Introduction

The Japanese scallop (Mizuhopecten yessoensis) is a large and old (dating back to ~350 Ma) group living on the cold and stable ocean bottoms of the northwestern Pacific Ocean. This species has been a major economic aquaculture species in Asian countries and is consumed worldwide1. However, the frequent outbreak of bacterial and parasitic diseases in M. yessoensis has caused severe economic losses. Spionid worms of the genus Polydora, which mainly inhabit the shells of many benthic bivalves2,3,4,5 including some commercially important species, have become a serious problem in recent scallop aquaculture. It was roughly estimated that over 90% of cultured M. yessoensis in the Dalian Zhangzidao Sea (Liaoning, China) had been infested by Polydora at various degrees, badly affecting the development of the scallop industry. Polydora mainly parasitize the left valve of the scallop by excavating U-type tubes, which badly damages the structure of the shells. To prevent further invasion, scallops accelerate the secretion of shell components to repair the damaged regions. As a result, many protuberances are formed on the inner surface of the shell (Fig. 1), which are signs of the shell repair process. When Polydora drill through the shell, they can directly infect the soft body (e.g. the mantle tissue) and the damaged shell exposes the scallop to various pathogens, activating the scallop’s immune modulation to combat the infection. At present, research about the Polydora disease mainly focuses on the biological characteristics and parasitic behaviour of Polydora6,7,8,9,10,11; however, the response of M. yessoensis, especially on the molecular level, has never been reported, limiting the understanding of disease resistance in scallops.

Figure 1
Figure 1
Full size image

The outer and inner shell surfaces of the healthy (A) and infested by Polydora (B) M. yessoensis.

The mantle is the main tissue involved in the shell formation, located midway between the shell and visceral mass. Different mantle regions are responsible for the formation of different shell layers with the edge mantle being responsible for the periostracal and prismatic layerand the central mantle for the nacreous layer12,13,14. The resistance of M. yessoensis to the Polydora disease is characterized by the shell repair process with protuberance formation on the inner surface. Mantle tissue is therefore essential in the defense against Polydora. In recent years, the molecular function of mantle tissue has been extensively studied. With the prevalence of next-generation sequencing, transcriptomes of the mantle tissue have been sequenced in many mollusks, such as Patella vulgate15, Mytilus edulis16, Hyriopsis cumingii17, Pinctada fucata18, P. margaritifera19,20, Chlamys farreri21, and M. yessoensis22,23, to explore the mechanisms of shell formation. A wide variety of candidate genes involved in biomineralization have been identified, such as calcium-binding protein, carbohydrate-binding protein, glycoprotein, collagen, nacrein, perlucin and others, and their functions in shell formation have been partially characterized15,16,17,18,19,20,21,22,23. However, due to the extreme complexity of shell formation and the differences among species, our understanding is still fragmented and lacks consensus. In addition, transcriptional changes in the mantle tissue of Ruditapes philippinarum caused by Brown Ring Disease (a shell disease caused by the bacteria Vibrio tapetis) not only include genes related to biomineralization but also many immunity genes24,25 (such as big defensin, lectin, C1q domain containing protein and so on), implying the important role of mantle tissue in the immune response to pathogens.

Like all invertebrates, scallops rely on an exclusively innate immune system to execute cellular and humoural (macromolecules) immune reactions to resist invaders26. As the classical immune tissue in scallops, hemocytes use encapsulation and phagocytosis to eliminate undesirable particles (e.g., pathogens and abnormal/dead cells) performing the cellular immune reactions and contribute to the secretion of immune macromolecules to execute humoural immune responses, such as immune recognition, signal transduction and destruction of invaders, as reviewed by Song et al.27. Recent studies have also demonstrated a key role of hemocytes in biomineralization due to their involvement in mineral transport, as well as production of the extracellular matrix of the shells28,29,30,31, indicating the multifunction of hemocytes. Ivanina et al., recognized that potential trade-offs existed between biomineralization and immunity in Crassostrea gigas and C. virginica29, but relevant studies have not been reported in M. yessoensis.

In the present study, a comprehensive and niche-targeted cDNA sequences database was constructed for M. yessoensis infested by Polydora, transcriptional changes in the edge mantle, central mantle and hemocytes of the diseased scallops were first detected, and genes related to shell formation and immunity were screened. The aim was to reveal the molecular response mechanism of M. yessoensis to the infestation of Polydora, and provide more valuable information for the mechanism studies of biomineralization and immunomodulation.

Results

Transcriptome sequencing, assembly and functional annotation

To identify differently expressed genes between healthy and diseased M. yessoensis infested by Polydora, 18 cDNA libraries of the edge mantle, central mantle and hemocytes of two groups were separately constructed for RNA-seq sequencing. After removing adaptors and low-quality reads, a total of 942,738,120 reads (~109.7 Gb) with an average of 52,374,340 (~6.1 Gb) for each sample were obtained, and the detailed sequencing information for each library is listed in Supplementary Table S1. De novo assembly for the transcriptome sequences was chosen to construct a comprehensive and niche-targeted cDNA sequence database for diseased scallops, and the Trinity method was used with all trimmed reads. Finally, 80,831 unigenes were obtained with an average length of 1607 bp (Supplementary Fig. S1), and the assembly information is summarized in Supplementary Table S2. The unigenes were annotated by searching the sequences against the Nr, SWISS-PROT, KOG, GO and KEGG databases using BLASTX with a cut-off of Evalue ≤ 1e-5 to describe their functions at different levels, and the annotation ratios were 31.32%, 23.75%, 20.27%, 21.73% and 8.97%, respectively (Supplementary Table S3).

For GO classification, 17,567 unigenes were annotated into 12,658 GO terms in three main GO categories: biological process (~65.67%), cellular component (~10.44%), and molecular function (~23.90%) (Supplementary Table S4 and Fig. S2). In the biological process category, most of the unigenes were assigned to the terms of ‘cellular process’ (~14.81%), ‘single-organism process’ (~12.65%) and ‘metabolic process’ (~11.58%), including ‘immune system process’ (~1.15%). In the cellular component category, most of the unigenes were assigned to the terms of ‘cell’ (~19.54%), ‘cell part’ (~19.49%) and ‘organelle part’ (~15.29%). In the molecular function category, most of the unigenes were assigned to the terms of ‘binding’ (~44.48%), ‘catalytic activity’ (~33.17%) and ‘transporter activity’ (~5.95%). A total of 16,387 unigenes were assigned to 25 ortholog groups in the KOG database (Supplementary Table S5 and Fig. S3), which were distributed mainly into three groups: (R) ‘General function prediction only’ (~19.30%), (T) ‘Signal transduction mechanisms’ (~17.82%), and (O) ‘Posttranslational modification, protein turnover, chaperones’ (~10.02%). The biological signaling pathways for the unigenes were predicted by the KEGG database, and 7253 unigenes were assigned to 342 pathways (Table S6) involving six KEGG categories with ‘cellular processes’ (~11.81%), ‘environmental information processing’ (~13.76%), ‘genetic information processing’ (~7.56%), ‘human diseases’ (~25.05%), ‘metabolism’(~18.38%) and ‘organismal systems’ (~23.43%), in which the subcategories of ‘infectious diseases’, ‘immune diseases’, ‘nervous system’, ‘immune system’, etc. were contained (Supplementary Fig. S4).

Overall gene regulation in diseased scallops

The expression levels of the assembled unigenes were estimated by the FPKM method. Gene differential expression analysis was conducted between healthy and infested scallops in the edge mantle, central mantle and hemocytes tissues (p ≤ 0.05). A total of 3670 unigenes were differentially expressed in at least one of the three tissues with 1729 in the edge mantle, 1992 in the central mantle and 1010 in the hemocytes. Among these DEGs, there were 822 DEGs (~22.4%) shared by at least two of the tissues with 623 between the edge and central mantles, 314 between the edge mantle and hemocytes and 363 between the central mantle and hemocytes. Moreover, 239 (~6.5%) DEGs were shared among all the three tissues (Fig. 2). It is noteworthy that these three-tissue-shared DEGs were significantly enriched in the functions and processes related to glycometabolism by GO and KEGG enrichment analysis (p ≤ 0.05) (Figs 3 and 4). For instance, the GO terms of the ‘fructose metabolic process’, ‘glucose 6−phosphate metabolic process’, ‘cellular glucose homeostasis’, ‘mannose metabolic process’, ‘glucose metabolic process’ and ‘glycolytic process’ were the most significantly enriched terms in the biological process category (Fig. 3A). Meanwhile, relevant molecular activity, such as ‘mannokinase activity’, ‘hexokinase activity’, ‘fructokinase activity’, ‘glucokinase activity’, ‘glucose binding’ and ‘ATP binding’ were also significantly overrepresented in the molecular function category (Fig. 3B). Additionally, KEGG enrichment analysis showed a similar trend with pathways related to glycometabolism being the most significantly enriched, including ‘carbohydrate digestion and absorption’, ‘galactose metabolism’, ‘fructose and mannose metabolism’, ‘starch and sucrose metabolism’, ‘glycolysis/gluconeogenesis’ and ‘carbon metabolism’ (Fig. 4). Moreover, all of the above relevant terms were found to be upregulated in all the three tissues (Tables 1–3), which likely implied very active energy metabolism in these tissues to meet the living demands of the diseased scallops.

Figure 2
Figure 2
Full size image

Distribution of differentially expressed unigenes between diseased and healthy M. yessoensis across different tissues (edge mantle, central mantle and hemocytes).

Figure 3
Figure 3
Full size image

GO terms significantly enriched among DEGs shared by the edge mantle, central mantle and hemocytes. (A) Biological process. (B) Molecular function. Only the top 20 are presented, which are sorted by the p-value (p ≤ 0.05) and displayed with the –log10(p-value). Terms or pathways related to ‘glycometabolism’ and ‘immune response’ are separately highlighted with green and orange dots.

Figure 4
Figure 4
Full size image

KEGG pathway enrichment analysis among DEGs shared by the three tissues. Only the top 20 are presented, which are sorted by the p-value (p ≤ 0.05). Terms or pathways related to ‘glycometabolism metabolism’ are highlighted with green triangles.

Table 1 GO terms significantly enriched in the edge mantle.
Full size table
Table 2 GO terms significantly enriched in the central mantle.
Full size table
Table 3 GO terms significantly enriched in the hemocytes.
Full size table

Gene regulation in different tissues

For the DEGs of the three tissues, there were 1162 (~67.21%) upregulated and 567 (~32.79%) downregulated in the edge mantle, 1040 (~52.21%) upregulated and 952 (~47.79%) downregulated in the central mantle, and 521 (~51.58%) upregulated and 489 (~48.42%) downregulated in hemocytes. GO enrichment analysis was separately carried out for the up and downregulated DEGs of each tissue. On the whole, a variety of functions were covered for the enriched GO terms, but many of them could be classified into one category according to their functions (Tables 1–3, Supplementary Fig. 5S), which were probably related to shell formation and immunomodulation, such as calcium ion binding and transport, shell organic matrix, nervous system development, epithelial cell motility and immune response, including the glycometabolism-related functions shared by the three tissues.

GO enrichment analysis for upregulated genes in the edge mantle are described in Table 1. First, some GO terms related to calcium ion activities, such as ‘calcium ion transport into cytosol’, ‘calcium ion binding’, ‘cytosolic calcium ion homeostasis’ and ‘calcium channel activity’ and shell organic matrix, such as ‘glycoprotein 3-alpha-L-fucosyltransferase activity’ and ‘glycosaminoglycan metabolic process’, were significantly enriched, which were closely related to the process of biomineralization. In addition, many GO terms related to cellular motility, especially the cilium movement of the epithelial cells, were also significantly enriched, and most of them ranked in the top 20 among all the enriched terms according to their significance level (Supplementary Fig. S5-A1, marked with red dots), such as the biological process for cilium assembly and movement with GO terms of ‘cilium movement’, ‘axoneme assembly’, ‘cilium movement involved in cell motility’, ‘outer dynein arm assembly’, ‘epithelial cilium movement’, ‘regulation of epithelial cell migration’, ‘microtubule-based movement’, the cellular component for the cilium with GO terms of ‘cilium’, ‘microtubule’, ‘motile cilium’, ‘axoneme’, ‘axonemal central apparatus’, ‘dynein complex’, ‘ciliary basal body’, and the molecular function for ‘microtubule motor activity’ and ‘microtubule binding’, suggesting exuberant activity of the epithelial cells. There was another notable enrichment related to the nervous system, including the biological process GO terms ‘ventricular system development’, ‘lateral ventricle development’, ‘nervous system development’, ‘positive regulation of neuron projection development’, ‘brain development’, ‘central nervous system development’, ‘sphingolipid metabolic process’ and ‘regulation of neuron apoptotic process’ (downregulated), cellular component GO terms ‘synaptic cleft’, ‘synapse’ and ‘neuron projection’, and molecular function GO term ‘neurotransmitter: sodium symporter activity’, implying the involvement of the nervous system in the modulation process after an infestation of Polydora. The final overrepresented GO terms were those related to energy metabolism and were similar to the other two tissues, including glycometabolism and ATP catabolism and anabolism related terms (e.g., ‘nucleoside diphosphate kinase activity’, ‘nucleoside diphosphate phosphorylation’, ‘nucleoside triphosphate biosynthetic process’ and ‘ATPase activity’). All the relevant GO terms are listed in Table 1, and those ranked in the top 20 are dotted with different colours in Supplementary Fig. 5S-A1.

In the central mantle, there was a similar distribution of the enriched GO terms for upregulated genes with terms related to calcium ion activity, shell organic matrix, energy metabolism, nervous system and cellular motility gathering, as well (Table 2; terms ranking in the top 20 are dotted with different colours in Supplementary Fig. 5S-A2). However, differences still existed between the two regions. For example, more terms related to the shell organic matrix were enriched in the central mantle and some showed upregulation (Cellular component: ‘proteinaceous extracellular matrix’, ‘extracellular region’ and ‘extracellular exosome’. Biological process: ‘protein O-linked glycosylation’ and ‘protein glycosylation’. Molecular function: ‘glycoprotein 3-alpha-L-fucosyltransferase activity’, ‘alpha-(1- > 3)-fucosyltransferase activity’, ‘fucosyltransferase activity’), while some also showed downregulation (Cellular component: ‘proteinaceous extracellular matrix’, ‘collagen trimer’ and ‘extracellular region’. Biological process: ‘glycosaminoglycan metabolic process’). For the terms related to the nervous system, fewer showed upregulation (‘neuropeptide signaling pathway’, ‘positive regulation of neuron projection development’, ‘neurotransmitter transport’ and ‘neuropeptide receptor activity’), but more showed downregulation (‘synapse’, ‘glycosphingolipid metabolic process’, ‘dendrite morphogenesis’ and ‘neural tube closure’). GO terms related to cellular motility were mainly focused on epithelial cell migration (‘positive regulation of epithelial cell migration’, ‘colon epithelial cell migration’ and ‘regulation of epithelial cell migration’), and terms related to cilium movement were not found in the central mantle. Different regulation processes probably occurred in the two mantle regions suggesting their different tissue functions in shell formation.

Among the downregulated DEGs, many immune-related GO terms were detected in all three tissues, especially in hemocytes (Tables 1–3), and most of them ranked in the top 20 among all the enriched terms (Supplementary Fig. 5S-B dotted with orange colour). For example, four related GO terms (‘inflammatory response’, ‘defense response to virus’, ‘innate immune response’ and ‘scavenger receptor activity’) were enriched in the edge mantle, and five (‘response to type I interferon’, ‘interferon-beta production’, ‘inflammatory response’, ‘defense response to virus’ and ‘scavenger receptor activity’) in the central mantle, while seven were enriched in hemocytes (‘activation of innate immune response’, ‘defense response to virus’, ‘positive regulation of defense response to virus by host’, ‘innate immune response’, ‘inflammatory response’, ‘immune response’ and ‘scavenger receptor activity’). The term ‘inflammatory response’ was also enriched for the three-tissue-shared DEGs (Fig. 3A, dotted with orange colour). However, no upregulated genes related to immunity were found to be significantly enriched, probably indicating the declining immune competence in diseased scallops. In addition, nervous system-related GO terms were also found enriched in hemocytes, with ‘neurotransmitter: sodium symporter activity’ being upregulated similarly with the edge and central mantle, while ‘neuropeptide signaling pathway’ and ‘neuropeptide receptor activity’ were downregulated, indicating the neuromodulation in the immune response of M. yessoensis.

Genes related to biomineralization, immune and nervous system

Many biomineralization-related genes were differentially expressed in mantle tissues (Table 4), and most of them showed upregulation, which suggested a more active shell formation in the infested scallops. The functions of these genes were mainly associated with calcium binding or transport, such as ‘calbindin’ (CL3204Contig1), ‘calmodulin (CL51277Contig1), ‘EF-hand calcium-binding domain-containing protein’ (CL3228Contig1, CL3161Contig1, CL8277Contig1), ‘sodium/calcium exchanger’ (CL36832Contig1) and ‘transient receptor potential cation channel’ (CL2149Contig2, CL1979Contig2, CL1618Contig1) and shell matrix proteins, such as ‘collagen’ (CL12529Contig1, CL2550Contig1), ‘laminin’ (CL27Contig2), ‘von Willebrand factor A domain-containing protein’ (CL41489Contig1, CL721Contig1), ‘sclerostin’ (CL12811Contig1), ‘glycoprotein’ (CL55247Contig1,CL4117Contig1), ‘fibronectin ‘(CL15874Contig1), ‘glutathione peroxidase’ (CL29217Contig1), ‘hephaestin’ (CL12821Contig1), ‘ferritin’ (comp87468_c1_seq2_1), ‘tenascin’ (CL27Contig7, CL1236Contig1), ‘chitotriosidase ‘(CL50207Contig1), ‘laccase’ (CL6551Contig1), ‘tyrosinase-like proteins’ (CL14061Contig1) and ‘carbohydrate sulfotransferase’ (CL6600Contig1, CL2673Contig1, CL55429Contig1). Most of these genes were modulated in only one region of the mantle, implying different gene activation in two regions to form the different shell layers.

Table 4 Genes related to biomineralization in the mantle tissue.
Full size table

Many immune-related genes were also modulated in the two mantle regions and hemoctyes of the infested scallops (Table 5). The functions of these genes could be classified into the following categories: I) Immune recognition, e.g., ‘lectin’ (CL33366Contig1), ‘toll-like receptor (TLR)’ (CL12334Contig1, comp147899_c0_seq5_3), ‘scavenger receptor’ (CL63572Contig1), ‘C1q domain containing proteins’ (CL50918Contig1) and ‘fibrinogen-related proteins’ (CL26062Contig1, comp144120_c0_seq2_3). II) Immune effectors, e.g., ‘g-type lysozyme’ (CL3126Contig2), ‘superoxide dismutase (SOD)’ (comp96557_c0_seq1_1, comp119797_c0_seq1_3), ‘catalase’ (CL10534Contig1), ‘big defensin’ (CL8776Contig1) and ‘heat shock proteins (HSPs)’ (CL1392Contig1, CL37312Contig1, CL1984Contig1, CL47124Contig1, CL2340Contig1). III) Signal transduction, e.g., ‘tumor necrosis factor (TNF) receptor’ (CL23037Contig1, comp110989_c0_seq6_2), ‘serine protease’ (CL61221Contig1) and ‘NF-kappa-B inhibitor’ (comp122698_c0_seq6_2). Other genes were included, such as ‘immunoglobulin’ (CL47987Contig1, CL19464Contig1), ‘integrin’ (comp127636_c0_seq5_2), ‘interleukin’ (CL13815Contig1), ‘interleukin receptor’ (CL13604Contig1), ‘stimulator of interferon genes protein’ (comp129146_c0_seq2_3) and ‘pathogen-related protein’ (comp52531_c0_seq2_1). Unexpectedly, many of these genes showed downregulation.

Table 5 Genes related to the immune response in the mantle tissue and hemocytes.
Full size table

In addition, some neural genes were differentially expressed in the three tissues (Table 6). Most of the genes showed upregulation, especially for genes for Na+/Cl− dependent neurotransmitter transporters, such as the ‘GABA transporter’(comp139403_c0_seq2_3), ‘glycine transporter’ (CL50297Contig1), ‘aurine transporter’(CL31867Contig1), ‘proline transporter’ (CL51047Contig1) and ‘creatine transporter’ (CL11023Contig1)in the edge mantle, ‘GABA transporter’ (CL6833Contig1) in the central mantle, and ‘glycine transporter’ (CL22349Contig1) in hemocytes. However, the gene for ‘neurotrypsin’ (CL8026Contig1) was downregulated in all the three tissues. Nevertheless, genes for some neurotransmitter receptors had different situations. For example, the gene for ‘neuronal acetylcholine (ACh) receptor’ (CL31305Contig1) was up modulated in the edge mantle, while downregulated in the central mantle (CL42803Contig1) and hemocytes (CL35487Contig1). In contrast to the ACh receptor, the gene for the ‘5-hydroxytryptamine (5-HT) receptor’ was upregulated in both the central mantle and hemocytes but was not detected in the edge mantle. Some neuropeptide and neuropeptide receptor genes, such as the ‘myomodulin neuropeptides’ (CL32152Contig1) and ‘neuropeptide S receptor’ (CL939Contig1), ‘FMRF-amide neuropeptides’ (CL65515Contig1) and ‘FMRF-amide receptor’ (comp119939_c1_seq1_2) showed upregulation in the central mantle, while ‘FMRF-amide receptor’ (CL24831Contig) was downregulated in the hemocytes. There were also other neural-related genes were modulated in the three tissues as listed in Table 6. Genes found in this part (Table 6) probably mediated the processes of shell repair or immune response through neuromodulation.

Table 6 Genes related to the nervous system in the mantle tissue and hemocytes.
Full size table

Discussion

The main objective of this study was to identify molecular changes in M. yessoensis caused by the infection of Polydora and explore the molecular response mechanism of the scallops. The study mainly focused on the transcriptional changes in the tissues of edge mantle (producing the periostracal and prismatic layers of the shell), central mantle (producing the nacreous layer of the shell) and hemocytes (classic immune tissue in scallops), which were directly related to the disease. The results showed significant alterations associated with the processes of biomineralization and immunomodulation in scallops to resist infection.

Changes related to biomineralization mainly occurred in the edge and central mantles. Genes related to calcium binding, calcium transporting and shell matrix proteins were significantly differentially expressed between healthy and diseased scallops (Table 4), and the relevant gene functions were also significantly enriched in the GO enrichment analysis (Tables 1 and 2). Calcium carbonate (95–99%) and shell matrix proteins (1–5%) are the main components of the shell32,33. Genes related to calcium binding and transporting regulate calcium concentration in the shell matrix and initiate calcium carbonate deposition, while the nucleation, growth and spatial orientation of calcium carbonate crystals are regulated by shell matrix proteins34,35,36,37,38,39. Notably, most of the involved genes, such as genes for calbindin, calmodulin, EF-hand calcium-binding domain-containing protein, collagen, laminin, tenascin, laccase, tyrosinase-like protein, carbohydrate sulfotransferase and others showed obvious upregulation in the present study, suggesting increased shell secretion activity in infested scallops to repair the damaged shells. This hypothesis was also proved by the extremely active epithelial cells, which secrete the organic matrix involved in shell formation40. As the results in Tables 1, 2 and Supplementary Fig. S5-A1 show, functions related to the activities of mantle epithelial cells were significantly upregulated and enriched, especially for those related to cilium motility, probably to accelerate the secretion and transport of matrix proteins needed for shell repair.

Differences existed between the two mantle regions with many genes regulated in only one region, indicating the different molecular functions of the two regions. For instance, genes for the calmodulin, sodium/calcium exchanger, EF-hand calcium-binding domain-containing protein, laminin, von Willebrand factor A domain-containing protein, fibronectin, glutathione peroxidase, etc. were only found to be differentially expressed in the edge mantle, while genes for the short transient receptor potential channel, hephaestin-like protein, tyrosinase-like protein, sclerostin domain-containing protein, ferritin, chitotriosidase, etc. were only in the central mantle, implying these genes possibly participate in different shell layers formation. In addition, some new shell matrix protein genes were first found in scallops in this study, i.e., genes for glutathione peroxidise and hephaestin, which were once thought to be nonfunctional or lost in the mollusk shell and independently arose in Lingula and corals41, but possibly function in scallops shell repair. Therefore, this work provides significant candidate genes for the mechanism study of shell formation, including genes involved in the different shell layers formation, and more research is needed to verify how these genes function.

Although hemocytes have been proved playing an important role in biomineralization in oyster due to their involvement in mineral transport as well as the production of the extracellular matrix of the shells28,29,30,31, the present study didn’t show obvious changes related to biomineralization in the hemocytes by gene differential expression analysis and GO and KEGG enrichment analysis. However, many genes related to immunomodulation were significantly differentially expressed in hemocytes and mantle tissue (Table 5) and the relevant GO terms were also significantly enriched (Tables 1–3, Supplementary Fig. 5SB). The results suggest that hemocytes may mainly play an immune function during Polydora infection in scallops and the mantle tissue was simultaneously involved in this process. As an invertebrate, scallops lack adaptive immunity and have evolved a series of sophisticated strategies to recognize and eliminate various invaders by employing a set of molecules to participate in immune recognition (e.g., lectins42, galectins43,44 and scavenge receptors45), signal transduction (e.g., serine protease cascades46,47,48, TLR49,50 and TNF51 signalling pathways) and incapacitation and elimination of invaders (e.g., lysozymes52, heat shock protein53,54, and antimicrobial peptides55,56). Immune-related genes detected in the present study (Table 5) are involved in all the processes mentioned above, which reflected an integrated immune response reaction in the scallops for the first time. Unexpectedly, a number of these genes showed significant downregulation, such as genes for toll-like receptor, scavenger receptor, complement C1q tumor necrosis factor-related protein, heat shock protein 70 and 90, etc. Furthermore, immune-related GO functions were significantly enriched in the downregulated but not upregulated genes, such as ‘inflammatory response’, ‘defense response to virus’, ‘innate immune response’, ‘scavenger receptor activity’, ‘activation of innate immune response’, etc., which indicated declining immunity of the infested scallops. Potential trade-offs between biomineralization and immunity has been demonstrated in C. gigas and C. virginica29. The expression of biomineralization-related genes was higher in C. virginica than in C. gigas, while expression was lower for immunity genes. Competition existed between these two processes. The contrasting regulation of biomineralization (upregulated on the whole) and immunity (downregulated) genes in the present study agreed with the above findings.

In addition, the nervous system likely played an important role in the response process. A number of neural genes were significantly differentially expressed in the two mantle regions and hemocytes (Table 6) and relevant GO functions were significantly enriched, as well (Tables 1–3, Supplementary Fig. 5S), suggesting the involvement of neuromodulation in the response process. Recently, neural-immune regulation in hemocytes was identified in mollusks; this regulation was conducted by neurotransmitters via a “nervous-hemocyte”-mediated neuroendocrine immunomodulatory axis (NIA)-like pathway57,58. In the present study, GO terms associated with neurotransmitter activity were significantly enriched in the mantle and hemocytes. Though genes for neurotransmitter were not detected among DEGs, genes for several neurotransmitter transporters and receptors were significantly differently expressed, such as Na+/Cl− dependent neurotransmitter transporters, ACh receptor and 5-HT receptor, which partly implied the existence of neural immunomodulation in M. yessoensis. Few studies about neuromodulation in shell formation have been reported. The neural-associated DEGs in the mantle tissues could not be excluded from the regulation process of shell formation in this study, which offers a new perspective on the mechanism of shell formation; however, further exploration is needed.

Finally, scallops likely needed more energy during the response process. As we know, glycometabolism is the main pathway to supply energy for the organism. The results showed that many glycometabolism-related GO terms and pathways, e.g., glucose metabolic, fructose metabolic and mannose metabolic, were significantly enriched in the upregulated genes among all three tissues (Tables 1–3, Figs 3, 4 and Supplementary Fig. 5SA), which indicated an exuberant energy metabolism process in the diseased scallops. The generated energy was probably used for shell repair, immune reactions or other activities. For example, the active activity of epithelial cells in mantle tissues inevitably consumed a lot of energy. Consequently, more energy was used for survival and less energy remained for growth.

Conclusions

To investigate the molecular response mechanisms of M. yessoensis infested by Polydora, a comprehensive and niche-targeted cDNA sequences database for diseased scallops was constructed and transcriptional changes in the edge mantle, central mantle and hemocytes were first detected in this study. The gene differential expression and enrichment analysis showed that the infestation of Polydora caused expression changes for genes or pathways involved in shell formation and immune response. Different mantle regions had different molecular functions probably responsible for the formation of different shell layers. An intact immune response process from immune recognition to pathogen elimination was first detected in scallops. Finally, neuromodulation and glycometabolism were involved in the infection response. These results provide a better understanding of shell formation and innate immune response in scallops and provide valuable resources for the genetic selection of disease-resistant scallops.

Materials and Methods

Sample collection and RNA extraction

Two-year-old healthy and infested M. yessoensis (Fig. 1) were collected from the Dalian Zhangzidao Sea. The two groups of scallops were acclimated in the laboratory for one week prior to the experiments. Filtered and aerated seawater was maintained at approximately 8 °C, which is within the optimum temperature range for their growth. Three scallops as biological replicates from each group were sampled, and tissues of the edge mantle, central mantle and hemocytes that had a direct correlation with the disease were immediately placed in liquid nitrogen and stored at −80 °C. Total RNA was isolated from each sample using an RNAprep pure tissue kit (Tiangent, China). The quantity and quality of total RNA were determined using the NanoDrop2000 spectrophotometer (Thermo Scientific, Wilmington, DE, USA) and agarose gel electrophoresis.

cDNA library constructionand Illumina sequencing

mRNA was purified from total RNA by oligo(dT) magnetic beads, and each paired-end cDNA library was generated using the TruSeq RNA Sample Preparation Kit (Illumina, Inc., USA), following the manufacturer’s protocol. The quality of the libraries was assessed using the Agilent 2100 Bioanalyzer (Agilent Technologies, USA). Finally, a total of 18 cDNA libraries were sequenced on the Illumina HiSeq2500 genomic sequencing platform.

Sequencing data processing, assembly and annotation

Raw sequencing data was processed to remove reads containing adapters or ambiguous “N” nucleotides (length exceeded 35 bp) and low quality reads (length of the bases with a quality score less than 20 exceeded 30%). The high-quality clean reads were combined and used for transcriptome assembly by the Trinity software (version: trinityrnaseq_r20131110)59 with default parameters. The generated contigs were clustered into unigenes by the TGICL software60, which were used as the reference sequences for subsequent analyses. The raw data have been deposited into the NCBI SRA database (accession number: SRP150161).

All the unigene sequences were aligned to the NCBI non-redundant protein (Nr) database, Swiss-Prot database, the Eukaryotic Orthologous Groups (KOG) protein database, Gene Ontology (GO) database and the Kyoto Encyclopedia of Genes and Genomes (KEGG) pathway database by BLASTX with an E-value of E ≤ 1e-5 to obtain the best putative functional annotations of each unigene.

Gene differential expression analysis

The expression levels of the unigenes were determined by mapping the reads from the samples to reference unigenes using Bowtie 2 software (http://bowtie-bio.sourceforge.net)61. The FPKM (fragments per kilobase per million mapped reads) method62 was used to calculate the expression levels of the unigenes for each sample, which eliminated the effects of gene length and sequencing depth on the calculation of gene expression. Significance tests for expression differences in the mantle and hemocytes between healthy and infested scallops were conducted with negative binomial distribution hypothesis-testing, and the false discovery rate (FDR) method was applied to multiple tested hypotheses to correct the significant levels (p-values) and eliminate the influence of random fluctuations and errors. After calibration, a fold change of 2-fold and p ≤ 0.05 was set as the threshold. Enrichment analyses were performed by mapping all DEGs (differentially expressed genes) to GO and KEGG databases. GO functions or pathways with the statistical significance (p-value) lower than 0.05, which were evaluated by the hypergeometric distribution test, were considered to be enriched. The DEGs, enriched GO terms and KEGG pathways related to shell formation and immunity were highlighted.