The comprehensive immunomodulation of NeurimmiRs in haemocytes of oyster Crassostrea gigas after acetylcholine and norepinephrine stimulation

Background Neural-endocrine-immune (NEI) system is a major modulation network among the nervous, endocrine and immune system and weights greatly in maintaining homeostasis of organisms during stress and infection. Some microRNAs are found interacting with NEI system (designated NeurimmiRs), addressing swift modulations on immune system. The oyster Crassostrea gigas, as an intertidal bivalve, has evolved a primary NEI system. However, the knowledge about NeurimmiRs in oysters remains largely unknown. Results Six small RNA libraries from haemocytes of oysters stimulated with acetylcholine (ACh) and norepinephrine (NE) were sequenced to identify neurotransmitter-responsive miRNAs and survey their immunomodulation roles. A total of 331 miRNAs (132 identified in the present study plus 199 identified previously) were subjected to expression analysis, and twenty-one and sixteen of them were found ACh- or NE-responsive, respectively (FDR < 0.05). Meanwhile, 21 miRNAs exhibited different expression pattern after ACh or NE stimulation. Consequently, 355 genes were predicted as putative targets of these neurotransmitter-responsive miRNAs in oyster. Through gene onthology analysis, multiple genes involved in death, immune system process and response to stimulus were annotated to be modulated by NeurimmiRs. Besides, a significant decrease in haemocyte phagocytosis and late-apoptosis or necrosis rate was observed after ACh and NE stimulation (p < 0.05) while early-apoptosis rate remained unchanged. Conclusions A comprehensive immune-related network involving PRRs, intracellular receptors, signaling transducers and immune effectors was proposed to be modulated by ACh- and NE-responsive NeurimmiRs, which would be indispensable for oyster haemocytes to respond against stress and infection. Characterization of the NeurimmiRs would be an essential step to understand the NEI system of invertebrate and the adaptation mechanism of oyster. Electronic supplementary material The online version of this article (doi:10.1186/s12864-015-2150-8) contains supplementary material, which is available to authorized users.


Background
Neural-endocrine-immune (NEI) system, defined firstly by Besedovsky and Sorkin in 1977, is a closely interacted network among the nervous, endocrine and immune system [1]. These three systems communicate actively with each other by releasing neurotransmitters, hormones and cytokines, respectively [2]. The well balanced NEI system and reciprocal regulation inside are essential for organisms to maintain homeostasis during the stress or infection [2]. Acetylcholine (ACh) and norepinephrine (NE) have been considered as representative neurotransmitters in parasympathetic and sympathetic nervous system, respectively [3,4], and they contribute an intensive modulation on the immune response of mammals during stress and bacteria challenge [5,6].
Recently, microRNAs (miRNAs) have been identified as an important mechanism to modulate the NEI system in mammals. As a class of endogenously encoded single-strand non-coding RNAs, miRNAs have been suggested as vital regulators in gene expression by binding to their 3'-UTR [7]. To date, a great number of miRNAs have been identified from diverse species, with striking conservation in their secondary structure and length (~22 nt) [8]. It has been estimated that miRNAs could be involved in the modulation of almost all biological processes, including growth, development, cell differentiation and immunity [7,9].
Accumulating evidences has recently demonstrated that some miRNAs (designated NeurimmiRs) could act as "negotiators" within the NEI system and function indispensably in maintaining the immune homeostasis during infection [10][11][12]. For example, miR-132, a wellknown NeurimmiR in mammals, could repress the expression of acetylcholinesterase after LPS stimulation, resulting in the attenuation of the vagus nerve and the release of tumor necrosis factor and interleukin-6 [11]. Other eight miRNAs were also identified as NeurimmiRs, which shared close connection with various neuropathologies [10]. However, less is known about NeurimmiRs in invertebrates who might own more preliminary NEI system than vertebrates.
The Mollusca is a large diverse phylum in invertebrates and possess the ancient neuroendocrine system, such as ancient cholinergic and catecholaminergic system, similar to that in mammals [13,14]. The ancient molluscan neuroendocrine system, can be activated by bacteria stimulation and modulate the immune response in turn [15]. As an intertidal bivalve, oysters Crassostrea gigas are continuously suffered by environmental stress, which attracts more attention to the balance of NEI system. The available genome information [16] and the knowledge of immune mechanism [17] as well as some previously identified miRNAs [18] make it suitable to investigate the modulation of NEI system in oyster [19]. The purposes of the present study were to (1) further identify neurotransmitter-induced NeurimmiRs from oyster, (2) characterize miRNA expression patterns after the neurotransmitter stimulation, (3) survey the potential modulation of NeurimmiRs on the immune response of oyster, hopefully providing new hints for the dynamic regulation within the NEI system in mollusc.  Table S1). Two peaks were observed in the length distribution (Additional file 2: Figure S1) of all the unique reads (8.6 M) obtained from these six libraries. A total of 715,372 reads remained with copy number more than 6, and were applied for the subsequent alignment with Rfam database and oyster mRNA. Finally, 519,571 reads were retained for further mapping to the genome (Additional file 3: Table S2).

The miRNAs identified from oyster haemocytes
To identify miRNAs expressed in oyster haemocytes after neurotransmitter stimulation, a total of 6,820 reads which have at least one alignment in genome, were subjected to miRDeep2 software for prediction of the precursor sequence and secondary structure. A total of 132 new miRNAs including 5 known ones and 127 novel ones were identified (Additional file 4: Table S3), among which, 27 miRNAs possessed at least two precursor-coding regions ( Fig. 1, Additional file 5: Table S4). Considering 199 miR-NAs (Additional file 4: Table S3) identified previously, a total of 331 miRNAs were discovered in oyster haemocytes. Within those miRNAs, 76 miRNAs shared the same seed sequence with miRNAs identified in other species (designated known miRNAs) and 255 ones possessed new seed sequence (designated novel miRNAs).

Differentially expressed miRNAs after neurotransmitter stimulation
Copy numbers of the 331 miRNAs identified in oyster haemocytes were counted and further converted to fragments per kilo base of transcript per million fragments mapped (FPKM) value to analyze their expression level (Additional file 6: Table S5). FPKM distributions of total miRNAs in those three groups were analyzed and depicted in box plot (Fig. 2). The bottom and top of the box represented the first and third quartiles of log 10 (FPKM + 1) values in corresponding group while the line insides the box stood for the median value. Though the median and third quartile in PBS, ACh and NE groups were similar, the first quartile of control group was significantly lower than that in neurotransmitter stimulation groups (Fig. 2).
Differentially expressed miRNAs after neurotransmitter stimulation were determined subsequently by edgeR software with FDR less than 0.05 (Additional file 7: Table S6). As a result, a sum of 21 miRNAs were found expressed differentially in the ACh stimulation group compared to that in the PBS control group, including 18 increased and three decreased ones (Fig. 3). Ten miRNAs were upregulated after NE stimulation while six were downregulated, compared with that in PBS group (Fig. 3). Five miRNAs (cgi-miR-125, cgi-miR-8, scaffold1144_2255, scaffold1711_599 and scaffold226_18954) were found responsive to both the ACh and NE stimulation. Furthermore, 21 miRNAs showed different expression pattern in ACh and NE stimulation groups, among which, eight miRNAs exhibited higher expression level in NE stimulation group and 13 miRNAs exhibited higher expression level in ACh stimulation group (Fig. 3).
Gene onthology (GO) terms of target genes of neurotransmitter-responsive miRNAs A total of 355 oyster genes were predicted as putative targets of those neurotransmitter-responsive miRNAs by miRanda software (Additional file 8: Table S7). The GO term distributions of all target genes were analyzed by Blast2GO software and displayed through Web Gene Ontology Annotation Plot (WEGO) (Fig. 5). Finally, immune-related biological process (death, immune system process and response to stimulus) and molecular function (antioxidant) were annotated from those targets of neurotransmitter-responsive miRNAs.

Changes of phagocytosis and apoptosis rate of oyster haemocytes after neurotransmitter stimulation
The phagocytosis and apoptosis rate of oyster haemocytes after neurotransmitter stimulation were evaluated to survey the immunomodulation role of ACh and NE. Both the phagocytosis rate in the haemocytes of ACh and NE group dropped significantly when compared with that in control group (P < 0.05) (Fig. 9a). No remarkable changes of early-apoptosis rate was found in the ACh and NE groups compared with control group (Fig. 9b) while the late-apoptosis and necrosis rate decreased significantly (P < 0.05) (Fig. 9c).

Discussion
As a class of gene regulator at post-transcription level, miRNAs are usually transcribed by RNA polymerase II and incorporated into RNA-induced silencing complex where they can binds to the 3'-UTR of target genes and induce their translational repression or degradation [20]. Though diverse miRNAs have been identified in regulating immune genes, fewer reports investigated their interactions with neurotransmitter [10]. In the present study, six miRNA libraries from oyster haemocytes after ACh and NE stimulation were sequenced to identify neurotransmitter-responsive miRNAs and further investigate their potential involvements in the modulation of immune response. A total of 132 new miRNAs were identified herein, including 5 known ones and 127 novel ones. In addition with 199 miRNAs identified previously, a total number of 331 miRNAs have been discovered in oyster C. gigas so far. Among those, 255 miRNAs failed to be mapped to any mature miRNAs in the miRBase database, suggesting that they might be oyster-specific. The total number of miRNAs in C. gigas, was larger than that in other molluscs such as Haliotis rufescens [21], Lotti agigantea [21] and Pinctada martensii [22], indicating that they might play indispensable roles in the thrive of oysters in intertidal environment.
ACh and NE have been proved to be vital neurotransmitters in the response against stress and infection [2,3,23,24] and the expression of diverse immunerelated genes could be fast switched by them (3-12 h) [13,14], indicating the participation of miRNAs during the modulation. Accordingly, in the present study, 38 of 331 oyster miRNAs depicted intensively expressional alteration at 6 h after ACh and NE stimulation. Among them, two ACh-responsive miRNAs (cgi-miR-125 and cgi-miR-8) and six NE-responsive miRNAs (cgi-miR-125, cgi-miR-133, cgi-miR-190, cgi-miR-278, cgi-miR-8 and cgi-miR-92d) were also found as immune-  [18]. Furthermore, some immune-related biological process (death, immune system process and response to stimulus) and molecular function (antioxidant) were annotated from the target genes of those 38 neurotransmitter-responsive miRNAs (Fig. 5), which collectively implied the potential immunomodulation role of the neurotransmitter-responsive miRNAs in oyster. Moreover, those 38 neurotransmitter-responsive miRNAs were clustered into eight distinct branches based on their FPKM value (Fig. 4), where miRNAs in the same branch shared similar expression pattern after the neurotransmitter stimulation. Those miRNAs with similar expression pattern were suspected to function synergistically to maintain the homeostasis of NEI system in oyster. Interestingly, most of the neurotransmitterresponsive miRNAs were newly discovered ones in this species, indicating the existence of novel miRNA regulation mechanism in the NEI system of oyster.
NeurimmiRs represent a class of miRNAs which could modulate the interactions in NEI system [10]. In the present study, multiple genes involved in the pathogen recognition, signaling transduction and immune function were predicted as targets of NeurimmiRs in oyster. PRRs are the first players to initiate the immune response by Fig. 4 Heat map and hierarchical clustering of differentially expressed miRNAs. Neurotransmitter-responsive miRNAs were clustered with converted FPKM value by Cluster 3.0 in default settings. Converted expressions of certain miRNA were shown with different colors. The "+" or "-" symbol after each miRNA ID indicates its significance state of expression difference in PBS/ACh, PBS/NE, ACh/NE comparison with "+" represents significant difference and "-" represents no significant difference recognizing pathogens [25]. As invertebrates, oysters have evolved a sophisticated innate immune system with large number of PRRs to survive from harsh environment full of pathogens [16,17,26]. It is found in the present study that NeurimmiRs from oyster could modulate some PRRs, such as mannose receptors [27], TLRs [28] and CEACAM5 [29]. Moreover, some intracellular receptors triggering immune response could also be targeted by oyster Neurim-miRs. For example, MEGF10, a bona fide homology of engulfment receptor cell death abnormality protein-1 which could promote cellular engulfment by cooperation with ATP binding cassette transporter [30], was annotated as a putative target of scaffold987_25421. Besides, CD63, which could act as the membrane receptor of metalloproteinase inhibitor (TIMP) and specially inhibit metalloproteinase from pathogenic bacteria Vibrio splendidus of oyster [31], was also found to be modulated by NeurimmiR scaffold444_16949. Collectively, those results suggest a dynamic regulation of NeurimmiRs in recognizing pathogens and triggering downstream immune signals in oyster.
Meanwhile, multiple signaling transducers and immune effectors were also proposed to be modulated by oyster NeurimmiRs. In the present study, two Rab genes (RAB-21 and RAB-5C), activators in integrin pathway [32], were found to be modulated by NeurimmiR scaf-fold625_3183 and scaffold987_25421, respectively. Three vital regulators in NF-κB pathway [33,34] (CREBBP [35], MIB2 [36] and TRAF3 [37]) were also predicated as targets of ACh-responsive NurimmiRs (scaffold987_25421, scaffold987_25421 and scaffold365_16162). However, the expression pattern of scaffold625_3183 and scaf-fold987_25421 was opposite, while CREBBP, MIB2 and TRAF3 conflicts with each other in activating NF-κB. It was speculated that contradictions in modulating integrin pathway and NF-κB pathway by those NeurimmiRs could render a flexible switch on the immune response of oysters. Moreover, some NurimmiRs could also target the immune effectors in oyster. For example, IFI44, a virus-responsive gene activated by TLR3 to inhibit virus proliferation [38,39], was annotated as putative target of scaf-fold1121_5128. Similarly, multiple NE-responsive Neur-immiRs were suggested to manipulate the expression of HSP70 [40], an important regulator in immune response.
Given the pathways and genes, the immunomodulation mechanism of ACh and NE might share certain similarities or was sometimes in an neurotransmitterspecific manner [41][42][43][44]. This can also be observed in those ACh-and NE-responsive miRNAs from oyster. For example, five out of 38 neurotransmitter-responsive miRNAs (cgi-miR-125, cgi-miR-8, scaffold1144_2255, scaffold1711_599 and scaffold226_18954) showed similar expression pattern after ACh or NE stimulation, and immune-related GO terms including response to stimulus and gene expression were annotated from their target genes (Fig. 10a). Those miRNAs confirmed the perspective that ACh and NE could partly share conserved immunomodulation pathway. On the other hand, 21 miRNAs exhibited different expression pattern after ACh or NE stimulation, with six of them (cgi-miR-182, cgi-miR-242, scaffold1600_212, cgi-miR-1989, scaf-fold1785_16815 and scaffold625_3183) in an opposite trend. Among them, cgi-miR-1989 was of the most interest as it decreased in ACh group while increased in NE group and was a putative regulator of TNFAIP3 and nAChR (Fig. 11). It was proved that TNFAIP3, a key factor in NF-κB pathway, could attenuate NF-κB activation during the infection [45] while nAChR, one of important ACh receptor, could also modulate the inflammatory responses through both NF-κB and JAK/ STAT pathway [46,47]. Those differentially expressed NeurimmiRs were supposed to result in the differences between immunomodulation of ACh and NE. Moreover, the modulation differences could also be augmented by scaffold1535_2758, a putative regulator of mAChR, which was significantly up-regulated after ACh stimulation and ascended modestly after NE stimulation. It was proposed that scaffold1535_2758 could promote the nAChR-initialed signaling after ACh stimulation while suppress the cholinergic signaling remarkably after NE stimulation by cooperating with cgi-miR-1989. Collectively, those NeurimmiRs synergistically render the host with a closely interacted NEI system in oyster. Phagocytosis of immunocytes is a highly integrated immune process, accomplished by multiple genes involved in pathogen recognition, signal transducing and immune effector expression [25,48,49]. In the present study, the phagocytic rate of oyster haemocytes against V. splendidus decreased markedly after ACh and NE stimulation (Fig. 9a), which was consistent with speculations proposed above. Comparatively, the apoptosis rate remained comparable in ACh and NE group (Fig. 9b), possibly owing to the modulatory conflicts in integrin pathway and NF-κB pathway, which were considered as two major players in maintaining cell survival [50,51]. Notwithstanding, late-apoptosis and necrosis rate decreased significantly after the neurotransmitter stimulation (Fig. 9c), indicating a tendency for ACh and NE in maintaining cell survival in oysters.

Conclusion
In conclusion, a total of 38 neurotransmitter-responsive miRNAs were identified in oyster with possible immunomodulation roles. A comprehensive NeurimmiRmediated network targeting PRRs, intracellular receptors, signaling transducers and immune effectors was proposed for oyster haemocytes to maintain the homeostasis of NEI system. Those findings for the first time depicted the interaction between invertebrate NeurimmiRs and the ancient NEI system, which would shed light on the understanding of the stress adaptation mechanisms of oyster.

Materials and methods
Oyster culture Oysters C. gigas (averaging 150 mm in shell length, 70 mm in width) were collected from a local farm in Qingdao, China. A narrowed notch was sawed for injection in the closed side of oyster shell, which was adjacent to the adductor muscle [52]. Oysters were maintained in aerated sea water (about 20°C) for two weeks before use. No specific permits are required for the present study and all experiments were conducted with approval

Neurotransmitter stimulation
A total of 30 oysters were employed for neurotransmitter stimulation experiment and randomly divided into three groups, including PBS control group, ACh stimulation group and NE stimulation group. Oysters in PBS control group received an injection of 50 μL phosphate buffered saline (PBS, 0.14 mol L −1 sodium chloride, 3 mmol L −1 potassium chloride, 8 mmol L −1 disodium hydrogen phosphate dodecahydrate, 1.5 mmol L −1 potassium phosphate monobasic, pH 7.4) at their adductor muscle while oysters in ACh and NE group were injected with 50 μL ACh (Sigma, 10 −7 M in PBS) and NE (Sigma, 10 −7 M in PBS), respectively. All oysters were returned to the aerated sea water immediately after the injection and sampled at 6 hour post injection. To reduce individual variations and improve result reliability, haemocytes from five individuals were harvested by centrifugation (800 g, 4°C for 10 minutes) and pooled into one sample. Two replicates were conducted for subsequent sequencing in each group.
Another 45 oysters were also employed and received the same treatment as described above for phagocytosis and apoptosis assay. At 6 h post the injection, haemocytes from five oysters in each group were collected aseptically from the posterior adductor muscle sinus with acid citrate-dextrose anticoagulant agent (22 g L −1 sodium citrate, 8 g L −1 citric acid, 24.5 g L −1 glucose, pH 7.4) at the ratio of 8:1 [53] and resuspended in L15 medium (Gibco) with additional saline solution to 2 x 10 6 cells mL −1 before subjected to phagocytosis and apoptosis assay. All trials were conducted with three replicates.

Construction and deep sequencing of oyster small RNA libraries
Total miRNAs in the haemocytes were extracted with PureLink miRNA Isolation Kit (Invitrogen) according to the manufacture's protocol. RNAs were quantified by

Identification of oyster miRNAs
Raw data obtained from Ion Torrent Proton was preprocessed by Fastx-toolkit pipeline to get the summary of sequencing quality, nucleotide distributions and length distributions. Then reads with high sequencing quality and ranging from 18-30 nt in length were collected by Fastx-toolkit and aligned to Rfam database [54] and oyster mRNAs. After abandoning reads mapped to either non-miRNAs in Rfam (such as rRNA, snoRNA etc.) or oyster mRNAs, the remaining was then mapped to oyster genome using bowtie-1.00 software [55] and analyzed by miRDeep2 [56] to identify known and novel miRNAs with mature miRNAs and precursor sequences from oyster and other species listed in miRBase [57] as references.

Expression analysis of miRNA
The copy numbers of oyster miRNAs were calculated by home-made Perl script. The differentially expressed miRNAs were determined by edgeR software using generalized linear models with FDR (false discovering rate) value less than 0.05 [58].The expression levels of miR-NAs in each group were estimated by FPKM method. A dendrogram of differentially expressed miRNAs was clustered by Cluster3.0 using FPKM value and displayed by Treeview software.

Target prediction and GO analysis
The 3'-UTR sequences of oyster genes were obtained based on the genome annotation information and subjected to miRanda [59] for target prediction of those neurotransmitter-responsive miRNAs. Genes were regarded as putative targets if they satisfied the criteria with free energy less than −25 kcal/mol and score value higher than 160. Fig. 11 Immunomodulation of ACh and NE were differentiated by ACh and NE responsive NeurimmiRs. NeurimmiR cgi-miR-1989 was found decreased after ACh and increased after NE stimulation and predicted to target oyster transcripts for nAChR and TNFAIP3. Scaffold1535_2758 was found up-regulated by ACh and NE and could modulate mAChR. Those two miRNAs were proposed to contribute largely to the immunomodulation differences between ACh and NE All oyster protein sequences were aligned to nonredundant database of NCBI by a local blastp algorithm (E-value < 1 x 10 −5 ) and further parsed by Blast2GO software [60] for assigning GO terms. GO term distribution of target genes was then calculated and exhibited by WEGO [61].

Phagocytosis and apoptosis assay after neurotransmitter stimulation
Phagocytosis rate of oyster haemocytes was determined using method modified from previous report [52]. In brief, Vibrio splendidus [62], isolated from lesions of moribund scallop Patinopecten yessoensis and cultured at 16°C overnight, were labeled with FITC (Sigma) and diluted into 10 8 cells mL −1 before use. A volume of 200 μL haemocytes (2 x 10 6 cells mL −1 ) obtained previously was then incubated with same volume of FITC-labeled bacteria for 60 min at room temperature in darkness. Haemocytes were subsequently recollected and resuspended in L15 medium to detect the phagocytosis rate using flow cytometry (BD biosciences) with 10, 000 events counted.
Apoptosis rate was detected using the Annexin V-FITC detection kit (Beyotime) according to the manufacturer's instructions. In brief, 195 μL diluted haemocytes were incubated firstly with 5 μL Annexin V-FITC for 10 min to label early-apoptotic cells and then stained with 10 μL propidium iodide for 5 min to mark the late-apoptotic or necrotic cells. After recollection and resuspension, the haemocytes were immediately subjected to flow cytometry for apoptosis rate detection (10,000 events countered).
Data obtained were analyzed by Statistical Package for Social Sciences (SPSS) 19.0, followed by one-way analysis of variance (ANOVA) and given as means ± S.D. Differences were considered significant at P < 0.05 and marked with different letters (a, b, c etc.).

Availability of supporting data
The raw sequencing data have been deposited in the Sequence Read Archive (SRA) under BioProject SRR1562260.