Article pubs.acs.org/JPCB
Energetic Basis for Inhibition of Calcium Phosphate Biomineralization by Osteopontin Meng Li,† Lijun Wang,*,† and Christine V. Putnis‡,§ †
College of Resources and Environment, Huazhong Agricultural University, Wuhan 430070, China Institut für Mineralogie, University of Münster, 48149 Münster, Germany § Department of Chemistry, Curtin University, Perth, Western Australia 6845, Australia ‡
S Supporting Information *
ABSTRACT: Calcium oxalate kidney stones form attached to Randall’s plaques (RP), calcium phosphate (Ca−P) deposits on the renal papillary surface. Osteopontin (OPN) suppresses crystal growth in the complex process of urinary stone formation, but the inhibitory role of active domains of OPN involved in the initial formation of the RPs attached to epithelial cells has yet to be clarified. Here we demonstrate the thermodynamic basis for how OPN sequences regulate the onset of Ca−P mineral formation on lipid rafts as a model membrane. We first quantify the kinetics of hydroxyapatite (HAP) nucleation on membrane substrates having liquid-condensed (LC) and liquid-expanded (LE) phases using in situ atomic force microscopy (AFM). We find that rates are sequence-dependent, and the thermodynamic barrier to nucleation is reduced by minimizing the interfacial free energy γ. Combined with single-molecule determination of the binding energy (ΔGB) of the OPN peptide segments adsorbed to the HAP (100) face, we show a linear relationship of γ and ΔGB, suggesting that the increase in the nucleation barriers correlates with strong peptide−crystal nuclei binding. These findings reveal fundamental energetic clues for inhibition of membrane-mediated nucleation by sequence motifs and subdomains within the OPN protein through spatial location of charged moieties and provide insight connecting peripheral cell membranes to pathological mineralization.
■
INTRODUCTION Kidney stone formation is a common chronic disease, with a complex crystalline mineral component of calcium salts (oxalate, phosphate, carbonate) and organic compounds (amino acids, proteins, and lipids).1 At the initial stage of stone formation, calcium phosphate (Ca−P) crystals are found within the apical surface of renal papillae and the urinary space,2−4 which efficiently allows the nucleation of calcium oxalate crystals.4 These calcium oxalate crystals develop attached to renal papillary subepithelial deposits of Ca−Ps, called Randall’s plaque (RP).5 The plaques originate inside the renal interstitia associated with the basement membranes of thin loops of Henle.3 Rapid adherence of calcium oxalate crystals to the apical surface of tubular epithelial cells could promote crystal retention in the kidney.6 Therefore, an interaction of crystals with the surface of renal epithelial cells could be a critical initiating event in nephrolithiasis,7,8 especially the initial formation of the RPs and the earliest crystal−cell membrane adhesion.9 Recent results showed that epithelial cells of hyperoxaluric kidneys acquire a number of osteoblastic features but without Ca−P deposition, perhaps as a result of upregulation of osteopontin (OPN).5 Plaque formation may additionally require localized increases in supersaturation with respect to hydroxyapatite (HAP) and brushite and a decrease in the mineralization inhibitory potential of osteogenic proteins. As a © 2017 American Chemical Society
urinary constituent, OPN in normal human urine (>100 nM) has been shown to inhibit the growth and modify the shape of calcium oxalate crystals.10,11 Formation and retention of calcium oxalate crystals at epithelial cell surfaces and in renal tubules are significantly suppressed through the in vivo secretion of OPN12 that exhibits an abundance of sequence domains rich in dicarboxylic acids and a functionally active cellbinding sequence.13 OPN also has been demonstrated to interfere with calcium oxalate attachment to renal epithelial cells.14 Adhesion processes to renal tubule surfaces are affected by specific adsorption of proteins on crystal surfaces because the acidic residues of OPN have been demonstrated to control mineral growth and aggregation.11,15,16 Molecular-scale investigations have further confirmed that the full OPN regulates calcium oxalate surface crystallization kinetics17 and the OPN peptide segments inhibit specific step growth of brushite.18 It has been established that OPN in urine can reduce crystal attachment to renal epithelial cells,19 and that phospholipid phase boundaries play an important role in calcium oxalate precipitation.20 Phase-separated assemblies containing lipid rafts are indeed present in cellular lipid membranes of Madin-Darby canine kidney (MDCK) epithelial cells.21 As a Received: May 2, 2017 Revised: June 5, 2017 Published: June 6, 2017 5968
DOI: 10.1021/acs.jpcb.7b04163 J. Phys. Chem. B 2017, 121, 5968−5976
Article
The Journal of Physical Chemistry B
membranes, individual constituents on mica, and bare mica were measured using amplitude modulated Kelvin probe force microscopy (AM-KPFM) with a lift height of 30 nm revealing the distribution of surface potential in the region of the applied voltage. Growth Solutions for In Situ AFM Nucleation. The relative supersaturation σ for HAP can be expressed by IAP σ = K − 1, where IAP is the actual ionic activity product, Ksp
pertinent model membrane, lipid rafts of membranes with highly dynamic assemblies contain sphingolipids and phospholipids,22−24 and they have been used to investigate mineral precipitation associated with stone formation.25 Although extensive investigations have confirmed the regulatory effect of OPN on nephrolithiasis, quantitative studies of the Ca−P nucleation kinetics in cell membranes have been lacking, and the energetic basis for OPN domains modulating the onset of Ca−P formation is not well established. The present study hypothesizes that the inhibitory activity of individual OPN active domains is different at the same membrane lipid rafts in supersaturated solutions that mimic physiological fluids. To test this hypothesis, we selected four pairs of peptides with highly conserved aspartic acid residues from the full OPN sequence (Figure S1).26 To mimic the Ca− P precipitation environments at cell membrane surfaces and evaluate the inhibitory role of the OPN peptide fragments in Ca−P nucleation, a series of in situ atomic force microscopy (AFM) experiments were performed to quantify the nucleation rates on the lipid membrane surface. Combined with the singlemolecule force spectroscopy determination, we reconciled OPN residue-regulated Ca−P nucleation kinetics with the free energy of OPN peptide fragments bound to the HAP (100) surface, and demonstrated that peptides with high-energy barriers to nucleation correlate with a strong peptide−crystal binding energy. The findings provide a thermodynamic baseline to link the Ca−P nucleation, the cell membrane, and active domains of inhibitory molecules.
sp
is the thermodynamic solubility product (−log(Ksp) = 116.8 for HAP at 25 °C).31 The thermodynamic database and software of SPEC 01 as a speciation model were used for the calculations of the activities and the interaction of the peptide with Ca and PO4 ions in solutions. A series of supersaturated solutions with respect to HAP (σHAP = 13.4−14.2, ionic strength (IS) = 0.15 M) in the absence and presence of 100 nM OPN peptides was prepared by the slow mixing of stock solutions of sodium chloride (NaCl) (1 M), potassium dihydrogen phosphate (KH2PO4) (0.04 M) (Sigma-Aldrich, St. Louis, MO), and calcium chloride (CaCl2) (Fluka, St. Louis, MO). The solution pH was adjusted to 7.4 with 0.01 M sodium hydroxide (NaOH) solution (Sigma-Aldrich, St. Louis, MO). Imaging Nucleation by AFM. All in situ heterogeneous nucleation experiments were conducted by collecting timelapse images of Ca−P nuclei that formed on the membrane substrate using a Nanoscope V- Multimode 8 AFM (Bruker, Santa Barbara, CA) in ScanAsyst mode. A high-precision syringe pump (Razel Scientific Instruments model R100-E) was used to continuously pump the supersaturated solutions into the flow cell at a constant rate of 10 mL/h at 25 °C to ensure steady-state conditions. The number of nuclei increased almost linearly with time, typically within the first 1−2 h of each experiment. Tip Decoration and Single-Molecule Force Spectroscopy Measurements. The Au-coated Si3N4 tips (Bruker, tip model SNL-10) were modified with heterobifunctional crosslinker LC-SPDP.32 Force spectroscopy measurements were performed in a liquid cell filled with freshly prepared phosphate buffering solutions. The inverse optical lever sensitivity (InvOLS) and the spring constant of the functionalized cantilever were calibrated at the beginning and at the end of the experiment to obtain the relative true force−distance curves. Measurements were done with a constant approach velocity of 200 nm s−1 and seven various pulling speeds of 20 nm/s, 200 nm/s, 601 nm/s, 1.04 μm/s, 1.90 μm/s, 2.6 μm/s, and 3.91 μm/s. To properly depict the complicated extension of oligopeptides, a nonlinear least-squares fitting method of the worm-like chain (WLC) model was used to analyze the number of tethers being stretched.33 In addition, the mean works W were plotted as a function of the loading rate. Fitting these data with the analytical approximation methods32,34 gave a single molecule binding free energy for the peptide adsorbed onto the HAP (100) surface.
■
EXPERIMENTAL SECTION OPN Peptide Synthesis. Peptides were synthesized utilizing 9-fluorenylmethoxycarbonyl amino acid according to standard procedures27 for solid phase peptide synthesis from Niusiter Biotech (Wuhan, China). The four peptide segments were detached from the solid support with trifluoroacetic acid and purified by reversed-phase high-performance liquid chromatography. The integrity and molecular weight of these peptides were verified by mass spectrometry. HAP Crystal Synthesis. The micron-sized HAP hexagonal prisms with (100) faces were synthesized according to a molten salt synthesis method and identified by Bruker D8 X-ray diffraction (XRD) (Billerica, MA, USA), as described previously.28,29 Preparation of Membrane Raft Substrates. We dissolved 1-palmitoyl-2-oleoyl-sn-glycero-3-phosphocholine (POPC), brain sphingomyelin (SM), and dihydrocholesterol (DChol) (Sigma-Aldrich, St. Louis, Missouri) into 9:1 chloroform/methanol (Sigma-Aldrich, St. Louis, Missouri) solutions at a molar ratio of 2:1:1.25 After mixing by a vortex shaker (IKA, Vortex Genius 3, Germany), about 100 μL of the mixtures was applied to a freshly cleaved muscovite mica and left in contact for more than 12 h under dry and dark conditions. Raman spectra were collected from 100 to 3600 cm−1 using a LabRAM HR Raman spectrometer operating with an excitation wavelength of 532 nm to identify the standard samples of POPC, SM, Dchol, and their mixtures (Figure S3). Micro-Raman mapping of lipid membranes supported on mica was imaged on the basis of characteristic peaks of three individual constituents from 2790 to 3020 cm−1,30 and the blue to red indicated that the peak intensity was weak and strong (Figure S3A). KPFM for Surface Potential Measurements. Surface potential measurements of mica-supported mixed lipid
■
RESULTS AND DISCUSSION Lipid raft membranes were prepared to contain a 2:1:1 mixture of 1-palmitoyl-2-oleoyl-sn-glycero-3-phosphocholine (POPC), sphingomyelin (SM), and dihydrocholesterol (DChol) (Figure S2)25 based on lipid composition analyses of MDCK epithelial cells and relevant membranes.21 The membrane constituents were identified by Raman spectra (Figure S3), and the supported membranes on mica were used to serve as a substrate for the Ca−P precipitation at lipid interfaces to mimic 5969
DOI: 10.1021/acs.jpcb.7b04163 J. Phys. Chem. B 2017, 121, 5968−5976
Article
The Journal of Physical Chemistry B
Figure 1. Representative time sequence of in situ AFM height images of surface features of membrane lipid rafts on mica recorded before (A) and after (B) addition of a supersaturated HAP solution (σ = 13.6) in the presence of 100 nM OPN peptide 93−106 (scale bar, 800 nm). Relatively dark LE domains and relatively bright LO islands in part A are POPC and SM/DChol, respectively. Green circles show the LO domain morphologies change from irregular margins in air (relative humidity