Article pubs.acs.org/Biomac

EF-Hand Mimicking Calcium Binding Polymer Hee Jung Chung,† Du Young Ko,† Hyo Jung Moon, and Byeongmoon Jeong* Department of Chemistry and Nanoscience, Ewha Womans University, 52 Ewhayeodae-gil, Seodaemun-gu, Seoul 120-750, Korea S Supporting Information *

ABSTRACT: There are four EF-hand polypeptides in calmodulin, a natural ubiquitous calcium binding protein that activates the enzymes involved in Ca2+-mediated signal transduction. An EF-hand polypeptide has six carboxylate functional groups in the middle loop region between two rigid polypeptides. In this study, a calcium binding polymer (CBP) with a structure of poly(Lalanine)-poly(L-alanine-co-L-glutamic acid)-poly(ethylene glycol)-poly(L-alanine-co-L-glutamic acid)-poly(L-alanine) (PA-PAE-PEG-PAE-PA; A 11.1A3.4E3.2-EG40.1-A3.4E3.2-A11.1) was synthesized by mimicking the EF-hand polypeptide. The 6−7 carboxylate functional groups from PAE are expected to form a binding site for Ca2+. As the Ca2+ bound to CBP, small changes in the circular dichroism spectra and 13C NMR spectra were observed, indicating that Ca2+ binding to CBP induced changes in the conformation of CBP. The binding constant of CBP to Ca2+ was investigated by using the competitive binding of 2,2′,2″,2‴-{ethane-1,2-diylbis[oxy(4-bromo-2,1-phenylene)nitrilo]} tetraacetic acid (5,5-Br2−BAPTA). The binding constant obtained with a CaLigator program by least-squares fitting of the absorbance profile as a function of Ca2+ concentration was 5.1 × 105 M−1, which was similar to that of calmodulin. The selectivity of CBP for metal ion binding was compared among Ca2+, Cu2+, and Zn2+. The binding constant was obtained through a similar competitive binding study with murexide. The binding constants for Ca2+, Cu2+, and Zn2+ were 7.0 × 105, 4.2 × 105, and 1.7 × 105 M−1, respectively, indicating 2−4-fold higher selectivity of CBP for Ca2+ compared to Cu2+ and Zn2+. The CBP has selectivity for Ca2+, and binding affinity for Ca2+ was similar to the biological Ca2+ binding motif of calmodulin.

■

INTRODUCTION Polypeptides are unique in that they can be designed as hydrophobic or hydrophilic, charged, or neutral by selecting appropriate amino acids. In addition, their secondary or higher hierarchical structures such as α-helix, β-sheet, random coil, ribbons, rod, and vesicles can be prepared through their selfassembling properties.1−5 The ring-opening polymerization of N-carboxy anhydrides of amino acids by using specific catalysts or an amino group containing initiators leads to the development of polypeptides with various compositions and structures in a simple way.6−9 In particular, metal binding polypeptides have been intensively investigated to understand the role of metal ions in a physiological system and to develop as biomimicking materials.10 For example, metal ions are involved in the β-amyloid formation in the early stage of Alzheimer’s disease.11 Metal binding ligands such as nitrilotriacetic acid, imidodiacetic acid, bipyridine, tripyridine, crown ether, histidine, catecholate, and so on have been introduced to polypeptide.10,12,13 The binding of metal ions induced conformational changes of the polypeptide and triggered the self-assembling process of the polypeptides. Calcium ions (Ca2+) play a significant role in biological functions such as signal transduction, bone formation, blood clotting, and platelet aggregation.14−16 Extracellular Ca2+ concentration is in the range of 1−3 mM, whereas intracellular Ca2+ concentration is 50−200 nM. Therefore, the Ca2+ concentration is about 10000 times lower in the cytosol than in the extracellular system.17,18 Camodulin (CaM) is expressed in all eukaryotic cells and acts as a calcium binding protein. © XXXX American Chemical Society

CaM has four EF-hand motifs. In each of the EF-hand polypeptides, two rigid polypeptides are connected to 12 residues that form a binding loop for Ca2+. Six carboxylate functional groups from Glu and Asp residues in positions 1, 3, 5, 7, 9, and 12 of the binding loop are involved in the Ca2+ binding. In particular, the residue at position 12 acts as a bidentate ligand, providing two oxygen atoms, and the six carboxylate functional groups form a pentagonal bipyramidal symmetry around the Ca2+.19 The binding constants of CaM for Ca2+ calculated by the NMR spectroscopy are reported to be 2.4 × 105, 1.1 × 106, 3.8 × 104, and 1.2 × 105 M−1.20 Therefore, the binding constant of the EF-hand polypeptide for Ca2+ is approximated to be 104−106 M−1. Binding of Ca2+ to the EF-hand polypeptide leads to a conformational change in the CaM. Recently, the change in conformation of CaM was visualized by forming a complex with fluorescent polythiophene to CaM.21 When the Ca2+ binding loop of the EF hand fragment, DADGNGTIDFPE, was connected to the elastin-like polypeptide, a Ca2+- and temperature-sensitive polymer was prepared.22 The polymer could be engineered to undergo unimer-to-micelle/nanoparticle formation by adding Ca2+. Most calcium binding polypeptides reported so far are based on the helix−loop−helix conformation of CaM.23−25 Replacement of amino acids at the calcium binding motif by varying hydrophobicity, charges, and number of acid residues have been Received: December 16, 2015 Revised: February 19, 2016

A

DOI: 10.1021/acs.biomac.5b01694 Biomacromolecules XXXX, XXX, XXX−XXX

Article

Biomacromolecules

°C under a dry nitrogen atmosphere, N-carboxy anhydrides of Lalanine (3.15 g, 27.37 mmol) and N-carboxy anhydrides of γ-benzyl Lglutamate (7.79 g, 29.59 mmol) were simultaneously added with anhydrous chloroform/N,N-dimethylformamide (29/1 v/v) cosolvent (100 mL). Copolymerization of the monomers was carried out for 24 h at 40 °C. The reaction mixtures were precipitated into diethyl ether. Then, the polymer was further purified by fractional precipitation by using methylene chloride/diethyl ether, and the residual solvent was eliminated under vacuum. The final yield of resulting poly(L-alanineco-γ-benzyl L-glutamate)-PEG-poly(L-alanine-co-γ-benzyl L-glutamate) (PAEz-PEG-PAEz) was 75%. 1 H NMR (CF3COOD, ppm): 1.45−1.75 (-NH-CO-CH(CH3)-), 2.00−2.58 (-CH2CH2COOCH2C6H5), 2.58−2.99 (-CH2CH2COOCH2C6H5), 3.60−4.45 (-CH2CH2O-), 4.54−5.00 (-NH-CO-CH(R)-), 5.10−5.50 (-CH2CH2COOCH2C6H5), 7.38− 7.58 (-CH2CH2COOCH2C6H5). N-Carboxy anhydrides of L-alanine (6.3 g, 54.74 mmol) were added to the PAEz-PEG-PAEz, and polymerization of the N-carboxy anhydrides of L-alanine was carried out for 24 h at 40 °C to prepare poly(L-alanine)-poly(L-alanine-co-γ-benzyl L-glutamate)-PEG-poly(Lalanine-co-γ-benzyl L-glutamate)-poly(L-alanine) (PA-PAEz-PEGPAEz-PA). The pentablock copolymer was purified by precipitation into diethyl ether, followed by evaporation of the residual solvent under vacuum. Then, the pentablock copolymer was dialyzed in water using a membrane with a molecular weight cutoff of 2000 Da, followed by freeze-drying of the polymer. Deprotection of the benzyl groups in PA-PAEz-PEG-PAEz-PA was carried out by using HBr/acetic acid solutions (33 wt %).30 The resulting PA-PAE-PEG-PAE-PA was dialyzed in water using a membrane with a molecular weight cutoff of 2000 Da, followed by freeze-drying. The final yield was 78%. 1 H NMR (CF3COOD, ppm): 1.45−1.90 (-NH−CO−CH(CH3)), terminal alanine (1.75−1.90 ppm), 2.00−2.54 (-CH2CH2COOH), 2.58−2.84 (-CH2CH2COOH), 3.60−4.45 (-CH2CH2O-), 4.54−5.00 (-NH-CO-CH(R)-). NMR Spectroscopy. The progress of reactions was confirmed by 1 H NMR spectroscopy (500 MHz NMR spectrometer; Varian, U.S.A.). CF3COOD was used as a solvent. In addition, changes in the 13C NMR spectra of the CBP (1.0 wt % in D2O) were investigated as a function of Ca2+ concentration. The molar ratio of Ca 2+ to polymer was varied between 0.0, 0.5, 1.0, 2.0, 3.0, and 4.0 at 15 °C. Gel Permeation Chromatography (GPC). The GPC system consisting of a pump (SP930D; Younglin, Korea) and a refractive index detector (RI750F; Younglin, Korea) was used to determine the molecular weight and molecular weight distribution of polymers. N,NDimethylformamide was used as an eluent and OHpak SB-803QH column (Shodex, Japan) was used for analysis. PEGs (Polysciences, Inc., U.S.A.) with molecular weights in a range of 200−20000 Da were used as molecular weight standards. Circular Dichroism (CD) Spectroscopy. CD spectra of the CBP aqueous solution (0.01 wt %) were measured by a CD instrument (J810, JASCO, Japan) as a function of Ca2+ concentration to study the change in the secondary structure of the polypeptide as a function of the metal ion concentration. The molar ratio of Ca 2+ to polymer was varied between 0.0, 0.5, 1.0, 2.0, 3.0, and 4.0 at 15 °C. Dynamic Light Scattering. The apparent sizes of the polymer in water (0.01 wt %, pH 7.5) were studied by a dynamic light scattering instrument (ALV 5000−60x0) as a function of Ca2+ concentration in a range of 0.0−4.0 equiv of CBP. A YAG DPSS-200 laser (Langen, Germany) operating at 532 nm was used as a light source. Measurements of the scattered light were made at an angle of 90° to the incident beam. The results of dynamic light scattering were analyzed by the regularized CONTIN method. The decay rate distributions were transformed to an apparent diffusion coefficient. From the diffusion coefficient, the apparent hydrodynamic size of a polymer can be obtained by the Stokes−Einstein equation. Dye Solubilzation. 1,6-Diphenyl-1,3,5-hexatriene solution in methanol (10 μL at 0.4 mM) was injected into a CBP aqueous solution (1.0 mL) to make the dye concentration of 4.0 μM. The UV− vis spectra (S-3100, SCINCO) were compared as a function of polymer concentration in a range of 0.0001−0.01 wt %.

investigated to understand the binding site information, specific affinity for Ca2+, and molecular mechanism of the binding.23−25 The EF-hand mimicking helix−loop−helix polypeptides were also investigated for binding of lanthanide.26 The polypeptide exhibited 10−6 M range of binding affinity (dissociation constant) for both Eu(III) and La(III). Interestingly, the lanthanide binding increased α-helicity of the polypeptide by 15−18%. In this study, we designed a polypeptide with β-sheet-loop-βsheet structure as a calcium binding polymer (CBP) with a structure of poly(L-alanine)-poly(L-alanine-co-L-glutamic acid)poly(ethylene glycol)-poly(L-alanine-co-L-glutamic acid)-poly(Lalanine) (PA-PAE-PEG-PAE-PA). Two rigid PA blocks were connected by 6−7 carboxylate functional groups from PAE to form a binding loop for Ca2+. A PEG was also introduced between two PA−PAE polypeptide blocks to provide flexibility around the Ca2+ binding site (Scheme 1). The PA with low Scheme 1. Schematic Presentation of the Binding of CBP (PA-PAE-PEG-PAE-PA) to Ca2+a

a Red lines, green curves, and blue curves indicate poly(L-alanine), Lpoly(L-alanine-co-L-glutamic acid), and PEG, respectively. Blue circles indicate the carboxylate functional groups of the glutamate residues and filled purple circles indicate the calcium ions.

molecular weight was reported to form β-sheet.27,28 By using the β-sheet forming sequences, such EF-hand mimicking analog was designed. The binding constant of CBP for Ca2+ was compared with CaM and poly(acrylic acid) by using a chelating chromophore that undergoes a change in absorbance by binding to metal ions. In addition, the selectivity of CBP for metal ion binding was compared among Ca2+, Cu2+, and Zn2+.

■

EXPERIMENTAL SECTION

Materials. γ-Benzyl L-glutamate (Fisher Science, U.S.A.), α,ωdiamino-poly(ethylene glycol) (PEG; Mn = 2000 Da; Pharmicell, Korea), N-carboxy anhydrides of L-alanine (Onsolution, Korea), and triphosgene (Sigma-Aldrich, U.S.A.) were used as received. Copper(II) chloride, zinc(II) chloride, and calcium chloride were used as received from Sigma-Aldrich, U.S.A.. Anhydrous ethyl acetate (Fisher Science, U.S.A.), α-pinene (Fisher Science, U.S.A.), and anhydrous N,Ndimethylformamide (Sigma-Aldrich, U.S.A.) were used as received. Calmodulin-1 (CaM; Phoenix Pharmaceuticals Inc., U.S.A.) and poly(acrylic acid) (PAA; Sigma-Aldrich, U.S.A.) were used as received. Murexide (Sigma-Aldich, U.S.A.) and 2,2′,2″,2‴-{ethane-1,2-diyl bis[oxy(4-bromo-2,1-phenylene)nitrilo]}tetraacetic acid (5,5-Br2− BAPTA; Alfa Aesar, England) were used as received. Chloroform (Daejung, Korea) was treated with anhydrous magnesium sulfate before use. Toluene (Daejung, Korea) was dried over sodium before use. Synthesis of PA-PAE-PEG-PAE-PA. First, N-carboxy anhydrides of γ-benzyl L-glutamate were synthesized by the reaction between γbenzyl L-glutamate and triphosgene (yield = 88%).29 α,ω-Diaminopoly(ethylene glycol) (PEG; M.W. = 2000 Da, 4.78 g, 2.39 mmol) was dissolved in dried toluene (50 mL), and the residual water in α,ωdiamino-PEG was removed by azeotropic distillation of the polymer solution to a final volume of 10 mL. After the system was cooled to 40 B

DOI: 10.1021/acs.biomac.5b01694 Biomacromolecules XXXX, XXX, XXX−XXX

Article

Biomacromolecules

Figure 1. Synthetic scheme of the pentablock copolymer of PA-PEA-PEG-PEA-PA (CBP). Determination of Chelator Concentration. First, 50 mM (4-(2hydroxyethyl)-1-piperazine ethanesulfonic acid (HEPES) buffer (pH 7.5) with 100 mM KCl was incubated with a dialysis bag filled with Chelex-100 resin (Sigma-Aldrich, U.S.A.) for 48 h at 15 °C to reduce free Ca2+. The exact concentration of 5,5-Br2-BAPTA was determined by withdrawing a 1.0 mL solution, adding 2 μL of 1.0 M Ca2+ aqueous solution, and recording the absorbance at λmax, 239.5 nm. The chelator concentration was calculated as A239.5/ε. The λmax and ε (1.4 × 104 M/ cm) were used from ref 31. Determination of Free Calcium Ion Concentration. The initial concentration of the Ca2+ (CaQ) was calculated as CaQ = CQ(A2 − A1)/(A2 − A3). CQ is the exact concentration of 5,5-Br2-BAPTA, A1 is the absorbance at 263 nm for 1.0 mL of the chelator solution, A2 is the absorbance at 263 nm for the chelator solution after adding 2 μL of 0.1 M ethylene diamnine tetraacetic acid (EDTA), and A3 is the absorbance at 263 nm for the chelator solution after adding 2 μL of 0.1 M Ca2+ aqueous solution. Binding Study for Calcium Ion. Calcium ion binding experiments were performed according to the published protocol.31−33 Aqueous solutions containing CBP or 5,5-Br2-BAPTA were prepared in decalcified buffer at a final concentration of 20 μM. Aliquots of 2 μL of 3.0 mM CaCl2 solution in the HEPES buffer were added to the aqueous solution (1.0 mL) containing CBP (20 μM) and 5,5-Br2BAPTA (20 μM). UV−vis spectra of 5,5-Br2-BAPTA were obtained while pipetting the external Ca2+ solution with 2 μL/step. The Ca2+ concentration was varied between 0.0, 6.0, 12.0 17.9, 23.8, 29.7, 35.6, 41.4, 47.2, 53.0, 58.8, 64.6, 70.3, 76.0, 81.7, 87.4, 93.0, 98.6, 104.2, 109.8, and 115.4 μM. Absorbance at 263 nm was fitted by the CaLigator program (downloaded from http://www.cmps.lu.se/ biostruct/people/ingemar-andre/caligator) to obtain the binding constant of CBP for Ca2+. The binding constants of CaM and PAA for Ca2+ were also analyzed with the same protocol. Selectivity among Metal Ions. To the murexide aqueous solution (50 μM), Cu2+, Zn2+, or Ca2+ were added as chloride salts for final concentration ranges of 0−400, 0−4000, and 0−60000 μM, respectively, and the UV−vis spectra of the murexide were recorded in free and metal ion bound states. To determine the binding constant between metal ion and CBP, the polymer aqueous solution was added to the murexide aqueous solution containing Cu2+, Zn2+, or Ca2+, and

murexide concentrations in free and bound states were calculated from the UV−vis spectra.34−37

■

RESULTS AND DISCUSSION CBP was synthesized by sequential ring-opening polymerization of N-carboxy anhydrides on α,ω-diamino-PEG, followed by deprotection of the γ-benzyl group from the polymer. The copolymerization of N-carboxy anhydrides of alanine and γbenzyl L-glutamic acid in the presence of α,ω-diamino-PEG led to the formation of the PAEz-PEG-PAEz triblock copolymer, followed by polymerization of the N-carboxy anhydrides of Lalanine, and resulting in the PA-PAEz-PEG-PAEz-PA pentablock copolymer. Elimination of the protecting group by using HBr/CH3COOH yielded the EF-hand mimicking PA-PAEPEG-PAE-PA pentablock copolymer (CBP; Figure 1). The PA block acts as a rigid block for the polymer and the carboxylic acid groups of the PAE block are supposed to bind to metal ions. 1 H NMR spectra of polymers (in CF3COOD) were monitored at each step (Figure 2). PEG has a large peak of (-CH2CH2O-) at 3.60−4.45 ppm and a small methylene peak of (-OCH2CH2-NH2) next to amine group at 3.40−3.60 ppm.38 In the PAEz-PEG-PAEz 1H NMR spectra, the phenyl peak (C6H5-) at 7.38−7.58 ppm and the methylene peak (-CH2-) at 5.10−5.50 ppm comes from the γ-benzyl L-glutamic acid moiety, while the methyl (CH3-) peak at 1.45−1.75 ppm comes from the alanine moiety of PAEz. The molecular weight of each block or the composition of the PAEz-PEG-PAEz was calculated by the phenyl peak (C6H5-) at 7.38−7.58 ppm, the PEG peak (-CH2CH2O-) at 3.60−4.45 ppm, and the methyl (CH3-) peak at 1.45−1.75 ppm using the following equation. x is assumed to be 44.1 due to the molecular weight of 2000 Da for PEG. y and z are the number of repeating units of γ-benzyl L-glutamic acid and alanine, respectively, in the PAEz block of C

DOI: 10.1021/acs.biomac.5b01694 Biomacromolecules XXXX, XXX, XXX−XXX

Article

Biomacromolecules

NMR spectroscopy at 15 °C. The molar ratio of Ca2+ to CBP was varied between 0.0, 0.5, 1.0, 2.0, 3.0, and 4.0. The CD spectra of the CBP aqueous solution (0.01 wt %) exhibited a typical β-sheet shape with the peak minimum at 216−218 nm.42−46 At the concentration (0.01 wt %), the CBP did not assemble to form hydrophobic domain such as micelles, and the secondary structure of polypeptide is distinguished (Figure S2). The spectra exhibited a blue-shift and the magnitude of the mean residue ellipticity decreased slightly as the Ca2+ concentration increased (Figure 3a). Only a small decrease

Figure 2. 1H NMR spectra (CF3COOD) of PEG, PAEz-PEG-PAEz, and PA-PAE-PEG-PAE-PA (CBP). Based on the 1H NMR spectra, x, y, z, and z′ are calculated to be 44.1, 3.2, 3.4, and 11.1, respectively. a′ at 1.75−1.90 ppm comes from the methyl group of terminal alanine of PA. As the reaction proceeds from PEG to PAEz-PEG-PAEz and PAPAEz-PEG-PAEz-PA, the new peaks corresponding PAEz and PA appear, whereas the peaks of protecting γ-benzyl group (Z) of PAEzPEG-PAEz at 5.10−5.50 and 7.38−7.58 ppm disappear by the deprotection reaction.

the PAEz-PEG-PAEz triblock copolymer, as shown in the scheme (Figure 1). A 7.38 − 7.58 /A3.60 − 4.45 = 10y/(4x + 2)

(1)

A1.45 − 1.75 /A3.60 − 4.45 = 6z /(4x + 2)

(2)

An increase in the intensity of the methyl peak at 1.45−1.75 ppm and the methine (-CH-) peak at 4.54−4.85 ppm, and appearance of the new methyl peak of terminal alanine at 1.75− 1.90 ppm in the1H NMR spectra of PA-PAE-PEG-PAE-PA relates to the formation of an additional PA block. During the deprotection step, the peaks at 7.38−7.58 and 5.10−5.50 ppm, corresponding to the γ-benzyl group, disappeared. The coupling constants of β- and γ-methylene protons of the γbenzyl L-glutamic acid moiety of PAEz decrease as the deprotection reaction of γ-benzyl group proceeds.39,40 Therefore, the six peaks at 2.0−3.0 ppm in the 1H NMR spectra of the PAEz-PEG-PAEz appeared as three peaks in the 1H NMR spectra of the PA-PAE-PEG-PAE-PA. The block length of PA was calculated with the following equation. z′ is the number of repeating units of alanine in PA-PAE-PEG-PAE-PA, as shown in the scheme (Figure 1). A4.54 − 5.00 /A3.60 − 4.45 = 2(y + z + z′)/(4x + 2)

Figure 3. (a) Circular dichroism spectra of CBP aqueous solution (0.01 wt % in H2O) as a function of Ca2+ concentration at 15 °C. (b) 13 C NMR spectra of CBP (1.0 wt % in D2O) as a function of Ca2+ concentration at 15 °C. The legend indicates the molar ratio of Ca 2+ to polymer ranging from 0.0, 0.5, 1.0, 2.0, and 3.0 to 4.0.

ellipticity in CD spectra was observed as the Ca2+ added up to 4.0 equiv. This fact suggests that the β-sheet structure is basically preserved as the main secondary structure of the polypeptide. The maximum change in CD spectra was observed at 3.0 equiv of Ca2+, which might be related to the charge neutralization between divalent Ca2+ and six carboxylates of CBP at the 3.0 equiv of Ca2+. 13C NMR spectroscopy (in 1.0 wt % of CBP in D2O) showed that the PEG peak shifted downfield by an amount equivalent to the Ca2+ added to the CBP aqueous solution. The peak then shifted slightly upfield when an excess amount of Ca2+ was added to the CBP aqueous solution (Figure 3b). The change in chemical shift in the 13C NMR as a function of Ca2+ concentration is open for discussion. The downfield shift of the peak at 72.4−72.6 ppm in the 13C NMR spectra indicates a decrease in electron density of the carbons of PEG. Binding the Ca2+ to CBP induced partial perturbation in β-sheet structure of the polypeptide as shown in the CD spectra. Ca2+ directly bound to PAE block through the Coulomb (ionic) interactions, at the same time, it might also affect PEG conformation through the ion-dipole interactions, salt-out or salt-in effect, and so on. The ethylene glycol unit of PEG in an anticonformation can bind two water molecules at

(3)

1

Based on the H NMR spectra at 4.54−5.00 (methine peak), 3.60−4.45 (PEG peak), and 1.45−1.90 ppm (methyl peak), the number of repeating units in each block of CBP were calculated. In the final structure in Figure 1, x, y, z, and z′ are 44.1, 3.2, 3.4, and 11.1, respectively. Therefore, the final molecular weight of the CBP is 4885 Da. The small peak at 1.75−1.90 ppm comes from the alanine end group of PA.41 In addition, gel permeation chromatography also confirmed the progress of the reaction (Figure S1). Retention time decreased from 8.0 to 7.6 and 6.4 min for PEG, PAEz-PEG-PAEz triblock copolymer, and PA-PAE-PEG-PAE-PA pentablock copolymer (CBP), respectively. The molecular weight (Mn) and molecular weight distribution of CBP determined by gel permeation chromatography against PEG standards are 4800 and 1.2, respectively. Changes in the conformation of CBP resulting from Ca2+ binding were investigated using CD spectroscopy and 13C D

DOI: 10.1021/acs.biomac.5b01694 Biomacromolecules XXXX, XXX, XXX−XXX

Article

Biomacromolecules room temperature.47,48 The carbons of ethylene glycol might experience electron deficiency in the anticonformation more than gauche conformation. The conformational change of PEG was also reported when temperature increased. The number of bound water molecules decreased and the content of gauche conformation of PEG increased, as the temperature increased.47−50 Ca2+ might affect the PEG conformation in a complex way and led to the change in chemical shift in the 13C NMR spectra of CBP. As for Cu2+ and Zn2+, the change in chemical shift was not as large as that of Ca2+ (Figure S3). This fact might reflect the higher binding affinity of CBP for Ca2+ than Cu2+ and Zn2+. The PEG chain tends to surround the hydrophobic PA to minimize the exposure of the hydrophobic PA to water. To estimate a conformation of CBP, contour length (lc) and rootmean square end-to-end (le) of hydrophilic PEG and hydrophobic PA were calculated. lc and le are defined by ln and l√n, respectively, for a linear polymer, where l is the length of a repeating unit and n is the number of repeating units of a polymer.51 Assuming that l = 0.36 nm and n = 44.1 for PEG, lc and le of PEG are 15.9 and 2.4 nm, respectively. The average axial length of an amino acid is 0.35 and 0.15 nm for polypeptides with β-sheet and α-helix conformations, respectively.52 Therefore, lc of PA with n = 11.1 is about 3.9 nm, which is the same length with the le of PA with a fully extended β-sheet conformation. Assuming that the two PA blocks form a fully extended β-sheet, the required contour length of PEG to surround the PA as a circle is 2πr = 2π(1.95 nm) = 12.2 nm. The analysis suggests that the two hydrophobic PA blocks, if they form a fully extended β-sheet, can be barely surrounded by one-to-two rounds in a three dimensional space by the hydrophilic PEG. Assuming that PA forms a α-helix, the axial length of PA is 0.5 nm. Then, the required contour length to surround the PA as a circle is 2πr = 2π(0.25 nm) = 1.57 nm. Therefore, the PEG with contour length of 15.9 nm can surround the PA blocks by 10 rounds. Apparent size of the CBP was investigated by dynamic light scattering of the CBP aqueous solution (0.01 wt %) as a function of Ca 2+ concentration. The apparent size of the CBP was less than 3.0 nm in the absence of externally added Ca2+. The size of CBP is larger than the axial length of α-helical PA (0.5 nm), whereas it is smaller than the axial length of β-sheet PA (3.9 nm). This fact suggests that the PA forms mainly β-sheet, as exhibited in the CD spectra, with some incorporation of other structures such as α-helix or random coil. As the concentration of Ca2+ increased from 0.0 to 4.0 equiv of CBP, the apparent size of CBP did not significantly change (Figure S4). The light scattering data also suggests that intermolecular binding is not significant under the current experimental conditions for binding studies of the CBP. If the intermolecular binding of CBP to Ca2+ occurred, the apparent size of CBP and the distribution of sizes should become much larger and broader when the Ca2+ added, compared with the CBP in the absence of externally added Ca2+. Binding of Ca2+ to the CBP was compared with calmodulin (CaM) and poly(acrylic acid). 5,5-Br2-BAPTA was used as a chelating chromophore in the Ca2+ binding assay by using UV− vis spectroscopy.31−33 5,5-Br2-BAPTA has four carboxylate functional groups and binds to two Ca2+ with a binding constant of about 6 × 105 M−1.32 The difference in the UV−vis spectra of the chromophore (5,5-Br2-BAPTA) between the free 5,5-Br2-BAPTA and Ca2+-bound 5,5-Br2-BAPTA were recorded in the presence of CBP during the stepwise addition of Ca2+.

Competitive binding of the CBP to Ca2+ lowers the concentration of free Ca2+ available for 5,5-Br2-BAPTA binding. The normalized absorbance values at 263 nm were fitted with a computer program (CaLigator) to obtain the binding constant of the CBP to Ca2+.33 The method is applied mostly to systems in which a significant population of free and bound Ca2+ is present. That is, a binding constant less than 106 M−1 is recommended for this method.31−33 The UV−vis spectra of 5,5-Br2-BAPTA (20 μM) dissolved in 50 mM HEPES buffer (pH = 7.5) containing 100 mM KCl was monitored in the presence of CBP (20 μM) as a function of Ca2+ concentration (Figure 4a). Ca2+ concentration was varied between 0.0, 6.0,

Figure 4. (a) UV−vis spectra of 5,5-Br2-BAPTA containing CBP as a function of Ca2+ concentration. Ca2+ concentration was varied between 0.0 (red), 6.0, 12.0 17.9, 23.8, 29.7, 35.6, 41.4, 47.2, 53.0, 58.8, 64.6, 70.3, 76.0, 81.7, 87.4, 93.0, 98.6, 104.2, 109.8, and 115.4 μM (black) at a fixed concentration of CBP (20.0 μM) and 5,5-Br2BAPTA (20.0 μM). (b) Normalized absorbance at 263 nm defined by (A − Amin)/(Amax − Amin). Amax and Amin correspond to the absorbance values at 263 nm in the absence of externally added Ca2+ and after saturation with Ca2+, respectively. Normalized absorbance at 263 nm for 5,5-Br2-BAPTA alone, 5,5-Br2-BAPTA in the presence of PAA (BAPTA/PAA), 5,5-Br2-BAPTA in the presence of CBP (BAPTA/ CBP), and 5,5-Br2-BAPTA in the presence of CaM (BAPTA/CaM) are compared. The concentration of 5,5-Br2-BAPTA was fixed at 20.0 μM.

12.0 17.9, 23.8, 29.7, 35.6, 41.4, 47.2, 53.0, 58.8, 64.6, 70.3, 76.0, 81.7, 87.4, 93.0, 98.6, 104.2, 109.8, and 115.4 μM. As the Ca2+ concentration increased, the absorbance at 263 nm decreased from 0.26 (in the absence of externally added Ca2+ or at 0.0 μM) to a saturating value of 0.11 (at 115.4 μM). The values were assigned as Amax and Amin, respectively. The normalized absorbance at 263 nm, defined by (A − Amin)/(Amax − Amin), was plotted as a function of Ca2+ concentration (Figure 4b). The plot exhibited S-band shaped curves for CaM and PAA, whereas single curved shape for CBP. The curve shape indicates that more than two binding sites are involved in CaM and PAA, and that Ca2+ binding to CBP can be described by a single binding site.33 CaM has four EF-hand polypeptides and each E

DOI: 10.1021/acs.biomac.5b01694 Biomacromolecules XXXX, XXX, XXX−XXX

Article

Biomacromolecules Table 1. Binding Constants of CBP, CaM, and PAA for Ca2+a K1 CBP CaM PAA a

5.1 (±3.5) × 10 3.6 (±2.2) × 105 2.5 (±1.1) × 105

K2

K3

K4

χ2

6.0 (±3.7) × 105 0.5 (±0.3) × 105

0.2 (±0.1) × 105

0.1 (±0.1) × 105

0.0025 0.0006 0.0025

5

n = 3. The binding constants were obtained with the CaLigator program. The unit of the binding constant is L/mol or M−1.

EF-hand polypeptide has one binding site for Ca2+.53,54 Therefore, there are four Ca2+ binding sites per CaM. The EF hand mimicking CBP is supposed to have one binding site and there are four binding sites per CaM. The CaLigator program is a nonlinear curve fitting program for frequently used for metal ion binding study. The absorbance of 5,5-Br2-BAPTA at 263 nm decreases when it binds to Ca2+; thus, the absorbance at the titration point of i can be expressed by the following equation.55 Ai = [A max − (A max − A min )Y /(Y + KDQ )]CQ i/CQ 1 (4) 2+

KDQ is the dissociation constant of the chelator and Ca . CQ1 and CQi are the initial chelator concentration and the chelator concentration at point i, respectively. Y is a parameter used in the optimization procedure. The best-fit was selected when the χ2 value (square sum of error between calculated values and measured values: goodness of fit) was minimized in the simulation by using the CaLigator program. χ 2 (a) =

∑ [Acalcd − A meas]2

(5)

To calculate the binding constant of CBP, a model with two binding sites was used and one site was blocked by assigning a very small binding constant of 10−10 M−1 because the computer program was developed for models with more than two binding sites.33 Least square fitting of the curve provides the binding constant of the CBP to Ca2+. The fact that a model with one binding site well-fits to the binding curve of CBP suggests that the first Ca2+ binding mode has much higher affinity than the second and third Ca2+ binding modes, and one-to-one binding mode of Ca2+ to CBP might be assumed. The binding constant of Ca2+ for CBP was comparable with CaM and greater than PAA (Table 1). The binding constant of the EF-hand polypeptide to Ca2+ varied over 105−107 M−1.55 For example, the binding constants of CaM for Ca2+ calculated by NMR were Ka1 = 2.4 × 105, Ka2 = 1.1 × 106, Ka3 = 3.8 × 104, Ka4 = 1.2 × 105 M−1, which are in a similar range as our values.20 To investigate the metal ion selectivity of CBP, the binding constants of CBP for Cu2+ and Zn2+, as well as Ca2+, were compared. Another chromophore, murexide, was used as a chelating agent because the chromophore strongly binds to above metal ions and the binding constant of Ca2+ to CBP can be compared by using the two different methods. Metal ion binding to murexide resulted in a blue shift in the UV−vis spectra from 520 to 525 to 476 nm (Cu2+), 456 nm (Zn2+), and 471 nm (Ca2+), respectively (Figure 5a and Figure S5a,b). UV− vis spectra of the dye (50 μM) are shown as a function of Cu2+ concentration (Figure 5a). In particular, the spectra of murexide (50 μM) at a fixed Cu2+ concentration (100 μM) before (thick deep blue) and after adding CBP (thick red curve) clearly demonstrate that the binding of the metal ion to CBP reduced the free metal concentration. An isosbestic point at around 490 nm was observed in the Figure 5a, indicating that there were two species contributed by the bound and unbound

Figure 5. (a) UV−vis spectra of murexide (50 μM) as a function of Cu2+ concentration. Cu2+ concentration was varied between 0 (thin sky blue), 5, 10, 20, 30, 40, 50, 60, 70, 80, 90, 100 (thick deep blue), 200, 300, 400, and 500 μM in the absence of CBP. The red curve is the UV−vis spectra of murexide (50 μM) at a fixed concentration of Cu2+ (100 μM) in the presence of CBP. b) Plot of normalized absorbance, defined by Log(A − Amin)/(Amax − A), at peak maxima of 521 (Cu2+), 523 (Zn2+), and 525 nm (Ca2+), as a function of metal ion concentration. Amax and Amin correspond to the absorbance values at each peak maxima in the absence and after saturation with externally added metal ions, respectively.

dyes in the system.56 After adding CBP to the aqueous solution containing both murexide and Cu2+, free murexide was liberated due to the binding affinity of the Cu2+ to CBP. For Zn2+ and Ca2+, the binding of the metal ions to CBP was similarly investigated with murexide. The detailed procedure used to calculate the binding constants of between the metal ions and the polymer was reported previously.37 Based on the plot of absorbance against metal ion concentration, the average binding constants (n = 3) between metal ions and CBP were calculated to be 4.2 (±0.3) × 105, 1.7 (±0.3) × 105, and 7.0 (±0.4) × 105 M−1 for Cu2+, Zn2+, and Ca2+, respectively (Figure 5b). The binding affinity of ethylene diamine tetra acetic acid (EDTA) is greater for Cu2+ than for Zn2+ or Ca2+.37,57 Pectin and alginate with carboxylate functional groups along the polysaccharides develop egg box−like structures by binding metal ions.58 The differences in the binding constant of CBP among Ca2+, Cu2+, and Zn2+ might come from the differences in size, polarizability, the involvement of d-orbitals of the metals, or the effective conformation of the CBP. The EF-hand mimicking CBP is F

DOI: 10.1021/acs.biomac.5b01694 Biomacromolecules XXXX, XXX, XXX−XXX

Article

Biomacromolecules proven to have 2−4-fold greater binding constant for Ca2+ than Cu2+ and Zn2+.

(12) Wruss, J.; Pollheimer, P. D.; Meindl, I.; Reichel, A.; Schulze, K.; Schöfberger, W.; Piehler, J.; Tampé, R.; Blaas, D.; Gruber, H. J. J. Am. Chem. Soc. 2009, 131, 5478−5482. (13) Albrecht, M.; Stortz, P. Chem. Soc. Rev. 2005, 34, 496−506. (14) Clapham, D. E. Cell 1995, 80, 259−268. (15) Herring, G. M. The Organic Matrix of Bone. The Biochemistry and Physiology of Bone; Academic Press: New York and London, 1972; Ch. 5, pp 128−184. (16) Epple, M.; Dorozhkin, S. V. Angew. Chem., Int. Ed. 2002, 41, 3130−3146. (17) Rossol, M.; Pierer, M.; Raulien, N.; Quandt, D.; Meusch, U.; Rothe, K.; Schubert, K.; Schöneberg, T.; Schaefer, M.; Krügel, U.; Smajilovic, S.; Bräuner-Osborne, H.; Baerwald, C.; Wagner, U. Nat. Commun. 2012, 3, 1329. (18) Barry, W. H.; Bridge, J. H. B. Circulation 1993, 87, 1806−1815. (19) Malmendal, A.; Evenas, J.; Forsen, S.; Akke, M. J. Mol. Biol. 1999, 293, 883−899. (20) Kitevski-LeBlanc, J. L.; Evanics, F.; Scott Prosser, R. J. Biomol. NMR 2010, 47, 113−123. (21) Yuan, H.; Xing, C.; An, H.; Niu, R.; Li, R.; Yan, W.; Zhan, Y. ACS Appl. Mater. Interfaces 2014, 6, 14790−14794. (22) Hassouneh, W.; Nunalee, M. L.; Shelton, M. C.; Chilkoti, A. Biomacromolecules 2013, 14, 2347−2353. (23) Procyshyn, R. M.; Reid, R. E. J. Biol. Chem. 1986, 269, 1641− 1647. (24) Topp, S.; Prasad, V.; Cianci, G. C.; Weeks, E. R.; Gallivan, J. P. J. Am. Chem. Soc. 2006, 128, 13994−13995. (25) Dooley, K.; Kim, Y. H.; Lu, H. D.; Tu, R.; Banta, S. Biomacromolecules 2012, 13, 1758−1764. (26) Welch, J. T.; Kearney, W. R.; Franklin, S. J. Proc. Natl. Acad. Sci. U. S. A. 2003, 100, 3725−3730. (27) Kim, J. Y.; Park, M. H.; Joo, M. K.; Lee, S. Y.; Jeong, B. Macromolecules 2009, 42, 3147−3151. (28) Choi, Y. Y.; Jang, J. H.; Park, M. H.; Choi, B. G.; Chi, B.; Jeong, B. J. Mater. Chem. 2010, 20, 3416−3421. (29) Knoop, R. J. I.; Habraken, G. J. M.; Gogibus, N.; Steig, S.; Menzel, H.; Koning, C. R.; Heise, A. J. Polym. Sci., Part A: Polym. Chem. 2008, 46, 3068−3077. (30) Ding, J.; Zhuang, W.; Xiao, C.; Cheng, Y.; Zhao, L.; He, C.; Tang, Z.; Chen, X. J. Mater. Chem. 2011, 21, 11383−11391. (31) Linse, S. Method and Techniques. In Calcium Binding Protocols; Vogel, H. J., Ed.; Humana Press Inc.: NJ, 2002; Vol 2, pp 15−24. (32) Yamniuk, A. P.; Gifford, J. L.; Linse, S.; Vogel, H. J. Biochemistry 2008, 47, 1696−1707. (33) Andre, I.; Linse, S. Anal. Biochem. 2002, 305, 195−205. (34) Mannel-Crois, C.; Meister, C.; Zelder, F. Inorg. Chem. 2010, 49, 10220−10222. (35) Grudpan, K.; Jakmunee, J.; Vaneesorn, Y.; Watanesk, S.; Maung, U. A.; Sooksamiti, P. Talanta 1998, 46, 1245−1257. (36) Izadmanesh, Y.; Ghasemi, J. B. Talanta 2014, 128, 511−517. (37) Joo, J. H.; Ko, D. Y.; Moon, H. J.; Shinde, U. P.; Park, M. H.; Jeong, B. Biomacromolecules 2014, 15, 3664−3670. (38) Xue, Y.; Tang, X.; Huang, J.; Zhang, X.; Yu, J.; Zgang, Y.; Gui, S. Colloids Surf., B 2011, 85, 280−288. (39) Perdih, P.; Č ebašek, S.; Možir, A.; Ž agar, E. Molecules 2014, 19, 19751−19768. (40) Kar, M.; Pauline, M.; Sharma, K.; Kumaraswamy, G.; Gupta, S. S. Langmuir 2011, 27, 12124−12133. (41) Patel, M.; Jung, B. K.; Jeong, B. Adv. Healthcare Mater. 2015, 4, 1565−1574. (42) Oh, H. J.; Joo, M. K.; Sohn, Y. S.; Jeong, B. Macromolecules 2008, 41, 8204−8209. (43) Hwang, J.; Deming, T. J. Biomacromolecules 2001, 2, 17−21. (44) Hamley, I. W.; Ansari, I. A.; Castelletto, V.; Nuhn, H.; Rosler, A.; Klok, H. A. Biomacromolecules 2005, 6, 1310−1315. (45) Park, S. H.; Choi, B. G.; Moon, H. J.; Cho, S. H.; Jeong, B. Soft Matter 2011, 7, 6515−6521.

■

CONCLUSIONS In this study, we designed an EF-hand mimicking Ca2+ binding polymer. The binding moieties of carboxylate functional groups were provided by the six glutamate residues between flexible PEG and rigid poly(L-alanine) blocks. Binding of Ca2+ to CBP induces changes in the polymer conformation and slight perturbation of the secondary structure of the polypeptide, as indicated by changes in the chemical shift of the PEG block and the ellipticity of the polypeptide. Based on the CaLigator program, the binding constant of CBP was calculated to be 5.1 × 105 M−1, which is similar to the EF-hand polypeptide of CaM and greater than the PAA. In addition, the CBP exhibited a 2− 4-fold greater binding constant for Ca2+ than Cu2+ and Zn2+.

■

ASSOCIATED CONTENT

* Supporting Information S

The Supporting Information is available free of charge on the ACS Publications website at DOI: 10.1021/acs.biomac.5b01694. Gel permeation chromatograms of polymers. UV−vis spectra of murexide (50 μM) as a function of Ca2+ and Zn2+ concentration (PDF).

■

AUTHOR INFORMATION

Corresponding Author

*E-mail: [email protected]. Fax: +82 2 3277 2384. Author Contributions †

These authors equally contributed to the paper (H.J.C. and D.Y.K.). Notes

The authors declare no competing financial interest.

■

ACKNOWLEDGMENTS This work was supported by the National Research Foundation of Korea Grant funded by the Korean Government (2012M3A9C6049835 and 2014M3A9B6034223).

■

REFERENCES

(1) Hartgerink, J. D.; Beniash, E.; Stupp, S. I. Science 2001, 294, 1684−1688. (2) Deming, T. J. Prog. Polym. Sci. 2007, 32, 858−875. (3) Sim, S.; Kim, Y.; Kim, T.; Lim, S.; Lee, M. J. Am. Chem. Soc. 2012, 134, 20270−20272. (4) Stupp, S. I.; Palmer, L. C. Chem. Mater. 2014, 26, 507−518. (5) Sathaye, S.; Zhang, H.; Sonmez, C.; Schneider, J. P.; MacDermaid, C. M.; Bargen, C. D. V.; Saven, J. G.; Pochan, D. J. Biomacromolecules 2014, 15, 3891−3900. (6) Deming, T. J. Nature 1997, 390, 386−389. (7) Chécot, F.; Lecommandoux, S.; Gnanou, Y.; Klok, H. A. Angew. Chem., Int. Ed. 2002, 41, 1340−1343. (8) Jeong, Y.; Joo, M. K.; Bahk, K. H.; Choi, Y. Y.; Kim, H. T.; Kim, W. K.; Lee, H. J.; Sohn, Y. S.; Jeong, B. J. Controlled Release 2009, 137, 25−30. (9) Zhang, S.; Alvarez, D. J.; Sofroniew, M. V.; Deming, T. J. Biomacromolecules 2015, 16, 1331−1340. (10) Zou, R.; Wang, Q.; Wu, J.; Wu, J.; Schmuck, C.; Tian, H. Chem. Soc. Rev. 2015, 44, 5200−5219. (11) Alies, B.; Hureau, C.; Faller, P. Metallomics 2013, 5, 183−192. G

DOI: 10.1021/acs.biomac.5b01694 Biomacromolecules XXXX, XXX, XXX−XXX

Article

Biomacromolecules (46) Jang, J. H.; Choi, Y. M.; Choi, Y. Y.; Joo, M. K.; Park, M. H.; Choi, B. G.; Kang, E. Y.; Jeong, B. J. Mater. Chem. 2011, 21, 5484− 5491. (47) Suh, J. M.; Bae, S. J.; Jeong, B. Adv. Mater. 2005, 17, 118−120. (48) Jeong, B.; Windisch, C. F., Jr.; Park, M. J.; Sohn, Y. S.; Gutowska, A.; Char, K. J. Phys. Chem. B 2003, 107, 10032−10039. (49) Yu, L.; Zhang, H.; Ding, J. Angew. Chem., Int. Ed. 2006, 45, 2232−2235. (50) Hwang, M. J.; Joo, M. K.; Choi, B. G.; Park, M. H.; Hamley, I. W.; Jeong, B. Macromol. Rapid Commun. 2010, 31, 2064−2069. (51) Stevens, M. P. Polymer Chemistry, An Introduction, 3rd ed.; Oxford University Press: New York, NY, 1999; pp 40−42. (52) Stryer, L. Biochemistry, 4th ed.; W. H. Freeman and Co.: New York, NY, 1999; p 30. (53) Lewit-Bently, A.; Rety, S. Curr. Opin. Struct. Biol. 2000, 10, 637− 643. (54) Grabarek, Z. J. Mol. Biol. 2006, 359, 509−525. (55) Linse, S.; Helmersson, A.; Forsén, S. J. Biol. Chem. 1991, 266, 8050−8054. (56) Vinokurov, I. A.; Kankare, J. J. Phys. Chem. B 1998, 102, 1136− 1140. (57) Skoog, D. A.; West, D. M. Fundamentals of Analytical Chemistry, 4th ed.; Saunders College Publishing: PA, 1982; pp 284−286. (58) Fang, Y.; Al-Assaf, S.; Phillips, G. O.; Nishinari, K.; Funami, T.; Williams, P. A. Carbohydr. Polym. 2008, 72, 2319−2327.

H

DOI: 10.1021/acs.biomac.5b01694 Biomacromolecules XXXX, XXX, XXX−XXX