Metagenomic analysis of orange colored protrusions from the muscle of Queen Conch Lobatus gigas (Linnaeus, 1758)

Assembly and metagenomic approach

The massive shotgun sequencing generated 515,368 reads with an average length of 279 bp that were cleaned and then assembled using MIRA software into 5,180 contigs. Taxonomic classification of the contigs was carried out using the software MEGAN. For this analysis, all the contigs were compared with the NCBI’s non-redundant protein database using the software BLASTX. With this strategy, 1,588 (30.7%) contigs were assigned to taxa (Bacteria: 412; Eukaryota: 1,157; other: 19), 866 (16.7%) were unassigned and 2,726 (52.6%) presented no hits. As expected, the Eukaryota group was dominant due to the origin of the sample. Furthermore, the group Gastropoda was frequently found in this analysis (186 hits), although many sequences remained unclassified due to the poor representation of these organisms in the public databases. Many bacterial sequences were also identified; 19 were assigned to the fungi group and 22 sequences were assigned to the Trematoda category. No viral or protozoa sequences were detected.

Following the MEGAN pipeline, with nucleotide comparisons using BLASTN and the nt/nr database, results were poorly classified. One contig was assigned to the root, 462 (8.9%) to particular taxa (Bacteria: 267; Eukaryota: 191; other: 4), 32 (0.6%) were unassigned and 4,685 (90.5%) had no hits. At the nucleotide level, most of the sequences were left unclassified. This reflects the lack of sequences in the databases of closely related organisms to the ones reported in this research.

Bacteria and fungi associated with orange colored protrusions

The ranges for the confidence criterion represented by the E-value, similarity and identity for protein comparisons are shown in Table 1. Bacteria assignments included the class Gammaproteobacteria and the phylum Firmicutes, which includes the orders Bacillales and Lactobacillales (Fig. 1). Specifically, the class Gammaproteobacteria showed 322 assignments for Psychrobacter, exhibiting hits to several types of proteins with strains of Psychrobacter sp. (Fig. 1, Table 1). The identity criterion for Psychrobacter sp. ranged from 67% to 100% (Table 1). Similarly, nucleotide sequences showed hits for different genomic regions of Psychrobacter sp. strains and congeners, displaying identities ranging from 72% to 99% and alignment lengths from 104 bp to 1,061 bp (Table 2).

On the other hand, 18 assignments for the order Lactobacillales (Fig. 1) showed hits for diverse proteins exhibiting similarities and identities up to 100% for Carnobacterium jeotgali and Carnobacterium sp. (Table 1). We also found hits for proteins of Lactobacillus jensenii and Enterococcus faecalis displaying identities above 68% and 44%, respectively (Table 1). Similarly, nucleotide analysis showed matches for genome regions and plasmids of Carnobacterium sp., displaying identities above 82% (Table 2). Additionally, the single hits for a plasmid and a genome fragment of Enterococcus casseliflavus and Enterococcus faecalis exhibited identities of 81% and 73%, respectively (Table 2).

A total of 28 contigs were assigned to different Bacillales bacteria within the phylum Firmicutes (Fig. 1); specifically, Brochothrix thermosphacta showed hits for several proteins exhibiting identities up to 100% (Table 1). Planococcus antarcticus, Bacillus cytotoxicus, Lactococcus lactis subsp. Lactis and Staphylococcus aureus showed identity ranges from 62% to 80% (Table 1). Furthermore, the nucleotide analysis showed hits for diverse bacteria belonging to genus Listeria, Bacillus and Paenibacillus (Table 2).

Only three out of 19 assignments to fungi clades satisfied the selection parameters; two hits supported the taxonomical levels of phylum and genus for Fusarium oxysporum and one hit classified to the genus taxonomical level for Neurospora tetrasperma (Table 1). In addition, nucleotide analysis (Table 2) showed three assignments for Mrakia frigida (rRNA genes, two hits) and Togninia minima (putative polyubiquitin protein mRNA, one hit).

A parasite associated with orange colored protrusions

The histological approach showed a tissue lesion characterized by the aggregation of hemocytes (cells endowed with phagocytic and immune-related functions) inside isolated foci surrounded by smooth muscle fibers and a basophilic tissue contiguous to a lamellated membrane (Fig. 2A). Additionally, some lesions exhibit interstitial immunocyte inclusions with morphology similar to a granulation process (Fig. 2B). Although the histological approach did not allow for detectection of key features for identification, the microscopic images showed structures around 0.55 mm in diameter, which are compatible with immature developmental stages of a trematode.

Furthermore, the metagenomic analysis assigned 22 contigs to the trematode parasites clade. Specifically, these contigs had hits to an endonuclease-reverse transcriptase of Schistosoma mansoni (17) and Schistosoma japonicum (4), showing identities above 46% and 42%, respectively. Similarly, in the nucleotide analysis, seven contigs showed identities above 71% for different regions of two chromosomes of S. mansoni (Table 3).

We successfully amplified and sequenced a 740 bp region that confirmed the presence of trematode DNA in the L. gigas tissue (GenBank accession number KR092371). Moreover, this sequence clustered in a basal position to the suborder Xiphidiata (Trematoda: Digenea), which encompasses Renicola and Helicometrina genera (posterior probability: 0.98; Fig. 3). Additionally, the BLASTN results showed hits for some members of Xiphidiata, such as Helicometrina labrisomid (query coverage: 89%; identity: 77%), Renicola cerithidicola (query coverage: 70%; identity: 78%), Synthesium pontoporiae (query coverage: 42%; identity: 83%) and Haematoloechus sp. (query coverage: 39%; identity: 77%).

Bacteria and fungi associated with orange colored protrusions

The scope of the massive sequencing approach allowed the detection of some bacteria previously reported as microbiota associated with L. gigas (Acosta et al., 2009; Pérez et al., 2014), as well as new reports. For instance, Psychrobacter sp. was found in the L. gigas muscle in both nucleotide and protein analyses (Tables 1 and 2). This outcome corroborates previous studies that found Psychrobacter sp. in environmental (Acosta et al., 2009; Pérez et al., 2014) and tissue (Pérez et al., 2014) samples from L. gigas.

However, this study also found bacteria and fungi that have not been reported so far in L. gigas. Specifically, homologous protein and nucleotide sequences of species (e.g., Carnobacterium jeotgali), family and genus of Carnobacterium sp. were detected in the L. gigas muscle (Tables 1 and 2). Carnobacterium strains have been reported to inhabit live fish and a variety of seafood, dairy and meat (Leisner et al., 2007).

In addition, this research found homologous protein and nucleotide sequences of genus Bacillus and Enterococcus in the affected tissue of L. gigas. Bacillus species have ubiquitous distribution, inhabiting different environments such as soils, rocks, vegetation, foods and waters (Nicholson, 2002). Similarly, the ubiquitous nature of enterococci determines their frequently being found in foods as contaminants, although their predominant habitat is human and animal gastrointestinal tracts (Giraffa, 2002). However, they also occur in soil, surface waters, vegetables and fermented foods such as sausages, meat and cheese (Giraffa, 2002; Foulquié et al., 2006).

Furthermore, another bacteria present in the sample was Brochothrix thermosphacta, as it was assigned to bacterial species or genus taxonomical levels according to Monzoorul Haque et al. (2009). This bacterium, closely related to Listeria, is a non-proteolytic food spoilage organism in prepacked meats and fish products (Gardner, 1981; Lannelongue et al., 1982; Pin, García de Fernando & Ordóñez, 2002). In addition, some Listeria hits were detected in the nucleotide analysis, although the identity values did not allow species identification. This result is concordant with studies that have isolated Listeria members from freshwater and marine environments (Colburn et al., 1990; El Marrakchi, Boum’handi & Hamama, 2005).

Metagenomic analysis also showed some fungi assignments related to Fusarium, Neurospora, Togninia and Mrakia. Both Fusarium and Neurospora exhibit wide distribution, including humid tropical and subtropical marine environments (Steele, 1967; Turner, Perkins & Fairfield, 2001; Babu et al., 2010; Summerell et al., 2010; Jebaraj et al., 2012; Saravanan & Sivakumar, 2013; Kumar, Gousia & Latha, 2015). Specifically, some Fusarium species are associated with infections in crustaceans and cultivated fishes (Hatai, 2012), whereas other species are endosymbionts of some seaweeds (Suryanarayanan, 2012), corals (Raghukumar & Ravindran, 2012) and some sea sponges (Höller et al., 2000; Wang, Li & Zhu, 2008; Liu et al., 2010; Paz et al., 2010).

In contrast, Togninia and Mrakia show more restricted distributions. For instance, Togninia comprises pathogenic fungi responsible for the development of wood diseases, and some strains have been isolated from submerged wood from streams, lakes, ponds, reservoirs and ditches (Hu, Cai & Hyde, 2012). Likewise, several Mrakia species have been isolated from icy environments, including meltwaters from glaciers and permafrost in Antarctica (Hua et al., 2010; Pathan et al., 2010; Carrasco et al., 2012; Zhang et al., 2012; Tsuji et al., 2013a; Tsuji et al., 2013b), Argentina (Brizzio et al., 2007; De Garcia, Brizzio & Broock, 2012), the Qinghai-Tibet Plateau (Su et al., 2016), Italy (Turchetti et al., 2008; Branda et al., 2010; Thomas-Hall et al., 2010) and the Arctic (Pathan et al., 2010).

Considering that several of the new bacteria reports are related to food microorganisms, we hypothesized that they might grow under environmental or freezing conditions instead of being native microbiota. Fungi findings suggest an environmental source; however, since some species of Fusarium and Neurospora produce orange spores (Davis & Perkins, 2002; Hatai, 2012), the colored protrusions found in L. gigas may be due to an opportunistic or primary fungal infection. Thus, the role of bacteria and fungi in the muscle of L. gigas and their relationship with the lesion, native microbiota or the environment remains to be explored.

A parasite associated with orange colored protrusions

Histology showed evidence of a membrane, that is consistent with a syncytium, enclosing a multicellular parasite, a mollusk inflammatory response elicited by hemocytes (De Vico & Carella, 2012). Moreover, such membranes are also compatible with the wall layers of the life cycle stage of Platyhelminthes, suggesting a possible infection by trematodes that infect other mollusks (Cake, 1976; Sorensen & Minchella, 2001). This finding was supported by the metagenomics analysis that showed sequences homologous to an endonuclease-reverse transcriptase of some species of trematodes like Schistosoma (Table 3). This result is expected since highly repetitive sequences, such non-LTR retrotransposons with an estimated copy number going up to 24,000, are more likely to be detected in whole genome shotgun amplification (DeMarco et al., 2005). Since databases of protein and nucleotide sequences are currently enriched with Schistosoma, but exhibit a poor representation of most members of Trematoda, these assignments require cautioned interpretation. Moreover, the lack of information in reference databases for most of the sequences from the studied sample (∼90% of nucleotide sequences and ∼52% of proteins) explains the relatively low number of hits for the parasite compared with Trematoda, bacteria and fungi taxa. Although these assignments are biased by the nucleotide and protein sequences available in the NCBI databases, it supports the histological finding that the protrusions may be caused by a trematode.

The Bayesian tree supported the last result due to clustering of the sample in a basal position to the suborder Xiphidiata (Trematoda: Digenea), which includes Renicola species that produce colored pigments (Stunkard, 1950; Galaktionov & Skirnisson, 2000). The phylogenetic relationships with Xiphidiata were consistent with the BLASTN analysis that revealed genetic similitudes between the sequence found in this study and Renicola, Helicometrina, Synthesium and Haematoloechus, although its genetic distance with other members of these genera remains to be determined due to the lack of information for cytochrome c oxidase I and endonuclease-reverse transcriptase sequences of these taxa.

The molecular findings let us hypothesize about the structures approximately 0.55 mm in diameter found in the snail muscle tissue, although histology did not allow detection of key features for its identification. According to the life cycle described for Xiphidiata, the microscopic life cycle stage found in the muscle of L. gigas could be sporocysts, which are described to preferentially infect gonads and digestive glands, but can also disperse to other tissues in the form of more sporocysts or rediae (Cribb et al., 2003). Based on the structure’s size, other stages such as metacercaria seems unlikely at least in Renicola, as they exhibit 0.12 to 0.16 mm in diameter (Stunkard, 1964). However, the evidence presented here is not enough to conclude which parasitic stage was observed within the colored lesions.

In conclusion, this study found evidence of a trematode infection, as well the presence of fungi and bacteria in the protruded muscle of L. gigas, which provides novel information for the parasitology and microbiology of this species. This first insight of a trematode infection in L. gigas is a baseline to expand the toolset to identify these organisms, the trematode life cycle, environmental conditions that trigger its appearance and epidemiological aspects regarding the host and possible effects on human health.