Insights into the innate immunity of the Mediterranean mussel Mytilus galloprovincialis

Abstract

Background: Sessile bivalves of the genus Mytilus are suspension feeders relatively tolerant to a wide range of environmental changes, used as sentinels in ecotoxicological investigations and marketed worldwide as seafood. Mortality events caused by infective agents and parasites apparently occur less in mussels than in other bivalves but the molecular basis of such evidence is unknown. The arrangement of Mytibase, interactive catalogue of 7,112 transcripts of M. galloprovincialis, offered us the opportunity to look for gene sequences relevant to the host defences, in particular the innate immunity related genes.

Results: We have explored and described the Mytibase sequence clusters and singletons having a putative role in recognition, intracellular signalling, and neutralization of potential pathogens in M. galloprovincialis. Automatically assisted searches of protein signatures and manually cured sequence analysis confirmed the molecular diversity of recognition/effector molecules such as the antimicrobial peptides and many carbohydrate binding proteins. Molecular motifs identifying complement C1q, C-type lectins and fibrinogen-like transcripts emerged as the most abundant in the Mytibase collection whereas, conversely, sequence motifs denoting the regulatory cytokine MIF and cytokine-related transcripts represent singular and unexpected findings. Using a cross-search strategy, 1,820 putatively immune-related sequences were selected to design oligonucleotide probes and define a species-specific Immunochip (DNA microarray). The Immunochip performance was tested with hemolymph RNAs from mussels injected with Vibrio splendidus at 3 and 48 hours post-treatment. A total of 143 and 262 differentially expressed genes exemplify the early and late hemocyte response of the Vibrio-challenged mussels, respectively, with AMP trends confirmed by qPCR and clear modulation of interrelated signalling pathways.

Conclusions: The Mytibase collection is rich in gene transcripts modulated in response to antigenic stimuli and represents an interesting window for looking at the mussel immunome (transcriptomes mediating the mussel response to non-self or abnormal antigens). On this basis, we have defined a new microarray platform, a mussel Immunochip, as a flexible tool for the experimental validation of immune-candidate sequences, and tested its performance on Vibrio-activated mussel hemocytes. The microarray platform and related expression data can be regarded as a step forward in the study of the adaptive response of the Mytilus species to an evolving microbial world.

Figure 1

Multiple alignment of the four AMP families identified in Mytibase. According to Muscle alignment and BioEdit processing, each mussel AMP family includes new sequences indicated by an asterisk (*). Identical and similar (PAM250) residues are reported in dark and light grey, respectively.

Figure 2

Diversity of mussel transcripts containing the C1q domain. Cladistic analysis of 65 C1q domains representing a subset of the 168 total sequences identified in Mytibase. This highly diversified non redundant subset was produced with NCBI BlastClust (-p F -L .5 -b T -S 40). The cladogram is based on the Neighbor-Joining method (1000 bootstrap replicates). The evolutionary distances were computed by using the JTT substitution matrix (number of amino acid substitutions for each of the 406 sites).

Figure 3

Multiple alignment of 87 Mytibase sequences identified as C-lectins (IPR001304). The SMART consensus terms for the CLECT domain (SM00034) are shown at the bottom. Four cystein residues are entirely conserved (positions 12, 105, 123 and 138 in the multiple alignment). Other conserved (75%) residues are W5, A8, L19, W51, G53, G64, W66, W68, W118, D120, I137, C138 and E139. Positions with at least 50% of conservation are also shadowed.

Figure 4

Hierarchical clustering of the Immunochip profiles referring to 3 and 48 h post-injection of live V. splendidus. Two mussel groups with 8 replicates per time point have been analyzed (R-package software). Scale of the expression values and probe ID are indicated. For space reasons, only instructive parts of the clustered data are reported.

Figure 5

Modelling the putative transcriptional hemocyte response to the Vibrio injection. The model outlines PRR, interrelated signalling pathways and cell processes significantly modulated at 3 and 48 h post-treatment. Various bacterial PAMPs and DAMPs are indicated in pink. The gene functions represented on the Immunochip are indicated in bold according to sequence similarities/identities: framed bold characters on red, green and yellow fields indicate over-expression, under-expression and heterogeneous trends, respectively, of specific Immunochip probes at 3 h (left half) and 48 h (right half). Other gene functions not represented in Mytibase have been introduced to recall specific molecular pathways. Similarities resulting from InterproScan Analysis are reported in brackets whereas an asterisk indicates annotations based on manual inspection of relevant similarities. All the acronyms with related extended names are reported in the List of Abbreviations.