but in the last decades it spread as one of the most invasive species to many European waters and to the western Asian region (Reusch et al. 2010). Its notorious expansion in the Black Sea, the Caspian Sea and North and Baltic Seas was associated with a decrease in the fish population due to fish larvae predation, and caused huge ecological and economic impact. Moreover, this species also became an important model species for basic biological research, in particular in the fields of bioluminescence (Schnitzler et al. 2012), body regeneration (Henry and Martindale 2000; Ramon-Mateu et al. 2019), axis patterning (Martindale et al. 2002), and population genetics (Ghabooli et al. 2011; Bolte et al. 2013; Verwimp et al. 2020; Jaspers et al. 2021). Furthermore, its easy collection in the field, the establishment and maintenance of cell cultures, and its successful reproduction in vitro have turned M. leidyi into a model species for evolutionary and developmental studies (Martindale 2022). Hence, obtaining a chromosome-scale genome will also provide crucial genomic resources for many other studies, along with our analysis on the immune gene repertoire.

Results

M. leidyi Chromosome-level Genome Assembly

High-molecular-weight DNA (total amount of 726.66 ng/μl; nanodrop A260/280:1.87; A230/260: 2.49) was extracted from one individual of M. leidyi collected in the Kiel Fjord (Fig. 1a). A draft assembly was initially generated using a PacBio Sequel II (PacBio, Menlo Park, CA). Data from two SMRT cells were collected, generating 728,870 CCS/HiFi reads and a total of 9.2 Gb (∼ 37 × coverage based on the final assembly size). Final assembly size of this draft assembly was 247.94 Mb, with N50 1.01 Mb. After mapping the Omni-C libraries to the draft genome assembly, a high-quality chromosome-level genome assembly was generated. This assembly contained 142 sequences and, based on length distribution and Hi-C contact maps, 13 chromosomes were identified (Fig. 1b, supplementary fig. S1, Supplementary Material online). Due to the chromosome-level conservation between ctenophores, we have labeled the 13 assembled chromosomes of M. leidyi according to the nomenclature introduced for other two ctenophore species (B. microptera and H. californiensis, see supplementary fig. S2, Supplementary Material online for details). From the 129 scaffold sequences that were not assigned to a chromosome, 127 had a high similarity blast hit (E < 1e-30) to one of the assembled chromosomes, indicating they are alternative haplotype regions. The remaining two sequences were mapped against NCBI nr database and shown to be a Marinomonas chromosome (CP025532.1 Marinomonas sp. A3A chromosome) and a tandem repeat (TACATGGAGTTACAATGT). Therefore, all 129 unassigned scaffolds were discarded from the chromosome-level assembly, and all the following analyses were performed on this clean assembly. To further confirm the absence of contamination, we fragmented the genome assembly into subsequences of 1,000 bp length and used the Kraken 2 taxonomic classification system to classify them. Out of 22,4,564 subsequences, 5,013 were classified by Kraken 2 with a confidence of 0.1 or lower. The low number of classified sequences is expected as the Kraken “standard” database we used does not include Mnemiopsis. To rule out spurious hits, all 5,013 subsequences were searched with the BLAST blastn program against the Genbank “nt” database. Only 24 had database hits of >100 length and >80% identity, 23 of those were matches to M. leidyi sequences. The remaining hit is a 800 bp 86% identity match to the genome of Haliclystus octoradiatus, a cnidarian jellyfish.

Fig. 1.

Comparative genome statistics. a) Sampling location of the individual Mnemiopsis leidyi used for genome sequencing and assembly (depicted in the image). b) Length of the 13 chromosomes found in the genome of M. leidyi compared to the analogous 13 chromosomes found in the published genomes of genome of Bolinopsis microptera and Hormiphora californensis. c) Comparison of BUSCO scores indicating the quality of the three genomes and their annotation. BUSCO scores were run for each genome and their annotation. The BUSCO analysis was run with BUSCO v.5 version with metazoa_odb10 database.

We are thus confident that the final assembly is not contaminated by other species. This final assembly had a total size of 247.97 Mb, with a chromosome N50 of 16.84 Mb and 39.1% GC content (Table 1). BUSCO scores indicated that 84.7% of single-copy metazoan orthologs were complete in the M. leidyi chromosome-level assembly (Fig. 1c). The mapping rates of the RNA-seq data from 84 samples of M. leidyi to the genome assembly were 91.4% on average, with the sample with the lowest rate having 86.4% mapped reads while the sample with the highest rate to have 94.3% mapped reads.

Table 1.

Genome assembly details and comparison to other recently characterized ctenophore genomes.

	Mnemiopsis leidyi	Bolinopsis microptera	Hormiphora californensis	Mnemiopsis leidyi (draft genome)
REFERENCE	This study	Schultz et al. (2023)	Schultz et al. (2021)	Ryan et al. (2013)
GENOME SIZE	248 Mb	265 Mb	111 Mb	156 Mb
NUMBER OF SCAFFOLDS	142	346	44	5,100
NUMBER OF CHROMOSOMES	13	13	13	Not detectable
LARGEST CHROMOSOME	20.0 Mb	26.2 Mb	10.5 Mb	na
NUMBER OF Ns	33.3 Kb	21.1 Kb	30.6 Kb	5.5 Mb
GC CONTENT	39.10%	38.50%	43.10%	37.50%
N50	16.8 Mb	20.8 Mb	8.5 Mb	187 Kb

After automatic annotation using the genomeannotator pipeline (https://github.com/marchoeppner/genomeannotator), a total of 18,972 protein coding genes were identified in the M. leidyi assembly (supplementary table S1, Supplementary Material online). After functional annotations of this gene set, KEGG pathways were assigned to 5,602 genes, GO terms to 7,404 genes and Pfam domains to 13,282 genes (supplementary table S1, Supplementary Material online). BUSCO scores showed that this annotation contains 76.2% complete single-copy metazoan orthologs (Fig. 1c).

Search of Protein Domains Common in Animal Immune Receptors

Upon genome annotation, we focused our search on 20 Pfam domains that are known to be part of conserved animal immune receptors (Buckley and Rast 2011). First, we looked for the presence of these domains in the predicted gene models of the genome of M. leidyi and of two other ctenophore species, B. microptera and H. californensis with available chromosome-level genomes (Fig. 2a). Out of 20 Pfam domains, 17 domains were detected in the genome of M. leidyi, based on the genome annotation and independent hmmersearch (Fig. 2a). Most of them were also detected in the other two species, with exception the SEF/IL-17R (SEFIR) domain that was absent in H. californensis but present in M. leidyi and B. microptera, and the Baculovirus IAP Repeat domain (BIR) domain that was absent in B. microptera but detected in M. leidyi and H. californensis (Fig. 2a). Counting how many genes are related with each domain, only one or very few genes were related with some of the Pfam domains (e.g. SEFIR, CARD, DED; Fig. 2a) while for other domains, hundreds of genes were related with (e.g. I-set, V-set; Fig. 2a). Interestingly, we did not detect the Toll/interleukin-1 (TIR) Pfam domain (accession number PF01582) in any of the ctenophore species, while the TIR2 Pfam domain was present (accession number PF13676).

Fig. 2.

Pfam domains commonly found in immune receptors. a) Presence/absence of the reported pfam domains in the genome of Mnemiopsis leidyi, and the other two ctenophores' published chromosome genomes, Bolinopsis microptera and Hormiphora caliphornensis. The evidence for presence or absence of a domain was taken after search in the annotation of the M. leidyi genome and independent hmmerscan for all the three species, based on default parameters. The number indicates the genes coding for each pfam domain in each species based on the hmmer search. Toll/interleukine-2 domain; Immunoglobilin I-set, Immunoglobilin V-set, Immunoglobilin C1-set, Immunoglobilin C2-set, Leucine-rich repeat 1 domain, Leucin-rich repeat 2 domain, Leucin-rich repeat 3 domain, Leucin-rich repeat N-terminal domain, Leucin-rich repeat C-terminal domain, Scavenger Receptor cystein-rich domain, Caspase recrutiment domain, DEAD/DEAH box helicase, Death effector domain, Lectin C-type domain, Fibrinogen beta and Gamma chains C-terminal globular domain, Inhibitor of Apoptosis domain, b) Genes containing the TIR2, NACHT, SRCR, and C-type Lectin pfam domains are depicted in which chromosome they are located in the genome. These pfam domains were chosen as they are part of the most conserved families of immune receptors in metazoans. The graph was created with RIdeogram.

Out of the above 17 detected domains in the genome of M. leidyi, we further focused on four of them because of their importance in the immune repertoire of early metazoans (e.g. Poole and Weis 2014; Riesgo et al. 2014; Neubauer et al. 2016; Pita et al. 2018; Schmittmann et al. 2021): The TIR2 domain, which is part of the canonical TLRs, the NACHT domain, which is component of the NLRs, the Scavenger Receptor Cystein-Rich domain (SRCR), part of the SRs and the C-type Lectin domain (CTLD) which is part of the CTLs. In total, 150 genes were coding for those domains in the M. leidyi genome (supplementary table S2, Supplementary Material online). Specifically, 15 genes were coding a TIR2 domain, mainly found in chromosomes 9 and 13 (Fig. 2b; supplementary table S2, Supplementary Material online), while 38 genes comprised the NACHT domain, spread among different chromosomes with many genes detected in chromosomes 3 and 7 (Fig. 2b; supplementary table S2, Supplementary Material online). Additionally, 42 genes contained CTLD, 13 of those were in chromosome 1 and 7 genes in chromosome 10. The remaining genes were distributed across all the other chromosomes (Fig. 2b; supplementary table S2, Supplementary Material online). Finally, 55 genes contained SRCR, most of those detected in chromosomes 7, 9, and 12 (Fig. 2b; supplementary table S2, Supplementary Material online).

M. leidyi proteins containing TIR2, NACHT, SRCR, or CTLD domains revealed a large diversity of architectures (Fig. 3). Above domains were combined with a large variety of other domains, motifs and signals, resulting in an overall diverse array of proteins. Among the TIR2 domain-containing proteins, canonical TLR-based gene models conserved across animals, invertebrates and vertebrates alike, could not be identified. One TIR2-containing protein included the death domain, and might be considered as Myd88-like (Fig. 3a). Similarly, among the NACHT-containing proteins, a single one contained the LRR domain which putatively could be a bona fide NLR, but we were unable to reconstruct the canonical structure of vertebrate NLRs (Fig. 3b). Furthermore, we did not find any sequence that corresponds to a NLR receptor in the eggNOGG annotation. The SRCR domain-containing proteins had a large expansion in M. leidyi. We found proteins with many SRCR repeats and one combined with a CTLD, which, in vertebrates, are common of SR-I and SR-E classes respectively. However, SRCR domains combined with collagen (class SR-A), which is also typical of the vertebrate SRs, were absent. Instead, we found a large number of SRCR-containing sequences with novel and unique domain combinations (Fig. 3c). Looking finally at the CTLD containing proteins, they were also highly diverse regarding their overall domain architecture, including combinations with several types of Immunglobulin (IG), Epidermal Growth Factor (EGF), and Sushi domain (CCP) (Fig. 3d).

Fig. 3.

Domain architecture of proteins potentially comprising the innate immune receptor repertoire of Mnemiopsis leidyi, based on the presence of conserved pfam domains a) TIR2, Toll/interleukine-1; znF, zink finger domain; Efh, EF hand domain; LTD, C-terminal lamin-tail domain; FN3, fibronectin III. b) NACHT, DHC_N, Dynein heavy chain; LRR, leucine-rich repeat; Vps4_c, Vacuolar protein sorting-associated protein 4WD40, repeat of 40 amino acids typically terminating in Trp-Asp; DUF4062, conserved domain of unknown function; RVT, reverse transcriptase; adh_short, short chain dehydrogenase; RLI, metal-binding domain in endoribonuclease RNase L inhibitor; TPR, tetratricopeptide repeat. c) SRCR, Scavenger Receptor Cysteine-Rich domain; CCP, complement control protein, known as Sushi; VWD, Von Willebrand factor type D domain, VWC, Von Willebrand factor type C domain, TRyp_SPc, Trypsin-like serine protease; TSP1, Thrombospondin-1; LDL, Low-Density Lipoprotein receptor domain; MIR, named after three of the proteins in which it occurs: protein Mannosyltransferase, Inositol 1,4,5-trisphosphate receptor and Ryanodine receptor; TNFR, tumor necrosis factor receptor; DM, DNA binding domain; IG, Immunoglobulin; IGc2, C2-set immunoglobulin; shKT, a 35-residue peptide toxin from sea anemone Stichodactyla helianthus, are potent inhibitors of K channels; ZnMc, Peptidase metallopeptidase; CUB, complement C1r/C1s, Uegf, Bmp1; STYKc, protein kinase. d) CTLD, C-Type Lectin Domain; TIL, trypsin inhibitor-like cysteine-rich domain; EGF, epidermal growth factor; EGF_ca, EGF-calcium-binding, epidermal growth factor; GPS, G-protein-coupled receptor proteolytic site; HormR, present in hormone receptors; LY, Ly-6 antigen/uPA receptor-like; TY, Thyroglobulin type I repeats; HX, hemopexin, TM, transmembrane region; 7TM, 7 helix transmembrane domain. The figure was created with BioRender.com.

The phylogenetic analysis of TIR containing proteins indicated that the TIR2 sequences of ctenophores including M. leidyi are closer to TIR sequences of nonmetazoans and early diverging metazoans, while they do not possess any TLR-like or Myd88-like proteins (Fig. 4). The analysis, including Filasterea, Placozoa, and different phyla of metazoans, divided those proteins into three main groups, one group with only TIR containing proteins, another group with TLR-like proteins and a third group with Myd88-like proteins. The TIR2-containing proteins of M. leidyi were exclusively found in the TIR containing group, together with sequences from Filasterea, Placozoa, the two other ctenophore species (B. microptera and H. californensis) and Porifera (Fig. 4). Eleven out of sixteen sequences were grouped together with other ctenophore TIR2-containing sequences, mainly of B. microptera, indicating the existence of unique structures of TIR2 sequences in ctenophores, while the other five sequences were spread in smaller groups together with the other ctenophore species (B. microptera and H. californensis) but also mixed with TIR sequences from other phyla. This shows first that there is an expansion of genes containing this domain in ctenophores, and second that there are also sequences with shared or similar structure with other early diverging metazoans (Fig. 4).

Fig. 4.

Phylogenetic analysis of the TIR2 domain based on GTR Bootstrap model and an estimated gamma shape parameter and 100 independent searches, generated with RAxMLv.8 (raxmlgui, v2.0.1). Sequences from TLR, Myd88 and TIR domain were extracted from NCBI for most of the phyla. Hmmer profiles were generated to extract sequences from the genomes of the ctenophores Mnemiopsis leidyi, Bolinopsis microptera and Hormiphora caliphornensis, and the genomes of the sponges Dysidea avara and Ephydatia muellieri.

Finally, we looked for any functional evidence among the 150 M. leidyi genes that contained TIR2, NACHT, SRCR or CTLD domains. We assessed their changes in relative expression after exposure of M. leidyi with the pathogenic bacterium Vibrio coralliilyticus (VB) and re-exposure with the same strain after 6d (VBVB), compared with the control condition (CO), in which animals were exposed only to water. Most of the candidate genes were constitutively expressed, independently of the condition (supplementary table S3, Supplementary Material online). While we did not detect any differential expression of these genes during the first exposure to the bacteria strain (VB), 32 genes were differentially expressed during the second bacterial exposure (VBVB vs. CO condition, 21% of all candidates) (Fig. 5). Four genes belonged to the TIR2 domain-containing group, 13 genes to the NACHT domain, 12 genes to the SRCR domain and 4 genes to CTLD (Fig. 5, supplementary table S4, Supplementary Material online). In total, 12 genes were downregulated (seven NACHT, four SRCR, two CTLD genes) and 20 were upregulated (four TIR2, six NACHT, eight SRCR, two CTLD genes) in VBVB versus CO condition. Almost half of the NACHT (seven genes) and CTLD (two genes) coding genes were downregulated while the other half were upregulated in VBVB versus CO condition. On the other hand, 8 out of 12 SRCR coding genes and all 4 TIR2 genes were upregulated upon exposure with VBVB (Fig. 5, supplementary table S4, Supplementary Material online), suggesting that they are involved in immune signaling and defense.

Fig. 5.

Comparative gene expression analysis. Heatmap with differential expression of, TIR2, NACHT, SRCR and CTLD containing genes identified in the genome of M. leidyi 6d after exposure with a pathogen bacterium Vibrio coralliilyticus (VBVB), compared to the control condition (CO). The gene expression in the heatmap increases from blue to red. The differential gene expression analysis was conducted with the edgeR package. TIR, Toll/interleukine-1; SRCR, Scavenger Receptor Cysteine-Rich; CTLD, C-type Lectin.

Discussion

We present the first chromosome-level genome for the ctenophore M. leidyi, obtaining first insights of the immune gene repertoire of this organism. The M. leidyi genome features almost the full diversity of conserved protein domains that are part of all other metazoan immune genes. Among the protein domains we further studied, they all revealed large expansions and some of them had also unique architectures in M. leidyi. We also observed that some of the genes, containing those immune-related conserved domains, were also differentially expressed after exposure of M. leidyi with the pathogen bacterium V. coralliilyticus. This indicates that these genes do play a role in host–microbe interaction, by potentially activating the immune system of the ctenophores in order to face potential infections by pathogens.

The genome presented in this study is of high quality and as such has a great potential to be used as a resource in future studies. As a case in point, it features a higher completeness (84.7%) than the previously published M. leidyi draft assembly (80.8%) (Ryan et al. 2013), and also more than the two chromosome-level ctenophore assemblies (80.8% in H. californensis (Schultz et al. 2021) and 81.6% in B. microptera (Schultz et al. 2023). Additionally, fewer metazoan BUSCO sequences were fragmented (7.8%) and the duplication rate (4.3%) is comparable to that found in the sister species B. microptera (4.7%) (Fig. 1c). The relatively low BUSCO score of completeness rate (< 90%) in the genome of ctenophores in general can be explained by the unique phylogenetic position of ctenophores as sister group to all other metazoans (Schulz et al 2023). This range is quite common for other early diverging metazoans (e.g. Kenny et al. 2020). Finally, the completeness for the genome annotation (76.2%) was also close to the 78.2% identified in the H. californensis annotation (Schultz et al. 2021) and higher than the 65.5% found in the B. microptera annotation (Schultz et al. 2023). Finally, our results indicate, along with the other two published genomes, that the chromosome number among ctenophores is conserved with N = 13 (supplementary fig. S2, Supplementary Material online). Size distribution of these 13 chromosomes in all three ctenophore species shows a constant pattern of B. microptera having larger homologous chromosomes than M. leidyi, in both cases much larger than the corresponding chromosomes of the more distantly related H. californensis species (Fig. 1b). The species M. leidyi and B. microptera belong to the same family of Bolinopsidae (Ctenophora: Lobata) with Mnemiopsis as a paraphyletic species within the Bolinopsis genus (Christianson et al. 2022), while H. californensis belongs to Cydippids which is phylogenetically far removed from the other two species (Whelan et al. 2017). Surprisingly, very few interchromosomal translocations of orthologous genes have been identified among the three studied species (supplementary fig. S2, Supplementary Material online).

Marine invertebrates lacking adaptive immune systems, rely exclusively on innate immune system to defend against pathogens they are continuously exposed to in the aquatic environments. Their defense mechanisms involve, among others, an extremely variant and complex repertoire of immune receptors. Most of the immunity-related protein domains, which make part of those receptors and examined in this study were present in all three ctenophore species, suggesting that they appeared in the common ancestor of the phylum Ctenophora. Notable absences were also shared, in particular one of the studied LRR domains (LRR3) and the LRCNT domain, which usually flanks LRR domains, were not detected in any of the three species. These domains are related with the NOD-like Receptors (NLRs) in vertebrates, similarly to BIR domain (Ting et al. 2008) which was lacking in B. microptera. However, their presence is not prerequisite in NLR formation, especially in invertebrate NLRs which do not contain any BIR domain (Lange et al. 2011; Hamada et al. 2013; Schmittmann et al. 2021). This indicates that Ctenophora already have most of the common domains present in potential immune receptors in other animals, and so we expect that these domains will execute similar functions in M. leidyi.

All TIR domain-containing sequences identified among all three ctenophore species were more similar to the TIR2 rather than the TIR domain. The latter is a very essential component of the innate immune system as it is part of the immune TLR receptor, the IL-1R family and the Myd88 immune modulator (Bowie and O’Neill 2000; Dunne et al. 2003; Leulier and Lemaitre 2008). Although the TIR2 domain was first detected in bacteria (Newman et al. 2006), it has been found in many metazoans (Huang et al. 2008; Zhang et al. 2011), including sponges and cnidaria (Poole and Weis 2014; Riesgo et al. 2014) and is considered to be acquired through horizontal bacterial transfer like other TIR adaptor sequences (Zhang et al. 2011). Undoubtedly, ctenophores have a relatively large expansion of TIR2-containing sequences, similar to other nonbilaterian invertebrates such as Cnidaria (Poole and Weis 2014). However, none of the TIR2-containing proteins in our study was combined with LRR domains, characteristic of canonical form TLR receptors in vertebrates (Takeda et al. 2003; Coscia et al. 2011). Traylor-Knowles et al. (2019) have previously mentioned the presence of putative TLR homologs in M. leidyi based on the presence of LRR domain-containing genes in the draft genome (Ryan et al. 2013); however, from our analysis, which is based on more extensive search in a complete genome assembly, we concluded that M. leidyi definitely lacks canonical TLR receptors. Furthermore, no TIR2-containing proteins from M. leidyi or the other two ctenophores, B. microptera and H. californensis, were grouped either with TLR-like (i.e. TLR homologs with noncanonical structure) or Myd88 proteins of other metazoans in our phylogenetic analysis (Fig. 4). This might suggest that noncanonical TLR-like receptors and/or Myd88 are also absent in ctenophores. Despite the phylogenetic divergence of ctenophore TIR2-containing genes, 4 of those genes were upregulated upon V. coralliilyticus challenge (Fig. 5). Interestingly, looking at the domain architecture of those 4 differentially expressed genes, 3 of them (gene1, gene2, gene4) are exclusively TIR2 domains while in the fourth gene (gene 3) the TIR2 appears combined with Fibronectin type III (FN3) extracellular repeats. FN3 is part of the extracellular matrix and has been shown to induce the pro-inflammatory cytokines via the activation of the TLR pathway (Kelsh et al. 2014). This suggests that TIR2-containing genes are activated after MAMPs recognition, potentially activating downstream immune pathways in ctenophores.

Another interesting finding was the large expansion of NACHT domain-containing proteins, combined with other domains, including LRRs but also other repeat domains such as WD40 or TPR domains. The NACHT domain is a major component of the NLRs (Proell et al. 2008; Ting et al. 2008) and based on the domain architecture of the NACHT domain-containing proteins, we suggest that M. leidyi has potential bona fide NLR receptors but it lacks the canonical NLR receptors of vertebrates. The complete canonical NLRs in vertebrates have an N-terminal with either CARD, Pyrin, Death or BIR domain, a NACHT domain and c-terminal repeats of LRR domains (Proell et al. 2008; Ting et al. 2008), while the bona fide NLRs are considered those with NACHT and LRR domains (Ting et al. 2008). For example, sponges and cnidarians have the complete structure of a canonical NLR receptor, such as CARD-NACHT-LRR, but also bona fide NLR receptors and other NACHT-containing proteins (Lange et al. 2011; Hamada et al. 2013; Yuen et al. 2014; Schmittmann et al. 2021) with similar domain architecture as found here. Our finding comes in contrast with findings in the study of Yuen et al (2014), in which they did not detect any bona fide NLR in the draft genome of M. leidyi. Similarly, Moroz et al (2014) did not detect any NLR sequence in the draft genome of the ctenophore Pleurobrachia bachei either. On the other hand, Traylor-Knowles et al. (2019) reported the presence of NLRs as well as TLRs in ctenophores, but based only on the presence of LRRs. Interestingly, challenging M. leidyi with V. coralliilyticus, revealed the differential expression of 13 NACHT-containing genes, including the up-regulation of one bona fide NLR (Fig. 5). Experimental evidence of NLR activation after MAMP exposure has been previously shown in the sponge Dysidea avara (Pita et al. 2018). Moreover, other noncanonical NLR genes were differentially expressed after LPS and flagellin stimulation, activating caspases in Hydra (Lange et al. 2011). Taking the fact that we find bona fide NLRs in ctenophores but not any bona fide TLR, we suggest that NLRs appeared prior to TLR receptors in the evolution of PRRs and possibly acquired an earlier role in the immune system in metazoans. Furthermore, a comparative genomic analysis by Yuen et al. (2014) showed no evidence for the existence of bona fide NLRs outside the Metazoa, suggesting that bona fide NLRs is a metazoan-specific feature with an important role in the origin and evolution of Metazoa.

A large expansion of CTLD and SRCR-containing proteins was detected in M. leidyi. The CTL and SR receptors are very diverse and play a role in host–microbe interaction and immunity, both in vertebrates and invertebrates (Drickamer and Taylor 2015; Pees et al. 2016). Extensive studies on CTLD containing proteins and their function in immunity have been conducted in crustaceans, nematodes and insects (Pees et al. 2016). In M. leidyi, different combinations of domains were found in CTLD proteins, similar to those found in the sponge Halichondria panicea (Schmittmann et al. 2021). We also found a large range of SRs, as has been previously described in other invertebrates (Pancer 2000; Huang et al. 2008; Rast and Messier-Solek 2008; Neubauer et al. 2016; Schmittmann et al. 2021). Interestingly, most of the domain combinations found in M. leidyi have also been described in both Porifera and Cnidaria (Neubauer et al. 2016; Ryu et al. 2016; Schmittmann et al. 2021). For example, M. leidyi also features the SR families that contain only SRCR domain (SR-I family), the SR family that contained the CUB domain, the SR family with the trypsin domain, and several SR containing proteins with the IG domain and LDL domain (Neubauer et al. 2016). In contrast, we did not detect SRCR-containing proteins that have a collagen or Lectin-like Oxidized Low-Density Lipoprotein Receptor 1 (LOX1) domains, characteristic of the human SRCRs, while we identified one sequence with a combined SRCR-CTL domain, which belong to the same SR family (SR-E) (Fig. 3c). Finally, we detected the cluster of differentiation 36 (CD36) domain in the genome of M. leidyi (data not shown), which is considered a separate SR family (SR-B family) (Silverstein and Febbraio 2009). However, previous phylogenetic analysis of metazoan CD36 had shown that the sponge and ctenophore CD36 is distinct to that of cnidarians and vertebrates (Neubauer et al. 2016), possibly meaning that CD36 has a different function in ctenophores. In any case, most of SRCRs were upregulated upon pathogen re-exposure, indicating their role in pathogen recognition as well (Fig. 5). Similarly, SRCR domain-containing genes were upregulated in sponges in response to MAMPs (Pita et al. 2018) and microbes (Yuen 2016). SRCR domain-containing genes were also activated in response to microbes in echinoderms and molluscs (Pancer 2000; Liu et al. 2011; Furukawa et al. 2012). One of the main functions of SRCRs is to recognise molecules from microbes and activate their elimination by phagocytosis (Hentschel et al. 2012). In mammals, deactivation or malfunction of those proteins has been linked with bacterial infections (Martínez et al. 2011). We conclude that ctenophores have CTLD and SRCR-containing proteins similar to those found in other early metazoans with function in immune response (Schmittmann et al. 2021).

Overall, most of the TIR2, NACHT, CTLD, and SRCR domain coding genes were expressed constitutively during our experiment, indicating their role in the maintenance of the normal physiological state of the host and microbe homeostasis (e.g. Pita et al. 2018, Schmittmann et al. 2021). However, genes that were upregulated upon exposure of the host with the pathogen V. coralliilyticus, can indicate their role in host–microbe interactions and in host immune defense. On the other hand, some of the NACHT, CTL, and SRCR-containing genes were downregulated upon pathogen exposure in our study (Fig. 5), which could mean that they are proteins related to other physiological and metabolic processes of the host, which are halted upon pathogen exposure, due to energetic costs. Another scenario could be that these genes participate in mutualistic recognition and enhancement of a healthy host–mutualist balance, and as such they are downregulated during pathogen exposure. For instance, SRs seem to also have a role in mutualism. Experimental analysis in the cnidarian A. pallida demonstrated that SRs increase host tolerance to its mutualistic dinoflagellate, while disfunction of these receptors elicits an immune response toward its mutualist S. minutum (Neubauer et al. 2016). It has also been previously suggested that a SRCR domain-containing protein in the sponge Petrosia ficiformis potentially plays a role in the recognition of their cyanobacterial mutualists (Steindler et al. 2007). Finally, it is also plausible that the downregulated genes we detected upon the bacterial challenge are the result of the pathogen regulating host gene repression in order to facilitate their invasion and further infection.

We conclude that the chromosome-level genome provided in this study allowed us to perform an accurate and extensive search of the immune repertoire of M. leidyi. Ctenophores have a large variety of common protein domains participating in invertebrate immunity, demonstrating that these domains appeared in the last common ancestor and are conserved through animal evolution. Even though their architectures do not describe canonical receptors in M. leidyi, it seems that they already have a functional role in sensing and recognizing microbial stimuli, as their expression level is regulated upon bacterial challenge. Their unique and large expansion in M. leidyi is consistent with its phylogenetic position as a sister group of all other metazoans in the tree of life. This study is a first step toward understanding the evolution of the immune system and the function of ancestral immune pathways in the animal tree of life. Good-quality genomes are crucial to future immunological studies on how ctenophores interact with microbial counterparts in their surroundings, how they distinguish their mutualists from pathogens, and how their interaction with microbes might allow them to become such a successful invasive species worldwide.

Materials and Methods

Specimen Collection

Two individuals of M. leidyi were collected in June 2020 from animals at Kiel fjord, Germany (54.329786N, 10.151005E) (Fig. 1a). One individual was immediately processed for DNA purification, while the other individual was flash frozen in liquid nitrogen and kept stored in −80 °C until further process.

High-molecular-weight DNA Purification

High-molecular-weight DNA was extracted with the NucleoBond® HMW DNA kit (Macherey-Nagel Gmbh & Co, Germany) according to manufacturer's instructions, with the following modifications: (i) the recommended volumes for the lysis steps was 4-fold increased to ensure the lysis of the whole individual and (ii) an extra wash step with solution H4 was added. The quality, quantity and size of the DNA extract was assessed by NanoDrop 2000c Spectrophotometer (peolab, Germany) and Qubit 2.0 (Life Technologies, Carlsband, CA).

Library Construction, Genome Sequencing and Scaffolding

Long-read sequencing was carried out with the Pacbio Sequel II platform at IKMB sequencing facility in Kiel. A first library was prepared using the Sequel II Binding Kit 2.0 and sequenced in CLR mode (insert size 16.5 Kb). This first SMRT Cell produced an output of 124 Gb, with a subread N50 of 11.9 Kb. Circular consensus reads were generated using the ccs command from the pbbioconda suite of tools (https://github.com/PacificBiosciences/pbbioconda), producing a total of 4.9 Gb with mean length of 10.0 Kb. A second library was prepared using the Sequel II Binding Kit 2.2 and sequenced in HiFi mode (insert size 15.4 Kb). This second SMRT Cell produced an output of 83 Gb, with a subread N50 of 33 Kb, where the HiFi yield was 4.3 Gb with mean length of 18.4 Kb. The assembler IPA v1.8 (https://github.com/PacificBiosciences/pbipa) was run with local mode (including polishing, purging haplotigs and phasing) to assemble together the reads from both libraries into a M. leidyi draft assembly. This assembly was used as input for Hi-C sequencing described below.

Dovetail Omni-C Library Preparation and Sequencing

Frozen material from the second individual of M. leidyi was sent to Dovetail Genomics for generating Omni-C® libraries. For each Dovetail Omni-C library, chromatin was fixed in place with formaldehyde in the nucleus. Fixed chromatin was digested with DNase I and then extracted, chromatin ends were repaired and ligated to a biotinylated bridge adapter followed by proximity ligation of adapter containing ends. After proximity ligation, crosslinks were reversed and the DNA purified. Purified DNA was treated to remove biotin that was not internal to ligated fragments. Sequencing libraries were generated using NEBNext Ultra enzymes and Illumina-compatible adapters. Biotin-containing fragments were isolated using streptavidin beads before PCR enrichment of each library. The library was sequenced on an Illumina HiSeqX platform to produce ∼ 30 × sequence coverage.

Scaffolding the Assembly With Omni-C HiRise

The input Pacbio draft assembly described above and all Dovetail Omni-C library reads were used as input data for HiRise, a software pipeline designed specifically for using proximity ligation data to scaffold genome assemblies (Putnam et al. 2016). Dovetail Omni-C library sequences were aligned to the draft input assembly using bwa (https://github.com/lh3/bwa). The separations of Dovetail Omni-C read pairs mapped within draft scaffolds were analysed by HiRise to produce a likelihood model for genomic distance between read pairs, and the model was used to identify and break putative misjoins, to score prospective joins, and make joins above a threshold. Additionally, the presence of known metazoan single-copy orthologs was analysed using BUSCO v5.3.0 to assess genome completeness. These values were compared to two previously published ctenophore genomes. H. californensis genome was downloaded from NCBI (BioProject PRJNA576068) and B. microptera genome was downloaded from Dryad repository (https://doi.org/10.5061/dryad.dncjsxm47.). Orthologous chromosomes were identified between M. leidyi, B. microptera, and H. californiensis based on synteny analysis of 954 metazoan BUSCO gene sequences using ChromSyn (Edwards et al. 2022). As an additional measure of genome completeness, we mapped a large RNA-Seq dataset consisting of 84 samples with 2369 M Illumina reads in total to the assembled genome.

Genome Annotation

The genomeannotator pipeline developed in-house (https://github.com/marchoeppner/genomeannotator) was used to annotate the newly generated chromosome-level genome of M. leidyi. This pipeline, which uses the Nextflow workflow language, performs automatic genome annotation based on ab-initio gene prediction as well as experimental and additional model hints from multiple possible sources (Di Tommaso et al. 2017). The main source of species-specific hints was RNA-seq data generated from an unpublished experiment described below. This dataset of three samples is a small subset of a larger dataset and the RNA-seq reads used for annotation can be found in supplementary table S6, Supplementary Material online. Briefly, an initial 20 to 30 mg of flash frozen tissue was used and the RNA was extracted with the AllPrep DNA/RNA mini kit Qiagen, according to the manufacturer's guidelines. The total RNA was eluted in 50 μl of RNase free water and stored at −20 °C until sequencing. The sequencing was carried out on an Illumina Hiseq platform in IKMB facilities in Kiel.

The genomeannotator pipeline was run as follows: first, RepeatModeler v2.0.4 (Smit and Hubley 2015) (was run on the genome sequence with default parameters and the option -LTRstruct. The resulting annotated consensus repeat sequences were used as input to run RepeatMasker v4.1.2 (Smit and Hubley 2015) as the next step. Once the genome was masked, three kinds of extrinsic hints were generated: (i) reviewed metazoan proteins were downloaded from UniProt (The UniProt Consortium 2017) and aligned to the masked genome using SPALN v3.0.0 (Gotoh 2008), (ii) raw RNA-seq reads, described in the previous section, were first trimmed using Fastp v0.23.2 (Chen et al. 2018) with default parameters and the additional parameter “–detect_adapter_for_pe”, and then aligned to the masked genome using STAR v2.7.10a (Dobin et al. 2013), and (iii) the same trimmed RNA-seq reads were assembled using Trinity v2.13.2 (Grabherr et al. 2011) and the resulting transcripts were aligned to the masked genome using Minimap2 v2.22 (Li 2018). All resulting alignments were used as extrinsic hints to run Augustus v3.4.0 (Stanke et al. 2008). In parallel, PASA v2.5.2 (Haas et al. 2003) was run using the RNA-seq derived hints (ii and iii before) to generate gene models containing UTRs and alternative isoforms. Gene models generated with Augustus and PASA were used as input to run EvidenceModeler v1.1.1 (Haas et al. 2008) to generate the final reconciled consensus set of gene models. Finally, eggNOG mapper v2.1.7 (Cantalapiedra et al. 2021) was run to functionally annotate the gene models.

Search of Specific Pfam Domains Related to Immunity

Specific domains of the metazoan immune repertoire were selected to be searched in the chromosome-scale genome assembly of M. leidyi (Fig. 2). In total, 20 Pfam domains were included to our analysis: Toll/interleukine-2 domain (TIR2; PF13676); SEFIR (PF08357), Caspase recruitment domain (CARD; PF00619), Death (PF00531), Death effector domain (DED; PF01335), NACHT (PF05729), DEAD/DEAH box helicase (DEAD; PF00270), Inhibitor of Apoptosis domain (BIR; PF00653), lectin C-type Domain (CTLD; PF00059), Fibrinogen beta and Gamma chains C-terminal globular domain (Fibrinogen C; PF00147); Immunoglobilin I-set (I-set; PF07679), Immunoglobilin V-set (V-set; PF07686), Immunoglobilin C1-set (C1-set; PF07654), Immunoglobilin C2-set (PF05790), Leucine-rich repeat 1 domain (LRR1; PF00560), Leucin-rich repeat 2 domain (LRR2; PF07723), Leucin-rich repeat 3 domain (LRR3; PF007725), Leucin-rich repeat N-terminal domain (LRRNT; PF01462), Leucin-rich repeat C-terminal domain (LRCNT; PF01463), SRCR domain (PF00530). First, candidate immune receptors were isolated based on their characteristic conserved domains based on Pfam ID from the automated eggNOG annotation. A further search for each domain was conducted with hmmerscan v.3.4 (http://hmmer.org/) with default parameters directly in the translated genome by using the respective Pfam database entry from the Interproscan page (https://www.ebi.ac.uk/interpro/entry/pfam/). The selected sequences were reverse blasted with Interproscan (https://www.ebi.ac.uk/interpro/result/InterProScan/) and SMART (http://smart.embl-heidelberg.de/) in order to reassure the presence of the searched domain in the respective protein sequences. A hmmerscan search for each Pfam domain of interest was also conducted in the two other published ctenophore genomes, H. californensis and B. microptera, with the same parameters, in order to identify and compare the presence or absence of these potentially immune-related Pfam domains. Based on the hmmerscan search, we also identified the copy of the genes per domain found in all the three species.

We further selected the proteins containing the Pfam domains related to common immune receptors (TIR2, NACHT, SRCR, CTL) to further explore their domain architecture. The structure of the domains in each protein sequence was checked with SMART in Genomic mode (http://smart.embl-heidelberg.de/) (Letunic et al. 2021). Proteins both obtained from annotation and hmmerscan were studied in each case. The relevant graph for these proteins was created with Biorender.com. To further confirm that the domain architectures are not an artifact of the genome annotation process, we blasted all immune genes studied against a transcriptome assembly created independently of the genome. Out of the 150 genes, 104 genes had matches of >95% identity over >80% query length, and the domain architecture of those transcripts was the same or similar in most cases (supplementary table S7, Supplementary Material online).

Phylogenetic Analysis of TIR2 Domaining-containing Proteins

We chose the TIR2 Pfam domain to estimate its divergence in M. leidyi compared to other phyla and understand its expansion in ctenophores. First, we retrieved different sequences that include this Pfam domain (TIR sequences, Myd88, TLR sequences) from a range of metazoans, the unicellular eukaryote species Capsaspora owczarzaki, and the placozoan Trichoplax adhaerens (supplementary table S5, Supplementary Material online) and we performed protein alignments (supplementary file S1, Supplementary Material online) with MUSCLE (Edgar 2004) using the SEAVIEW software (Gouy et al. 2010). We used the alignments to construct a HMM profile for the TIR domain and we further used HMMER to extract protein sequences using this HMM profile from the genomes of M. leidyi, H. californesis (Schultz et al. 2021) and B. microspora (Schultz et al. 2023) and from the published genome of the sponge E. muelleri (Kenny et al. 2020) and the genome of Dysidea avara (our unpublished data).

The extracted sequences from each species were aligned consecutively to the database created above using MAFFT (Katoh and Standley 2013) in the Geneious software (Kearse et al. 2012) with default parameters. After each alignment, the consensus sequence was kept and re-blasted in SMART to verify its identity. A maximum likelihood phylogenetic tree was generated with analysis in RAxML v.8 (Stamatakis 2014) with the raxmlGUI platform v2.0.1 (Edler et al. 2021), using GTR Bootstrap model for proteins and an estimated gamma shape parameter. The node support was calculated with the algorithm via bootstrapping and 100 independent searches. The tree was further processed with figtree v.1.4.4. and the online tool iTOL (https://itol.embl.de/).

Differential Gene Expression

We tested the relative expression of the genes belonging to the above categories (TIR2, NACHT, SRCR, CTLD) after two consecutive immune challenges of M. leidyi with the pathogen bacterium V. coralliilyticus (DSM ID 19607). This bacteria was chosen as it has been described to be a potential pathogen for marine invertabrates and has also been used in an immune challenge experiment in M. leidyi by Bolte et al. (2013). The bacteria was grown in MB medium (10 g Peptone, 10 g meat extract [both Carl Roth, Karlsruhe, Germany] per 1 litre of 30 PSU ASW) at 30 °C and 120  × g. The cell number of the strain was determined using optical density measurements at 600 nm (OD600) and cell number counting using the Neubauer count chamber (Assistant, Sondheim vor der Röhn, Germany). Therefore, 250 µL of an overnight culture was inoculated into 25 mL fresh MB medium and incubated at 30 °C and 120  × g over a time period of 10 h. Sample for cell number determination was taken each 30 min. Exponential growth phase was reached after 5.5 h. Bacterial stock was generated in culture medium with 10% glycerol from exponentially growing bacterial cell culture as 105 cells/mL and immediately frozen in liquid nitrogen and stored at −80 °C. Before use, the stock (1 mL) was thawed and the bacterial cells were pelleted for 3 min at 9,000  × g. The medium with glycerol was removed and the bacterial pellet was resolved in sterile ASW (105 cells/mL). For the experiment, animals reared in filtered ambient seawater were transferred into a light and temperature-controlled chamber (18 °C, 8-h day–night cycle), and subsequently submerged in the bacterial solution (10⁵ cells mL⁻¹). During the first challenge, samples were taken 1 h and 3 h post infection (VB). After a recovery period of 6 days without any bacterial exposure, individuals were exposed to the same bacterial strain in a second bacterial challenge (VBVB). Samples were taken again at one and 3 h. As control groups (CO), individuals were submerged in filtered (0.22 µm) seawater without introduction of bacteria, and animal samples were taken at the same time points. RNA-sequencing was conducted with the Illumina Hiseq platform and the resulting transcripts were mapped to the current genome. Differential gene expression analysis was conducted with the edgeR package (Robinson et al. 2009; McCarthy et al. 2012), and genes presenting FDR ≤ 0.01 and 2-fold change differences were identified as differentially expressed. The heatmap was generated in R v. 4.3.2.

Supplementary Material

Supplementary material is available at Genome Biology and Evolution online.

Author Contributions

V.K. and M.T.O. contributed to data analysis, graph generation, writing of the manuscript, T.B. contributed to analysis of RNA-seq experiment, editing manuscript, J.F. contributed to sequencing, N.G. contributed to conduction of RNA-seq experiment, D.G. contributed to high-molecular-weight DNA extraction, A.M.G. contributed to troubleshooting and high-molecular-weight DNA extraction; N.J. contributed to bacteria culture and preparation for the bacterial injection of M. leidyi during the RNA-seq experiment. R.A.S. contributed to study design, funding, L.P. contributed to study design, funding, interpretation of results, editing of the manuscript, T.B.H.R. contributed to study design, funding, interpretation of results, editing of the manuscript.

Funding

This work was funded by the Deutsche Forschungsgemeinschaft (DFG) Sequencing Call proposal IMMUBASE [417981041 to L.P, R.S.S. and T.B.H.R]; and by the Collaborative Research Centre of Deutsche Forschungsgemeinschaft “Origin and Function of Metaorganisms (CRC1182)” (B2) [261376515 to R.S.S. and T.B.H.R]. Additional funding support was provided by “la Caixa” Foundation, Spain (to L.P.), [ID 10010434 to L.P.], co-financed by the European Union Horizon 2020 research and innovation program under the Marie Sklodowska-Curie Action grant agreement No 847648 [104855]; and by the “Severo-Ochoa Centre of Excellence” accreditation [CEX2019-000928-S]. This is a contribution from the Marine Biogeochemistry and Global Change research group Grant 2021SGR00430, Generalitat de Catalunya, Spain (to L.P.).

Data Availability

Chromosome-level genome assembly of M. leidyi generated in this study is available at the European Nucleotide Archive (ENA) under the study ID PRJEB71361 and accession GCA_963919725.