Report

An Acidic Matrix Protein, Pif, Is a Key Macromolecule for Nacre Formation

See allHide authors and affiliations

Science  11 Sep 2009:
Vol. 325, Issue 5946, pp. 1388-1390
DOI: 10.1126/science.1173793

Abstract

The mollusk shell is a hard tissue consisting of calcium carbonate crystals and an organic matrix. The nacre of the shell is characterized by a stacked compartment structure with a uniformly oriented c axis of aragonite crystals in each compartment. Using a calcium carbonate–binding assay, we identified an acidic matrix protein, Pif, in the pearl oyster Pinctada fucata that specifically binds to aragonite crystals. The Pif complementary DNA (cDNA) encoded a precursor protein, which was posttranslationally cleaved to produce Pif 97 and Pif 80. The results from immunolocalization, a knockdown experiment that used RNA interference, and in vitro calcium carbonate crystallization studies strongly indicate that Pif regulates nacre formation.

Pearls made by pearl oysters are structurally identical to the nacre of oyster shell. The nacreous layer forms inside the mollusk shell to protect the soft body (Fig. 1A). The nacre has a laminated composite structure of calcium carbonate crystals in the aragonite polymorph and biological macromolecules such as the polysaccharide chitin and proteins, which provide high mechanical strength and unusual optical properties. The layered compartment structure reflects light and causes interference, which imparts luster (1). The shell of the Japanese pearl oyster Pinctada fucata, the common “mother-of-pearl” shell, consists of two layers. The outer layer, the prismatic layer, comprises most stable calcite crystals, whereas the inner layer, the nacreous layer, is made up of thermodynamically metastable aragonite crystals (2). The nacreous layer has an orderly, planar compartment structure, and the single aragonite crystal in each compartment has almost a uniform orientation of the c axis perpendicular to the plane of the nacreous layer (3, 4). Although crystal alignment may be attained by using an organic template or heteroepitaxy, how it is achieved in nacre remains unclear.

Fig. 1

Structure of the nacreous layer and the schematic representation of Pif. (A) The shell and pearl of P. fucata. (B) SDS-PAGE of the samples from the calcium carbonate crystal-binding experiment. Lane 1 shows the acid-insoluble, SDS-soluble organic matrices from the nacreous layer; lanes 2 to 4 show samples from the aragonite-binding experiment; and lanes 5 to 7 show samples from the calcite-binding experiment. In lanes 2 and 5, the washings were with DW; in lanes 3 and 6, the washings were with 0.1 M NaCl; and in lanes 4 and 7, the washings were with 0.5 M NaCl. The red arrow indicates the 80-kD band (Pif 80) specifically bound to aragonite. The green arrowhead indicates the 97-kD band (Pif 97). (C) Schematic representation of the deduced amino acid sequence of Pif (DNA Data Bank of Japan number AB236929). The orange box is a signal peptide. The blue box indicates the VWA domain. The yellow box is the Peritrophin A-type chitin-binding domain. RMKR is a Kex2-like proteinase cleavage site. The green box is the aragonite-binding protein (Pif 80). Scale bars, 1 cm (A) and 100 amino acids (C).

In 1960, Watabe and Wilbur reported that the whole organic matrices extracted from the nacreous layer induced aragonite crystal formation (5). Later, it was suggested that Asp residues in the organic matrices interacted with calcium atoms in the calcium carbonate in order to regulate the crystal polymorph (68). Recent studies imply that a few macromolecules in the nacreous layer play important roles in the formation of aragonite crystals (912). Although a number of matrix proteins have been identified from various mollusk shells (13), no proteins that induce aragonite crystal formation characteristic of nacre have been characterized, nor has any Asp-rich acidic macromolecule been identified.

To identify the aragonite-inducing factor in the nacreous layer, we searched for an aragonite-specific binding protein from the nacre using a calcium carbonate–binding assay (14). Almost all the matrix proteins were recovered in the distilled water (DW) washings in both incubations with calcite and aragonite crystals, and no matrix proteins were detected in the NaCl washings (Fig. 1B and fig. S1). In the DW washing, an 80-kD band disappeared after incubation with aragonite crystals. This 80-kD protein, Pif 80, also bound to calcite crystals to some extent but bound more specifically to aragonite crystals.

The complete cDNA sequence of Pif 80 was determined with conventional cDNA cloning methods (Fig. 1C and fig. S2) (14). The deduced amino acid sequence revealed that this cDNA encoded an additional protein (named Pif 97). The N-terminal sequence of an SDS–polyacrylamide gel electrophoresis (SDS-PAGE) band around 97 kD agreed well with that of Pif 97 deduced from the cDNA sequence. The intervening sequence between Pif 80 and Pif 97, RMKR, is a cleavage site often observed in posttranslational processing (15). Both Pif 80 and Pif 97 are acidic with calculated isoelectric point (pI) values of 4.99 and 4.65, respectively. Pif 97 consists of 525 amino acid residues and has two conserved domains, a von Willebrand type A (VWA) for protein-protein interaction domain (16, 17) and a chitin-binding domain similar to that of Peritrophin A (1820). Pif 97 contains a high proportion of charged amino acid residues, Asp (14.9%), Glu (6.5%), Lys (11.1%), and Arg (5.0%), and many Cys residues (23 residues) that might form disulfide bridges to confer a rigid three-dimensional conformation. In contrast, Pif 80 consists of 460 amino acid residues and has no conserved domains. Pif 80 has more charged amino acid residues [Asp (28.5%), Glu (4.1%), Lys (18.7%), and Arg (10.9%)] than does Pif 97. Pif 80 has 17 repeats of a four-amino-acid motif, Asp-Asp-Arg (Lys)–Lys (Arg), scattered throughout its sequence and a cluster of acidic amino acid residues (Asp2-Glu-Asp7) near the center of the molecule. The high ratio of Asp in Pif 80 may play a role in aragonite-binding, considering that acidic amino acid residues are associated with the regulation of crystal polymorph (7, 8) and that poly-(Asp) has been shown to induce aragonite formation (21).

In order to study the localization of Pif 80 in the nacreous layer, we examined immunohistochemical localization using a scanning electron microscope (SEM). A cross-sectional surface of the nacreous layer was first incubated with a polyclonal antibody raised against Pif 80 in rabbit and then treated with an antibody to rabbit immunoglobulin G (IgG) conjugated with gold nanoparticles. We checked the specificity of this antibody by means of Western blot (fig. S3). The gold nanoparticles were observed with back-scattered electron imaging. Gold nanoparticles could be observed throughout the nacreous layer (137 ± 15/100 μm2; n = 5 areas) after incubation with the antibody to Pif 80 (Fig. 2A). Conversely, only a small number of gold nanoparticles were observed in the nacreous layer (18 ± 15/100 μm2; n = 8 areas) after treatment with preimmune IgG (Fig. 2B). The organic boundary area between the nacreous and prismatic layers is also recognized as a place for initiation of aragonite crystallization. A large number of gold nanoparticles were observed only at the nacreous side of the organic boundary (Fig. 2C), whereas almost no gold nanoparticles were observed after incubation with preimmune IgG (Fig. 2D). This is consistent with the organic boundary as the initiation site of aragonite formation in the nacreous layer (22, 23). Thus, Pif is an important component in the nacreous layer and takes part in the initiation of aragonite crystallization as well as subsequent stacking of aragonite tablets in the nacreous layer.

Fig. 2

Localization of Pif 80 by means of immunohistochemical SEM image analysis using gold nanoparticles. (A and C) The nacreous layer incubated with an antibody to Pif 80. (B and D) The nacreous layer incubated with the preimmune IgG fraction. Arrows indicate the organic boundary between the nacreous and prismatic layers. P, the prismatic layer; N, the nacreous layer. Scale bars, 1 μm.

To clarify the function of Pif in vivo, knockdown of Pif gene expression by means of RNA interference (RNAi) was performed. Double-stranded RNA (dsRNA) designed from the Pif cDNA sequence was injected into the muscle of P. fucata (which had a shell length of 5 to 6 cm), and the expression levels of Pif mRNA in the mantle were measured with real-time quantitative polymerase chain reaction (PCR) 7 days after injection (Fig. 3A). The Pif expression levels decreased with increasing injection doses of Pif dsRNA (5, 15, and 30 μg). The expression level of the group injected with 30 μg of Pif dsRNA was suppressed to approximately 40% of that of the phosphate-buffered saline (PBS) or green fluorescent protein (GFP) dsRNA–injected group.

Fig. 3

Knockdown of the Pif gene by means of RNAi. (A) The expression levels of Pif mRNA in the mantle were determined with real-time quantitative PCR 7 days after injection. The Pif mRNA expression level of the PBS group is attributed a relative value of 1.0. Nine oysters (n = 9) were used in each experiment. Statistically significant differences were analyzed by means of one-way analysis of variance. Asterisk indicates a significant reduction (P < 0.01) as compared with PBS-injected oysters. (B and C) SEM images of the surface of the nacreous layer of the oysters injected with PBS and 30 μg of Pif dsRNA, respectively. (D and E) Cross sections of the nacreous layer of the oysters incubated in seawater containing a high concentration of strontium on the day of injection of PBS and 30 μg of Pif dsRNA, respectively, followed by returning to normal seawater for 6 days and observed by using SEM. The arrowed circle area was analyzed with EDX, and strontium was detected. Scale bars, 10 μm [(B) and (C)] and 2 μm [(D) and (E)].

The surface structure of the nacreous layer in each injection group was observed with SEM. The normal orderly structure of the nacreous layer was observed in the PBS and GFP dsRNA–injected groups (Fig. 3B and fig. S4A, respectively), whereas a disordered growth of the nacreous layer was observed in the 15 and 30 μg of Pif dsRNA–injected groups (fig. S4B and Fig. 3C, respectively). To observe the growth of the nacreous layer after injection, the pearl oysters were incubated in seawater containing a high concentration of strontium for 24 hours after injection and then returned to normal seawater, which was followed by rearing for 6 more days. Strontium was successfully incorporated into the shell, which was detected with a bright contrast in the back-scattered electron image and by an energy dispersive x-ray detector (EDX). The growth rate of the shell after injection could be estimated by the thickness of the nacreous layer from the surface to the position of the strontium marker. The strontium marker in the PBS or GFP dsRNA–injected groups was detected at the 10th layer from the surface, indicating that the compartment structures grew normally after injection (Fig. 3D and fig. S4, C and D). On the other hand, the strontium marker in the 30 μg of Pif dsRNA–injected group was detected at the surface of the nacreous layer, indicating that growth of the nacreous layer was considerably arrested (Fig. 3E). Thus, the injection of 30 μg of Pif dsRNA stopped regular crystallization and allowed random growth, suggesting that Pif also takes part in lamellar sheet formation, which is indispensable for compartment structure formation, and therefore Pif is an essential component of the organic matrix for normal growth of the nacreous layer.

To clarify how Pif is associated with calcium carbonate crystallization, we performed in vitro crystallization experiments (24). The acid-insoluble, SDS-soluble (without dithiothreitol) organic matrices from the nacreous layer were separated into four fractions by means of gel filtration high-performance liquid chromatography (HPLC) (fig. S5), and each fraction was subjected to SDS-PAGE. SDS-PAGE under reducing conditions revealed that Pif 80 and Pif 97 were recovered as major components in fraction 2, which also contained a part of N16, a major protein in the nacreous layer (25), whereas SDS-PAGE under nonreducing conditions revealed that these proteins formed a high-molecular-weight complex (Fig. 4A). The presence of the chitin-binding domain in Pif 97 and solubilization of Pif from the acid-insoluble chitin complex suggests that the Pif complex binds to chitin in vivo. Thus, the materials of fraction 2 were applied to a chitin-coated glass plate and incubated in a calcium carbonate–supersaturated solution. Optical microscope, polarizing microscope, and SEM observations showed that a number of flat crystals were formed (fig. S6, A to D). Analyses of the crystal polymorphs performed with micro-Raman spectroscopy revealed that these crystals were aragonite or vaterite (fig. S6, E and F). A thin cross section of one of the aragonite crystals was prepared with a focused ion beam (FIB) system. Analysis of the inner structure with a transmission electron microscope (TEM) (Fig. 4, B and C) showed that the crystal was formed between the chitin membrane and the glass plate and that the monocrystal had its c axis perpendicular to the glass plate (Fig. 4D). Thus, fraction 2 induced aragonite crystal formation with characteristics close to those observed in the nacreous layer. Although fraction 3, obtained by means of gel filtration, contained more N16 than did fraction 2, it did not induce characteristic aragonite crystal formation. However, we do not exclude that it could be involved in the context of a Pif complex. In negative control experiments, only rhombohedral calcite crystals were observed on the chitin-coated glass plate with or without bovine serum albumin (fig. S6, G to J). These results suggest that Pif 97 and Pif 80 are key molecules in the induction of aragonite crystal formation, though they could be part of a larger protein complex on which the proteins are also essential for the formation of aragonite crystals.

Fig. 4

The effect of Pif on calcium carbonate crystallization in vitro. (A) Each fraction separated by means of gel filtration HPLC (fig. S5) was subjected to SDS-PAGE under reducing or nonreducing conditions. (B) A piece of calcium carbonate crystal in fig. S6A was observed by use of SEM. Scale bar, 50 μm. (C) A cross section of the white box in (B) prepared by means of FIB was observed with TEM. The crystal was formed in the space between the chitin membrane and the glass plate. (D) The electron diffraction pattern of the white circle area in (C), indicating a single aragonite crystal and its c axis perpendicular to the glass plate. 2ME, 2-mercaptoethanol; C, chitin; G, glass plate.

Because the lamellar sheet may be constructed by a complex process that includes the self-assembling of chitin and proteins, Pif 80 and Pif 97 probably participate in this important process. The complex of Pif 97 and Pif 80 appears to be a candidate molecule for the putative molecule that interacts with chitin and nucleates aragonite crystals in the model of the nacreous layer previously proposed by Levi-Kalisman et al. (26). On the basis of our findings, we propose a model for the molecular mechanism of nacreous layer formation (fig. S7). A complex of Pif 97 and Pif 80 is formed in the mantle epithelial cells and secreted into the extrapallial fluid. Pif 97 may bind to chitin microfibrils through the Peritrophin A-type chitin-binding domain and make a bigger aggregate with N16 and other proteins, contributing to the formation of the lamellar sheet. In this sheet, the acidic matrix protein, Pif 80, may concentrate calcium carbonate and induce aragonite crystal formation as well as regulate the orientation of the c axis. Chitin sheets, Pif complexes (Pif 97, Pif 80, N16, and other proteins), and aragonite crystals then stack repeatedly to form the nacreous layer.

Supporting Online Material

www.sciencemag.org/cgi/content/full/1173793/DC1

Materials and Methods

Figs. S1 to S7

References

References and Notes

  1. Materials and methods are available as supporting material on Science Online.
  2. We are grateful to M. Hayashi and H. Aoki of the Fisheries Research Institute, Mie Prefecture, Japan, for a kind gift of live pearl oysters and to S. Akera and E. Okamoto of Tasaki Shinju Co. for providing us Japanese pearl oyster shells. We are also grateful to H. Kagi of the Geochemical Laboratory, Graduate School of Science, the University of Tokyo, Japan, for measuring Raman spectra and to H. Tanaka of Okutama Kogyo Co. for a kind gift of aragonite crystals. We also thank V. Jayasankar of the Japan International Research Center for Agricultural Sciences for critical reading of this manuscript. This work was supported by a Grant-in-Aid for Scientific Research (17GS0311) from the Japan Society for the Promotion of Science (JSPS). M.S. was supported by a Research Fellowship of JSPS for young scientists.
View Abstract

Navigate This Article