Perna canaliculus - American Chemical Society

Oct 24, 2013 - Exposed and Buried Biomineral Interfaces in the Aragonitic Shell of. Perna canaliculus Revealed by Solid-State NMR. Ira Ben Shir,. †...
37 downloads 3 Views 2MB Size
Article pubs.acs.org/cm

Exposed and Buried Biomineral Interfaces in the Aragonitic Shell of Perna canaliculus Revealed by Solid-State NMR Ira Ben Shir,† Shifi Kababya,† Itai Katz,† Boaz Pokroy,‡ and Asher Schmidt*,† †

Schulich Faculty of Chemistry and Russell Berrie Nanotechnology Institute, Technion − Israel Institute of Technology, Technion City, Haifa 32000, Israel ‡ Department of Materials Science and Engineering, Technion − Israel Institute of Technology, Technion City, Haifa 32000, Israel S Supporting Information *

ABSTRACT: A comprehensive molecular description of the inorganic− bioorganic interfaces and internal structure of the aragonitic shells of Perna canaliculus is derived by employing solid-state NMR spectroscopy. The primary component of the shell, the highly ordered aragonite polymorph of CaCO3, is shown to possess a small fraction of disordered carbonates whose average chemical-structural identity is similar to that of aragonite. These disordered carbonates were found to interact with bioorganics, bicarbonates, and water molecules and are denoted as interfacial. Characterization of the bleached and of the annealed shells enables the distinguishing of two classes of interfacial carbonates: exposed, solvent accessible, which interact primarily with bioorganics, and buried, solvent inaccessible, which interact exclusively with spatially separated water and bicarbonates. Shell annealing shows that the decomposition of the buried bicarbonate defects correlates with removal of lattice distortions, as detected by XRD, a phenomenon often found in biogenic calcium carbonates. The solid-state NMR investigation exposes the molecular bioorganic−inorganic interfaces in a mollusk shell and demonstrates the unique capability of NMR to determine comprehensively the structure of biogenic composite materials. KEYWORDS: solid-state NMR, biomineralization, biogenic calcium carbonate, mollusk shell, interfaces, bicarbonate



INTRODUCTION Biomineralization is the selective means by which organisms extract and uptake elements from their local environment and incorporate them into biomineralized functional structures under strict biological control.1,2 The evolution of biomineralization resulted in materials with strength and both microscopic and morphologic properties which are not yet attainable synthetically. Moreover, utilizing biochemical environments, organisms are capable of producing biominerals under ambient conditions in aqueous environments under mild conditions. Special characteristics of biomineral structure and properties are achieved via the interaction of inorganic matter with specialized bioorganic molecules.1 Such molecules are often entrapped in the biomineral during the biomineralization process, forming a composite and affecting its resulting properties. The accumulated studies report on a diverse variety of inter- and intracrystalline bioorganics. Yet, direct characterization of the mineral−bioorganic interfaces is still lacking. A detailed, molecular-level understanding of biomineral structure and inorganic−bioorganic interfaces, which control its properties, is essential to produce advanced biomimetic materials and to improve further materials-engineering capabilities. In this article, we aim to devise solid-state NMR techniques to unravel the molecular (chemical/structural) details of such bioorganic− inorganic interfaces and to shed light on their role specifically for the aragonitic shell of the Perna canaliculus mollusk. © 2013 American Chemical Society

Biomineralized structures are based mostly on three major minerals: calcium carbonate, calcium phosphate, and silica. Calcium carbonate has high lattice energy and low solubility and therefore its two most thermodynamically stable polymorphs, calcite and aragonite, are widespread in calcareous biominerals. Most of the shells consist of calcium carbonate mineral that is intimately associated with a complex assemblage of organic macromolecules (the organic matrix). This matrix and its interactions with the mineral are of fundamental importance.1,2 The shell of the bivalves usually consists of an inner aragonitic nacreous layer and an outer prismatic layer. In general, the nacreous layer is composed of continuous lamellae built of polygonal tablets. The lamellae are separated by sheets of interlamellar organic matrices, and each tablet is surrounded by intertabular organic matrices.3−5 The organic matrix within the nacreous layer was reported to consist of three major components: (1) acidic glycoproteins,6 some of which serve as nucleating centers for aragonite,6,7 (2) Ala and Gly-rich silklike fibril proteins8,9 present in a hydrated gel-like state,10,11 and (3) β-chitin, positioned in the interlamellar matrix, which serves as a scaffold.8 In contrast to the nacre’s conserved structure, not all mollusc-shell prismatic layers have the same structure nor the Received: August 21, 2013 Revised: October 10, 2013 Published: October 24, 2013 4595

dx.doi.org/10.1021/cm4028226 | Chem. Mater. 2013, 25, 4595−4602

Chemistry of Materials

Article

purchased from Sigma-Aldrich. All chemicals were used as purchased. A 0.5 M EGTA stock solution was prepared by dissolving 9.50 g of EGTA in 50 mL of H2O. Samples Preparation. P. canaliculus mollusk shells were cleaned by sonication in methanol and double-distilled water (DDW) and then air-dried. The shells were powdered by crushing with a mortar and pestle followed by sieving through a 25 μm sieve; such samples are denoted as BU, biogenic untreated. BU samples that were bleached for 2 weeks by vigorous stirring in 5% sodium hypochlorite are denoted as BB, biogenic bleached. BU samples that were heated at 250 °C overnight are denoted as BA, biogenic annealed. Sample of aragonite from geological origin (Sefrou, Morroco) was used as the control sample and denoted GU. Decalcification. A 50 mg BU sample was immersed in a known volume of 0.5 M EGTA solution (∼2 mL/mg), adjusting the pH to 7 to 8 with NaOH and HCl solutions throughout decalcification (2 days). The dissolved CaCO3 and matter within it created a solution that was subjected to solution 31P NMR characterization. Solid-State NMR. NMR spectroscopy measurements were carried out with as follows. (a) A Chemagnetics/Varian 300 MHz CMXinfinity solid-state NMR spectrometer equipped with three radiofrequency channels and double- and triple-resonance APEX Chemagnetics probes with 7.5 and 5 mm (o.d.) Zirconia rotors. Temperatures were actively regulated to 15 ± 2 °C, and samples were spun at 5000 ± 2 and 6666 ± 2 Hz. (b) A 300 MHz AVANCE III (Bruker) solid-state NMR spectrometer equipped with three radiofrequency channels and triple-resonance probes using 4 mm zirconia rotors, spinning at 10 000 ± 2 and 5000 ± 2 Hz. Cross-polarization (CP) magic-angle spinning (MAS) echo experiments (indirect excitation) were carried out with a 5.0 μs π/2, 10.0 μs π pulse widths, an echo interval τ (200 μs) identical to the rotor period TR, a 1H decoupling level of 100 kHz, and a relaxation delay of 1.5, 2, and 4 s; Hartmann−Hahn rf levels were matched at 50 kHz, with contact times (ct) of 2 and 8 ms for 13C and 1 and 2 ms for 31P. With Bruker spectrometers, ramped CP (70−100 kHz) and SPINAL 1H decoupling were used. Direct 13C excitation echo experiments, DE, were carried out with 5.0 μs π/2, 10.0 μs π pulse widths, an echo interval τ equal to the rotor period TR (200 or 150 μs), a 1H decoupling level of 100 kHz, and a relaxation delay of 15 and 2400 s. (c) A 300 MHz AVANCE III (Bruker) equipped with four radio-frequency channels and tripleresonance probe; 4 mm zirconia rotors were spun at 10 000 ± 2 Hz for 1 H, 13C, 31P and 2D HETCOR NMR experiments. 1H-X HETCOR experiments were run employing wPMLG5 homonuclear decoupling scheme during the 1H chemical-shift evolution period (t1), with 25− 130 increments spaced by 11 μs followed by CP (using the same parameters as described above) with 0.2, 2, and 8 ms contact times and 100 kHz SPINAL 1H decoupling, collecting 1024, and 4096 transients per t1 increment for 1H−13C/31P HETCOR, respectively. Zero filling was 4 k for X channel and 1 k for H channel prior to Fourier transformation. Time-proportional phase incrementation (TPPI)29 was used to obtain purely absorptive 2D NMR spectra with frequency sign discrimination along the indirect dimension. Solution NMR. The spectra were acquired on AV-III 600 Bruker spectrometer at room temperature with resonance frequencies for 1H and 31P of 600.55, 243.11 MHz, respectively, with two-channel, directdetection probes with automatic tuning and matching and equipped with z-gradients. Bruker Topspin 2.1 was used on a PC (windows-XP) for spectral acquisition and processing. 13C and 1H chemical shifts are reported relative to TMS, and 31P, relative to 85% H3PO4.

same polymorph, which could be either aragonite or calcite. The individual prisms may be single crystals or polycrystalline.12 Since the early 1970’s,13 many attempts have been made to study mineral-associated organics and organic/inorganic interfaces in shells. These works employed a variety of characterization methods, including single-crystal diffraction,14 which exposed synthetic calcite interactions with biomineral isolated proteins,15−17 small-angle X-ray scattering (SAXS) on Pinna nobilis prisms,18 a combination of high-resolution and energy-filtering TEM,19 dark-field TEM imaging of nacre,20 annular dark-field scanning transmission electron microscopy (STEM), and electron tomography on calcite prisms in Atrina rigida.21 Moreover, the presence of an amorphous calcium carbonate layer surrounding the crystallites was reported in the aragonite shells of Haliotis laevigata22,23 by HRTEM and solidstate NMR. This work also concluded that there are no interfacial interactions between the protein/polysaccharide matrix and this amorphous layer. The detected disordered aragonite coating of the single-crystalline aragonite platelets was interpreted as clear evidence against an epitaxial match between the organic matrix and the formed aragonite. P. canaliculus mollusk shell shows greatly enhanced fracture toughness, possibly because of crack deflection and arrest at the inorganic−organic interfaces.24 It was previously shown that the lattice of the biogenic aragonite derived from this shell is anisotropically distorted.25,26 Moreover, the nacreous structure that comprises most of this shell (layered structure of ∼200 nm thick aragonitic tablets separated by ∼20 nm of organic matter) resembles a mechanically strained multilayered structure and is elastically bent because of the forces evolving at the organic− inorganic interfaces.26 These works classified the presence of two kinds of organic molecules in biogenic crystals: intracrystalline (entrapped within individual crystallites) and intercrystalline (filling the interlamellar space and so located between crystallites). In addition, these works suggested that the intracrystalline organics are responsible for the structural distinctions between biogenic and geological crystals. In this study, we employ multinuclear solid-state NMR to investigate the bioorganic−inorganic interfaces of the composite aragonitic shell of P. canaliculus at the molecular level. In particular, we qualitatively and quantitatively characterize the chemical composition using carbon, hydrogen, and phosphorus NMR. Furthermore, we elucidate the different structural environments that arise from the shell’s biocomposite nature with its high complexity architecture, determine the short-range structural order within each class of environment (molecular level), and identify the interfacial interactions of the calcium carbonate with bioorganic and water molecules. Finally, the effects imparted by bleaching (removal of solvent accessible organics) and annealing (partial, nonselective destruction/ removal of organic and water content) to the structure and composition of the mineral are investigated while aiming to combine the NMR-derived insights with the previously published25,27,28 X-ray diffraction investigation as a means to understand better the complex structure of the P. canaliculus shell.





RESULTS AND DISCUSSION Carbon Content of the P. canaliculus Shell (13C DE MAS NMR). Solid-state 13C MAS NMR spectroscopy is a wellestablished technique to identify calcium carbonate forms (crystalline polymorphs or amorphous).22,30−32 The carbonate chemical shift identifies the local order, whereas the peak width is as a measure of the uniformity/heterogeneity of the local

EXPERIMENTAL SECTION

Materials. Green mussels were purchased frozen from New Zealand, and their shells were cleaned, crushed, treated (bleached or annealed), and used in a powdered form. X-ray diffraction studies of analogous samples were published separately.25,27,28 Chemicals were 4596

dx.doi.org/10.1021/cm4028226 | Chem. Mater. 2013, 25, 4595−4602

Chemistry of Materials

Article

S1), the peak (position and width) of the geological aragonite, GU, is unchanged (Table 1 and Figure S1, inset). The observed interfacial fraction of BU reports carbonates with a spread of local order (structural heterogeneity) around that of aragonite. The observed difference between the interfacial environments of the geological versus the biogenic aragonite must originate in their pure versus composite nature. In the latter, interfacial interactions with bioorganics and water contribute to the dispersion of the local structures; lattice edges alone, as in the GU, contribute only to the reduced rigidity.35 The characteristics of the biogenic interfacial sites will be further substantiated and refined below. In addition to the fast-relaxing interfacial carbonates, this 13C DE MAS spectrum exposes the presence of bioorganic carbon content (peaks in the 20−60 ppm region, Figure S2), consistent with its much shorter relaxation than that of bulk aragonite. Bioorganic Content and Interfacial Carbonates (13C CP-MAS NMR). 13C CPMAS NMR experiments utilize polarization transfer from hydrogen atoms to the rare 13C nuclei, resulting in the enhancement of hydrogen-rich environments (bioorganics and interfacial aragonite carbonates). The 13 C CPMAS spectrum (Figure 2a, 1 ms contact time) clearly

chemical environments and hence reports crystalline order versus disorder (Figure 1).32,33

Figure 1. Fully relaxed 13C DE MAS NMR spectra of different CaCO3 forms, synthetic and biogenic, that were recorded in our lab.36,37 The 13 C chemical shift is indicative of the specific structural form, crystalline polymorph or amorphous, whereas the heterogeneous peak width (δν, fwhm) represents the degree of crystalline order/ disorder, namely, structural uniformity vs heterogeneity. ACC, amorphous calcium carbonate.

Table 1. Carbonate Chemical Shift and Peak Width Obtained from the Different Excitation Schemes for the P. canaliculus Shell (BU) and for the Reference Geological Aragonite (GU)a sample BU

GU a

experimental NMR technique

chemical shift (ppm, ±0.1)

δν (ppm, ±0.02)

DE (pd =2400 s) DE (pd =15 s) CP (ct = 8 ms; pd=2 s)

171.0 170.8 170.8

0.45 0.85 0.97

DE (pd = 2400 s, 15 s)

171.0

Figure 2. 13C CP MAS spectra of the aragonitic shell of P. canaliculus, BU, with cross-polarization time (contact time) of (a) 1 and (b) 8 ms. The spectra expose the bioorganic content, mainly proteinaceous, and partially resolve the ∼171 ppm peak of the interfacial carbonates. (c) Expansion of the carbonate-peak region of the 13C MAS spectra of the aragonitic shell of P. canaliculus (BU): DE with 2400 s relaxation delay, red; DE with 15 s relaxation delay, blue; and CP with 8 ms contact time and 2 s relaxation delay, black. Schematic models: On the left, highly ordered, bulk crystalline aragonite, as manifested by the DE 2400 s spectrum. On the right, buried and exposed disordered interfacial environments in contact with hydrogen-bearing moieties (e.g., organics, water, and bicarbonates).

0.32 13

Peak widths were obtained from 75.4 MHz C MAS spectra.

The 13C DE MAS spectrum of the untreated shell (Table 1 and Figure S1, BU) shows only the 171.0 ppm (δν ∼0.45 ppm) aragonite carbonate peak, indicating that it is the major carbon component present.34 The peaks of the 0.2−5 wt % bioorganic content25,27,28 are not detectable here; however, they will be selectively exposed using other NMR excitation schemes (vide infra). For comparison, the analogous 13C DE MAS NMR spectrum of the reference aragonite, GU, exhibits a carbonate peak at the same chemical shift as the BU sample (Table 1 and Figure S1, inset, top); however, it has narrower peak width of δν ∼0.28 ppm. Together, the two spectra show that the main carbon component of the P. canaliculus shell is the highly ordered bulk crystalline aragonite, yet its structural uniformity is slightly lower compared to that of geological aragonite. Further examination of the resulting peaks in the 15 s DE MAS spectrum shows that although the shell (BU) exhibits a significantly broader carbonate peak, δν ∼0.85 ppm, at a slightly reduced chemical shift, 170.8 ppm (Table 1 and Figure

exposes the proteinaceous content22,23 of the untreated mollusk shell (BU) of P. canaliculus. The spectra are dominated by the 45 and 53 ppm peaks of Cα of Gly and Ala, consistent with silk fibril proteins content.8,9 Carboxylic acid residues of the acidic proteins6 are also visible around 190−195 ppm. Interestingly, although the occurrence of chitin in bivalave shells was reported, its characteristic peaks8 (∼ 75, 85, and 105 ppm) are not detectable, indicating that if chitin is present, then it is of very low abundance. We note that the bioorganic content that comprises only a minor fraction (≤2 wt %)38 of the shell was undetectable in the quantitative DE MAS spectrum (pd 4597

dx.doi.org/10.1021/cm4028226 | Chem. Mater. 2013, 25, 4595−4602

Chemistry of Materials

Article

Hydrogen Content and H-Bearing Moieties within the Shell (1H MAS NMR, 2D 1H−13C HETCOR). As a first step, we use 1H MAS NMR to identify the hydrogen species present in the shell. The 1H MAS NMR spectrum of the untreated shell (Figure 3, bottom), BU, shows a major 5 ppm peak assigned to

2400s) (Figure S1), whereas CPMAS yields high-quality spectra. Close examination of the 13C CPMAS spectrum in Figure 2a reveals a relatively narrow ∼171 ppm (δν