A total of 250 healthy H. cumingii with an average weight of 15 g, ~1 year old, were purchased from an aquaculture farm in Wuhu City, Anhui Province, China. The freshwater mussels were maintained in aerated freshwater and allowed to acclimate at room temperature for 7 days before processing.

LPS (E. coli serotype 055:B5) and PGN (Staphylococcus staphylolyticus) were purchased from Sigma (St. Louis, MO, USA). Staphylococcus aureus, Micrococcus luteus, Bacillus subtilis, Vibrio anguillarum, E. coli (maintained and available in our laboratory), and Vibrio parahaemolyticus (ATCC 17802, Microbial Culture Collection Center, Beijing, China) bacteria were grown in Luria–Bertani (LB) at 37°C. Aeromonas hydrophila (ATCC 7966, Microbial Culture Collection Center, Beijing, China) was grown in LB broth and incubated at a temperature of 28°C.

A total of 100 μl of S. aureus (3 × 10⁷ cells) or V. parahaemolyticus (3 × 10⁷ cells) was injected into the adductor muscles of H. cumingii using a 1-mL sterile syringe. Phosphate-buffered saline (PBS; 0.14 M NaCl, 3 mM KCl, 8 mM Na₂HPO₄, and 1.5 mM KH₂PO₄; pH 7.4) was used as a control. At 2, 6, 12, and 24 h post-bacterial or PBS injection, the hepatopancreas or mantles from three H. cumingii were mixed for RNA extraction. Hemocytes, hepatopancreas, gills, and mantles from three healthy mussels were extracted and mixed for RNA extraction and further tissue distribution analysis. All experiments were repeated three times.

Pure RNA High-Purity Total RNA Rapid Extraction Kit (Spin-Column; Bioteke, Beijing, China) was used to isolate the total RNA from above samples based on the manufacturer's protocol. First-strand cDNA synthesis for quantitative real-time (qRT) polymerase chain reaction (PCR) analysis was conducted using the PrimeScript® First-Strand cDNA Synthesis Kit (Takara, Dalian, China) with the Oligo-d(T) Primer. Approximately 5 μg of the total RNA obtained from the hepatopancreas were reverse transcribed via the Clontech SMARTer™ RACE cDNA Amplification Kit (Takara, Dalian, China) with 5′-CDS Primer A and SMARTerIIA oligos (5′-RACE Ready cDNA) and 3′-CDS Primer A (3′-RACE-Ready cDNA). The detailed procedures were performed based on the manufacturer's instructions.

The data analysis of the hepatopancreas transcriptome of H. cumingii (unpublished data) revealed a sequence of C1qDC gene (HcC1qDC6). Then, the 5′ and 3′ fragments of HcC1qDC6 were amplified using gene-specific primers (HcC1qDC6-F: 5′-GCGTGCGAATGCCAAATAGCGGAG-3′ and HcC1qDC6-R: 5′-CCACTCCGCTATTTGGCATTCGCAC-3′) and a Universal Primer A Mix (UPM). The PCR was performed in accordance with the following procedures: 1 cycle at 94°C for 2 min; 30 cycles at 94°C for 30 s, 68°C for 30 s, and 72°C for 3 min; and 1 cycle at 72°C for 2 min using the Clontech Advantage 2 PCR Kit (Takara, Dalian, China). The full-length HcC1qDC6 cDNA was obtained by overlapping the 5′ and 3′ fragments.

The cDNA sequence of HcC1qDC6 was analyzed using BLAST (http://blast.ncbi.nlm.nih.gov/Blast.cgi), and the deduced protein sequences were obtained with ExPASy (http://web.expasy.org/). Signal peptide and domain organization were predicted with SMART (http://smart.embl-heidelberg.de/). The theoretical isoelectric point (pI) and molecular weight (MW) were determined using ExPASy (http://web.expasy.org/compute_pi/). MEGA 5.05 and GENDOC were used to align HcC1qDC5 and HcC1qDC6. MEGA 5.05 was utilized to phylogenetically analyze the relationship of HcC1qDC6 and other C1qDC proteins from H. cumingii and other species (Kumar et al., 2008).

First, genomic DNA was extracted from the gills of healthy mussels by using NucleoSpin Tissue (Clontech). Next genome walking was performed using the Universal GenomeWalker 2.0 Kit (Clontech) to amplify the genome sequences of HcC1qDC6. The first round of PCR was done using the gene-specific forward primer HcC1qDC6-walk-F1: 5′-GTGCCCTGCCGAAGGAGTTTACAAGTT-3′ and primer AP1 (5′-GTAATACGACTCACTATAGGGC-3′). The digestive genomic DNA was used as template in the first round PCR. The second round of PCR was then conducted using primers HcC1qDC6-walk-F2: 5′-CACAGAAGGAGATGCGATTTTCGTCAA-3′ and AP2: 5′-ACTATAGGGCACGCGTGGT-3′. In order to walk toward 5′ end, nested PCRs were performed using sequence-specific reverse primers (HcC1qDC6-walk-R1: 5′-CCAGAGAGAACCCAAAGCAACATCCAA-3′, HcC1qDC6-walk-R2: 5′-CGTGTCTTGACGAAAATCGCATCTCCT-3′) and AP1 and AP2 primers. Finally, a pair of gene-specific primers (gHcC1qDC6-F: 5′-TTCCAGGATGATTGCTTTCTCTGTCGG-3′, gHcC1qDC6-R: 5′-TTAAGGTGTGGACGAGACAAGGTAACC-3′) were designed to amplify partial genome sequences using undigested genomic DNA as template. Above 3 fragments were overlapped to obtain the genome sequences of HcC1qDC6. The above procedures were performed according to the manufacturer's protocols.

Based on the sequence of HcC1qDC6, the siRNA was specifically designed by BLOCK-iT™ RNAi Designer to target HcC1qDC6 and then synthesized in vitro with a commercial kit based on the manufacturer's instructions (Takara, Japan). The siRNA used was HcC1qDC6-siRNA (5′-GGUAUGGAUCCACAGGUAA-3′), and the siRNA sequence was scrambled to generate the control HcC1qDC6-siRNA-scrambled (5′-UGUGAAGAUGGAACGUACC-3′).

RNA interference (RNAi) assay was performed by injecting 40 μg of siRNA into the adductor muscles of each H. cumingii. Approximately 20 μg of siRNA or siRNA-scrambled cells and V. parahaemolyticus (3 × 10⁷ cells) were injected into a mussel at a volume of 100 μL per mussel. At 16 h after the injection, 20 μg of siRNA or siRNA-scrambled cells (100 μL/mussel) were injected into the same mussel. At the same time, 100 μL of V. parahaemolyticus was injected into the mussel as a positive control. For each group, 27 mussels were used. At 24, 36, and 48 h after the last injection, the gills and mantles from three mussels were collected. The experiment was repeated three times on each sample.

The tissue distribution and expression patterns of HcC1qDC6 gene at an mRNA level were analyzed through qRT-PCR using the primers HcC1qDC6-RT-F: 5′-GCGTGCGAATGCCAAATAG-3′ and HcC1qDC6-RT-R: 5′-ATGGACAGGGGCCGTAAAA-3′. The gills and mantle tissues of HcC1qDC6 knockdown mussels were tested to identify the transcriptional levels of tumor necrosis factor HcTNF and whey acidic protein HcWAP. The primers HcTNF-RT-F: 5′-ATTCCTTCCTGACATACAAAACCGA-3′ and HcTNF-RT-R: 5′-AGAAAGTGAAGAGTGGAGGCGAGAT-3′; HcWAP-RT-F: 5′-TGTAATGTTGACGGGAGTG-3′, and HcWAP-RT-R: 5′-CTGTTTTGTTTTGATGGCT-3′ were used in the test. β-Actin was amplified as an internal control using Hc-actin-RT-F: 5′-GTGGCTACTCCTTCACAACC-3′ and Hc-actin-RT-R: 5′-GAAGCTAGGCTGGAACAAGG-3′. The methods utilized were based from the previous study (Huang et al., 2013). Each sample used for qRT-PCR analysis was prepared in triplicates. The 2^−ΔΔCT method was used to calculate the expression levels of genes (Livak and Schmittgen, 2001). All data were given as mean ± standard error (S.E). Unpaired sample t-test was applied, in which P < 0.05 was considered statistically significant.

Three pairs of primers with EcoR I sites in the forward primers and Xho I sites in the reverse primers (rC1q1-ex-F: 5′-TATCGGATCCGAATTCGCAAATACACAAGATGATGTC-3′ and rC1q1-ex-R: 5′-GGTGGTGGTGCTCGAGCATTCGCACGCCGCTGAACG-3′; rC1q2-ex-F: 5′-TATCGGATCCGAATTCTCAAATGGATATAAGGACGAC-3′ and rC1q2-ex-R: 5′-GGTGGTGGTGCTCGAGCTCTCTGATTATTAGGGCAC-3′; rC1q3-ex-F: 5′-TATCGGATCCGAATTCAACATACCGCTGAGTTCCAGG-3′ and rC1q3-ex-R: 5′-GGTGGTGGTGCTCGAGTGTGGACGAGACAAGGTAAC-3′) were used to amplify the cDNA fragments of three C1q domains of HcC1qDC6. Both fragments and pET-30a(+) vector were digested with EcoR I and Xho I, and the purified DNA fragments were ligated into the linearized pET-30a(+) vector. The plasmids were transformed into E. coli BL21 (DE3) competent cells for recombinant expression. The recombinant C1q proteins with an N-terminal His tag were purified using His-Bind resin chromatography (Novagen, USA). The procedure was done in accordance with the manufacturer's instructions (Huang et al., 2014). Moreover, sodium dodecyl sulfate polyacrylamide gel electrophoresis (SDS-PAGE, 12.5%) is the method used to detect the purified protein.

The binding activity of recombinant proteins to bacteria was analyzed through a microbial binding assay with slight modifications (Huang et al., 2014). Three Gram-positive bacteria, namely, S. aureus, M. luteus, and B. subtilis, and four Gram-negative bacteria, namely, A. hydrophila, V. parahaemolyticus, V. anguillarum, and E. coli, were utilized in the bacterial binding assay. These bacteria were cultured overnight in LB broth, washed three times with tris-buffered saline (TBS), centrifuged at 6000 rpm for 5 min at room temperature, and then resuspended in TBS. The purified recombinant proteins (600 μg/mL, 500 μL) were added, incubated with microorganisms (2 × 10⁸ cells/mL), and gently rotated for 1 h at room temperature. Microorganisms were washed four times with TBS, subsequently resuspended in 50 μL of TBS, and analyzed using 12.5% of SDS-PAGE.

An enzyme-linked immunosorbent assay (ELISA) was performed to detect the direct binding of rC1q proteins to LPS and peptidoglycan (PGN; Zhang et al., 2011). In summary, sugar was initially dissolved in water at a concentration of 80 μg/mL. Each well of a microtiter plate was then coated with 25 μL of sugar solution. Afterwards, the microtiter plate was incubated at 37°C overnight and heated at 60°C for 30 min. Each well was blocked with 200 μL of bovine serum albumin (BSA, 600 μg/mL) at 37°C for 2 h and then washed four times with tris-buffered saline with Tween 20 (TBST). The purified proteins with gradient dilution (0–30 μg/mL in TBS contains 0.1 mg/mL BSA) were added to the wells and incubated at 28°C for 3 h. The wells were washed four times; a total of 100 μL of anti-His antiserum (1/2,000 diluted in 0.1 mg/mL BSA) was then added and incubated at 37°C for 1 h. Each well was washed four times and then incubated with 100 μL of peroxidase-conjugated goat anti-rabbit IgG (1/10,000 diluted) for 1 h at 37°C. The plate was washed four times with TBST and developed with 0.01% 3,3′,5,5′-tetramethylbenzidine (Sigma). Approximately 2 M H₂SO₄ was utilized to prevent the reaction. The absorbance was read and noted at 450 nm. His-tag protein ADP ribosylation factor (rArf) was used in this assay as control. The assays were performed in triplicates.

The full sequence of HcC1qDC6 (GenBank accession no. MF038002) was 1782 bp in length. This sequence contains 1335 bp of open reading frame encoding a 444-amino-acid protein, 6 bp of 5′-untranslated terminal region (UTR), and a 441-bp of 3′ UTR with a poly(A) tail (Figure 1A). HcC1qDC6 was analyzed using the SMART program. Results show that the gene contains a signal peptide of 19-amino-acid residues and three C1q domains (amino acids 24–159, 167–302, and 305–443). The MW and pI of HcC1qDC6 were 49.5 kDa and 5.76, respectively. The MWs of rC1q1 to rC1q3 were 15.2, 15.4, and 15.5 kDa, respectively, and rC1q1 to rC1q3 have pI values of 4.70, 5.71, and 9.21, respectively.

The partial genome sequence of HcC1qDC6 contained at least five exons interrupted by four introns. Because 5′ end of the HcC1qDC6 genomic sequence was not fully amplified, the length of the four exons (337, 426, 223, and 190 bp) and three introns (476, 295, and 2,098 bp) were only determined. The exon–intron boundaries of HcC1qDC6 genomic sequence are GT and AG at 5′ and 3′ splice sites (Figure 1B and Figure S1).

BLASTP analysis showed that HcC1qDC6 shares 49% identity with complement C1q-like protein 2 from C. gigas and 45% identity with putative complement C1q-like protein 4 from Biomphalaria glabrata. The C1q1, C1q2, and C1q3 domains of HcC1qDC6 showed higher identity with three corresponding C1q domains of HcC1qDC5 (Figures 2A–C). The full length amino acid sequences of HcC1qDC6 and HcC1qDC5 were highly conserved (Figure 2D). The constructed phylogenetic tree presents the HcC1qDC6 protein initially clustered together with HcC1qDC5 from H. cumingii and then joined with C1qDC4 from B. glabrata and C1qDC3 from Aplysia californica (Figure 3A). Another phylogenetic tree using C1q domains of different C1qDCs (20 single C1q domain-containing C1qDCs and 2 multiple C1q domain-containing C1qDCs) from H. cumingii was also constructed. Results showed that C1q1, C1q2, and C1q3 of HcC1qDC5 were clustered with C1q1, C1q2, and C1q3 of HcC1qDC6, respectively. Only a single C1q domain-containing C1qDC (HcC1qDC15) were grouped with C1q domains of HcC1qDC5 and HcC1qDC6 (Figure 3B).

Tissue distribution analysis was conducted through qRT-PCR. Results revealed that HcC1qDC6 was expressed in all tested tissues, such as, mantles, hepatopancreas, hemocytes, and gills (Figure 4).

qRT-PCR was also used to monitor the mRNA expression of HcC1qDC6 transcripts in the hepatopancreas and mantles of mussels stimulated by two bacteria (Figure 5). In S. aureus-stimulated group, the HcC1qDC6 transcription level in the hepatopancreas was initially upregulated at 2 h, slightly downregulated from 6 to 12 h, and finally increased to the highest level at 24 h. HcC1qDC6 in the mantles was significantly upregulated at 6 h after stimulation and then gradually decreased to the original level. In the V. parahaemolyticus stimulated group, HcC1qDC6 mRNA in the hepatopancreas was slowly increased at 2 and 6 h and then quickly upregulated from 12 to 24 h. After a significant increase at 6 h post-stimulation, the mRNA expression of HcC1qDC6 in the mantles decreased at 12 h and then gradually increased at 24 h. In the control group, no significant change in HcC1qDC6 expression was observed during the entire experiment after PBS injection.

The siRNA interference group showed a reduced HcC1qDC6 expression levels after 24, 36, and 48 h of V. parahaemolyticus challenge as compared with that of the control groups (V. parahaemolyticus only and siRNA-scrambled group; Figure 6). C1qDC protein functions as a PRR in innate immunity. The expression levels of effector genes, such as, TNF and WAP, were analyzed after knockdown of C1qDC6. The expression levels of HcTNF and HcWAP decreased in the mussels with HcC1qDC6 knockdown as compared with that of the control groups (V. parahaemolyticus only and siRNA-scrambled group; Figure 6).

The plasmids pET-30a-rC1q1, pET-30a-rC1q2, and pET-30a-rC1q3 were transformed into E. coli BL21 (DE3). Cell lysates were then analyzed using SDS-PAGE after IPTG induction. Results showed that rC1q1 to rC1q3 were successfully expressed and purified (Figure 7).

A bacterial binding assay was performed to determine the binding capacity of three recombinant C1q domains (Figure 8). Results suggested that rC1q1 to rC1q3 can bind to all tested microorganisms, including Gram-positive bacteria (S. aureus, M. luteus, and B. subtilis) and Gram-negative bacteria (A. hydrophila, V. parahaemolyticus, V. anguillarum, and E. coli). rC1q1 exhibits a strong binding activity against A. hydrophila and V. parahaemolyticus, whereas rC1q2 strongly binds to S. aureus and V. parahaemolyticus but has a weak binding activity toward M. luteus. rC1q3 shows stronger binding activity toward Gram-positive bacteria than Gram-negative bacteria.

Optical density (OD) value is used to measure the binding activity of rC1q1, rC1q2, and rC1q3 to different PAMPs. Figure 9 shows an OD value of 450 nm. The recombinant rC1q proteins exhibited different affinity to LPS and PGN and a dose-dependent binding ability. The rArf group showed an OD value of 0.1, thereby indicating that rArf could not bind any tested PAMPs (data not shown).

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.