Integration of Transcriptomics and Proteomics Improves the Characterization of the Role of Mussel Gills in a Bacterial Waterborne Infection

fmars-08-735309 September 29, 2021 Time: 16:23 # 1 ORIGINAL RESEARCH published: 05 October 2021 doi: 10.3389/fmars.2021.735309 Edited by: Ciro Rico, Institute of Marine Sciences of Andalusia (ICMAN), Spain Reviewed by: Alberto Pallavicini, University of Trieste, Italy Ana Varela Coelho, Universidade Nova de Lisboa, Portugal *Correspondence: Angel Pérez Diz [email protected] Antonio Figueras [email protected] Specialty section: This article was submitted to Marine Molecular Biology and Ecology, a section of the journal Frontiers in Marine Science Received: 05 July 2021 Accepted: 09 September 2021 Published: 05 October 2021 Citation: Saco A, Panebianco A, Blanco S, Novoa B, Diz AP and Figueras A (2021) Integration of Transcriptomics and Proteomics Improves the Characterization of the Role of Mussel Gills in a Bacterial Waterborne Infection. Front. Mar. Sci. 8:735309. doi: 10.3389/fmars.2021.735309 Integration of Transcriptomics and Proteomics Improves the Characterization of the Role of Mussel Gills in a Bacterial Waterborne Infection Amaro Saco1, Antonella Panebianco2,3, Sofía Blanco2,3, Beatriz Novoa1, Angel P. Diz2,3* and Antonio Figueras1* 1Instituto de Investigaciones Marinas (CSIC), Vigo, Spain, 2Centro de Investigación Mariña da Universidade de Vigo (CIM-UVigo), Vigo, Spain, 3Department of Biochemistry, Genetics and Immunology, University of Vigo, Vigo, Spain In recent years, the immune response of mussels (Mytilus galloprovincialis) has been studied at the transcriptomic level against several bacterial infections. As a result, different immune mechanisms have been revealed, including both conserved essential innate pathways and particularities of the mussel immune response according to its nature and environment. However, there is often a lack of functional verification because mussels are a non-model species and because transcriptomic and proteomic information is not always well correlated. In the current study, a high-throughput quantitative proteomics study coupled to LC-MS/MS analysis using isobaric tandem mass tags (TMTs) for protein labeling was employed to study the mussel gill immune response to a Vibrio splendidus bath (waterborne) infection at a functional protein level. A total of 4,242 proteins were identified and quantified, of which 226 were differentially expressed (DEPs) after infection, giving to the study a depth that was lacking in previous proteomic studies of the bivalve immune response. Modulated proteins evidenced an important cytoskeletal disruption caused by bacterial infection. A conserved network of associated proteins was modulated, regulating oxidative stress and NF-kB inflammatory responses and leading to innate immunity effectors. Proteomic results were submitted to an integrated analysis with those obtained in a previous transcriptomic approach with the same infection. Half of all the quantified proteins had a concordant transcriptomic expression trend, but this concordance increased when focusing on the DEPs. The correlation was higher within the immune-related DEPs, and the activation of the conserved NF-kB pro-inflammatory pathway was the main response in both approaches. The results of both techniques could be integrated to obtain a more complete vision of the response. Keywords: proteomic, transcriptomic, immune, mussel, bacterial infection, mollusc Frontiers in Marine Science | www.frontiersin.org 1October 2021 | Volume 8 | Article 735309 fmars-08-735309 September 29, 2021 Time: 16:23 # 2 Saco et al. Mussel Gill Proteomic Immune Response INTRODUCTION Mussels (Mytilus galloprovincialis) have great importance in scientific research for different reasons. These bivalves are among the most produced species in aquaculture, having a nonnegligible impact on the European economy (Avdelas et al., 2020). Mussels are filter feeders in contact with large amounts and diverse potential pathogens, and despite this, they do not commonly suffer severe mortality episodes. Mussels can use their efficient innate immune system to deal with different pathogenic agents in a specific way (Costa et al., 2009;Balbi et al., 2013). This attribute is supported by the vast genetic variability that characterizes this species, with several expanded immune gene families as immune recognition C1q factors (Gerdol et al., 2011) or pro-inflammatory cytokines as IL-17 (Saco et al., 2021). The recent discovery of mussel genomic organization as a pangenome revealed presence/absence phenomena in the genetic repertoires of these animals, where a core set of genes was accompanied by great amounts of dispensable and variable genes (Gerdol et al., 2020). Gene families with immune importance are enriched in the dispensable and variable genomic fraction, such as the antimicrobial peptide myticins (Rey-Campos et al., 2020). To reveal how mussels use their immune system to deal with infections, transcriptomic approaches have been quite successful (Figueras et al., 2019). Classical approaches have studied the transcriptome of the main immune cells, hemocytes, against in vitro and in vivo stimulation with different pathogenic bacteria or pathogen-associated molecular patterns (PAMPs) (Rey-Campos et al., 2019a,b;Moreira et al., 2020). In addition to hemocytes, mucosal immunity should have great importance in filtering bivalves such as mussels. Transcriptomic studies indeed revealed interesting immune expression patterns in more tissues beyond hemocytes (Moreira et al., 2015). A transcriptomic study performed in mussel gills against a bacterial bath (waterborne) infection tried to mimic how mussels recognize an infection from the surrounding water as in natural conditions. The modulation of different pathogen recognition and immune triggering genes revealed that gills may play a natural first step in the activation of the immune response in these animals (Saco et al., 2020). These transcriptomic studies described both essential key immune mechanisms that would therefore be conserved during evolution up to higher vertebrates, as well as particularities adapted to the mussel environment, such as the high variability of certain genes (Rey-Campos et al., 2020), to survive in the ever-changing marine environment. In addition to transcriptomic studies, more functional approaches have been used to improve our understanding of the mussel immune system. Traditional proteomic analyses consist of two-dimensional electrophoresis (2-DE)-based methods and have proven to be useful in several analyses on mussels and other non-model species. However, high-throughput and more effective techniques emerged, allowing the identification of many more proteins in each analysis and a more accurate relative quantification, such as iTRAQ (isobaric tag for relative and absolute quantitation) and TMT (tandem mass tag) approaches, both coupled to liquid chromatographytandem mass spectrometry (LC-MS/MS). They allow much larger identification rates, higher precision and accuracy in relative quantification than traditional 2-DE gel assays (Aslam et al., 2017). Different proteomic studies have been conducted with mussels and other bivalves subjected to different stimuli. Mussel and other bivalve gills have been used in proteomic studies on environmental stress or in studying the bioaccumulation and toxic effects of different water pollutants. Traditional 2-DE analyses identified a limited number of differentially expressed proteins (DEPs) after the analysis of all the protein spots (Manduzio et al., 2005;Ji et al., 2013b;Song et al., 2016;Chen H. et al., 2018). A big leap in the magnitude of the analyses and the number of identified and quantified proteins has been reported with the most recent LC-MS/MS technological advances (Zhang et al., 2015;Sánchez-Marín et al., 2021). Proteomics has also been employed for the study of the immune response in bivalves, as in the scallop hepatopancreas infected by Vibrio (Huan et al., 2011), for characterization studies of the immune proteins expressed in hemocytes from different bivalves (Campos et al., 2015;Leprêtre et al., 2020), for the study of the immune response in mussel gills to the intramuscular injection of bacteria (Ji et al., 2013a), among other studies (Castellanos-Martínez et al., 2014;Dinguirard et al., 2018; Jiang et al., 2018;Smits et al., 2020). Most likely, the most studied infection model in bivalve proteomics is oyster infection by herpesvirus and virus-like stimuli (Corporeau et al., 2014; Masood et al., 2016;Leprêtre et al., 2021). Few immune proteomic responses to bacterial infections have been reported in mussels or in other bivalves, and the evolution of proteomic techniques since the 2DE analyses caused some past studies to have severe limitations. The objective of the current work was to carry out a high-throughput proteomic study in gills of the marine mussel Mytilus galloprovincialis subjected to a waterborne infection with Vibrio splendidus, simulating natural infection conditions by a bloom of certain infective bacteria. The infection model followed the same design as a previous transcriptomic experiment, which had revealed key mechanisms in gills in terms of pathogen recognition and activation of the immune response (Saco et al., 2020). Multiplexing of TMTlabeled protein samples coupled to liquid chromatographytandem mass spectrometry (LC-MS/MS) was applied to this proteomic study, and the results were also complemented with those obtained by the previous transcriptomic approach with the objective of gaining a more complete understanding of the marine mussel response to this infection model. MATERIALS AND METHODS Animals, Experimental Infection, and Sampling Adult mussels (Mytilus galloprovincialis) were obtained in May from a mussel farm (Vigo, Galicia, NW Spain) and maintained in open-circuit filtered seawater (FSW) tanks at 15◦C with aeration. The mussels were fed daily with Phaeodactylum tricornutum and Isocrhysis galbana. The mussels were used for the experiments after at least 1 week of acclimatization. Frontiers in Marine Science | www.frontiersin.org 2October 2021 | Volume 8 | Article 735309 fmars-08-735309 September 29, 2021 Time: 16:23 # 3 Saco et al. Mussel Gill Proteomic Immune Response Vibrio splendidus (reference strain LGP32) was cultured in tryptic soy agar (TSA) plates at 22◦C for 24 h before use. Bacterial suspensions were prepared in FSW before infection. Serial dilutions were made to calculate the number of colony forming units (CFU)/mL 24 h after inoculation. Waterborne infections were performed by treating mussels with a concentration of 108CFU/mL V. splendidus in tank water. Control mussels were maintained in another tank with only filtered seawater (FSW). Gill samples from 5 control and 5 infected mussels were taken 24 h after the infection. To perform the sampling, gills were cautiously separated from gonads and mantle and frozen directly at −80◦C. Protein Quantification Gill samples were homogenized with several ultrasonication cycles (5-s pulses, with 30-s ice stops and 20% amplitude) in lysis buffer consisting of 7 M urea, 2 M thiourea and 4% CHAPS. Once the tissue was correctly disaggregated and proteins released and solubilized in the buffer, samples were centrifuged (14,000 g, 60 min, 4◦C), and the supernatants were collected in 50-µL aliquots and frozen at −80◦C. Protein quantification was performed following the Bradford method in microplates. Both BSA standards and samples were analyzed in triplicate, and the absorbance was read at 595 nm in a BIO-RAD iMarkTM Microplate Reader. After quantification, samples were diluted to an equal concentration of 0.5 µg/µL. Protein Precipitation To perform the precipitation step, 600 µL of prechilled acetone (−20◦C) was added to 100 µL of each sample. Samples were incubated for 2 h at −20◦C to ensure precipitation. After the incubation step, the samples were centrifuged (8,000 g, 10 min, 4◦C). The supernatant consisting of acetone was discarded, and the pellet was dried for a maximum of 10 min before adding 100 µL of 100 mM TEAB (triethylammonium bicarbonate). A brief bath sonication was performed to resuspend the pellet. Then, 5µL of TCEP (Bond-Breaker TCEP Solution) was added, and samples were incubated under soft shaking conditions (300 rpm) for 1 h at 55◦C. To alkylate the samples, 5 µL of 375 mM iodoacetamide (diluted in 100 mM TEAB) was added, and the samples were incubated for 30 min at room temperature and under dark conditions. Finally, 10 µL of trypsin (0.1 µgµL−1 in 100 mM TEAB) was added, and the samples were incubated at 37◦C overnight with soft shaking at 300 rpm. Because we began protein precipitation with 100 µL of homogenized protein sample at 0.5 µg/µL, we expected to precipitate a maximum of 50 µg of proteins per sample. Isobaric Tagging Labeling of peptides was performed using the TMT 10plexTM Isobaric Mass Tagging Kit (Thermo Fisher Scientific, Waltham, MA, United States). Each of the 10 TMT tags was used for each of the 10 samples: 5 control (C1–C5) and 5 infected (I1–I5) samples (126 label: C1; 127N label: C2; 127C label: C3; 128N label: C4; 128C label: C5; 129N label: I1; 129C label: I2; 130N label: I3; 130C label: I4; 131 label: I5). Tags (0.8 mg) were dissolved in 90 µL of anhydrous acetonitrile, and 41 µL of each equilibrated tag was added to their corresponding sample after incubation with trypsin. After incubation for 1 h at room temperature, 8 µL of 5% hydroxylamine was added. Samples contained 50 µg of peptides in a volume of 169 µL. Peptide Fractionation The 10 already labeled samples were combined in the same tube at equal concentrations, for which 10.3 µL was added from each sample, and then the final volume was raised to 300 µL with TFA (trifluoroacetic acid) 0, 1%. The mix consisting of 300 µL with an estimated peptide amount of 30 µg was submitted to peptide fractionation using the Pierce High pH Reversed-Phase Peptide Fractionation Kit (Thermo Fisher Scientific) following the kit protocol. After conditioning and washing the column, a 300-µL sample mix was applied to the column. Elution solutions consisted of a series of 8 solutions with increasing acetonitrile concentrations from 12.5 to 50% supplemented with 0.1% triethylamine. Therefore, 8 fractions or elutions were obtained, and each one of them was divided in 2 aliquots of 150 µL with approximately 2 µg of peptides. Liquid Chromatography-Tandem Mass Spectrometry Analysis Each 150 µL aliquot was dried using a SpeedVack (Thermo Fisher Scientific), and 2 µg of peptides from each aliquot was suspended in 13 µL of 0.5% formic acid. A brief bath sonication for 10 min was followed by centrifugation (14,000 rpm, 1 min, 10◦C). From each aliquot, 10 µL corresponding to 1.5 µg of peptides was analyzed with LC–MS/MS using a Proxeon EASYnLC II liquid chromatography system (Thermo Fisher Scientific) coupled to an LTQ-Orbitrap Elite mass spectrometer (Thermo Fisher Scientific). Protein Identification To identify the proteins using the MS/MS spectra, a customized protein database was generated from the six-frame translation of the 162,167 contig sequences from the mussel gill transcriptome performed under the same bacterial bath infection stimulus (Saco et al., 2020). MS/MS spectra were searched against this database using PEAKS Studio v8.0 software (Bioinformatics Solutions Inc., Waterloo, Canada). A contaminant database (CRAPome) was included, as well as a decoy sequence database to calculate the false discovery rate (FDR) in the peptide identification, which was set to 1% FDR for PSMs (Peptide-Spectrum Matches). Trypsin was indicated to be the enzyme for protein cleavage, and only one missed cleavage was allowed. For peptide identification, the precursor and fragment ion mass tolerances were set to 10 ppm and 0.02 Da, respectively. Carbamidomethylation of cysteine and TMT 10 plex were set as fixed modifications, and oxidation on methionine was set as a variable modification. For protein identification, the number of matched peptide sequences was set to ≥2, the number of unique peptides to ≥1 and the protein identification PEAKS score to ≥20. In addition to the mentioned self-built protein database, for comparative purposes, the same procedure (both the identification step and the downstream analyses) was repeated Frontiers in Marine Science | www.frontiersin.org 3October 2021 | Volume 8 | Article 735309 fmars-08-735309 September 29, 2021 Time: 16:23 # 4 Saco et al. Mussel Gill Proteomic Immune Response using the NCBI non-redundant protein database restrained to Mollusca (23/05/2019; 749,599 sequences). Protein Quantification For protein quantification, only unique peptides were considered. The spectrum filter for the reporter ions was set to FDR 1% and quality ≥10. Intra-experiment normalization was performed by calculating an equalizer factor to normalize the reporter ion intensities for each TMT 10plex label using the total ion counts per channel. After intra-normalization, a logarithmic transformation was applied to the normalized protein abundance data to meet the normality and homoscedasticity assumptions necessary to carry out parametric statistics, which is a standard procedure for proteomic datasets. Differential expression analysis was performed with a t-test followed by a multi-test correction. These corrections were calculated using SGoF + software v.3.8 (Carvajal-Rodriguez and de Uña-Alvarez, 2011) following the strategy previously described in Diz et al. (2011). Adjusted p-values for false discovery rate (q-values, Storey, 2003) were obtained and set to ≤0.05 as the criterion to define DEPs. Proteomic Data Analysis Annotation of the identified and quantified proteins was derived from the annotation of the transcriptome used to construct the identification database. This annotation was performed using a Blastx approach on OmicsBox (BioBam Bioinformatics, Valencia, Spain) (Götz et al., 2008) against UniProt/Swiss-Prot with an e-value of 10−3. The annotation of DEPs was manually verified with Blastp against the NCBI non-redundant protein database. Two enrichment analyses, both of which were performed with generic GO slim terms and with specific GO terms, were performed with Fisher’s exact test and FDR ≤0.05. The first analysis considered all the identified and quantified proteins as the test set against the reference set (identification database) to analyze which types of proteins were identified in the proteome (a sort of functional clustering). The other enrichment analysis was performed with the modulated proteins against the reference (all the identified proteins) to associate enriched terms with the response to the infection. Enrichment analyses were represented using a ratio between the weight of each term in the test set and the weight in the reference set. Thus, processes with better test/reference ratio will be of interest as they are more enriched in the studied test set than in the reference set. A protein interaction network was built using the STRING algorithm (Szklarczyk et al., 2015) and the modulated proteins. Default parameters were employed, including the use of human homologs to the mussel proteins as a reference, a common procedure with non-model species that are not included in the server. Only high confidence interactions were shown (0.7 minimum required interaction score) and disconnected nodes were removed. Enriched biological processes in the network were also analyzed with the STRING algorithm. A second STRING network was constructed, following the same criteria (0.7 minimum score and no disconnected nodes), using homolog proteins from a phylogenetically closer species, the marine invertebrate Strongylocentrotus purpuratus. Both analyses were then compared to check which interactions are recovered using the homologs from each species and which ones are shared by both analyses. Heatmaps and PCA were constructed using ClustVis (Metsalu and Vilo, 2015). For the PCA, singular value decomposition was used, and for the heatmaps, distances were calculated with the unit variance scaling method. Clustering was performed based on average correlation. Schemes were created with BioRender.com. Proteomic and Transcriptomic Data Analysis Transcriptome analysis performed on gills from the same marine mussel species (M. galloprovincialis) after waterborne V. splendidus infection (Saco et al., 2020) was employed both as a database for proteomic analysis and to retrieve transcriptomic expression of the detected proteins. The experimental design was the same for both approaches, with the same concentration of bacteria in the infection (108CFU/mL V. splendidus) and the same sampling time (24 h). The transcriptomic analysis was performed with 3 biological replicates for each control and infected group instead of the 5 biological replicates employed in the proteome analysis. Reads accessible from Bioproject PRJNA638821 under the accessions SRR11996464, SRR11996686, and SRR11996723 for control samples and SRR11996734, SRR11996735, and SRR11996743 for infected samples were assembled with the parameters indicated in Saco et al. (2020). Expression of transcriptomic contigs equivalent to the modulated proteins in the proteome was retrieved in transcripts per million (TPMs). Log2 transformation of the fold change values was used for the correlation analyses between proteomic quantification and transcriptomic expression. Linear regression algorithm analyses were made in R. RESULTS Characterization of the Mussel Gill Proteome The use of the specific mussel gill transcriptome (Saco et al., 2020) allowed us to achieve better protein identification rates in the analysis than the NCBI database (Table 1). After the LCMS/MS analysis, 245,047 MS/MS spectra were obtained, and using the mentioned mussel gill transcriptome as a database, 35,217 peptide spectrum matches (PSMs) were retrieved. A total of 4,388 proteins were identified from 28,141 peptide sequences based on the parameters used. Considering only unique peptides that contained signals from reporter (TMT) ions, 4,242 proteins could be quantified. Protein relative abundance data could be retrieved for those proteins. Enrichment analyses were performed with all the identified and quantified proteins to reveal a general characterization of what type of proteins dominate the mussel gill proteome. For this purpose, GO Slim terms were employed. This broad overview of the proteome is shown in Figure 1A for the molecular functions, Figure 1B for the cellular components and Figure 1C for the biological processes. More than half of the proteins were related to binding functions, both inside the cell and in the plasma Frontiers in Marine Science | www.frontiersin.org 4October 2021 | Volume 8 | Article 735309 fmars-08-735309 September 29, 2021 Time: 16:23 # 5 Saco et al. Mussel Gill Proteomic Immune Response TABLE 1 | Protein identification and determined DEPs using the transcriptome database and the mollusc NCBI nr protein database. Database Peptide-spectrum matches (FDR 1%) Peptide sequences Identified proteins Quantified proteins Differentially expressed proteins DEPs (q-value <0.05) Mollusc NCBI 9,259 7,093 3,497 1,459 188 Gill transcriptome 35,217 28,141 4,388 4,242 226 Information about the transcriptome database is highlighted in gray as this database allowed better results and was the one employed for the whole downstream analysis. FIGURE 1 | General characterization of gill proteome from the marine mussel (M. galloprovincialis) based on the confidently identified proteins. Identified proteins in the proteomic analysis were submitted to an enrichment analysis with GO-Slim terms to obtain a generic view of the type of proteins that were identified in the gill proteome. The top 15 groups of GO-Slim terms with the highest number of representative proteins are shown for (A) “molecular function”; (B) “cellular component”; and (C) “biological processes.” (D) The most present specific biological processes are also represented without GO-Slim generalization. membrane. These proteins are implicated in functions such as response to stress, signal transduction and immune processes, among others. Biological processes of the general proteome were also studied with specific GO terms, without GO slim generalization, specifying the importance of gill proteins in the recognition of exogenous antigens and in the transduction of C-type lectin, Fc-epsilon receptor, interleukin or Wnt signaling pathways, among others (Figure 1D). Modulated Proteomic Response in Mussel Gills Against Bacterial Bath Infection Based on the quantification data, after applying the statistical methodology explained, a final number of 226 DEPs were obtained (q-value ≤0.05). The DEPs and the rest of the quantified proteins are displayed in Supplementary Table 1. Modulated proteins displayed a clear difference between the two conditions (Figure 2A), and no great variability was observed between biological replicates within the control and infected groups of samples (Figure 2B). The biological processes of the modulated proteins were first revealed with enrichment analyses of the upregulated and downregulated DEPs. Functions such as activation of innate immune cells/response and response to oxidative stress are among the most common upregulated proteins against infection (Figure 3A). Other apparently less-related immune terms, more focused on regulatory functions, response to ions, and regulation of gene expression, are also directly related to the response to the infection. Cilia movement, as expected in the gills, involved in cell motility was the dominant process in the downregulated proteins (Figure 3B). Immune processes downregulated with the infection were response to cytokine stimulus, response to external stimulus and oxidation-reduction processes. Frontiers in Marine Science | www.frontiersin.org 5October 2021 | Volume 8 | Article 735309 fmars-08-735309 September 29, 2021 Time: 16:23 # 6 Saco et al. Mussel Gill Proteomic Immune Response FIGURE 2 | Proteomic modulated response to a bacterial bath infection in mussel gills. (A) Heatmap showing the 226 differentially expressed protein DEPs (q-value <0.05). Upmodulated and downregulated proteins are indicated with different colors in the legend. (B) Principal component analysis (PCA) shows clear differences between the control (C1–5) and infected (I1–5) groups for these 226 DEPs. FIGURE 3 | Enrichment analysis showing the top 30 enriched biological processes in the upregulated (A) and downregulated (B) differentially expressed proteins (DEPs). The 30 most overrepresented biological processes (Fisher’s exact test, FDR <0.05) were represented for both the upregulated (A) and downregulated (B) DEPs. Biological processes are represented in accordance with their test/reference proportion (test = presence of a particular process in the studied DEPs; reference = presence of a particular process in the whole proteome). Focusing on specific functions of the modulated proteins, many of them were related to the cytoskeleton or to immune pathway regulation, as revealed by the GO Slim biological processes represented in Figure 4A.Table 2 also displays the identity of the DEPs specifically related to the mentioned cytoskeleton-related and immune-related functions Frontiers in Marine Science | www.frontiersin.org 6October 2021 | Volume 8 | Article 735309 fmars-08-735309 September 29, 2021 Time: 16:23 # 7 Saco et al. Mussel Gill Proteomic Immune Response FIGURE 4 | Modulated proteins (DEPs) enriched in cytoskeletal and immune functions could conform to a conserved protein interaction network. (A) GO-Slim Enrichment Analysis of the most present biological processes in the modulated proteins. Biological processes are represented in accordance with their test/reference proportion (test = presence of a particular process in the studied DEPs; reference = presence of a particular process in the whole proteome). Cytoskeletal and immune functions are highlighted because of their importance in the proteomic response. (B) Significant protein interaction net conformed by human homologs to the modulated proteins in the analysis (DEPs). The interaction network is enriched in cytoskeletal and immune-related proteins, which are highlighted in the same color legend as in (A). The thickness of the lines represents the confidence of each interaction. The legend that indicates to which proteins the symbols of the net correspond is in Supplementary Table 2. Frontiers in Marine Science | www.frontiersin.org 7October 2021 | Volume 8 | Article 735309 fmars-08-735309 September 29, 2021 Time: 16:23 # 8 Saco et al. Mussel Gill Proteomic Immune Response that dominated the response. All the cytoskeletal proteins that were modulated (main components such as actin and tubulin, transport proteins such as dynein and many regulators) evidenced general cytoskeletal disruption and reorganization with the infection, which will be discussed in the following sections. DEPs with immune implications were involved mainly in the regulation of the NF-kB pro-inflammatory response and the response to oxidative stress. An interaction net was constructed with human homolog proteins to the mussel DEPs (Figure 4B). The net was significant (PPI enrichment p-value <1.0e-16), meaning that proteins modulated in the analysis showed a significantly higher degree of interactions within each other than random proteins. The core of the interaction network was enriched in cytoskeletal and immune DEPs (Figure 4B and Supplementary Table 2). The comodulation of this net of associated proteins in the gill mussel proteome could suggest a conservation of these protein interactions in the mussel innate immune response with respect to the higher vertebrates in which protein interactions are usually described. Using homolog proteins from a phylogenetically closer species, the sea urchin (Strongylocentrotus purpuratus), another significant interaction net was constructed (Supplementary Figure 1A). Interactions detected using sea urchin were mostly shared with the human homologs analysis, since sea urchin interactions information mainly relied on data from homologs in the more studied model species (Supplementary Figure 1B and Supplementary Table 3). An effort was made to summarize how the modulated proteins regulate the NF-kB immune response according to their function (Figure 5). The NF-kB inflammatory response is initiated after the recognition of pathogen molecular patterns with pro-inflammatory cytokine signaling. Then, an intracellular pathway liberates NF-kB from its inhibitory factor and allows it to express pro-inflammatory and innate immune effectors. The trend was clear: proteins that can positively regulate the pathway were mainly upregulated, while inhibitory proteins were all downregulated. Figure 5 also displays the response related to oxidative stress, comprising ROSand NOS-related proteins, as well as the calcium-regulatory response that affects them. This redox stress response caused by the infection would also affect the NF-kB regulatory net, as both responses are associated, mainly through the Pirin protein. Other immune-related proteins were also modulated, affecting different functions than previously seen NF-kB and ROS. These proteins are also listed in Table 2 and display more effector roles against infection: Tubby protein homolog stimulates phagocytosis, alanine and aspartate aminotransferases, which are related to apoptosis under bacterial infection, Kinesin-like proteins related to antibacterial lytic granule secretion, Histone deacetylase 6, and granulin, among others. Gene expression was retrieved from the equivalent transcriptomic analysis (Saco et al., 2020) for all the proteins detected in the proteome. The expression of the equivalent contigs is represented for immune-related DEPs in Figure 6A. There was a general correlation between these modulated immune proteins and their transcriptomic expression trends. Furthermore, the transcriptomic modulated response agreed with the proteomic study in key processes, as will be discussed later. Transcriptomic and proteomic comparison can shed some light on the putative reasons why some expected terms do not appear in the proteomic study, i.e., the great transcriptomic response related to pathogen recognition and cytokine signaling (Saco et al., 2020) was not modulated in the proteomic experiment. Figure 6B shows how contigs corresponding to the detected proteins have higher average expression than the overall transcriptome. The depth of both approaches is a factor to consider as well since the proteome retrieved 4,388 proteins and the transcriptome retrieved 162,167 contigs. Correlation analyses between the quantitative proteomic and RNA-Seq transcriptomic experiments revealed that 139 out of 226 DEPs (60%) were concordant gene-protein interactions (gene expression and protein quantification fold change values with the same trend). Fold change values for these concordant DEPs had a significant correlation of R2= 0.4192 (Figure 6C). The correlation was more noticeable within the modulated immune proteins (Figure 6C), which is especially important since DEPs related to other functions, such as the cytoskeleton, showed null correlation (R2= 0.0641). When focusing on the entire proteome characterization instead of the modulated response, the agreement decreased, with half the terms concordant (2,137 out of 4,242 quantified proteins) and with less correlation in the magnitudes of their fold changes (Figure 6D). The complementarity between proteomic and transcriptomic experiments for the immune implicated DEPs supposes a crossvalidation for the implication of certain biological responses to the infection, as NF-kB and the oxidative responses. Other groups of proteins/transcripts did not show complementarity, as the cytoskeletal DEPs in the proteome or the modulated immune recognition response in the transcriptome. Immune recognition proteins of interest were identified and quantified in the proteome but they were not modulated in the study (listed in Table 2). Relative quantification of some of these important but non-modulated proteins is shown in Supplementary Figure 2, along with their corresponding transcriptomic expression: AIF-1, Myd88 and a toll-interacting protein. DISCUSSION After the LC-MS/MS analysis, a mussel gill proteome with high protein identification and quantification rates was obtained. Identification rates in proteomic analyses, especially in nonmodel species, are improved by using customized databases from appropriate and specific transcriptomes, as we did, rather than using general databases such as NCBI (Romero et al., 2019). The number of identified proteins is similar to the number of proteins reported in other studies in bivalves following high-throughput proteomic approaches based on ITRAQ or TMT sample labeling (Zhang et al., 2015;Sánchez-Marín et al., 2021)., far above the numbers obtained in traditional 2-DE assays (Manduzio et al., 2005;Ji et al., 2013a;Song et al., 2016;Chen H. et al., 2018). General characterization of the mussel gill proteome revealed functional information on terms that are typically found in gill transcriptomes, in particular the biological processes related to Frontiers in Marine Science | www.frontiersin.org 8October 2021 | Volume 8 | Article 735309 fmars-08-735309 September 29, 2021 Time: 16:23 # 9 Saco et al. Mussel Gill Proteomic Immune Response TABLE 2 | Differentially expressed proteins (DEPs, q-value <0.05) and non-modulated proteins of interest. Cytoskeleton-related DEPs FC Inflammatory and Defensive DEPs FC Inflammatory and Defensive DEPs FC Non-modulated proteins of interest Cytoskeleton main components NF-kB signaling regulation proteins Other immune/defense-related proteins Allograft inflammatory factor 1 Dynein light chain, cytoplasmic 6.53 Stimulator of interferon genes STING 4.27 Tumor susceptibility gene 101 protein 1.30 Toll interacting protein B Cytoplasmic dynein 1 intermediate chain 1.29 E3 ubiquitin-protein ligase TRIM32-like domain 1.84 Regulator complex LAMTOR5 homolog 1.80 MyD88 Dynein light chain, axonemal 1.24 Ubiquitin-conjugating enzyme E2 N 1.38 Transcription intermediary factor 1-B (TRIM28) −2.82 TRAF 2 Dynein beta chain, ciliary 1.33 E3 ubiquitin-protein ligase TRIM39 −1.88 Synaptosomal-associated protein 29 (Snap29) −2.73 TRAF 3 Dynein beta chain, ciliary −1.53 Parkin coregulated gene protein homolog PACRG 1.30 Histone deacetylase 6 (HDAC6) −1.85 Nuclear factor NF-kB p105 subunit Cytoplasmic dynein heavy chain, axonemal −2.10 Importin subunit alpha-4 1.30 von Willebrand factor A domain protein 3B −1.81 Tumor necrosis factor liganf TNF14 Dynein regulatory complex protein 1 −1.43 Protein NipSnap 1.23 U5 nuclear ribonucleoprot. 200 kDa helicase −1.75 L rhamnose binding lectin ELEL1 Cytoplasmic dynein 2, heavy chain −1.27 14-3-3 protein epsilon 1.20 Histone deacetylase complex subunit SAP18 −1.54 Collectin 11 Dynein beta chain, flagellar outer arm −1.26 SAMHD1 −1.84 60 kDa SS-A/Ro ribonucleoprotein −1.65 Collectin 12 Tubulin beta chain 2.16 Methylcrotonoyl-CoA carboxylase MCCC1 −1.81 Endoplasmin (HSP90b1) −1.29 Galectin 4 Tubulin polymerization-promoting protein 1.49 Pirin −1.70 Ras-related protein Rab-11 A −1.27 Galectin 6 Tubulin alpha 1 chain −1.94 Filamin C −1.50 Armadillo repeat-containing protein 6 −1.26 Galectin 8 Tubulin beta 2A chain −2.43 Filamin B −1.67 Spectrin beta chain −1.19 D galactoside specific lectin Actin cytoplasmatic −4.17 Ras association domainprotein 2 RASSF2 −1.67 Glutamate/proline-tRNA ligase (EPRS1) −1.19 Malectin Cytoskeleton regulators Fibulin 1 −1.65 Ependymin 1.69 Endoplasmatic reticulum lectin F-actin-capping protein subunit beta −1.46 Beta arrestin 1 −1.55 Heavy metal-binding protein HIP −2.34 Plectin Nesprin-1 1.83 Ubiquitin carboxyl-term-hydrolase (USP9X) −1.42 Proteasome subunit alpha type-1 1.32 Techylectin 5A Thymosin beta 1.80 50–30exoribonuclease 2 (XRN2) −1.40 26S proteasome non-ATPase regulatory subunit −1.15 Fibrinogen C domain containing proteins 4.1 protein 1.79 Armadillo repeatprotein 4 (ARMC4/ALP) −1.39 Stress related proteins: ROS, NOS, Ca2+ Fibrinogen like protein A Nephrocystin-4 1.64 Gelsolin-like protein 1 −1.38 Cytochrome c oxidase subunit 12 2.30 Complement C1r-B subcomponent Profilin 1.53 Gelsolin-like protein 2 −1.22 Methionine-R-sulfoxide reductase B2 1.54 Complement C1q subcomponent-binding prot Titin 1.31 HSPG2 −1.15 Calmodulin 1.75 Complement component C9 PDZ and LIM domain protein 3 1.33 Other immune/defense-related proteins Kinase suppressor of Ras −2.18 Complement C1q TNF related prot3 Nck-associated protein 1 −1.61 Aspartate aminotransferase 1.28 Dual oxidase (DUOX) −1.75 Myticin A WD repeat-containing protein 63 −1.46 Alanine aminotransferase 2 1.47 Hydroperoxide glutathione peroxidase −1.48 Interferon induced protein 44 like IF44L (Continued) Frontiers in Marine Science | www.frontiersin.org 9October 2021 | Volume 8 | Article 735309 fmars-08-735309 September 29, 2021 Time: 16:23 # 16 Saco et al. Mussel Gill Proteomic Immune Response Hatada, E. N., Krappmann, D., and Scheidereit, C. (2000). NF-kappaB and the innate immune response. Curr. Opin. Immunol. 12, 52–58. doi: 10.1016/s09527915(99)00050-3 Huan, P., Wang, H., and Liu, B. (2011). Comparative proteomic analysis of challenged Zhikong scallop (Chlamys farreri): a new insight into the anti-Vibrio immune response of marine bivalves. Fish Shellfish Immunol. 31, 1186–1192. doi: 10.1016/j.fsi.2011.10.014 Ji, C., Wu, H., Wei, L., Zhao, J., Wang, Q., and Lu, H. (2013a). Responses of Mytilus galloprovincialis to bacterial challenges by metabolomics and proteomics. Fish Shellfish Immunol. 35, 489–498. doi: 10.1016/j.fsi.2013.05.009 Ji, C., Wu, H., Wei, L., Zhao, J., and Yu, J. (2013b). Proteomic and metabolomic analysis reveal gender-specific responses of mussel Mytilus galloprovincialis to 2,2’,4,4’-tetrabromodiphenyl ether (BDE 47). Aquat. Toxicol. 14, 449–457. doi: 10.1016/j.aquatox.2013.07.009 Jiang, S., Qiu, L., Wang, L., Jia, Z., Lv, Z., Wang, M., et al. (2018). Transcriptomic and quantitative proteomic analyses provide insights into the phagocytic killing of hemocytes in the oyster Crassostrea gigas.Front. Immunol. 9:1280. doi: 10. 3389/fimmu.2018.01280 Kausar, S., Asif, M., Bibi, N., and Rashid, S. (2013). Comparative molecular docking analysis of cytoplasmic dynein light chain DYNLL1 with pilin to explore the molecular mechanism of pathogenesis caused by Pseudomonas aeruginosa PAO. PLoS One 8:e76730. doi: 10.1371/journal.pone.0076730 Leprêtre, M., Almunia, C., Armengaud, J., Le Guernic, A., Salvador, A., Geffard, A., et al. (2020). Identification of immune-related proteins of Dreissena polymorpha hemocytes and plasma involved in host-microbe interactions by differential proteomics. Sci. Rep. 10:6226. doi: 10.1038/s41598-020-63321-z Leprêtre, M., Faury, N., Segarra, A., Claverol, S., Degremont, L., Palos-Ladeiro, M., et al. (2021). Comparative Proteomics of Ostreid Herpesvirus 1 and pacific oyster interactions with two families exhibiting contrasted susceptibility to viral infection. Front. Immunol. 11:621994. doi: 10.3389/fimmu.2020.621994 Liang, P., Zhang, H., Wang, G., Li, S., Cong, S., Luo, Y., et al. (2013). KPNB1. XPO7 and IPO8 mediate the translocation ofNF-κB/p65 into the nucleus. Traffic 14, 1132–1143. doi: 10.1111/tra.12097 Licciulli, S., Cambiaghi, V., Scafetta, G., Gruszka, A. M., and Alcalay, M. (2010). Pirin downregulation is a feature of AML and leads to impairment of terminal myeloid differentiation. Leukemia 24, 429–437. doi: 10.1038/leu.2009.247 Liu, F., Rehmani, I., Esaki, S., Fu, R., Chen, L., de Serrano, V., et al. (2013). Pirin is an iron-dependent redox regulator of NF-κB. Proc. Natl. Acad. Sci. U.S.A. 110, 9722L–9727. doi: 10.1073/pnas.1221743110 Liu, R., Cheng, Q., Song, X., Wang, H., Wang, X., Wang, L., et al. (2019). A vital ubiquitin-conjugating enzyme CgUbe2g1 participated in regulation of immune response of Pacific oyster Crassostrea gigas.Dev. Comp. Immunol. 91, 132–142. doi: 10.1016/j.dci.2018.10.014 Liu, T., Zhang, L., Joo, D., and Sun, S.-C. (2017). NF-κB signaling in inflammation. Signal Transduct. Target. Ther. 2:17023. doi: 10.1038/sigtrans.2017.23 Liverman, A. D. B., Cheng, H.-C., Trosky, J. E., Leung, D. W., Yarbrough, M. L., Burdette, D. L., et al. (2007). Arp2/3-independent assembly of actin by Vibrio type III effector VopL. Proc. Natl. Acad. Sci. U.S.A. 104, 17117L–17122. doi: 10.1073/pnas.0703196104 Lu, Q., Pallas, D. C., Surks, H. K., Baur, W. E., Mendelsohn, M. E., and Karas, R. H. (2004). Striatin assembles a membrane signaling complex necessary for rapid, nongenomic activation of endothelial NO synthase by estrogen receptor α.Proc. Natl. Acad. Sci. U. S. A 101, 17126L–17131. doi: 10.1073/pnas.0407492101 Luk, S. C., Ngai, S. M., Tsui, S. K., Fung, K. P., Lee, C. Y., and Waye, M. M. (1999). In vivo and in vitro association of 14-3-3 epsilon isoform with calmodulin: implication for signal transduction and cell proliferation. J. Cell. Biochem. 73, 31–35. doi: 10.1002/(sici)1097-4644(19990401)73:1<31::aid-jcb4<3.0.co;2-x Ma, W., Smigel, A., Tsai, Y.-C., Braam, J., and Berkowitz, G. A. (2008). Innate immunity signaling: cytosolic ca elevation is linked to downstream nitric oxide generation through the action of calmodulin or a calmodulin-like protein. Plant Physiol. 148, 818L–828. doi: 10.1104/pp.108.125104 Manduzio, H., Cosette, P., Gricourt, L., Jouenne, T., Lenz, C., Andersen, O.-K., et al. (2005). Proteome modifications of blue mussel (Mytilus edulis L.) gills as an effect of water pollution. Proteomics 5, 4958–4963. doi: 10.1002/pmic. 200401328 Masood, M., Raftos, D. A., and Nair, S. V. (2016). Two oyster species that show differential susceptibility to virus infection also show differential proteomic responses to generic dsRNA. J. Proteome Res. 15, 1735–1746. doi: 10.1021/acs. jproteome.5b00615 McDonagh, B., Tyther, R., and Sheehan, D. (2005). Carbonylation and glutathionylation of proteins in the blue mussel Mytilus edulis detected by proteomic analysis and Western blotting: actin as a target for oxidative stress. Aquat. Toxicol. 73, 315–326. doi: 10.1016/j.aquatox.2005.03.020 Meschede, J., Šadiæ, M., Furthmann, N., Miedema, T., Sehr, D. A., Müller-Rischart, A. K., et al. (2020). The parkin-coregulated gene product PACRG promotes TNF signaling by stabilizing LUBAC. Sci. Signal. 13:eaav1256. doi: 10.1126/ scisignal.aav1256 Metsalu, T., and Vilo, J. (2015). ClustVis: a web tool for visualizing clustering of multivariate data using principal component analysis and heatmap. Nucleic Acids Res. 43, W566–W570. doi: 10.1093/nar/gkv468 Mikenberg, I., Widera, D., Kaus, A., Kaltschmidt, B., and Kaltschmidt, C. (2007). Transcription Factor NF-κB Is transported to the nucleus via cytoplasmic dynein/dynactin motor complex in hippocampal neurons. PLoS One 2:e589. doi: 10.1371/journal.pone.0000589 Mimuro, H., Suzuki, T., Suetsugu, S., Miki, H., Takenawa, T., and Sasakawa, C. (2000). Profilin is required for sustaining efficient intraand intercellular spreading of Shigella flexneri. J. Biol. Chem. 275, 28893–28901. doi: 10.1074/ jbc.M003882200 Mitta, G., Hubert, F., Noël, T., and Roch, P. (1999). Myticin, a novel cysteinerich antimicrobial peptide isolated from haemocytes and plasma of the mussel Mytilus galloprovincialis.Eur. J. Biochem. 265, 71–78. doi: 10.1046/j.1432-1327. 1999.00654.x Moreira, R., Pereiro, P., Canchaya, C., Posada, D., Figueras, A., and Novoa, B. (2015). RNA-Seq in Mytilus galloprovincialis: comparative transcriptomics and expression profiles among different tissues. BMC Genomics 16:728. doi: 10. 1186/s12864-015-1817-5 Moreira, R., Romero, A., Rey-Campos, M., Pereiro, P., Rosani, U., Novoa, B., et al. (2020). Stimulation of Mytilus galloprovincialis hemocytes with different immune challenges induces differential transcriptomic, mirnomic, and functional responses. Front. Immunol. 11:606102. doi: 10.3389/fimmu.2020. 606102 Muscolini, M., Sajeva, A., Caristi, S., and Tuosto, L. (2011). A novel association between filamin A and NF-κB inducing kinase couples CD28 to inhibitor of NF-κB kinase αand NF-κB activation. Immunol. Lett. 136, 203–212. doi: 10. 1016/j.imlet.2011.01.011 Oyanedel, D., Gonzalez, R., Flores-Herrera, P., Brokordt, K., Rosa, R. D., Mercado, L., et al. (2016). Molecular characterization of an inhibitor of NF-κB in the scallop Argopecten purpuratus: first insights into its role on antimicrobial peptide regulation in a mollusk. Fish Shellfish Immunol. 52, 85–93. doi: 10.1016/ j.fsi.2016.03.021 Perez-Riverol, Y., Csordas, A., Bai, J., Bernal-Llinares, M., Hewapathirana, S., Kundu, D. J., et al. (2019). The PRIDE database and related tools and resources in 2019: improving support for quantification data. Nucleic Acids Res. 47, D442–D450. doi: 10.1093/nar/gky1106 Rajan, B., Lokesh, J., Kiron, V., and Brinchmann, M. F. (2013). Differentially expressed proteins in the skin mucus of Atlantic cod (Gadus morhua) upon natural infection with Vibrio anguillarum.BMC Vet. Res. 9:103. doi: 10.1186/ 1746-6148-9-103 Rey-Campos, M., Moreira, R., Gerdol, M., Pallavicini, A., Novoa, B., and Figueras, A. (2019a). Immune tolerance in Mytilus galloprovincialis hemocytes after repeated contact with Vibrio splendidus.Front. Immunol 10:1894. doi: 10.3389/ fimmu.2019.01894 Rey-Campos, M., Moreira, R., Valenzuela-Muñoz, V., Gallardo-Escárate, C., Novoa, B., and Figueras, A. (2019b). High individual variability in the transcriptomic response of Mediterranean mussels to Vibrio reveals the involvement of myticins in tissue injury. Sci. Rep. 9:3569. doi: 10.1038/s41598019-39870-3 Rey-Campos, M., Novoa, B., Pallavicini, A., Gerdol, M., and Figueras, A. (2020). Comparative genomics reveals a significant sequence variability of myticin genes in Mytilus galloprovincialis.Biomolecules10:943. doi: 10.3390/ biom10060943 Rocher, B., Bultelle, F., Chan, P., Foll, F. L., Letendre, J., Monsinjon, T., et al. (2015). 2-DE Mapping of the Blue Mussel Gill Proteome: the Usual Suspects Revisited. Proteomes 3, 3–41. doi: 10.3390/proteomes3010003 Frontiers in Marine Science | www.frontiersin.org 16 October 2021 | Volume 8 | Article 735309 fmars-08-735309 September 29, 2021 Time: 16:23 # 17 Saco et al. Mussel Gill Proteomic Immune Response Romero, M. R., Pérez-Figueroa, A., Carrera, M., Swanson, W. J., Skibinski, D. O. F., and Diz, A. P. (2019). RNA-seq coupled to proteomic analysis reveals high sperm proteome variation between two closely related marine mussel species. J. Proteomics 192, 169–187. doi: 10.1016/j.jprot.2018.08.020 Rother, S., Bartels, M., Schweda, A. T., Resch, K., Pallua, N., and Nourbakhsh, M. (2016). NF-κB-repressing factor phosphorylation regulates transcription elongation via its interactions with 50→30exoribonuclease 2 and negative elongation factor. FASEB J. 30, 174–185. doi: 10.1096/fj.15-270256 Saco, A., Rey-Campos, M., Novoa, B., and Figueras, A. (2020). Transcriptomic response of mussel gills after a vibrio splendidus infection demonstrates their role in the immune response. Front. Immunol. 11:3273. doi: 10.3389/fimmu. 2020.615580 Saco, A., Rey-Campos, M., Rosani, U., Novoa, B., and Figueras, A. (2021). The evolution and diversity of interleukin-17 highlight an expansion in marine invertebrates and its conserved role in mucosal immunity. Front. Immunol. 12:692997. doi: 10.3389/fimmu.2021.692997 Sánchez-Marín, P., Vidal-Liñán, L., Fernández-González, L. E., Montes, R., Rodil, R., Quintana, J. B., et al. (2021). Proteomic analysis and biochemical alterations in marine mussel gills after exposure to the organophosphate flame retardant TDCPP. Aquat. Toxicol. 230:105688. doi: 10.1016/j.aquatox.2020.105688 Smits, M., Artigaud, S., Bernay, B., Pichereau, V., Bargelloni, L., and Paillard, C. (2020). A proteomic study of resistance to brown ring disease in the manila clam, Ruditapes philippinarum.Fish Shellfish Immunol. 99, 641–653. doi: 10. 1016/j.fsi.2020.02.002 Somboonwiwat, K., Chaikeeratisak, V., Wang, H.-C., Fang Lo, C., and Tassanakajon, A. (2010). Proteomic analysis of differentially expressed proteins in Penaeus monodon hemocytes after Vibrio harveyi infection. Proteome Sci. 8:39. doi: 10.1186/1477-5956-8-39 Somboonwiwat, K., Supungul, P., Rimphanitchayakit, V., Aoki, T., Hirono, I., and Tassanakajon, A. (2006). Differentially expressed genes in hemocytes of vibrio harveyi-challenged shrimp penaeus monodon. J. Biochem. Mol. Biol. 39, 26–36. doi: 10.5483/bmbrep.2006.39.1.026 Song, H., Kim, H., Lee, K., Lee, D.-H., Kim, T.-S., Song, J. Y., et al. (2012). Ablation of Rassf2 induces bone defects and subsequent haematopoietic anomalies in mice. EMBO J. 31, 1147–1159. doi: 10.1038/emboj.2011.480 Song, Q., Chen, H., Li, Y., Zhou, H., Han, Q., and Diao, X. (2016). Toxicological effects of benzo(a)pyrene. DDT and their mixture on the green mussel Perna viridis revealed by proteomic and metabolomic approaches. Chemosphere 144, 214–224. doi: 10.1016/j.chemosphere.2015.08.029 Storey, J. D. (2003). The positive false discovery rate: a Bayesian interpretation and the q-value. Ann. Stat. 31, 2013–2035. doi: 10.1214/aos/107429 0335 Suzuki, M., Watanabe, M., Nakamaru, Y., Takagi, D., Takahashi, H., Fukuda, S., et al. (2016). TRIM39 negatively regulates the NFκB-mediated signaling pathway through stabilization of Cactin. Cell. Mol. Life Sci. 73, 1085–1101. doi: 10.1007/s00018-015-2040-x Szklarczyk, D., Franceschini, A., Wyder, S., Forslund, K., Heller, D., Huerta-Cepas, J., et al. (2015). STRING v10: protein-protein interaction networks, integrated over the tree of life. Nucleic Acids Res. 43, D447–D452. doi: 10.1093/nar/ gku1003 Tagliazucchi, D., and Conte, A. (2005). Calcium/calmodulin dependence of nitric oxide synthase from Viviparus ater immunocytes. Invertebr. Surviv. J. 2, 54–59. Tanguy, M., McKenna, P., Gauthier-Clerc, S., Pellerin, J., Danger, J.-M., and Siah, A. (2013). Functional and molecular responses in Mytilus edulis hemocytes exposed to bacteria, Vibrio splendidus.Dev. Comp. Immunol. 39, 419–429. doi: 10.1016/j.dci.2012.10.015 Uotila, L. M., Guenther, C., Savinko, T., Lehti, T. A., and Fagerholm, S. C. (2017). Filamin a regulates neutrophil adhesion, production of reactive oxygen species, and neutrophil extracellular trap release. J. Immunol. 199, 3644–3653. doi: 10.4049/jimmunol.1700087 Wang, T., Zhao, M., Liang, D., Bose, U., Kaur, S., McManus, D. P., et al. (2017). Changes in the neuropeptide content of Biomphalaria ganglia nervous system following Schistosoma infection. Parasit. Vectors 10:275. doi: 10.1186/s13071017-2218-1 Witherow, D. S., Garrison, T. R., Miller, W. E., and Lefkowitz, R. J. (2004). betaArrestin inhibits NF-kappaB Activity by Means of Its Interaction with the kappaB Inhibitor IkappaBalpha. Proc. Natl. Acad. Sci. U.S.A 101, 8603–8607. Wu, H., Ji, C., Wei, L., Zhao, J., and Lu, H. (2013). Proteomic and metabolomic responses in hepatopancreas of Mytilus galloprovincialis challenged by Micrococcus luteus and Vibrio anguillarum. J. Proteomics 94, 54–67. doi: 10. 1016/j.jprot.2013.09.001 Wu, S.-H., Lin, H.-J., Lin, W.-F., Wu, J.-L., and Gong, H.-Y. (2018). A potent tilapia secreted granulin peptide enhances the survival of transgenic zebrafish infected by Vibrio vulnificus via modulation of innate immunity. Fish Shellfish Immunol. 75, 74–90. doi: 10.1016/j.fsi.2018.01.044 Xiong, Q., Feng, J., Li, S., Zhang, G., Qiao, Z., Chen, Z., et al. (2015). Integrated transcriptomic and proteomic analysis of the global response of Synechococcus to high light stress. Mol. Cell. Proteomics 14, 1038–1053. doi: 10.1074/mcp. M114.046003 Xu, W., and Faisal, M. (2008). Putative identification of expressed genes associated with attachment of the zebra mussel (Dreissena polymorpha). Biofouling 24, 157–161. doi: 10.1080/08927010801975345 Xu, Y., Shi, J., Hao, W., Xiang, T., Zhou, H., Wang, W., et al. (2018). iTRAQbased quantitative proteomic analysis of Procambarus clakii hemocytes during Spiroplasma eriocheiris infection. Fish Shellfish Immunol. 77, 438–444. doi: 10.1016/j.fsi.2018.04.005 Yamamoto, S., Ogasawara, N., Yamamoto, K., Uemura, C., Takaya, Y., Shiraishi, T., et al. (2017). Mitochondrial proteins NIP-SNAP-1 and -2 are a target for the immunomodulatory activity of clarithromycin, which involves NF-κBmediated cytokine production. Biochem. Biophys. Res. Commun. 483, 911–916. doi: 10.1016/j.bbrc.2016.12.100 Yoshida, S., and Sasakawa, C. (2003). Exploiting host microtubule dynamics: a new aspect of bacterial invasion. Trends Microbiol. 11, 139–143. doi: 10.1016/S0966842X(03)00023-4 Zhang, P., Li, C., Zhang, P., Jin, C., Pan, D., and Bao, Y. (2014). iTRAQ-Based proteomics reveals novel members involved in pathogen challenge in sea cucumber Apostichopus japonicus.PLoS One 9:e100492. doi: 10.1371/journal. pone.0100492 Zhang, Y., Li, J., Yu, F., He, X., and Yu, Z. (2013). Allograft inflammatory factor-1 stimulates hemocyte immune activation by enhancing phagocytosis and expression of inflammatory cytokines in Crassostrea gigas.Fish Shellfish Immunol. 34, 1071–1077. doi: 10.1016/j.fsi.2013.01.014 Zhang, Y., Li, X., Carpinteiro, A., Goettel, J. A., Soddemann, M., and Gulbins, E. (2011). Kinase suppressor of Ras-1 protects against pulmonary Pseudomonas aeruginosa infections. Nat. Med. 17, 341–346. doi: 10.1038/nm.2296 Zhang, Y., Sun, J., Mu, H., Li, J., Zhang, Y., Xu, F., et al. (2015). Proteomic basis of stress responses in the gills of the Pacific Oyster Crassostrea gigas.J. Proteome Res. 14, 304–317. doi: 10.1021/pr500940s Zhou, Y., Li, X., Yuan, X., and Hao, L. (2019). Role of allograft inflammatory factor1 in the regulation of inflammation and oxidative stress in primary peritoneal mesothelial cells. Cell Biol. Int. 43, 495–505. doi: 10.1002/cbin.11115 Zuo, S., Xue, Y., Tang, S., Yao, J., Du, R., Yang, P., et al. (2010). 14-3-3 epsilon dynamically interacts with key components of mitogen-activated protein kinase signal module for selective modulation of the TNF-alpha-induced time coursedependent NF-kappaB activity. J. Proteome Res. 9, 3465–3478. doi: 10.1021/ pr901137 Conflict of Interest: The authors declare that the research was conducted in the absence of any commercial or financial relationships that could be construed as a potential conflict of interest. Publisher’s Note: All claims expressed in this article are solely those of the authors and do not necessarily represent those of their affiliated organizations, or those of the publisher, the editors and the reviewers. Any product that may be evaluated in this article, or claim that may be made by its manufacturer, is not guaranteed or endorsed by the publisher. Copyright © 2021 Saco, Panebianco, Blanco, Novoa, Diz and Figueras. This is an open-access article distributed under the terms of the Creative Commons Attribution License (CC BY). The use, distribution or reproduction in other forums is permitted, provided the original author(s) and the copyright owner(s) are credited and that the original publication in this journal is cited, in accordance with accepted academic practice. No use, distribution or reproduction is permitted which does not comply with these terms. Frontiers in Marine Science | www.frontiersin.org 17 October 2021 | Volume 8 | Article 735309