Cloning, characterization and functional analysis of an Alveoline-like protein in the shell of Pinctada fucata

Shell matrix proteins (SMPs) have important functions in biomineralization. In the past decades, the roles of SMPs were gradually revealed. In 2015, our group identified 72 unique SMPs in Pinctada fucata, among which Alveoline-like (Alv) protein was reported to have homologous genes in Pinctada maxima and Pinctada margaritifera. In this study, the full-length cDNA sequence of Alv and the functional analysis of Alv protein during shell formation were explored. The deduced protein (Alv), which has a molecular mass of 24.9 kDa and an isoelectric point of 11.34, was characterized, and the functional analyses was explored in vivo and in vitro. The Alv gene has high expression in mantle and could response to notching damage. The functional inhibition of Alv protein in vivo by injecting recombinant Alv (rAlv) antibodies destroyed prism structure but accelerated nacre growth. Western blot and immunofluorescence staining showed that native Alv exists in the EDTA-insoluble matrix of both prismatic and nacreous layers and has different distribution patterns in the inner or outer prismatic layer. Taken together, the characterization and functional analyses of matrix protein Alv could expand our understanding of basic matrix proteins and their functions during shell formation.

Sample preparation. The adult pearl oyster, P. fucata (with shells 5.5-6.5 cm in length and 30-40 g of wet weight at approximately 2 years of age) were cultured in a pearl farm (Zhanjiang, Guangdong Province, China). The oysters were then raised in the laboratory at approximately 20 °C in a fish tank that contained aerated artificial seawater of 3% salinity. Table 1 for primer details.

Full-length acquisition and analysis of Alveoline-like (Alv).
The full-length cDNA sequence of the Alv was obtained with rapid-amplification of cDNA ends protocol (RACE; SMARTer TM RACE cDNA Amplification Kit, Clontech Laboratories, Inc., Japan). The primers GSP1 and NGSP1 were used for 5′RACE. GSP2 and NGSP2 were applied to 3′RACE. The PCR product was purified using an EasyPure Quick Gel Extraction Kit (TransGen Biotech, China) and was ligated into a pMD18-T vector (TaKaRa, Japan). The recombinant plasmid was sequenced by Beijing Ruibio Biotech Co., Ltd (China). The 5′RACE product sequence and 3′RACE product sequence were then connected together to obtain the full-length cDNA sequence of Alv. The Alv gene sequence was submitted to the ORF finder website (https://www.ncbi.nlm.nih.gov/orffinder/) to obtain the opening reading frame sequence. The SignalP 4.1 Server website (http://www.cbs.dtu.dk/services/SignalP/) could indicate whether the Alv protein had a signal peptide. The Phyre2 website (http://www.sbg.bio.ic.ac.uk/phyre2/html/page. cgi?id=index) could predict the secondary structure of Alv protein. The Jalview procedure was used to align the amino acid sequences of Alv proteins from different species (P. fucata, P. margaritifera and P. maxima).

Production of polyclonal antibody of rAlv protein.
The production and purification of rAlv protein was produced with hexahistidine (His 6 ) tag in the N-terminus but lacking in the signal peptide with primers Alv-p-F and Alv-p-R, and then was applied to 12% SDS-polyacrylamide gel. The target stripe was cut off from the gel, which was stained with Coomassie brilliant blue before. After the standard immunization procedure against New Zealand rabbit, the polyclonal antibodies were raised. The specificity of the antibodies was tested by Western blot against the purified protein with an enhanced HRP-DAB chromogenic substrate kit (TIANGEN Biotech Co., Ltd, China).
Gene expression pattern analysis by real-time quantitative PCR (qPCR). There are five different tissues (mantle, foot, gonad, gill and muscle) in P. fucata in which the Alv gene has different expression levels.
The pair of primers used for qPCR include qAlv-F/R that was designed for Alv and qβ-actin-F/R which was designed for actin 34  In vivo Alv function interference. Polyclonal antibodies against rAlv were used to inhibit the function of native Alv during shell formation. The purified antibodies were injected into the extrapallial fluid through the byssal foramen at the dosage of 1 µg per g per day using a 0.8 mm × 50 mm syringe needle. Three samples in each group were collected six days later. The shells were separated and washed with Milli-Q water by ultrasonic cleaning. The clean shells were then observed by scanning electron microscope (SEM, 15 kV; FEI Quanta 200, USA) or stained with DyLight ® 594-conjucted goat-anti-rabbit antibody for detection of immunofluorescence.
In situ hybridization. The primers anti-F/R were designed to synthesize antisense probes. And the primers of sense-F/R were used to make sense probes. The single-stranded RNA (ssRNA) was produced following the protocol of Promega T7RiboMAX TM with DIG-labeled uracil (Roche Applied Science, USA). Frozen sections of the mantle tissue were obtained using a freezing microtome (LEICA CM1900, Germany). The procedures of hybridization were executed according to the protocol of the Enhanced Sensitive ISH Detection Kit II (AP) (Boster Biological Technology Co., Ltd, China).
Detection of native Alv in the shell. Cleaned shells of P. fucata were immersed in 5% sodium hydroxide for 12 h and then washed by diluted water and dried at room temperature (RT). Nacreous layers and prismatic layers were ground using a grinding mill, and then shell powders were sieved with a 100 µm sifter. The calcite prismatic layer and the aragonitic nacreous layer were decalcified with 0.8 M sodium ethylenediaminetetraacetic acid (EDTA; pH 8.0) for 60 h in 4 °C for chromatography with continuous stirring. The supernatant was collected by centrifugation and was then desalted by 3 K ultrafiltration device. The insoluble part was washed with water and was denatured using a denaturing solution (30 mM Tris-HCl, pH 8.0, 1% sodium dodecyl sulfate (SDS), 10 mM dithiothreitol) at 100 °C for 20 min and was then centrifuged for a short duration. The EDTA-insoluble matrix (EISM) and EDTA-soluble matrix (ESM) of the nacreous and prismatic layers were applied to 12% SDS-polyacrylamide gels (SDS-PAGE) which were stained with Coomassie brilliant blue. The existence of the native Alv protein in the ESM and/or EISM of nacre and/or prism was tested by Western blot with the polyclonal antibodies against rAlv as the first antibody (1:1000) and HRP-Goat-anti-Rabbit IgG Fc as the second antibody (1:2000; EarthOx, USA). Immunofluorescence staining was also employed according to Nudelman et al. 36 with minor modification. First, the shell was cut into pieces of approximately 0.5 cm 2 with both the prismatic layer and nacreous layer. After being washed and sonicated in distilled water for 10 min, some shell pieces were immersed in 0.5 M EDTA (pH 8.0) for 3 d at 4 °C with gentle shaking. After complete decalcification, some shell pieces were sliced to 30 μm thickness with freezing microtome. Decalcified shell sections, shell squares, and undecalcified shell sections were blocked with 10% bovine serum albumin (BSA) at RT for 2 h. The experimental group and control group were then incubated in PBS containing 10% BSA along with polyclonal antibodies against rAlv (1:50) or preimmune serum (1:50) at RT for 1 h, respectively. Unbounded antibodies were removed by PBS containing Tween 20 (0.05% w/v) of twice for 5 min. Then, shell pieces and sections were incubated with DyLight ® 594-conjucted goat-anti-rabbit antibody (1:1000; Agrisera, Sweden) for 30 min at RT in the dark. Unbounded antibodies were removed by PBS containing Tween 20 (0.05% w/v) of twice for 5 min. The shells were then rinsed in water in the dark and observed by inverted laser confocal microscope (Zeiss, LSM780, Germany) within 2 h.

Results
Sequence analyses of the Alveoline-like (Alv) gene. The full-length cDNA sequence of Alv was obtained by RACE. We blasted the Alv gene with the NCBI database and confirmed the existence of the Alv protein which has homologous amino acid sequence in P. maxima and P. margaritifera 26 . Additionally, we confirmed part of the Alv gene in the genome of P. fucata with the transcript ID pfu_aug1.0_89.1_5796-7.t1. The nucleotide sequence of Alv obtained by RACE is 993 bp with an open reading frame of 702 bp in length (GenBank TM Accession No. KR872410) (Fig. 1a). It contains a 5′-untranslated sequence of 121-base, a 224-base 3′-untranslated sequence, an ATG in-frame start codon and a TAA in-frame stop codon (Fig. 1a).
The deduced mature Alv protein possesses 217 amino acid residues, and the calculated molecular mass is 24.9 kDa. The Alv protein is characterized with VP repeats. The amino acid composition of Alv is VP-rich, with 23.0% Val and 12.9% Pro, but has low Asp and Glu content which are rich in many other shell matrix proteins (Fig. 1a) 29,37 . In addition, there are 12.0% lysine residues, 8.3% arginine residues, and 7.8% histidine residues, which are basic amino acids ( Table 2). In the extrapallial fluid (pH 7.4), these residues are liable to be positively charged; therefore, they are able to interact with negative ions, such as carbonate ions, and negatively charged amino acid residues from acidic proteins 19,37 . The isoelectric point of the Alv protein is 11.34, which is different from most other matrix proteins that are acidic. The Alv protein has five predicted phosphorylation sites, including one Ser and four Thr, which are related to protein functions during biomineralization. The first 16 amino acids are predicted to be a signal peptide, and there is no transmembrane domain predicted in the Alv protein, which is consistent with the secretion character of matrix proteins. The predicted secondary structure of the Alv protein without signal peptide has almost 80% beta strands (Fig. S1), which is unique compared with other proteins. There are Alv homologous proteins in P. maxima and P. margaritifera with high conservation, indicating that Alv protein might serve important roles in biomineralization (Fig. 1b). The deduced Alv amino acids were confirmed by comparing with liquid chromatography-mass spectrometry/mass spectrometry (LC-MS/MS) results (Fig. S2a). However, there has been no reports of any known functions.
Production of the polyclonal antibody anti rAlv protein. The recombinant Alv (rAlv) protein was expressed with a his-tag in the N-terminus by E. coli (Fig. S2b). The apparent mass of rAlv is around 35 KDa which is larger than the predicted mass about 27.5 KDa. MS analysis has confirmed the existence of Alv protein of P. fucata in purified rAlv protein. Anti-rAlv antibody was used as first antibody during western blot to test whether the polyclonal antibody work. According to the result of western blot, the specificity of anti-rAlv antibody was confirmed (Fig. S2b).  In vivo investigation of Alv functions in shell formation. We analyzed the expression of Alv in different tissues of P. fucata using qPCR (Fig. 2a). The relative expression of Alv was detected in the muscle, foot, gonad, mantle and gill, whereas the expression level of Alv in the mantle was highest with 287 times higher than expression level in the muscle (the control).
To investigate the Alv functions in shell formation, we conducted assays called "notching experiments", in which shells were slightly notched to induce shell repair. Gene expression levels of Alv at 0 h, 6 h, 12 h, 24 h, 36 h, The function of the Alv protein in vivo was inhibited by the polyclonal antibody against Alv protein, which was injected into the extrapallial fluid of P. fucata. As shown in Fig. 3a, the prismatic layer of the shells from the polyclonal antibody-injected group was destroyed and many holes occurred on the surface. On the other side, the nacreous tablets of shell from the polyclonal antibody-injected group overgrew (Fig. 3c). In the meantime, the prismatic and nacreous layers of the preimmune serum-injected group were still intact (Fig. 3b,d). To confirm that the effect of the Alv has been suppressed by polyclonal antibody, we detected the Alv location (Fig. 3e,f) and fluorescence intensity of the antibody-injected group and the preimmune serum-injected group, respectively (Fig. S3). The highest intensity of antibody-injected group is 72, which is less than the highest intensity of control group at 125; in addition, the average intensity of the control group is 40.57, while the average intensity of the antibody-injected group is 8.56 (Fig. S3).

Expression location details of Alv in P. fucata.
To explore the precise expression sites of Alv mRNA in the mantle, we conducted in situ hybridization with sense and antisense ssRNA probes of Alv, in which the sense ssRNA probe was used as a negative control. We found that Alv mRNA was located in the outer epithelial cells of the outer fold (OF) and outer epithelial cells of the middle fold (MF); no hybridization signals were found in the inner fold (IF) and mantle pallial (MP) (Fig. 4a,b). There are no hybridization signals in negative control group (Fig. 4c).
To investigate whether the Alv protein exists in the prismatic and/or nacreous layer of P. fucata, matrix proteins were extracted from two layers, respectively. The SDS-PAGE observations of the EISM (EDTA-insoluble matrix) and ESM (EDTA-soluble matrix) matrix proteins from both layers (Fig. S4) were transferred to the PVDF membrane and western blotting followed. The gel bands indicated different compositions of the EISM and ESM matrix proteins from different shell layers (Fig. S4). Western blot results showed that native Alv was detected in the EISM extracted from both the prismatic and nacreous layers but not in the ESM from the two layers (Fig. 4d), which was consistent with the previous study in proteomic analysis of P. fucata 27 . What's more, the biochemical location pattern of Alv is consistent with the previous gene expression analyses by qPCR in different tissues. Additionally, the molecular weight of native Alv is almost 10 kDa larger than the predicted molecular weight of Alv, which corresponds to the post-translation modification of Alv in P. fucata.
To reveal the microstructure location of native Alv in shell, undecalcified shells, full-faces of decalcified shells and frozen sections of the cross-section of decalcified shells were incubated by DyLight ® 594-conjucted goat-anti-rabbit antibody 18,19 . As Fig. 5 shows, the immunofluorescence location of Alv in decalcified shells of P. fucata was existed in both the prismatic sheath and nacreous layer, which is consistent with the gene expression analysis and biochemical location 38,39 . To confirm the location of Alv in prismatic sheath and nacreous layer, we also explored cross-section of decalcified shells (Fig. 6). We found that Alv exist in both the decalcified prismatic (black arrows) and nacreous layer (white arrow). A small amount of background staining was visible in the control group but was much weaker than that of the experimental group. What's more, we also observed the full-face of the non-demineralized prismatic layer and nacreous layer, which were only cleaned by sonicating. As Fig. 7 shows, the immunofluorescence location of Alv was existed in the prismatic and nacreous lamellae, which is consistent with the gene expression analyses in mantle and EISM location by Western blotting. A small amount of background staining was also visible in the control group but was much weaker than the experimental group. Surprisingly, around the boundary of the prismatic and nacreous layers, the fluorescence intensity of Alv protein is stronger and does not fill the surface of the lamellae, but rather the edge of the location is parallel to the margin of the lamellae. While the location of Alv in the inner prismatic layer which is away from the border of the prismatic and nacreous layers is uniform on the prismatic lamellae and prismatic sheath, and the fluorescence intensity is weaker than that of lamellae around the border. More dramatically, the Alv location in the outer and mature prismatic layer concentrates in the prismatic sheath.

Discussion
The shell matrix protein of P. fucata was first obtained by classical fractionation and then identified by microarray in previous study. Alv, which exists in the prismatic layer of P. margaritifera and P. maxima with high homology, was first reported by Marie B et al. 26 . In 2015, Chuang Liu found that Alv also exists in the prismatic layer of P. fucata using a proteomic approach 27 . In Liu's research, Alv was the third most abundant matrix protein in the prismatic layer. Currently, Alv is the first studied matrix protein that was obtained by a proteomic analysis. There are Alv homologous proteins in P. maxima and P. margaritifera with high conservation, indicating that the Alv protein might play important role in biomineralization. Although the function of VP repeated sequences found in the Alv protein is unclear, it may be related to arranging the repeated structure of polymers, such as chitin according to study before 40 . Additionally, the predicted secondary structure of the Alv protein is composed of almost 80% β-strand. The Rebers-Riddiford chitin binding motif 41,42 , which is the most spread motif in chitin-binding proteins, indicating that the chitin-binding domains of chitinases are characterized by forming β-strand structures 39,43 . The mantle was reported to be important for biomineralization of which the mantle pallial is responsible for the nacreous layer formation and the mantle edge contributes to the prismatic layer formation 6 . Real-time quantitative PCR result of Alv had highest expression in the mantle suggests the contribution of Alv to shell formation, which is consistent with the expression model of other matrix proteins. Expression of Alv was also detected in the outer epithelial cells of the outer fold (OF), outer epithelial cells of the middle fold (MF), and a part of the mantle pallial by in situ hybridization experiments, implying the functions of Alv in both the prismatic and nacreous layers 6 . To confirm confirm whether native Alv exists in both layers, western blot was executed, and the results showed that the native Alv protein exists in the EISM from both layers. It is reported that EISM proteins are responsible for the construction of the organic framework of the shell; therefore, it is logical to presume that Alv is a part of the frame component according to its location and amino acids composition 33,44 .
The immunofluorescence location details of native Alv in the shell of P. fucata showed that native Alv was located in both layers. Before this study, details of matrix proteins in the whole shell had never been identified and characterized in mollusks. Hence, the detection of Alv in the whole shell could enrich matrix protein detection and the location study. The discovery of different immunofluorescence distributions of Alv in the inner and outer prismatic layers provides clues on the formation of the prismatic layer. Prior to this, scientists proposed many hypotheses about the relationship between the prismatic sheath and prismatic lamellae. According to our study, the distribution pattern supported Erben's "crush" model, in which the sheath was crushed into the margin of the prismatic lamellae 45 . According to the finding of different location pattern in different prismatic layers, we assume the Alv protein was first secreted onto the center of the lamellae to offer nucleation sites and was then distributed following prismatic lamellae growth. As the prismatic layer grew, the Alv protein was squeezed and was distributed into the prismatic sheath (Fig. 8). This presume could be supported by Erben's model and Checa's model and could enrich the assembly of macromolecules and the formation of the prismatic layer 45,46 . However, the distribution of more matrix proteins needs to be studied to either confirm the theory or to modify the theory, as many questions, such as why sheaths have the same width, cannot be explained.  antibody injection assay. The inhibition of functions of Alv in P. fucata led to opposite results of a destroyed prismatic layer and an overgrown nacreous layer which is rare in functions of matrix proteins and need to be further studied. According to the location of Alv in different prismatic layers in shell, three stages of prismatic layer were divided. At the first stage of prism formation, Alv was secreted from mantle tissue and adhere at the prismatic layer to influence calcite crystallization. Along with the growth of calcite lamella, the Alv was dispersed uniformly in second stage of prism formation. At mature stage of prism, Alv was crushed by growing calcite and distributed in the in-between organic sheath exclusively.
In conclusion, Alv, as a matrix protein exists in prism and nacre of P. fucata, plays important roles in shell formation. In this study, we cloned full-length cDNA sequence of Alv and explored the function of Alv in vivo. What's more, the finding of different location of Alv in prismatic layer provides insights into the formation of the prismatic layer of P. fucata. Figure 8. Schematic illustration of the proposed roles of Alv in the prismatic layer formation in P. fucata. Three stages from immature prism to mature prism. (i) Immature prism. Alv was secreted from mantle tissue and adhere at the immature prismatic layer to support nucleation sites and stimulate calcite crystallization. (ii) Transition prism. Alv was dispersed uniformly in second stage of prism formation as the calcite lamella grows. (iii) Mature prism. Alv was crushed by growing calcite and located in the in-between organic sheath exclusively. P, prism; N, nacre; I, in-between organic sheath.