Ethics Statement

Mussels were handled in accordance with the guidelines on the care and use of animals for scientific purposes set by the Institutional Animal Care and Use Committee (IACUC) of Zhejiang Ocean University, Zhoushan, China.

Sample preparation

M. galloprovincialis was purchased from The Mussel Farm of Gouqi Island (30°42′48″N 122°46′42″E) of Zhejiang Province, China. The adult mussels (two-year old and 6~7 cm in shell length) were cut open using a razor blade and the soft parts were removed. The shells were freshly collected and the superficial organic contaminants (including adductor muscle) were removed by incubating each intact shell in 200 mL NaOH (5%, v/v) for 1 h at 25°C. The shells were then washed with de-ionized water six times and dried. Powdered samples were collected using a scalpel from shell interior surface at the regions corresponding to nacre, myostracum, and fibrous prism, respectively.

SEM

The shell was fractured at the region around adductor muscle scar and the prepared samples were sputter-coated with gold. The section of shell samples were examined with a VEGA-3 TCSCANER scanning electron microscope at 20 kV accelerator voltage. Shell layers were identified by the sharp contact between the two types of shell microstructures, and by the change in their mineralogy.

Protein extraction and purification

Step 1, the powdered samples (50mg) from each shell layer were suspended in 100mL acetic acid (5%, v/v) for 12 h at 4°C with continuous stirring. The suspension was then centrifuged (10,000×g, 20 min, 4°C). The supernatant comprising the acid-soluble matrix (ASM) was filtered (0.22 μm) and dialyzed against de-ionized water before being freeze-dried and weighed.

Step 2, the acid-insoluble matrix was recovered in 5% acetic acid (v/v,) containing 8M urea and 0.1 mM tri (carboxyethyl)-phosphine. After homogenization at 4°C using a small hand-held tissue grinder., the homogenate was centrifuged (12,000 ×g, 20 min, 4°C) and the supernatant was dialyzed against 4 liters of 5% acetic acid twice. The supernatant, corresponding to the urea-soluble matrix (USM), was then freeze-dried and weighed.

Step 3, after extracted by acetic acid and urea, the resulting pellets, corresponding to insoluble matrix, was rinsed six times with de-ionized water, freeze-dried and weighed.

HPLC purification was carried out on a Waters Delta 600 HPLC system equipped with a Waters 2487 absorbance detector (Waters USA). About 1 mg of ASM and USM from each of the shell layers were subjected to a Sunfire analytic C8 column (4.6×250 mm, Waters) that equilibrated with deionized water containing 0.1% trifluoroacetic acid (TFA). Gradient of 5–20% acetonitrile containing 0.1% TFA over 5 min, followed by 20–60% gradient over 60 min and 60–95% gradient over 5 min were performed at a flow rate of 1.0 mL/min. The eluted fractions (monitored at 280 nm) were collected manually and freeze-dried for LC-MS/MS analysis.

SDS-PAGE was done on a Mini-PROTEAN Tetra Electrophoresis System (8.3×7.3cm, Bio-Rad) with a 4% stacking gel and 13% resolving gel. Samples (100μg) of ASM, USM, and insoluble matrices were suspended in 20 μL loading buffer and heated to 95°C for 5 min. Loading buffer-insoluble material was removed by centrifugation in an Eppendorf bench top centrifuge for 5 min at 13000 rpm. Gels were loaded with 15 μL of matrix sample supernatant per lane and stained with Coomassie Brilliant Blue R250 after electrophoresis.

MS analysis

Before mass spectrometry (MS) analysis, the protein sample (about 100 μg) was reduced with 10 mM dithiothreitol in 50 mM NH₄HCO₃ at 57°C for 1 h and alkylated by 50 mM iodoacetamide in 50 mM NH₄HCO₃ for 45 min at room temperature in the dark. After treated with 2 μg of trypsin in 10 μL 50 mM NH₄HCO₃ for 18 h at 37°C, the sample was dried in a vacuum concentrator and re-suspended in 30 μL of 0.1% (TFA) and 4% acetonitrile.

The LC-MS/MS analysis was performed using a QSTAR-ELite system (Applied Biosystems, USA) interfaced with a 20 AD HPLC system (Shimadzu, Japanese). The trypsin digested mixture of protein sample (about 20 μg) was separated on a Zorbax 300SB-C18 column (0.1×150 mm, 5μ, 300 Å; Microm, Auburn, CA). The HPLC gradient was 5~35% buffer B (95% ACN, 0.1% formic acid) in buffer A (5% ACN, 0.1% formic acid) at a flow rate of 0.3 mL/min over 90 min. MS data were acquired automatically using Analyst QS 1.1 software (Applied Biosystems). Following a MS survey scan over m/z 350~1 500, the MS/MS spectra were sequentially and dynamically acquired for the eight most intense peptide molecular ions over m/z 100~2 000.

The LTQ analysis of insoluble fraction was performed on a LC-20AD system (Shimadzu, Tokyo, Japan) connected to an LTQ Orbitrap mass spectrometer (Thermo Electron, Bremen, Germany) equipped with an online nano-electrospray ion source (Michrom Bioresources, Auburn, CA). The trypsin digested mixture of insoluble fraction was separated on a reverse phase column (100 μ, MICHROM Bioresources Inc., Auburn, CA). The HPLC gradient was 5–45% of solvent B (95% ACN containing 0.1% formic acid) over 90 min at a flow rate of 60 μL/min. Eluted peptides were analyzed by MS and data-dependent MS/MS acquisition, selecting the 8 most abundant precursor ions for MS/MS with a dynamic exclusion duration of 1 min.

Data analysis

Protein identification was performed using the MASCOT search engine (version 2.1, Matrix Science, London, UK) against a protein database comprising 68209 sequences derived from the EST libraries of Mytilus spp. (Taxid: 6548, mainly represented by 42354 sequences from M. californianus, 19617 from M. galloprovincialis, 5300 from M. edulis, and 719 from M. coruscus) downloaded (August, 2014) from the NCBI server (http://www.ncbi.nlm.nih.gov). Methionine oxidation and carbamidomethylation were set as variable and fixed modifications, respectively. The mass tolerance was set to 20 ppm for the precursor and 0.3 Da for the product ions. The peptide hits were manually confirmed by the inspection of the MS/MS spectra. Ions scores are defined as (-10*Log(p)), where p is the probability that the observed match is a random event. Only peptide matches with an expectation score ≤0.05 were accepted.

Sequence analysis was performed using Lasergene software (version 7.1). Sequence homologies were determined by searching non-redundant protein databases via the BLAST server (http://www.ncbi.nlm.nih.gov/BLAST/). Signal peptides were predicted using the SignalP 4.0 online tool (http://www.cbs.dtu.dk/services/SignalP/) and conserved domains were predicted using SMART online tool at http://smart.embl-heidelberg.de/. The amino acid composition was calculated by ProtParam tool at http://web.expasy.org/protparam/. Sequence alignments were performed with ClustalW2 at http://www.ebi.ac.uk/Tools/msa/clustalw2/.

The microstructure of M. galloprovincialis shell

As shown in Fig 1, on the interior surface of adult M. galloprovincialis shell, three regions with different color and texture can be observed optically, including the adductor muscle scar (AMS), white color region (designated as AMS-A, anterior side of AMS), a black region located on the growing margin of the shell (designated as AMS-P, posterior side of AMS). The SEM images of the section of AMS, AMS-A, and AMS-P were shown in the Fig 1C–1F. Three layers, nacre, myostracum, and fibrous prism, were observed by SEM with different morphology. The top layer of ASM-A, AMS, and AMS-P, corresponding to the nacre, myostracum, and fibrous prism, respectively, was scraped off and the shell matrices were then extracted for proteomic analysis.

Download:

• PPT
  PowerPoint slide
• PNG
  larger image
• TIFF
  original image

Fig 1. The photograph of the whole shell and the SEM images of cross-section of M. galloprovincialis shell.

A: the exterior side of M. galloprovincialis shell; B: the interior side of M. galloprovincialis shell, the white line represents the cutting plane; “AMS”, “AMS-A”, and “AMS-P” represent the area of adductor muscle scar (AMS), anterior side of AMS, and posterior side of AMS, respectively. C: the cross-sectional SEM image of AMS-A, showed nacre (N) on the top layer and fibrous prism (FP) at the bottom of this area; D: enlarged image of nacre from AMS-A; E: the cross-sectional SEM image of central AMS, showed myostracum (M) at the top, nacre (N) at the middle, and fibrous prism (FP) at the bottom of this area; F: the cross-sectional SEM image of AMS-P, showed fibrous prism at this area.

https://doi.org/10.1371/journal.pone.0133913.g001

Isolation of shell matrices from different shell layers

The amount of total organic matrix from shell layers represents an average of 3.4% by weight of the shell powder (wt/wt), and the highest content of organic material was presented in nacre (5.3%), followed by myostracum (3.5%) and fibrous prism (1.9%). The amount of ASM was 1.2%, 0.7%, and 0.5% for the nacre, myostracum and fibrous prism, respectively. Further extracted USM yielded additional 1.0%, 0.6%, and 0.8% for the nacre, myostracum and fibrous prism, respectively. The soluble matrix (ASM plus USM), formed more than 40% of the total organic matrix in each shell layer, was further fractioned by HPLC. The insoluble fractions were directly analyzed by LTQ.

HPLC purification of ASMs from three shell layers showed a similar profile with minor differences (Fig 2). The chromatogram obtained on a C8 column for ASMs showed a bell-shaped elution profile at the region of 25–35% acetonitrile and several peaks at the region of 35–45% acetonitrile, indicating the presence of multiple proteins with similar chromatographic properties. For nacre, an additional fraction was eluted at the region of 45–55% acetonitrile, indicating more hydrophobic proteins of nacre than that of myostracum and fibrous prism. The HPLC chromatogram of USMs showed a better elution profile with several sharp peaks presented in the region of 40–45% of acetonitrile and these individual peaks were observed with different retention time for USMs of three layers. In order to investigate the largest part of M. galloprovincialis shell proteome, the HPLC eluted fractions were collected according to the order of the eluted peaks (denoted by “I”, “II”, “III”, etc. in S1 Fig) and analyzed by LC–MS/MS.

Download:

• PPT
  PowerPoint slide
• PNG
  larger image
• TIFF
  original image

Fig 2. HPLC elution profiles and SDS-PAGE analysis of shell matrices from nacre, myostracum, and fibrous prism, respectively.

A: HPLC elution profiles of acid-soluble matrices from nacre (N-A), myostracum (M-A), and fibrous prism (FP-A). The bold line represents the concentration of acetonitrile. B: HPLC elution profiles of urea-soluble matrices from nacre (N-U), myostracum (M-U), and fibrous prism (FP-U). The bold line represents the concentration of acetonitrile. C: PAGE comparison of acid-soluble (ASM), urea-soluble (USM), and insoluble matrices (IM) of nacre (N), myostracum (M), and fibrous prism (FP).

https://doi.org/10.1371/journal.pone.0133913.g002

Comparison of the protein band pattern of matrices from different shell layers showed comparable differences (Fig 2C). More protein bands were observed in insoluble matrices than that in soluble matrix (ASM and USM). Further, the samples of nacre and myostracum apparently showed more complexity of protein composition than that of fibrous prism. These results suggest that the proteome variables were not only in different shell layers but also in different extraction protocols.

Protein identification of soluble matrices from different shell layers

By combining the LC-MS/MS analysis with Mytilus EST database interrogations, a set of 18 proteins were identified from the ASMs of three layers, including fifteen, seven, and five proteins yielded by nacre, myostracum and fibrous prism, respectively (Table 1 and S1 Table). We observed some proteins with high mascot score and high sequence identity (more than 80%) with previously reported Mytilus SMPs, including MUSPs, Shell Matrix Protein, Perlucin-like protein, MSI60-like protein, Fibronectin-like protein, and EP-protein [20, 21], In addition, many novel shell proteins were identified from the three layers. From nacre, a protein containing chitin-binding domain was identified with low sequence identity (34%) to BMSP (blue mussel shell protein), a calcium carbonate-binding protein from M. galloprovincialis shell [22]. EST gi|154348940 was identified as a 216-AA long protein with a 16-AA long signal peptide and showed 71% sequence identity with Apextrin-like protein of M. galloprovincialis, a potential immune-related protein [23]. Five proteins showed neither homology nor putative domains, but with unusual amino acid composition. For example, EST gi|58308563, the most abundant protein in the ASM, encodes an AGL-rich protein containing abundant Ala (41.5%), Gly (21.6%), and Leu (8.8%), and several poly-Ala blocks (consist of 11–12 alanines per block) (Fig 3 and S1 Table). The other novel proteins were AGS-rich protein (43.7% of Ala, 22.8% of Gly, and 10.2% of Ser), GRS-rich protein (22.8% of Gly, 9.5% of Arg, and 8.2% of Ser), and GA-rich protein (33.4% of Gly and 30.8% of Ala). For myostracum, three novel proteins were identified. Including collagen-like protein, protein with THY domains (thymosin beta actin-binding motif), and AGL-rich protein. For the fibrous prism, only five proteins were identified, including MSI60-like protein.

Download:

• PPT
  PowerPoint slide
• PNG
  larger image
• TIFF
  original image

Fig 3. A representative MS/MS spectrum for the two peptides matching with EST gi|58308563, an AGL-rich protein of M. galloprovincialis.

A: protein sequence derived from EST gi|58308563. The matched peptides from MS/MS spectra were underlined; B: MS/MS spectrum of the peptide “-AAAAAGASAAAGGSGGTLR-” with m/z of 1486.90 Da; C: MS/MS spectrum of the peptide “-LISRIIAR-” with m/z of 940.71 Da.

https://doi.org/10.1371/journal.pone.0133913.g003

Download:

• PPT
  PowerPoint slide
• PNG
  larger image
• TIFF
  original image

Table 1. The proteome of nacre (N), myostracum (M), and fibrous prism (M) of M. galloprovincialis shell.

https://doi.org/10.1371/journal.pone.0133913.t001

A set of 16 proteins were identified from the USMs of three shell layers, including nine, seven, and six yielded by nacre, myostracum, and fibrous prism, respectively (Table 1 and S2 Table). Comparing to the ASMs, nine proteins were additional. For nacre, only two additional proteins, a GYS-rich protein and a SRAV-rich protein, were identified from the USM. For myostracum, five additional proteins were identified from the USM and four of them exhibited homology with SMPs previously detected in other mollusk shell, such as Calponin-like proteins [24, 25], KU domain-containing protein [26], and Whirlin-like protein [27]. For fibrous prism, only one additional protein was identified as FGV-rich protein with abundant Phe (23.9%), Gly (17.8%), and Vla (17.4%).

Protein identification of insoluble matrices from different shell layers

The shotgun proteomic approach allowed us to identify 43, 53, 43 proteins from nacre, myostracum, and fibrous prism, respectively, yielded a set of 97 proteins from insoluble matrices (Table 1 and S3–S5 Tables). Of the 43 proteins identified from the insoluble matrix of nacre, the most abundant were a series of calponin-like proteins (8 matched ESTs). Additionally, two proteins were identified as calponin homology (CH) domain containing proteins. Six proteins were identified as homologues of reported Mytilus SMPs, including MUSP, Perlucin-like protein, Shell Matrix Protein, and Fibronectin-like protein. Five proteins were identified as Chitinase-like proteins with significant sequence identity (> 50%) to chitinase from other species. Four Collagen-like proteins were identified, including homologues of precollagen D and proteins containing von Willebrand factor type A (VWA) domain. Two proteins with filament domain were identified with high sequence identity (>50%) to filament proteins of other species. One proteinase inhibitor-like protein was identified with homology (32% sequence identity) to nacre serine protease inhibitor 2 (NSPI2) of P. maxima shell [26]. The insoluble matrix of nacre also contained A-rich, G-rich, M-rich, L-rich, R-rich, and S-rich proteins (Table 1 and S3 Table).

For myostracum, 53 proteins were identified from the insoluble matrix, of which, 19 proteins contained calponin domain or CH domain and seven were Collagen-like proteins, including three collagen domain-containing and four VWA domain-containing proteins. Five proteins were identified as arginine kinase-like protein with ATP-guanido phosphotransferases domain and 67–78% sequence identity to arginine kinase of other species. Three proteins were identified as filament domain-containing protein and showed sequence identity of 65–88% with filament protein of other species. One protein was identified as Nacrein-like protein with sequence identity of 42% to the Nacrein-like protein from P. vulgate shell. The other proteins included adenylate kinase-like protein, C1q domain containing protein, glycolytic domain containing protein, protease inhibitor-like protein, and various proteins with unusual amino acid composition, including S-rich, L-rich, M-rich, A-rich, and G-rich proteins (Table 1 and S4 Table).

For fibrous prism, 43 proteins were identified and the most abundant were proteins with calponin domain and/or CH domain, and Alveoline-like proteins with homology to Alveoline of pearl oyster P. margaritifera shell [28]. Five proteins were identified as Cathepsin-like proteins containing cathepsin propeptide inhibitor domain and/or cysteine protease domain. Moreover, proteins with filament domain, Collagen-like proteins, protein with C1q domain, Gigasin-like protein, and chitinase-like protein were also identified. In addition, 11 proteins were identified as uncharacterized protein, including four M-rich proteins, three G-rich proteins, two R-rich proteins, one A-rich protein, and one S-rich protein (Table 1 and S5 Table).

Myostracum-related proteins

The proteins identified from myostracum is more in number than those from nacre and fibrous prism. Abundant of calponin-like proteins, transgelin-like proteins, collagen-like proteins, filamin-like proteins, and enzyme-like proteins were detected exclusively to myostracum (Table 1). Calponin domain and/or calponin-homology domain were detected in some of these proteins, including calponin-like proteins, transgelin-like proteins, and a filamin-like protein. The potential actin-binding function of calponin domain [32, 33] confers these proteins with possible roles in calponin-actin interaction, which may be involved in myostracum-muscle attachment. In addition, a protein containing thymosin beta actin-binding motif (THY domain) was also identified from myostracum, providing another candidate for actin-binding. Besides of these possible actin-binding proteins, abundant of collagen-like proteins were detected in myostracum. Earlier work revealed that collagenase can significantly reduce the adherence of muscle-shell of C virginica [34]. Therefore, the collagen-like proteins from myostracum may also participate in shell-muscle attachment.

Protease inhibitor-like, Gigasin-like and Alveoline-like proteins

A potential protease inhibitor was identified from both of the nacre and myostracum. This 149-AA long sequence, with a 15-AA long signal peptide and KU domain in its sequence, exhibited 32% sequence identity with NSPI2, a KU domain-containing protein from the oyster P. maxima shell (Fig 5A) [28]. Protease inhibitors have also been identified in other mollusks shell [16, 35] and the protease inhibitor domain has been detected in other SMP, such as lustrin A, an insoluble protein identified from the nacre of abalone H. rufescens [36]. These results suggested a protecting system that precludes extracellular proteolysis during the shell formation. An EGF domain-containing protein, with 241-AA long sequence including an 18-AA long signal peptide, was identified from the insoluble matrix of fibrous prism. This protein showed 29% sequence identity with gigasin-2, an EGF domain-containing protein of C. gigas shell [18] (Fig 5B). The EGF domain has already been found in several reported SMPs from other Mollusca, such as perlustrin from H. laevigata [37] and SMPs from L. gigantea [38]. Abundant of val-rich proteins were identified from the insoluble matrix of fibrous prism. These proteins, characterized by abundant Val, Pro, and Lys residues and showed 30~40% of sequence identity with alveoline-like protein, a prismatic SMP of P. margaritifera [28] (Fig 5C).

Download:

• PPT
  PowerPoint slide
• PNG
  larger image
• TIFF
  original image

Fig 5. Domain organization and sequence alignment of representative novel proteins of M. galloprovincialis shell with their homologous from other shell.

A: protease inhibitor-like protein of M. galloprovincialis (gi|238643545) and NSPI2 of P. maxima (sp|P86964.1); B: EGF-domain containing protein of M. galloprovincialis (gi|145890922) and Gigasin-2 of C. gigas (sp|P86785.1); C: alveoline-like protein of M. galloprovincialis (gi|212821271) and alveoline-like protein of P. margaritifera (sp|H2A0K6.1); D: chitinase-like proteins of M. galloprovincialis (gi|58308904 and gi|58307961) and acidic mammalian chitinase isoform X4 of Saimiri boliviensis (gi|72558135); E: EP proteins of M. galloprovincialis (gi|145893971 and gi|34304719) and EP protein of M. edulis (gb|AAQ63463.1); F: Cathepsin-like protein of M. galloprovincialis (gi|238641522) and Cathepsin of C. gigas (gi|405966499). Signal peptides are underlined; “*”represents identical amino acids; “:” and “.” represent similar amino acids; “-”represents the gaps inserted in the sequence.

https://doi.org/10.1371/journal.pone.0133913.g005

Other shell proteins of M. galloprovincialis

Chitin metabolic-related proteins were identified in this study, including 5 chitinase/chitotriosidase-like proteins with Glyco_18 domain from N layer (Fig 5D) and a chitin synthase-like protein from FP layer. Chitin is a major non-protein component of mollusc shells [39–41] and involved in constitution a framework binding silk-like proteins and acidic proteins [42]. The identification of chitin metabolic-related proteins in mussel shell suggests a possible roll of these proteins in shell construction or repair by modifying the chitin framework. Two histidine-rich proteins were identified with high homology to M. edulis EP protein (Fig 5E). Both of the two sequences contain a C1Q domain in the C-terminus and abundant His (14.3%). The function of EP protein in shell formation is unknown. However, an earlier work identified a histidine-rich calcium-binding protein of the sarcoplasmic reticulum of the rabbit muscle with 13% of His [43], suggesting that the positively charged His residue may also play a functional role in the binding of calcium ions.

A set of cathepsin-like proteins were identified mainly from fibrous prism. These proteins showed high sequence identity (70~80%) with cathepsin L, a protease previously identified from C.gigas shell [27] (Fig 5F). Together with the protease inhibitor-like identified in the nacre and myostracum, a protein degradation/protection system has been found in mussel shell and may play a role in regulating the catabolic of shell protein during shell formation and growing.

Uncharacterized M. galloprovincialis shell proteins

Many identified proteins contain unusual amino acid composition, or short tandem repeats, or blocks of identical amino acids. Most of these proteins exhibited neither homology nor domains except for low complexity regions (LCRs). We observed that LCRs are different among the proteins of three layers. For example, GRS-rich, AGS-rich, and GYS-rich types are only presented in the proteins of nacre; SGC-rich and LKD-rich in the proteins of myostracum; FGV-rich, GFS-rich, and GYDK-rich in the proteins of fibrous prism. Some of these primary sequence features are also found in other SMPs that have been proposed to adopt a specific structure when binding to the calcium carbonate crystal surface [26, 44].

In addition, we also observed some intracellular proteins were presented in M. galloprovincialis shell, especially in insoluble matrix, such as actin/paramyosin, ubiquitin/polyubiquitin, and tubulin (data not shown). Proteins of previously known intracellular location were also found in other mollusk shell proteome [27, 28, 38]. In the present study, before LC-MS/MS analysis, the shell samples were processed by sodium hydroxide bleaching to eliminate potential organic contaminants on the shell, thus, it is puzzling and hard to explain the presence of these intracellular proteins in the shell, however, we cannot exclude the possibility that these intracellular protein may be transported into extrapallial fluid via special pathway and may bind with the shell or SMPs. As true intra-crystalline components, although probably without any function, intracellular proteins may not be removed even by rigorous sodium hydroxide cleaning. Furthermore, the insoluble matrices contained more of these intracellular components, one may conclude that many of them were already structurally modified and aggregated before incorporation into the shell.