Molecular Dynamic ... - ACS Publications

Article pubs.acs.org/IC

A Combined Spectroscopic/Molecular Dynamic Study for Investigating a Methyl-Carboxylated PEI as a Potential Uranium Decorporation Agent Florian Lahrouch,† Anne Christine Chamayou,† Gael̈ le Creff,† Magali Duvail,‡ Christoph Hennig,§,⊥ Marisol Janeth Lozano Rodriguez,§,⊥ Christophe Den Auwer,*,† and Christophe Di Giorgio*,† †

Institut de Chimie de Nice, Université Côte d’Azur, CNRS, 06108 Nice, France Institut de Chimie Séparative de Marcoule, UMR 5257, CEA-CNRS-Université Montpellier-ENSCM, Site de Marcoule, BP 17171, 30207 Bagnols-sur-Cèze, France § Institute of Resource Ecology, HZDR, 01314 Dresden, Germany ⊥ Rossendorf Beamline, ESRF, 38043 Grenoble, France ‡

S Supporting Information *

ABSTRACT: Natural uranium has a very limited radioactive dose impact, but its chemical toxicity due to chronic exposure is still a matter of debate. Once inside the human body, the soluble uranium, under its uranyl form (U(VI)), is quickly removed from the blood system, partially excreted from the body, and partially retained in targeted organs, that is, the kidneys and bone matrix essentially. It is then crucial to remove or prevent the incorporation of uranium in these organs to limit the longterm chronic exposure. A lot of small chelating agents such as aminocarboxylates, catecholamides, and hydroxypyridonates have been developed so far. However, they suffer from poor selectivity and targeting abilities. Macromolecules and polymers are known to present a passive accumulation (size related), that is, the so-called enhanced permeability and retention effect, toward the main organs, which can be used as indirect targeting. Very interestingly, the methyl carboxylated polyethylenimine (PEI-MC) derivative has been described as a potent sequestering agent for heavy metals. It would be therefore an interesting candidate to evaluate as a new class of decorporation agents with passive targeting capabilities matching uranium preferential sequestering sites. In the present work, we explored the ability of a highly functionalized (89% rate) PEI-MC to uptake U(VI) close to physiological pH using a combination of analytical and spectroscopic techniques (inductively coupled plasma optical emission spectrometry (ICP-OES); extended X-ray absorption fine structure (EXAFS); and Fourier transformed infrared (FT-IR)) together with molecular dynamics (MD) simulation. A maximum loading of 0.47 mg U(VI) per milligram of PEI-MC was determined by ICP-OES measurements. From FT-IR data, a majority of monodentate coordination of the carboxylate functions of the PEI-MC seems to occur. From EXAFS and MD, a mix of mono and bidentate coordination mode was observed. Note that agreement between the EXAFS metrical parameters and MD radial distribution functions is remarkable. To the best of our knowledge, this is the first comprehensive structural study of a macromolecular PEI-based agent considered for uranium decorporation purposes.

■

between kidneys (12%) and bones (15%).1 On the one hand, the U(VI) ions are filtered in the kidneys through the glomerulus, fixed by the anionic sites in the proximal tubular epithelial brush border, and may enter the cells through endocytosis.2 On the other hand, U(VI) ions are retained in bones by incorporation into the hydroxyapatite matrix, under the coprecipitate form Ca(UO2)2(PO4)2. As a consequence of the successive destruction/reconstruction cycles of the bone surface, U(VI) is released in the blood system or burrows into the amorphous hydroxyapatite matrix.3−5 In addition to the

INTRODUCTION

Uranium is found in the earth’s crust at a mixing secular equilibrium of U-238, U-235, and U-234. It is a heavy element, and all its isotopes are radioactive. Although the radioactive dose impact of natural uranium is very limited (its massic activity is very low), chemical toxicity due to chronic exposure is much less clear. In atmospheric conditions, uranium is mostly found in its oxocationic form {U(VI)O22+} that is the most stable form of uranium at formal oxidation state +VI (called uranyl form and referred to simply as “U(VI)” below). Once inside the human body, soluble U(VI) is quickly removed from the blood system and partially excreted from the body. The retained part of the absorbed U(VI) is almost one-third, shared © XXXX American Chemical Society

Received: October 9, 2016

A

DOI: 10.1021/acs.inorgchem.6b02408 Inorg. Chem. XXXX, XXX, XXX−XXX

Article

Inorganic Chemistry modification of the biological mechanisms the oxidation stress linked to the uranyl presence leads to the death of kidney and bone cells. Its chemical toxicity makes it a nephrotoxic and a carcinogenic agent.6−9 It is then crucial to remove or prevent incorporation of uranium in the target organs (bone matrix or kidneys essentially) to limit chronic exposure in the long term. U(VI), as the other elements of the actinide series, is a hard acid (in Pearson’s classification10,11) known to exhibit a high affinity for hard basis such as deprotonated carboxylic and phosphonic acids. For plutonium contamination (Pu(IV)), for instance, diethylenetriamine pentaacetic acid (DTPA) is currently the recommended treatment in case of contamination. However, DTPA was shown to be quite ineffective toward U(VI). Even worse, the use of DTPA in case of U(VI) contamination was proved to increase its nephrotoxicity.12 Ligands based on sulfocatecholate (TIRON) were the first chelating agents showing a modest reduction of acute U(VI) toxicity.13−15 More recently promising ligands for actinide(IV) and (III) decorporation have been developed based on catecholate (CAM) and hydroxypyridonate (HOPO) attached to spatially suitable molecular backbones.16 K. Raymond et al. explored the uranyl chelating ability of CAM and HOPO derivatives in in vivo conditions.17 In this study, the most efficient adduct (4LI(Me-3,2-HOPO)) was shown to remove the total uranium amount in mice with almost total kidney decorporation when injected less than 3 min after contamination. But the same authors underlined the difficulties to decorporate U(VI) once out of the bloodstream (e.g., in skeleton) and have observed the toxicity of the chelating agents that can lead to a renal failure.18,19 Instead of molecular chelates, the use of macromolecules such as polymers has never been reported for actinide decorporation. It presents several advantages over a classical molecular decorporation strategy such as a higher capacity (greater abundance of chelating sites per area unit that could enhance the complexation efficiency) and indirect vectorization properties correlated to the particular biodistribution of polymers (which are usually mechanically retained in the main organs due to their size). It could also allow extra blood system complexation. The highly soluble poly(ethylene imine) (PEI) has been extensively studied in medicinal chemistry, and its functionalization is rather easy to implement.20 For instance, the functionalized methyl carboxylate poly(ethylene imine) (PEI-MC) has been shown to be efficient for the treatment of water containing heavy metals.21 Some studies have also reported the use, in in vivo conditions, of phosphonate-functionalized poly(ethylene imine) (PEI-MP) for bone cancer imagery and scintigraphy.22 To the best of our knowledge, however, the PEI-MC has never been reported for uranyl decorporation. In the present report, we explore the ability of PEI-MC to uptake U(VI) close to physiological pH using a combination of analytical and spectroscopic techniques together with molecular dynamics (MD) simulation. Inductively coupled plasma optical emission spectrometry (ICP-OES) was used to estimate the U(VI) uptake curves of PEI-MC, extended X-ray absorption fine structure (EXAFS), and Fourier transformed infrared (FTIR) spectroscopies were performed to characterize the coordination environment of uranium. Finally MD was performed on a simplified model of PEI-MC to further define the structural parameters obtained with EXAFS data analysis. PEI-MC (monomer shown in Scheme 1) is the first representative of the PEI family, and this investigation is a

Scheme 1. Representation of the Monomeric Unit of PEIMC as if Functionalization Were Equal to 100%

first step toward the development of more elaborate macromolecular candidates for decorporation purposes. This work on a model representative is a preliminary step before envisioning further biological evaluation.

■

EXPERIMENTAL SECTION

Thin-layer chromatography was performed on aluminum sheets coated with Kiesel gel 60 F254 (Merck Millipore), with visualization by charring with H2SO4 (10%) or spraying with ninhydrin (0.1%; SigmaAldrich) in EtOH. Size exclusion chromatography was performed on Sephadex LH-60 (Amersham pharmacia). Dialysis experiments were performed with regenerated cellulose (RC) membranes Spectra/Por7 (spectrum laboratories.com) with a molecular weight cut off (MWCO) of 1 kDa. 1H NMR spectra were recorded at 200 MHz on an AC 200 spectrometer (Bruker). Chemical shifts (δ, ppm) were measured relative to CDCl3 (7.26 ppm), CD3OD (3.31 ppm), or D2O (4.79 ppm) for 1H. Synthesis and Polymer Characterization. Trifluoroacetic acid (TFA), branched poly(ethylene imine) (bPEI, 25 kDa MW), and tertbutyl bromoacetate 98% were purchased from Sigma-Aldrich; N,Ndiisopropylethylamine (DiPEA) was purchased from Iris Biotech GmbH. Polyethylenimine Methyl Carboxylate. Branched PEI 25 kDa (250 mg, 1.45 mmol of monomeric units) was dissolved in dichloromethane (DCM) with DiPEA (1.75 mL, 10 mmol). Then tert-butyl bromoacetate (1.7 g, 8.7 mmol) diluted into DCM was added dropwise. The reaction mixture was stirred at room temperature overnight. The resulting solution was then concentrated in vacuo and washed with distilled water. The remaining organic fraction was purified through size exclusion chromatography (sephadex LH60) to yield 380 mg (64%) of PEI functionalized with tert-butyl acetate groups. PEI tert-butyl ester (PEI-MCOOtBu): 1H NMR(200 MHz; CDCl3): 1.42 ppm (36H, s, C-(CH3)3), 2.25−5.00 ppm (25H, m, −CH2−CH2−N−CH2−COOtBu). Integration of the methyl signal from tert-butyl ester allows to determine the functionalization level, 89%, which corresponds to ∼3.5 tert-butyl functions per monomer (maximum would have been 4). tert-Butyl ester functions of PEI-MCOOtBu were then cleaved with a TFA/DCM (1:1) solution containing 1% of triisopropylsilane. After it was stirred for 1 h, the mixture was concentrated and submitted to several evaporation cycles with HCl (1 N) to fully remove TFA and displace negative counterions with chloride. Finally, the residue was washed with water, purified by dialysis (1 kDa MWCO, RC membrane), lyophilized, and then dried over P2O5 to give the final product (quantitative yield) as a yellow powder. The water content was estimated to be 1% with ATG. Cl− content, as ammonium counterions had been measured by Mohr salt titration after mineralization. [Cl−] = 1.08 mol/mol of monomer. 1 H NMR (200 MHz; D2O): 2.63−4.35 ppm (25H, m, −CH2− CH2−N−CH2−COOH). IR analysis: 1681 cm−1 (CO stretching vibrations), 1129 cm−1 (C−O stretching vibrations). ICP-OES Analysis. PEI-MC, aliquots of 100, 50, 25, and 20 μL (10 mg/mL) were added to a solution of 5 μL of uranyl nitrate (UO2(NO3)2·6H2O from pro-labo) [U] = 0.1M, pH 1, respectively. Tris-buffered saline (TBS; 400 μL) was added to each aliquot, and the pH was adjusted to 5.0 by addition of sodium hydroxide (1 M). If one considers that PEI-MC formula could be written as [C8H20N4]n ≈ 145 B


Article

Inorganic Chemistry

ensemble. To compensate for the negative charge due to the presence of negatively charged PEI-MC (−3 e for each), counterion Na+ (10) was added to the solution. Periodic boundary conditions were applied to the simulation box. Long-range interactions were calculated using the particle-mesh Ewald method.27 Equations of motion were numerically integrated using a 1 fs time step. Systems were previously equilibrated at 298.15 K over at least 100 ps, and production runs were subsequently collected for 9 ns. Water molecules were described by the rigid POL3 model,28,29 which takes into account the polarization. The van der Waals energy is described here by a 12−6 Lennard-Jones (LJ) potential. The polarizable force field developed by Åqvist et al. was used to describe the Na+ cation.30 The polarizable force field for describing the hydration properties of the uranyl cation was adjusted to provide consistent results with the EXAFS results. PEI-MC polymer was modeled here as a monomer, and the optimized potentials for liquid simulations all-atom force field was used, except for the carboxyl functional group COO−. Indeed, this functional group was modeled using the force field developed by Kamath et al.31 Atomic partial charges on the PEI-MC monomer were calculated with the restricted electrostatic potential procedure developed by Bayly et al.32 As generally done, to take into account the explicit polarization in the model, we rescaled the partial atomic charges on the complexing functions.33 Therefore, a scaling factor of 0.80 was applied on the partial atomic charges of the carbon atoms of the carboxyl functional groups, and the partial atomic charges of the oxygen atoms were adjusted to get the correct total charge of the monomer, that is, −3 e. Atomic partial charges on PEI-MC used for MD simulations are presented in Supporting Information Figure S1. The theoretical EXAFS signals were calculated directly from snapshots issued from MD simulations, typically on the order of 900 (i.e., each 10 ps) using the FEFF9 program.25 The EXAFS spectra were then analyzed by means of the IFFEFIT software.24

(monomer structure is derived from a tetra-amino scaffold resulting from the ratio 25:50:25, corresponding, respectively, to the percentage of primary, secondary, and tertiary amine in its structure23), the U/ monomer ratio corresponds to 0.2, 0.4, 0.7, 0.9, and 1.1, respectively. The solutions of the complexes PEI-MC-U were purified by ultracentrifugation on Microcon filters (Millipore, MWCO 3 kDa) at 10 000 g and washed twice with 200 μL of TBS buffer at pH 5.0. This pH was selected to avoid hydrolysis of free U in our conditions. The Microcon filters (containing the PEI-MC-U complex) were then digested with nitric acid (65%) into a microwave furnace during 2 h (200 °C). Each digested solution was dried and dissolved into a nitric acid solution at pH 1.5 before being measured by ICP-OES (Perkins Elmer Optima 8000). Each experiment was repeated three times to obtain 15 independent values for each aliquot. FT-IR. FT-IR was performed in attenutated total reflection (ATR) mode, using a 1 reflection diamond ATR device coupled to a Bruker tensor 27 equipped with a nitrogen-cooled MCT detector. Spectra of the free PEI-MC polymer were performed by deposition of droplets of polymer solution (10 mg/mL) on the ATR in the mid-infrared range, between 4000 and 600 cm−1 (resolution: 2 cm−1, 64 scans per spectrum, scan velocity: 10 kHz). Spectra of the PEI-MC-U complex were obtained after mixing 5 μL of 0.1 M uranium nitrate solution or 20 μL of 0.025 M uranium chloride solution with 1 mg of PEI-MC, which corresponds to a value of U/monomer ratio of 0.2. The pH of all samples was fixed to 7.0 by addition of sodium hydroxide (1 M). Spectra of PEI-MC-U were recorded after drying droplets of their solutions on the ATR diamond. EXAFS. Sample Preparation. Functionalized PEI-MC, (5 mg) was dissolved in 150 μL of TBS buffer at pH 7.0. A solution (50 μL) of uranyl nitrate (UO2(NO3)2·6H2O), [U] = 0.1 M, pH 1 was added to obtain a final uranium concentration of 2.5 × 10−3 M. This corresponds to a stoichiometry of U/monomer = 0.04. The pH was adjusted to 7 by addition of sodium hydroxide (1 M). The solution was purified with ultracentrifugation onto Microcon (MWCO 3 kDa) at 10 000 g and washed twice with TBS buffer at pH 7.0 as described above. Data Recording. EXAFS experiments at the U LIII edge were performed on the ROBL beamline of ESRF synchrotron facility. The ROBL beamline is devoted to the investigation of radioactive materials in the hard X-ray range. Because of the low concentration of the samples, EXAFS measurements were performed in fluorescence mode using a 13-element high-purity germanium detector with specifically designed 200 μL cells (Ets CANAPLE). The ROBL beamline is equipped with a Si(111) water-cooled monochromator in the channelcut mode. Two Pt-coated mirrors were used for harmonic rejection. Data processing was performed using the ATHENA code.24 The eo energy was identified at the maximum of the absorption edge. Fourier transformation (FT) in k2 was performed between 2.8 and 12.2 Å−1 with Hanning windows using the ARTEMIS code.24 Only one global amplitude factor S02 and one energy threshold correction factor e0 were used for all paths. The agreement factor R (%) and the quality factor (QF = reduced χ2) of the fit were provided from ARTEMIS. Data Fitting. Phases and amplitudes were calculated using the FEFF9 simulation code.25 The model for phases and amplitude calculation was taken from one of the most representative snapshots of the MD calculations (vide infra). The total coordination number was fixed to 5 + 2, as it is a usual coordinate for uranyl. Single scattering paths were considered for the oxo oxygens (Oax), the equatorial oxygens, and the carbon atoms of the carboxylates. The number of monodentate and bidentate carboxylate functions was linked according to the relation monodentate = (5-bidentate/2). Triple U−O−Cmon and quadrupole U−Oax−U−Oax scattering paths were also considered and linked to the corresponding single paths. The addition of a single scattering path involving a fraction of chlorine atom improved significantly the fit of the imaginary part of the FT. This is discussed in the Results Section. Molecular Dynamics Simulations. MD simulations of one UO22+ cation in the presence of four PEI-MC residues in aqueous solutions (3022 water molecules) were performed with SANDER14, a module of AMBER1426 using explicit polarization in the NPT

■

RESULTS AND DISCUSSION Uptake uranium curves, FT-IR spectroscopy, and EXAFS together with MD were combined to decipher the uranium interaction mechanisms with PEI-MC. Uptake curves were performed below physiological pH (TBS buffer at pH 5.0) to avoid competition with U hydrolysis, especially when the U/ monomer molar ratio is close to or above 1. For FT-IR and EXAFS experiments, pH was set to 7, because the U/monomer molar ratio is kept far below 1, and this precludes the presence of free hydrolyzable U. Note that in a more realistic case of contamination, full U hydrolysis is also not likely to occur, because several competitors are present (metabolites, peptides, proteins, etc.34). FT-IR spectra of PEI-MC at pH 5 and 7 (Supporting Information Figure S2) shows no significant differences, indicating that protonation of the polymer is not largely modified between those two values. Therefore, the use of pH values between 5 and 7 is not irrelevant for a simplified system, even if physiological conditions are ultimately targeted. Uptake Curves of Uranium. The uptake curve of U(VI) by PEI-MC is shown in Figure 1. It exhibits a linear increase up to ca. 1.0−1.5 U/monomer ratio. After this zone a plateau is reached, meaning that full complexation has occurred. Note that after U/monomer ratio = 1.5−2.0, the uptake curve is impossible to determine, because the excess of free U(VI) precipitates at this pH and introduces a strong bias in the measurement. For instance, the decrease of the two last points at U/monomer = 2 is a consequence of the partial precipitation of the excess of free U(VI). Considering the 1.0−1.5 ratio zone as the beginning of a plateau, it corresponds to a maximum uranium load of 0.47 mg of {UO22+} per milligram of PEI-MC. As a point of comparison, sodium bicarbonate, which may be used in case of acute uranyl contamination, is injected with a C


Article

Inorganic Chemistry

cross-linked amines.38 Complexation with uranyl was performed from uranyl nitrate (to compare with the EXAFS samples) and with uranyl chloride to fully avoid the strong absorption bands of the possible remaining nitrate ions around 1350 cm−1.39 In both cases, PEI-MC-U (nitrate) and PEI-MCU (chloride), a distinct band appears at 909 cm −1 corresponding to the antisymmetric stretching mode of the transdioxo unit of U(VI). The band is red-shifted compared to its value for aqueous uranyl ion [UO2(H2O)5]2+, reported at 962 cm−1.40 According to the relation derived by Bartlett et al.,41 linking the energy of the vibration band to the UO bond length, the value at 909 cm−1 corresponds to U−Oax = 1.78 Å, in perfect agreement with both MD and EXAFS data. With uranyl nitrate, an extra band appears at 1043 cm−1, which corresponds to a vibrational mode of some remaining NO3− anions. Major modifications upon complexation come from the carboxylate bands: the band of carbonyl stretching mode at 1683 cm−1 is less intense after complexation; the antisymmetric stretching mode of −COO− at 1596 cm−1 is red-shifted to 1571 cm−1 and merged with the stretching mode of the amine functions, which becomes more intense; the weak shoulder at 1402 cm−1 corresponding to the symmetric stretching mode of the carboxylate is red-shifted to 1391 cm−1 (nitrate) and to 1396 cm−1 (chloride) (this difference may come from the remaining nitrate absorption band contribution). It is wellknown that the energy gap Δ(νAS − νS) between the antisymmetric and symmetric modes of the carboxylate function is indicative of the complexation mode. If Δ(νAS − νS) increases after complexation, monodentate binding mode is favored; if Δ(νAS − νS) decreases after complexation bidentate binding mode is favored.42 On the one hand, in our situation Δ(νAS − νS) is difficult to identify because of the shoulder shape with the band at 1683 cm−1 after complexation. On the other hand νS is shifted to the lower energy (from 1402 to 1391 cm−1) indicating that Δ(νAS − νS) would increase after complexation, corresponding to mainly monodentate coordination mode. But this trend is difficult to assess without any ambiguity. EXAFS Data and Molecular Dynamics. The experimental EXAFS spectrum of PEI-MC-U at the U LIII edge is shown in Figure 3a, and the corresponding FT is shown in Figure 3b. The modulus of the FT exhibits for the first coordination sphere the characteristic short-range axial shell of the two oxo bonds of UO22+. The equatorial shell of the first coordination sphere is also well-resolved and exhibits a single contribution. A second coordination sphere is also visible between R + Φ = 2.2−3.2 Å with much weaker signal. Comparison between experimental EXAFS data and MD simulations was performed to bring additional insight into the uranyl cation coordination mode. In the MD simulation, one uranyl cation in the presence of four PEI-MC monomers in aqueous solution was modeled. However, because of the steric hindrance of the monomer, it seems impossible to have the four monomers together in the UO22+ coordination sphere. Therefore, only three monomers remained during all the simulation time (9 ns) and are coordinated to UO22+ thanks to their −COO− groups. Figure 4a,b shows the snapshots obtained from MD simulations. Although only three PEI-MC are present in the UO22+ coordination sphere, four carboxyl functional groups are coordinated to UO 22+ implying thus those two COO − functions of the same PEI-MC are involved. For the sake of clarity, discussion on metrical parameters below is further divided in two parts.

Figure 1. Uptake curve of U(VI) by PEI-MC at pH 5 (see Experimental Section). Data are reported in triplicate.

single dose of 3.5 g per adult.35 Also, for plutonium decorporation the standard treatment is Ca-DTPA injections with a dose of 30 μmol/kg/d,36 and this corresponds to ∼1 g per adult per day. Infrared Spectroscopy. The FT-IR spectra of PEI-MC and PEI-MC-U are compared in Figure 2, where they are

Figure 2. FT-IR spectra in absorption mode of free PEI-MC, PEI-MCU(nitrate), and PEI-MC-U(chloride), pH 7. Normalization was performed on band at 1208 cm−1, and spectra were shifted in ordinates.

presented normalized on the 1208 cm−1 band. The spectrum of free PEI-MC exhibits a series of expected strong absorption bands corresponding to the carboxylate and amine functions. Carboxylate functions present an absorption band at 1683 cm−1 characteristic of the carbonyl stretching mode ν(CO);37 a band at 1596 cm−1 and a weak shoulder at 1402 cm−1 could be attributed to the antisymmetric and symmetric stretching mode of the COO− group engaged in hydrogen bonding with the protonated tertiary amine functions; a peak at 1183 cm−1 is associated with the vibration band of the C−O bond of the carboxylate. The amine functions show a peak at 1627 cm−1 that could correspond to a bending mode attributed to the D


Article

Inorganic Chemistry

Figure 3. (a) EXAFS data at the U LIII edge of PEI-MC-U and (b) corresponding FT. Experimental = black line; fit = black dots.

Figure 4. Snapshots issued from MD simulations showing the UO22+ first coordination shell: (a) overview of uranyl surrounded by three PEI-MC monomers and (b) enlarged view of the first shell. For the illustrations, UO22+ is colored in green and red, Na+ is colored in purple, oxygen atoms in red, hydrogen atoms in white, nitrogen atoms in blue, and carbon atoms in gray. For clarity, hydrogen atoms of the PEI-MC monomer are not presented.

Figure 5. RDFs (solid line) calculated for the (a) U−Oax, (b) U−OPEI‑MC, and (c) U−CPEI‑MC distances. Corresponding coordination numbers (dashed line) are also shown.

times lower than the one determined by EXAFS (0.0028 Å2). This difference originates in the force field we developed for the UO22+ cation and has already been observed for UO22+ in pure water. In the equatorial oxygen shell, two U−OPEI‑MC distances were determined from the RDFs calculated from MD simulations, one corresponding to monodentate coordination and one to bidentate coordination. This calculation resulted in a total number of 5.05 O atoms in the first coordination shell. To

First Coordination Sphere. From the radial distribution functions (RDFs) presented in Figure 5a−c the mean U−Oax, U−OPEI‑MC, and U···CPEI‑MC distances were calculated (Table 1) and can be further compared to the EXAFS best-fit metrical parameters described in Table 2. From MD, the mean distance calculated for the axial oxygen shell U−Oax is equal to 1.78 Å with a Debye−Waller factor of 0.0010 Å2. This distance compares very well with the EXAFS data: U−Oax = 1.78 Å. Note that from MD, the calculated Debye−Waller factor is ∼3 E


Article

Inorganic Chemistry

shell was fixed to 5 (close to 5.05 from MD), the relative number of each binding mode was allowed to vary as explained in the Experimental Section. Thus, the EXAFS fitting procedure led to 2.9 O at 2.31 Å for monodentate and to 2.1 O at 2.45 Å for bidentate coordination, the average distance being equal to 2.37 Å. These data (MD and EXAFS) are in remarkably good agreement relative to each other and also with the literature.43−45 Debye−Waller factors are also in the same order, although the value for the monodentate O is 3 times smaller with MD than with EXAFS. For monodentate O: σ2 = 0.0023 Å2 for MD, σ2 = 0.0079 Å2 for EXAFS; for bidentate O: σ2 = 0.0052 Å2 for MD, σ2 = 0.0059 Å2 for EXAFS. Second Coordination Sphere. To improve the fit in the second coordination sphere (more specifically the imaginary part of the FT in the R + Φ = 2.2−2.6 Å range), an additional shell of Cl atom was added. The presence of chlorine in the vicinity of uranyl would not be surprising, since the latter might be associated with the positive charge of the amines. Indeed the experimental conditions (fixed NaCl ionic force and TBS buffer) could explain the high amine protonation rate and the presence of chloride ions as counterions. The addition of 0.4 Cl atoms in the EXAFS fit is significant, since it decreases the quality factor from 1.17 to 0.54 (see imaginary part of the FT in Supporting Information Figure S3). The number of additional Cl was not fitted, but a value lower than 0.5 gives the best agreement. Occurrence of both coordination modes for the carboxylates and comparison between the relative numbers of mono and bidentate carboxylates in EXAFS and MD must be considered with care. The presence of bidentate carboxylate functions is not confirmed by the IR data analysis as described above. It is possible that upon drying on the ATR spectrum, the complexation conditions differ significantly from the fixed condition in the EXAFS cell. On the one hand, this may explain the absence of a clear fingerprint of bidentate coordination mode in the IR spectra. On the other hand, fitting the EXAFS data with only monodentate carboxylate functions also led to a good fit (quality factor equal to 0.70 compared to 0.58 for the fit in Table 2). This is not really surprising, since Cbid does not contribute to the fit as explained in the Experimental Section. In that case, five O at 2.37(1) Å (σ2 = 0.0126 Å2) were

Table 1. Structural Properties of UO22+ in the Presence of 3 PEI-MC Monomers in Water Calculated from Molecular Dynamics Simulations da

atom Oax OPEI‑MC OPEI‑MC (bi) OPEI‑MC (mono) CPEI‑MC (bi) CPEI‑MC (mono)d

1.78 2.36 2.40 2.32 2.85 3.39 (3.34/3.52)

σ2 c

CNb 2 5.05 2.49 2.56 1.05 2.95 (2.14/0.81)

0.0010 0.0034 0.0052 0.0023 0.0029 0.0153 (0.0197/0.0037)

Distances in the UO22+ first coordination shell (in Å). bCoordination number. cDebye−Waller factor (Å2). dThe values in parentheses correspond to the two U−CPEI‑MC distances calculated for the monodentate configuration.

a

Table 2. EXAFS Best-Fit Metrical Parameters sample PEI-MCU

first coordination sphere

second coordination sphere

fit parametersa

2 Oax at 1.78(1) Å σ2 = 0.0028 Å2 2.9(3) Omon at 2.31(2) Å σ2 = 0.0079 Å2 2.1(3) Obid at 2.45(2) Å σ2 = 0.0059 Å2

2.9 Cmon at 3.41(2) Å σ2 = 0.0033 Å2 0.4 Cl at 3.00(2) Å

S02 = 1.1 e0 = 1.75 eV R factor =

σ2 = 0.0049 Å2

0.45% CHI2r = 0.58

a Uncertainties are in brackets, and numbers in italics were fixed or linked. S02 is the global amplitude factor; e0 is the energy threshold factor. The R factor is the agreement factor of the fit in percentage, and CHI2r is the reduced quality factor of the fit.

determine the average U−OPEI‑MC distance for both coordination modes, the U−OPEI‑MC RDF peak was fitted using two Gaussian functions, each representing one coordination mode. Accordingly, the calculation lead to 2.56 O atoms located at 2.32 Å and 2.49 O atoms at 2.40 Å, corresponding, respectively, to the monodentate and bidentate configurations. The average U−OPEI‑MC distance is equal to 2.36 Å. For the fit of the EXAFS spectrum, two types of carboxylate functions (monodentate and bidentate) were also introduced for consistency with MD. Although the total number of oxygen atoms in the equatorial

Figure 6. Comparison between experimental EXAFS data and EXAFS data calculated form MD. (a) Experimental (black) and theoretical (red) k2weighted EXAFS spectra and (b) FT obtained from k2-weighted EXAFS spectra for a k range from 3 to 12.55 Å−1. F


Article

Inorganic Chemistry

confirm the monodentate coordination mode of carboxylate functions but also show that bidentate carboxylate functions may also be implied in the complexation mode. Agreement between the EXAFS metrical parameters and MD RDF is remarkable, resulting in a coordination mode of mixed bidentate and monodentate carboxylate functions, that is, 2.9(3) Omon at 2.31(2) Å and 2.1(3) Obid at 2.45(2) Å. These data suggest that functionalized PEI polymers may be considered as a promising new strategy to decorporate actinides in the human body and, in particular, uranyl, for which no treatment is currently available. For uranyl decorporation, the molecular strategy (DTPA, HOPO, carbonate) is solely based on affinity constants and has a less clear effect on uranyl once outside the blood system (e.g., in skeleton and kidneys). The strategy presented here, based on macromolecular objects, directly targets the retention organs of uranyl. Future studies will be devoted to the in vitro behavior of the PEI-MC with kidney culture cells and hydroxyapatite as models for the uranium natural and preferential retention sites.

obtained. This distance compares remarkably well with the average of the equatorial shell in Table 2 (2.37 Å) and the average of MD RDF in Table 1 (2.36 Å), but with a significant increase in the Debye−Waller factor. The Debye−Waller factor of the U···Cmon contribution also increases significantly (σ2 = 0.0102 Å2) with regard to the value in Table 2 (σ2 = 0.0033 Å2). In conclusion, although the presence of bidentate coordination mode cannot be fully asserted by EXAFS, it corresponds to slightly better-fitting parameters. From MD, one must now keep in mind that the PEI-MC monomers change their coordination modes (from bi- to monodentate, or reciprocally) every 2−3 ns in the MD calculation. Consequently, some configurations correspond to a mix of mono- and bidentate coordination modes. Therefore, since both distances Omon and Obid are relatively close (2.32 and 2.40 Å), it appears complicated to distinguish one configuration from the other when only looking at the U−O distances. Figure 5c shows that three U···CPEI‑MC distances were calculated, one corresponding to the bidentate coordination mode and two to the monodentate coordination mode. Quite surprisingly, it seems that the monodentate coordination mode of the PEI-MC monomer depends on the position of the noncoordinated oxygen atom of the functional group COO−. Indeed, it was observed that in the case of nonbonding O atom far from uranyl (U−Ononcoord ≈ 4.2 Å) the mean U···Cmon distance is equal to 3.52 Å, whereas it is equal to 3.34 Å when the same Ononcoord atom is closer to the uranyl (in that case U···Cmon ≈ U−Ononcoord ≈ 3.3 Å). The average U···Cmon distance of the two monodentate conformations is then equal to 3.39 Å (2.14 C at 3.34 Å and 0.81 C at 3.52 Å). With regard to the bidentate conformation, one distance was obtained with MD: 1.05 C at 2.85 Å. From the EXAFS point of view, only the monodentate C atoms are visible because of the relative alignment of the U− O−Cmon atoms (focusing effect). The fitted distance (3.41 Å) compares very well to the average (3.39 Å) obtained by MD, and again this favors the assumption that both coordination modes occur. The very good agreement between MD calculations and EXAFS data is also observed when comparing the theoretical EXAFS signal resulting from MD simulations and that of the experiment in Figure 6a,b. Moving from theory to application, the changes in the EXAFS signal magnitude, as well as in the height of the FT, was the impetus for how the U−Oax distance is described by MD simulations. Last, it must be underlined that the U environment described here may result from an average of conformations, although the relatively low values of the Debye−Waller factors associated with the first coordination sphere suggest that all U centers are relatively similar.

■

ASSOCIATED CONTENT

* Supporting Information S

The Supporting Information is available free of charge on the ACS Publications website at DOI: 10.1021/acs.inorgchem.6b02408. PEI-MC FT-IR spectra, atomic partial charges on PEIMC used for MD simulations, and imaginary part of the FT of the EXAFS spectrum of PEI-MC-U (PDF)

■

AUTHOR INFORMATION

Corresponding Authors

*E-mail: [email protected]. (C.d.G.) *E-mail: [email protected]. (C.d.A.) ORCID

Magali Duvail: 0000-0003-1586-260X Christophe Den Auwer: 0000-0003-2880-0280 Notes

The authors declare no competing financial interest.

■

ACKNOWLEDGMENTS MD simulations were performed using high-performance computing facilities of TGCC/CCRT and the computing center of CEA Marcoule. EXAFS data were recorded at ESRF synchrotron, the European synchrotron facility, on the ROBL beamline. F.L. is very grateful to the Région Provence-AlpesCôte d’Azur for Ph.D. grant and to the Commissariat à l’Energie Atomique under the ToxNuc program for financial support.

■

CONCLUSION We have explored in this work the uptake mechanism of uranyl (U(VI)) with a functionalized PEI-MC considered here as a potential new class of uranyl decorporation agent. The maximum load of U(VI) was determined by ICP-OES measurements and is equal to 0.47 mg per milligram of PEIMC. Structural details about the complexation mode of uranyl with the polymer chain have been further obtained by a combination of FT-IR and EXAFS spectroscopies and MD calculations. From FT-IR data, monodentate coordination of the carboxylate functions seems to be preferred, although the complexity of the IR spectra and different complexation conditions for IR makes it difficult to conclude without any ambiguity. Both EXAFS measurements and MD calculations

■

REFERENCES

(1) Ansoborlo, E.; Prat, O.; Moisy, P.; Den Auwer, C.; Guilbaud, P.; Carriere, M.; Gouget, B.; Duffield, J.; Doizi, D.; Vercouter, T.; Moulin, C.; Moulin, V. Actinide speciation in relation to biological processes. Biochimie 2006, 88, 1605−1618. (2) Shaki, F.; Hosseini, M. J.; Ghazi-Khansari, M.; Pourahmad, J. Toxicity of depleted uranium on isolated rat kidney mitochondria. Biochim. Biophys. Acta, Gen. Subj. 2012, 1820, 1940−1950. (3) Vidaud, C.; Bourgeois, D.; Meyer, D. Bone as Target Organ for Metals: The Case of f-Elements. Chem. Res. Toxicol. 2012, 25, 1161− 1175. (4) Safi, S.; Creff, G.; Jeanson, A.; Qi, L.; Basset, C.; Roques, J.; Solari, P. L.; Simoni, E.; Vidaud, C.; Den Auwer, C. Osteopontin: A

G


Article

Inorganic Chemistry Uranium Phosphorylated Binding-Site Characterization. Chem. - Eur. J. 2013, 19, 11261−11269. (5) Kurzbach, D.; Platzer, G.; Schwarz, T. C.; Henen, M. A.; Konrat, R.; Hinderberger, D. Cooperative Unfolding of Compact Conformations of the Intrinsically Disordered Protein Osteopontin. Biochemistry 2013, 52, 5167−5175. (6) Finkel, M. P. Relative biological effectiveness of radium and other alpha-emitters in cf no-1 female mice. Exp. Biol. Med. 1953, 83, 494− 498. (7) Hodge, C. V. a. H. Pharmacology and toxicology of uranium compounds, with a section on the pharmacology and toxicology of fluorine and hydrogen fluoride, 1st ed.; McGraw-Hill Book Co.: New York, 1949−1953. (8) Miller, A. C.; Blakely, W. F.; Livengood, D.; Whittaker, T.; Xu, J. Q.; Ejnik, J. W.; Hamilton, M. M.; Parlette, E., St; John, T.; Gerstenberg, H. M.; Hsu, H. Transformation of human osteoblast cells to the tumorigenic phenotype by depleted uranium uranyl chloride. Environ. Health Persp. 1998, 106, 465−471. (9) Tannenbaum, A. Toxicology of Uranium; McGraw-Hill Book Co.: New York, 1951. (10) Parr, R. G.; Pearson, R. G. Absolute hardness - companion parameter to absolute electronegativity. J. Am. Chem. Soc. 1983, 105, 7512−7516. (11) Pearson, R. G. Hard and soft acids and bases. J. Am. Chem. Soc. 1963, 85, 3533−3539. (12) Muller, D.; Houpert, P.; Henge Napoli, M. H.; Métivier, H.; Paquet, F. Synergie potentielle entre deux toxiques rénaux: le DTPA et l’uranium. Radioprotection 2006, 41, 413−420. (13) Basinger, M. A.; Jones, M. M. Tiron (sodium 4,5dihydroxybenzene-1,3-disulfonate) as an antidote for acute uranium intoxication in mice. Res. Commun. Chem. Pathol. 1981, 34, 351−358. (14) Lusky, L. M.; Braun, H. A. Sodium catechol disulphonate protection in experimental uranium nitrate poisoning. Fed. Proc. 1950, 9, 297−297. (15) Stradling, G. N.; Gray, S. A.; Moody, J. C.; Hodgson, A.; Raymond, K. N.; Durbin, P. W.; Rodgers, S. J.; White, D. L.; Turowski, P. N. The efficacy of dfo-hopo, dtpa-dx and dtpa for enhancing the excretion of plutonium and americium from the rat. Int. J. Radiat. Biol. 1991, 59, 1269−1277. (16) Sturzbecher-Hoehne, M.; Deblonde, G. J. P.; Abergel, R. J. Solution thermodynamic evaluation of hydroxypyridinonate chelators 3,4,3-LI(1,2-HOPO) and 5-LIO(Me-3,2-HOPO) for UO2(VI) and Th(IV) decorporation. Radiochim. Acta 2013, 101, 359−366. (17) Gorden, A. E. V.; Xu, J. D.; Raymond, K. N.; Durbin, P. Rational design of sequestering agents for plutonium and other actinides. Chem. Rev. 2003, 103, 4207−4282. (18) Durbin, P. W.; Kullgren, B.; Ebbe, S. N.; Xu, J. D.; Raymond, K. N. Chelating agents for uranium(VI): 2. Efficacy and toxicity of tetradentate catecholate and hydroxypyridinonate ligands in mice. Health Phys. 2000, 78, 511−521. (19) Durbin, P. W.; Kullgren, B.; Xu, J. D.; Raymond, K. N. New agents for in vivo chelation of Uranium(VI): Efficacy and toxicity in mice of multidentate catecholate and hydroxypyridinonate ligands. Health Phys. 1997, 72, 865−879. (20) Jager, M.; Schubert, S.; Ochrimenko, S.; Fischer, D.; Schubert, U. S. Branched and linear poly(ethylene imine)-based conjugates: synthetic modification, characterization, and application. Chem. Soc. Rev. 2012, 41, 4755−4767. (21) Masotti, A.; Giuliano, A.; Ortaggi, G. Efficient ComplexationUltrafiltration Process for Metal Ions Removal from Aqueous Solutions Using a Novel Carboxylated Polyethylenimine Derivative (PEI-COOH). Curr. Anal. Chem. 2010, 6, 37−42. (22) Jarvis, N. V.; Zeevaart, J. R.; Wagener, J. M.; Louw, W. K. A.; Dormehl, I. C.; Milner, R. J.; Killian, E. Metal-ion speciation in blood plasma incorporating the water-soluble polymer, polyethyleneimine functionalised with methylenephosphonate groups, in therapeutic radiopharmaceuticals. Radiochim. Acta 2002, 90, 237−246.

(23) Von Harpe, A.; Petersen, H.; Li, Y. X.; Kissel, T. Characterization of commercially available and synthesized polyethylenimines for gene delivery. J. Controlled Release 2000, 69, 309−322. (24) Ravel, B.; Newville, M. ATHENA, ARTEMIS, HEPHAESTUS: data analysis for X-ray absorption spectroscopy using IFEFFIT. J. Synchrotron Radiat. 2005, 12, 537−541. (25) Rehr, J. J.; Kas, J. J.; Vila, F. D.; Prange, M. P.; Jorissen, K. Parameter-free calculations of X-ray spectra with FEFF9. Phys. Chem. Chem. Phys. 2010, 12, 5503−5513. (26) Case, D. A.; Babin, V.; Berryman, J. T.; Betz, R. M.; Cai, Q.; Cerutti, D. S.; Cheatham, III, T. E.; Darden, T. A.; Duke, R. E.; Gohlke, H.; Goetz, A. W.; Gusarov, S.; Home- yer, N.; Janowski, P.; Kaus, J.; Kolossváry, I.; Kovalenko, A.; Lee, T. S.; LeGrand, S.; Luchko, T.; Luo, R.; Madej, B.; Merz, K. M.; Paesani, F.; Roe, D. R.; Roitberg, A.; Sagui, C.; Salomon-Ferrer, R.; Seabra, G.; Simmerling, C. L.; Smith, W.; Swails, J.; Walker, R. C.; Wang, J.; Wolf, R. M.; Wu, X.; Kollman, P. A. AMBER 14; University of California: San Francisco, CA, 2014. Online: http://ambermd.org/. (27) Darden, T.; York, D.; Pedersen, L. Particle mesh ewald - an n.log(n) method for ewald sums in large systems. J. Chem. Phys. 1993, 98, 10089−10092. (28) Caldwell, J. W.; Kollman, P. A. Structure and properties of neat liquids using nonadditive molecular-dynamics - water, methanol, and n-methylacetamide. J. Phys. Chem. 1995, 99, 6208−6219. (29) Meng, E. C.; Kollman, P. A. Molecular dynamics studies of the properties of water around simple organic solutes. J. Phys. Chem. 1996, 100, 11460−11470. (30) Aqvist, J. Ion water interaction potentials derived from freeenergy perturbation simulations. J. Phys. Chem. 1990, 94, 8021−8024. (31) Kamath, G.; Cao, F.; Potoff, J. J. An improved force field for the prediction of the vapor-liquid equilibria for carboxylic acids. J. Phys. Chem. B 2004, 108, 14130−14136. (32) Bayly, C. I.; Cieplak, P.; Cornell, W. D.; Kollman, P. A. A wellbehaved electrostatic potential based method using charge restraints for deriving atomic charges - the resp model. J. Phys. Chem. 1993, 97, 10269−10280. (33) Duvail, M.; Villard, A.; Nguyen, T. N.; Dufreche, J. F. Thermodynamics of Associated Electrolytes in Water: Molecular Dynamics Simulations of Sulfate Solutions. J. Phys. Chem. B 2015, 119, 11184−11195. (34) Van Horn, D. J.; Huang, H. Uranium (VI) bio-coordination chemistry from biochemical, solution and protein structural data. Coord. Chem. Rev. 2006, 250, 765−775. (35) Guide National, Intervention médicale en cas d’évènement nucléaire ou radiologique, V3.6, 2008. www.asn.fr. (36) Fritsch, P.; Serandour, A. L.; Gremy, O.; Phan, G.; Tsapis, N.; Fattal, E.; Benech, H.; Deverre, J. R.; Poncy, J. L. Structure of a single model to describe plutonium and americium decorporation by dtpa treatments. Health Phys. 2010, 99, 553−559. (37) Deng, S. B.; Ting, Y. P. Characterization of PEI-modified biomass and biosorption of Cu(II), Pb(II) and Ni,(II). Water Res. 2005, 39, 2167−2177. (38) Ghoul, M.; Bacquet, M.; Crini, G.; Morcellet, M. Novel sorbents based on silica coated with polyethylenimine and crosslinked with poly(carboxylic acid): Preparation and characterization. J. Appl. Polym. Sci. 2003, 90, 799−805. (39) Goebbert, D. J.; Garand, E.; Wende, T.; Bergmann, R.; Meijer, G.; Asmis, K. R.; Neumark, D. M. Infrared Spectroscopy of the Microhydrated Nitrate Ions NO3-(H2O)(1−6). J. Phys. Chem. A 2009, 113, 7584−7592. (40) Lucks, C.; Rossberg, A.; Tsushima, S.; Foerstendorf, H.; Scheinost, A. C.; Bernhard, G. Aqueous Uranium(VI) Complexes with Acetic and Succinic Acid: Speciation and Structure Revisited. Inorg. Chem. 2012, 51, 12288−12300. (41) Bartlett, J. R.; Cooney, R. P. On the determination of uraniumoxygen bond lengths in dioxouranium(v1) compounds by raman spectroscopy. J. Mol. Struct. 1989, 193, 295−300. H


Article

Inorganic Chemistry (42) Palacios, E. G.; Juarez-Lopez, G.; Monhemius, A. J. Infrared spectroscopy of metal carboxylates - II. Analysis of Fe(III), Ni and Zn carboxylate solutions. Hydrometallurgy 2004, 72, 139−148. (43) Denecke, M. A.; Reich, T.; Bubner, M.; Pompe, S.; Heise, K. H.; Nitsche, H.; Allen, P. G.; Bucher, J. J.; Edelstein, N. M.; Shuh, D. K. Determination of structural parameters of uranyl ions complexed with organic acids using EXAFS. J. Alloys Compd. 1998, 271, 123−127. (44) Groenewold, G. S.; Jong, W. A.; Oomens, J.; Stipdonk, M. J. Variable Denticity in Carboxylate Binding to the Uranyl Coordination Complexes. J. Am. Soc. Mass Spectrom. 2010, 21, 719−727. (45) Zhang, Y. J.; Karatchevtseva, I.; Bhadbhade, M.; Tran, T. T.; Aharonovich, I.; Fanna, D. J.; Shepherd, N. D.; Lu, K.; Li, F.; Lumpkin, G. R. Solvothermal synthesis of uranium(VI) phases with aromatic carboxylate ligands: A dinuclear complex with 4-hydroxybenzoic acid and a 3D framework with terephthalic acid. J. Solid State Chem. 2016, 234, 22−28.

I


Molecular Dynamic ... - ACS Publications

Recommend Documents