Role of Sacrificial ProteinâMetal Bond Exchange in Mussel Byssal

Article pubs.acs.org/Biomac

Role of Sacrificial Protein−Metal Bond Exchange in Mussel Byssal Thread Self-Healing Clemens N. Z. Schmitt, Yael Politi, Antje Reinecke, and Matthew J. Harrington* Department of Biomaterials, Max Planck Institute of Colloids and Interfaces, Potsdam 14424, Germany

Downloaded by CENTRAL MICHIGAN UNIV on September 10, 2015 | http://pubs.acs.org Publication Date (Web): August 28, 2015 | doi: 10.1021/acs.biomac.5b00803

S Supporting Information *

ABSTRACT: Marine mussels tether to seashore surfaces with byssal threads, proteinaceous fibers that effectively dissipate energy from crashing waves. Protein−metal coordination bonds have been proposed to contribute to the characteristic mechanical and self-healing properties of byssal threads; however, very little is understood about how these crosslinks function at the molecular level. In the present study, combined Raman and X-ray absorption spectroscopy (XAS) measurements were employed to confirm the presence of protein−Zn2+ coordination bonds in the mussel byssus and to monitor transitions in the coordination structure during thread deformation and self-healing. Results indicate that Zn2+ coordination bonds, primarily mediated via histidine, are ruptured during thread yield and reformed immediately following thread relaxation. Mechanical healing, on the other hand, is correlated with the transition toward shorter coordination bond lengths. Calculation of the healing activation energy suggests that protein−Zn bond exchange provides a primary rate-limiting step during healing.

■

INTRODUCTION Sacrificial bonds (SBs) provide a successful material toughening strategy prevalent in a wide range of biological materials including bone, wood and silks.1−5 Reversible SBs are typically noncovalent interactions that break prior to covalent cross-links in the material and are often combined with the unfolding of folded protein chains, leading to so-called hidden length (i.e., increased extensibility).2 Mussel byssal threads have been proposed to utilize a specialized SB approach employing strong, but reversible metal coordination cross-links to achieve tough and self-healing properties.6,7 The primary goal of the present study is the detailed investigation of the role of metal coordination in mussel byssal thread mechanical performance. Marine mussels (Mytilus sp.) populate intertidal regions of the rocky seashore, where enormous wave forces provide a major selective pressure. The key to mussel dominance in this hazardous habitat is a holdfast anchor known as the byssus that establishes a secure attachment to the underlying substrate (Figure 1A).8 Byssal threads are protein-based fibers that consist of an adhesive interface (the plaque) and a tethering fiber (the thread; Figure 1B). The distal region of the thread dominates the mechanical performance of the thread under high strain9 and exhibits high stiffness, a yield point at ∼15% strain and a notable hysteresis during cyclic loading (Figure 1C).9 Tensile loading of byssal threads following yield results in a loss of up to 70% of the initial stiffness and strain energy in subsequent loading cycles (Figure 1C). However, if threads are allowed to rest for several hours, initial properties are significantly recovered in an intrinsic and autonomic selfhealing process (up to 90% recovery in 1 day; Figure 1C).9,10 Because byssal threads are acellular and composed almost © XXXX American Chemical Society

entirely of protein building blocks, they have emerged as an exciting biological archetype for the design of self-healing polymers.11−13 However, before this emergent behavior can be successfully mimicked, a deeper understanding of the structure−function relationships must be established. Byssal threads are composed almost entirely of protein, and the distal region of the thread consists primarily of a protein family known as the preCols (Figure 1B).14 There are three preCol variants, each with a similar modular structure consisting of a central triple helical collagen domain, variable flanking domains at either end of the collagen that possess βsheet secondary structure in the thread distal region15,16 and terminal histidine-rich domains (HRDs) at both ends of the protein containing at least 20 mol % histidine14,17 (Figure 1B). Notably, histidine has a known propensity for coordinating divalent metal ions such as Zn, Cu, and Ni, which are also present in threads.18 Previous X-ray-based studies revealed that preCols are arranged end-to-end in series in a semicrystalline framework that can deform elastically, recovering the initial nanoarchitecture instantaneously (Figure 1D).19 However, this is at odds with thread mechanical behavior, which is neither elastic nor immediately reversible. During thread deformation, it was found that the collagen domains do not unfold; rather, the large extensibility of the threads arises from the reversible unfolding of the flanking domains and HRDs.10,19 In situ SAXS measurements suggested that during thread self-healing, two processes set in, namely, a fast elastic recovery of structural Received: June 16, 2015 Revised: August 17, 2015

A

DOI: 10.1021/acs.biomac.5b00803 Biomacromolecules XXXX, XXX, XXX−XXX

Article


Biomacromolecules

Figure 1. Structure−function relationships in the mussel byssus. (A) Mussels attach to surfaces with a byssus. (B) Distal region of byssal threads is comprised of triple helical preCol protein organized into 6 + 1 hexagonal bundles and assembled into a semicrystalline nanoarchitecture. His-rich domains (HRDs) are proposed to form coordination bonds with metals such as Zn2+ and Cu2+, but the structure is unknown. (C) Cyclic mechanical loading of the distal byssal thread, highlighting reduced mechanical performance following relaxation and mechanical recovery following healing. (D) Structure and length of semicrystalline preCol framework recovers immediately following unfolding of flanking and His-rich domain; however, stiffness of preCols recovers slowly.

order and a slow recovery of mechanical properties.19 The former was attributed to the rapid refolding of protein structure, while the latter was ascribed to the slower reformation of broken sacrificial bonds.19 It was first proposed by Waite and co-workers14 that metal coordination bonds mediated primarily by histidine residues in the HRDs might function as reversible sacrificial bonds in the byssus. Current support for the presence and mechanical role of His-metal bonds is based on numerous pieces of indirect, but compelling, evidence: (1) Sequences of the His-rich preCol terminal domains from three different marine mussel species are highly conserved.14,17,20 (2) The concentration of transition metal ions known to coordinate with histidine (e.g., Zn2+ and Cu2+) is elevated in byssal threads.18 (3) EDTA removal of metal ions reduces thread stiffness and perturbs healing ability.7 (4) Thread stiffness shows a sigmoidal pH dependence with a halfway point at the pKa of histidine and low pH treatment eliminates self-healing.10,17 (5) In vitro soft colloidal probe (SCP) experiments on byssal thread preCol his-rich domain (HRD) peptides showed a reversible metal-dependent selfinteraction.21 While these data provide support for the mechanical importance of metal coordination in the thread core, there is as of yet no direct evidence for the presence of these cross-links or their role in mechanics. The aims of the current study are 2-fold: (i) the direct identification of metal coordination bonds within the core of M. californianus byssal threads and (ii) the tracking of changes in the metal coordination structure throughout tensile loading and healing of the threads. To achieve these aims, X-ray absorption

spectroscopy (XAS), including analysis of X-ray absorption near-edge structure (XANES) and extended X-ray absorption fine structure (EXAFS) was performed on byssal threads in order to characterize the coordination sphere around Zn ions in the thread at different time points throughout thread stretching and healing. The results of this spectroscopic investigation provide new insights into the complex interplay of metal coordination sacrificial bonds with hierarchical structure in the intricate self-healing mechanism of the byssus, which will aid current efforts to design the next generation of advanced selfhealing metallopolymers.13,22,23

■

EXPERIMENTAL SECTION

Sample Preparation. Mussel byssal threads from Mytilus californianus were harvested from mussels grown in a tank of flowing seawater from the Santa Barbara channel. Threads were washed, and the distal portions of threads were dissected from the rest of the thread and stored in distilled water prior to further experimentation. XAS spectra from byssal threads were compared to inorganic and organic standards containing Zn. These standards were finely powdered samples of ZnO, ZnCl2, insulin, Nereis virens jaws, and mixtures of imidazole and polyhistidine with Zn2+ in the ratio 3:1. All standards except for the worm jaws were acquired from Sigma-Aldrich (St. Louis, MO, U.S.A.). Inductively Coupled Plasma−Optical Emission Spectrometry (ICP-OES). For each ICP-OES measurement, five threads were freeze-dried, weighed, and then dissolved in 500 μL of aqua regia (167 μL of HNO3 + 333 μL of HCl) at 40 °C for 24 h. The resulting solutions were diluted at a 1:10 ratio with Millipore H2O and emission spectra were recorded with an ICP-OES analyzer (Optima 8000, PerkinElmer). Each sample was measured two times. B


Article

Biomacromolecules


Figure 2. Comparison of XANES spectra of byssus at the Zn K-edge (9659 eV) with inorganic and organic standards. (A) Comparison with elemental Zn and Zn minerals with byssal threads. (B) Comparison with organic materials coordinating Zn by histidine or imidazole. dried after being unclamped. “Healed threads” were strained to 25%, brought back to 0% strain, rested in water for 5 days at room temperature after being unclamped, and then dried. An assessment of the effects of drying and (wet) relaxation on the stress in distal byssal threads stretched to 10, 25, and 40% strain is given in Figure S1. XAFS measurements were carried out at the Zn K-edge (9659 eV) on the KMC-2 beamline at the BESSY II synchrotron (BerlinAdlershof, Germany), using a SiGe(111) double-crystal monochromator. The ring current was between 150 and 300 mA during measurements. Beam intensity was stabilized using MOSTAB electronics. The samples were measured in fluorescence geometry using an energy-dispersive detector (Röntec X-Flash). The beam intensity, before (I0) and after (It) the sample, was recorded with ionization chambers and a zinc standard located behind the sample was simultaneously measured in transmission with a Si-PIN photodiode. The beam size on the sample was approximately 0.5 × 0.5 mm2. The XAS data was collected in the range from 9550 to 10400 eV with varying step size (with 0.2 eV maximum energy resolution near the edge). The exposure time was 6 s, and at least six spectra were recorded for each sample and averaged. The XAFS spectra were deglitched, energy calibrated, aligned, averaged, and background corrected with the Athena module of the Demeter XAS software package.25 The edge position of each spectrum was set to the energy at the first maximum of the first derivative. Energy calibration was performed according to the simultaneously acquired spectrum of a Zn foil standard. After conversion to k-space, the usable data range was determined to be from 2 to 9.5 Å−1. EXAFS. The EXAFS data was analyzed using the Artemis module of the Demeter XAS software package.25 Due to the high content of His and Asp in the byssal thread core proteins, the protein database (PDB) was searched for proteins known to coordinate Zn by these two amino acids. Using Artemis, theoretical models were constructed from the crystallographic data in the protein database (PDB) of the Zn sites in several proteins. One zinc site in human collagenase 3 (MMP-13, PDB entry 4fu4), containing 3 His and 1 Asp, was selected for further analysis. The coordinates of all atoms within 6 Å from the Zn atom were extracted from PDB crystallographic data with PyMol,26 using a previously described procedure.27 The program FEFF (within Artemis) was used to calculate the theoretical scattering amplitudes f(k) and the phase-shifts δ(k). The theoretical model was built from the most important scattering paths contributing to the r-range of interest. All fits were performed using Artemis, applying a Hanningshape window function in the r-range of 1.05−2.6 Å and k-range of 2− 9.5 Å−1. Typical measurement uncertainty in R is 0.2 Å. The value of the amplitude factor S20 = 0.926 was determined from the fit of a ZnO standard. The total coordination number of the first shell was fixed between 4 and 5 according to information in the XANES region. The relative number of the first shell N and O atoms (i.e., the degeneracies of the corresponding scattering paths) was fit. The number of second shell C atoms was defined relative to the number of N and O atoms in the first shell based on the assumption that first shell N and O atoms are provided by His and Asp residues, respectively. The fit parameters

X-ray Fluorescence (XRF). XRF measurements were carried out on whole distal threads and microtomed sections of the thread core at the KMC-2 beamline at BESSYII (Berlin-Adlershof, Germany) with an X-ray energy of 11000 eV, using a SiGe(111) double-crystal monochromator. The fluorescence of the samples was detected with an energy-dispersive detector (Röntec X-Flash). In order to determine XRF spectra of just the thread core, the cuticle was removed from some threads using a cryo-microtome (HM560, Microm) and a microlaser dissection device (P.A.L.M. Microlaser Technologies GmbH, Germany, with a pulsed UV-A laser at 355 nm, CryLas GmbH, Germany). Raman Spectroscopy. Raman spectra were obtained with a Confocal Raman Microscope (alpha300, WITec, Germany) equipped with a piezo-scanner (P-500, Physik Instrumente, Karlsruhe, Germany). The laser (λ = 532 nm) was focused on thread samples through a 60× water-immersed objective (Nikon, NA = 1.0). The laser power on the sample was set to 15 mW. The spectra were acquired using a thermoelectrically cooled CCD detector (DU401A-BV, Andor, Belfast, North Ireland) behind a 600 g mm−1 grating spectrograph (UHTS 300, WITec, Ulm, Germany) with a spectral resolution of 3 cm−1. The ScanCtrlSpectroscopyPlus software (Version 1.38, WITec, Ulm, Germany) was used for measurement setup. Raman measurements were made on 10 μm cryo sections of the byssal thread core. At least three measurements were taken from different spots for each setting of the polarizer and analyzer (0°−0°, 90°−90°, 90°−0°, and 0°−90° (polarizer- analyzer)). A polynomial background was subtracted from the spectrum with OPUS software version 7.0 (Bruker Optik GmbH). The calculation of the isotropic spectrum was performed with OriginPro Software version 8.6 (OriginLab Corporation, Northampton, MA, U.S.A.), according to a method described previously.24 The depolarization ratio was determined to be 0.21 by measurements of gelatin. Synthetic peptides based on the sequence of the N-terminal HRD domain from preCol-NG20 (NH2-GHGGGHGGGHGGGHGGSASAAAHAAAG-COOH) were dissolved in distilled water and mixed with ZnCl2 solution in a ratio of 3 His/1 Zn2+ (Genbank accession number AF043944). The solution was evaporated on a glass surface, and Raman measurements were made using a 20× objective (Nikon NA = 0.4) with an acquisition time of 10 s and five accumulations. X-ray Absorption Spectroscopy (XAS). Prior to the XAFS measurements, threads were dried, and for each sample, the ends of five threads were glued side by side between two plastic strips and allowed to dry for 24 h under a fume hood. Before tensile stretching, samples were allowed to rehydrate in distilled water for at least 1 h. Tensile testing was carried out in the hydrated state on a custom-built motorized tensile tester, equipped with a 50 N load cell. The applied strain rate was 5 × 10−4 s−1, and the strain was monitored by video extensiometry. The samples for measurements on strained threads were dried (at least 3 days at room temperature), keeping the strain fixed at 10, 25, and 40%, and were glued to plastic frames to maintain the strain state during the XAFS measurements. “Relaxed threads” were strained to 25%, brought back to 0% strain, and then immediately C


Article

Biomacromolecules


[i.e., shift in energy origin (ΔE0), coordination number (N), bond distance (R), and the Debye−Waller factor (σ2)] were alternately floated or stepped and refined until the best fit was achieved. The stability of the fit was further examined by varying the initial conditions of the fit while monitoring statistical parameters. Throughout the whole fitting procedure, the maximum allowed number of variables was restricted to two-thirds of the number of independent points of the data in the fitting window. All obtained parameters were tested for stability by variation of the initial guess values. After establishing a good fit of the unstretched sample, this fit was applied to the stretched samples allowing small modification by floating fit parameters; if this procedure did not result in a good fit, larger changes were allowed. This was done to test the null hypothesis that the Zn-coordination environment is not changing in the course of tensile deformation. Many alternative models were tested, in particular, one based on the His-Zn2+ metal coordination in the Nereis jaws in which Cl was a first shell ligand;28 however, this fit was found to be unsatisfactory.

Figure 3. EXAFS fitting of native thread Zn K-edge spectra. r-Space (magnitude and real part of the Fourier transform) EXAFS of the XAFS data (blue line) and the fit (red line) of byssal threads in the native state. The inset shows the corresponding k-space data (blue line) and the fit (red line).

■

RESULTS Zn Coordination in the Native State. XRF measurements detected the presence of Zn, Cu, and Fe in the thread distal regions. Of these metals, only Cu and Zn were detected in isolated sections of the thread core, in addition to Cl (Figure S2). Consistent with previous reports,18 ICP-OES measurements detected Cu and Zn in the distal thread at concentrations of 0.37 ± 0.17 and 1.04 ± 0.12 μmol/g dry thread, respectively. Based on the higher concentration, Zn was chosen as the main focus for further spectroscopic investigation with XAS. XANES. Figure 2 shows the XANES region of the Zn K-edge XAS spectrum from byssal threads compared with standard Zncontaining materials, clearly illustrating the similarity of the byssal thread XANES region with other organic materials, which coordinate Zn2+ by histidine or imidazole vis-à-vis inorganic Zn standards. The intensity of the Zn K-edge spectra white line (centered here at E0 = 9662.4 eV) has been shown to be indicative of the coordination number of Zn, with values of 1.5 or greater associated with octahedral coordination and values less than 1.5 associated with Zn-coordination by 4 or 5 ligands.29 The white line intensity of the byssal thread spectrum was about 1.4, whereas the intensities of the other measured samples were 1.3 or below, including insulin, which is known to exhibit tetrahedral Zn2+ coordination.28 Thus, Zn2+ ions in the byssal threads likely possess a coordination number of 4 or 5 atoms on average in the first coordination shell. The byssal thread spectrum is also less structured than the other spectra in which two well-defined peaks can be seen, suggesting a less organized coordination sphere. EXAFS. Based on the XANES white line intensity and the most probable Zn-binding amino acids found in the byssal thread distal core, EXAFS models were fitted with both 4 (Table S1) and 5 (Figure 3, Table 1) first shell N/O ligands, corresponding to a mixture of His and Asp amino acid ligands. Models with 4 first shell ligands were not preferred, as they necessitated a change in the amplitude reduction factor S20 obtained from fitting of the ZnO standard. In contrast, satisfactory fitting models with 5 first shell N/O ligands were achieved while restricting the value of S20 to the one obtained from the standard. After screening numerous models based on crystallographic data of Zn binding sites with 5 ligands from the PDB, the most reasonable fit was obtained with a model based on the collagenase, MMP-13 (PDB entry 4fu4). This model possessed on average 2.4 N and 2.6 O atoms in the first coordination shell and 6.1 C atoms in the second shell at

Table 1. EXAFS Fitting Results for Zn K-Edge Spectra of M. californianus Byssal Threadsa ligand First Shell N/O long N/O short Second Shell C long C short

R (Å)

N

σ2

2.03 ± 0.01 1.94 ±

Role of Sacrificial ProteinâMetal Bond Exchange in Mussel Byssal

Recommend Documents

Role of Sacrificial ProteinâMetal Bond Exchange in Mussel Byssal