Probing SilicaâBiomolecule Interactions by Solid-State NMR and

Article pubs.acs.org/Langmuir

Probing Silica−Biomolecule Interactions by Solid-State NMR and Molecular Dynamics Simulations Stephan Ingmar Brückner,† Sergii Donets,‡ Arezoo Dianat,‡ Manfred Bobeth,‡ Rafael Gutiérrez,‡ Gianaurelio Cuniberti,*,‡,§,∥ and Eike Brunner*,† †

Chair for Bioanalytical Chemistry, Department of Chemistry and Food Chemistry, TU Dresden, 01062 Dresden, Germany Institute for Materials Science and Max Bergmann Center of Biomaterials, TU Dresden, 01062 Dresden, Germany § Dresden Center for Computational Materials Science (DCMS), TU Dresden, 01062 Dresden, Germany ∥ Center for Advancing Electronics Dresden, TU Dresden, 01062 Dresden, Germany ‡

S Supporting Information *

ABSTRACT: Understanding the molecular interactions between inorganic phases such as silica and organic material is fundamental for chromatographic applications, for tailoring silica−enzyme interactions, and for elucidating the mechanisms of biomineralization. The formation, structure, and properties of the organic/inorganic interface is crucial in this context. Here, we investigate the interaction of selectively 13C-labeled choline with 29Si-labeled monosilicic acid/silica at the molecular level. Silica/choline nanocomposites were analyzed by solid-state NMR spectroscopy in combination with extended molecular dynamics (MD) simulations to understand the silica/organic interface. Crosspolarization magic angle spinning (CP MAS)-based NMR experiments like 1 H−13C CP-REDOR (rotational-echo double resonance), 1H−13C HETCOR (heteronuclear correlation), and 1H−29Si−1H double CP are employed to determine spatial parameters. The measurement of 29Si−13C internuclear distances for selectively 13C-labeled choline provides an experimental parameter that allows the direct verification of MD simulations. Atomistic modeling using classical MD methodologies is performed using the INTERFACE force field. The modeling results are in excellent agreement with the experimental data and reveal the relevant molecular conformations as well as the nature and interplay of the interactions between the choline cation and the silica surface. Electrostatic interactions and hydrogen bonding are both important and depend strongly on the hydration level as well as the charge state of the silica surface.

■

propyl units, finally resulting in polymers with sometimes more than 20 aminopropylene repeating units.5,12,15,16 The number of repeating units, as well as the molecular structure of the positively charged LCPAs, strongly influences the silicaprecipitation behavior.17−20 Tertiary and quaternary amines are found in addition to the primary and secondary amines. The quaternary amines exhibit a permanently positive charge. They preferentially occur at the ends of LCPAs and modified lysine residues of silaffins.15,21−23 In contrast to LCPAs, the hydrated silica surface is negatively charged in the relevant pH regime.24 As a consequence, electrostatic interactions between the silica surface and the positively charged amines are likely to occur25,26 and may be one key mechanism for biosilica formation.27 On the other hand, functional groups allowing the formation of hydrogen bonds are also suggested as important.28

INTRODUCTION Interactions between silica and organic molecules are fundamental for chromatographic applications, for tailoring silica−enzyme interactions, and for elucidating the mechanisms of biomineralization. Biomineralization, i.e., the ability of organisms to form mineralized structures, is based on elaborate cellular processes that are not yet fully understood. One example is the silicification of cell walls of diatoms, unicellular algae. Diatoms create an amorphous nanocomposite material consisting of silica and various organic molecules in contrast to other biomineralization processes where crystalline phases are formed.1,2 Strongly associated biomolecules remain after sodium dodecyl sulfate (SDS)/ethylenediaminetetraacetic acid (EDTA) treatment of the cells.3,4 Several compounds were identified in the past: long-chain polyamines (LCPAs);5 peptides like silaffins,6 silacidins,7 cingulins,8 frustulins,3 and pleuralins;9 and carbohydrates like mannose-phosphate10 and chitin.11 The silica-precipitating effect of LCPAs and silaffins has been studied extensively based on in vitro studies.12−14 LCPAs are formed by elongation of putrescine with amino© XXXX American Chemical Society

Received: September 8, 2016 Revised: October 13, 2016 Published: October 19, 2016 A

DOI: 10.1021/acs.langmuir.6b03311 Langmuir XXXX, XXX, XXX−XXX

Article

Langmuir

ultrapure water to quantitatively transfer all choline molecules to the sodium metasilicate solution. After this, the pH value was adjusted to 7.0 and the solution was filled up with ultrapure water to ensure a reproducible sample volume of 3 mL. The evolving gel was centrifuged after 24 h at room temperature at 4000 rpm (2576g) for 5 min, and the centrifugate was dried for 6 days at 37 °C. The xerogel was put into a 2.5 or 3.2 mm zirconia rotor and measured by solid-state NMR. Solid-State NMR. The samples were measured using a 2.5 mm 1 H−13C/15N−29Si triple-resonance probe and a 3.2 mm HX doubleresonance probe from Bruker Biospin (Karlsruhe) on a Bruker AVANCE 800 spectrometer operating at a magnetic field of 18.8 T with a proton resonance frequency of 800.18 MHz as well as frequencies of 201.21 and 158.97 MHz for 13C and 29Si, respectively. A sample spinning rate of 16 kHz was used. During signal acquisition, proton decoupling was applied using SPINAL6457,58 at 100 kHz decoupling field strength. For REDOR experiments, the radio frequency(rf) field strengths of hard pulses were set to 55.6 kHz for carbon and 41.7 kHz for silicon, respectively. CP was accomplished using a rf field of 63.3 kHz in the carbon channel and 78.5 kHz in the proton channel with a 100 to 90 ramp on the proton channel. For this experiment, 40 spectra were collected for the reference and for the dephasing experiment with 1024 scans using the XY-8 phase cycle method with interleaved acquisition of Sr and S0 signals and 3 s interscan delay for each row. H−X−H double CP experiments were performed under MAS at 8 kHz sample spinning rate. Proton 90° pulses and spin-lock pulses with fields of 58.1 and 51.3 kHz for proton and 41.7 kHz for silicon were applied, respectively. Measurements were done using a variable contact time for the second CP step. The contact time for the first CP step was set to 5 ms; 1600 scans per increment were acquired. A line broadening of 50 Hz was applied before deconvolution of the spectra with the dmfit program “dm2011vs/rel/release #20111221”.59 The single-pulse 1H spectrum was acquired by an rf field of 58.1 kHz with an interscan delay of 3 s. An exponential window function resulting in a line broadening of 2 Hz was applied. MD Simulation. MD simulations were performed by using the molecular dynamics package NAMD 2.10.60 MD simulations were conducted in the isothermal−isobaric (NPT) ensemble for the solvated case and in the canonical (NVT) ensemble in the “dried” states. All simulations were performed with an integration time step of 2 fs at 298 K and 1 atm. The temperature was kept constant using a Langevin thermostat with a damping coefficient of 1 ps−1. To maintain a constant pressure, a Langevin piston was used as implemented in NAMD. Bond parameters and van der Waals parameters (included in the AMBER force field library61,62) as well as atomic charges for choline were obtained by the use of the RESP ESP charge derive (R.E.D.) server.63,64 For the silica substrate, we used the parameters introduced recently in the INTERFACE force field.54 Water was represented using the TIP3P model. Nonbonding interactions were cutoff at 12 Å with the SWITCH function turned off.65 Long-range electrostatic interactions were evaluated using the particle-mesh Ewald method implemented in NAMD.66 Bonds involving hydrogen and any other atom were kept rigid using the SHAKE algorithm.67 In this study, amorphous Q3 silica surfaces with an area density of silanol groups of 4.7 per nm2, and with 0% and 9% ionization, were considered.54 For hydrated samples, 9% ionization is expected at pH 7.54 The initial size of the simulation box was 43.2 × 42.5 × 100.0 Å3 for the ionized substrates and 42.7 × 42.8 × 100.0 Å3 for the nonionized substrates. Periodic boundary conditions were applied along the x-, y-, and z-directions of the simulation box. To keep the solid silica substrate during the simulation, the silica atoms 7 Å from the bottom (roughly one-third of the total thickness) were restrained to their original positions by applying a harmonic force with a force constant of 1 kcal/mol/Å2. The MD results were analyzed mainly by using the software VMD,68,69 including the preparation of the figures for this work.

Here, we have chosen choline as a simple model substance to evaluate the role of these two interactions on a less complicated system than LCPAs. The ability of phosphatidylcholine to interact with silica is well-known.29−32 The electrostatic interaction is to some extent shielded by the methyl groups, and the formation of hydrogen bonds of the phosphate (ester) group toward the surface seems to be energetically preferred.31,32 Choline is even simpler. Apart from the positively charged quaternary amine, it also contains a hydroxy group (C−OH) capable of forming hydrogen bonds with the silica surface, thus competing with the electrostatic interaction. Hence, the interactions of choline salts with silica are supposed to be affected by the two functional groups: the quaternary amine and the C−OH. In contrast to phosphatidylcholine, it is a small cation. It does not contain the fatty acids bound to the corresponding glycerol ester. These fatty acids may, however, introduce further, unwanted interactions, e.g., hydrophobic interactions resulting in self-assembly. Therefore, use of the choline cation reduces the complexity compared to phosphatidylcholine and allows us to probe the surface interactions of closely neighbored quaternary amines and hydrogen-bondforming C−OH. Solid-state NMR spectroscopy is well-suited to study biomolecule−surface interactions33−35 and allows for determining the orientation of adsorbed molecules with respect to the silica surface. One important solid-state NMR experiment for internuclear distance measurements under magic angle spinning (MAS) is the REDOR (rotational-echo double resonance) technique invented by Gullion and Schaefer.36−39 The double CP (double cross-polarization) technique40 is another distance-sensitive experiment. REDOR and double CP are well-suited to determine the orientation of biomolecules like proteins, sugars, or lipids with respect to inorganic surfaces as demonstrated previously.41−51 The NMR-derived geometry parameters can then be used to validate the results of MD simulations and vice versa.46,52,53 Such simulations are greatly facilitated by the recently developed thermodynamically consistent silica parameters compatible with the AMBER force field.54,55 This so-called INTERFACE force field reproduces bulk, surface, and interfacial properties of amorphous silica and was validated for the description of biomolecule−silica interactions in aqueous solution.55 The force field is used in our MD simulations to model the interactions between choline and silica at an atomic level of description under relevant conditions with respect to the hydration state and pH, i.e., the charge state of the involved molecular species.

■

EXPERIMENTAL SECTION

Materials. Isotopically labeled 29SiO2, 1−13C−choline chloride (99% 13C), 1−15N−choline chloride (98% 15N), and choline bromide−(methyl−13C) (99% 13C) were purchased from Cortecnet (France). Hydrochloric acid and anhydrous sodium carbonate were obtained from Sigma-Aldrich. Sample Preparation. The synthesis of isotopically labeled sodium metasilicate was performed according to the established protocol.50,56 The product was tested for purity by 29Si solid-state NMR using direct excitation. Choline halogenic salt (210 μmol) was weighed into a 1 mL tube and mixed with 250 μL of ultrapure water. For the preparation of silicic acid, 210 μmol of sodium metasilicate was inserted into a 15 mL falcon tube. After the addition of 2.1 mL of ultrapure water, the solution was vortexed at 2500 rpm. To dissolve the metasilicate completely, an ultrasonic bath was used for 3 min. Then, 150 μL of 2.4 M hydrochloric acid and the choline halogenic solution were added to the metasilicate solution. The empty tube was washed with 250 μL of B


Article

Langmuir

■

RESULTS AND DISCUSSION The 1H-MAS NMR spectra of the air-dried gels (xerogels) consisting of choline salts associated with amorphous silica (CSX) are well-resolved (see Figure 1). The signal assignment

unequivocal conclusion due to the superimposition with the signals of strongly adsorbed water. Additionally the surface ionization may also play a crucial role in context of the protonation of silica, as indicated in Figure 1. It should be noted that the signals observed by 1H−29Si−1H double CP are broader than the signals in the directly detected 1H MAS NMR spectrum. This can be explained by the fact that the double CP experiment mainly detects strongly adsorbed choline cations (see below). This would also explain the minor shift of the 1H NMR signals observed in these two different experiments. An exchange of the silanol protons with the protons from water cannot be modeled with MD, but it might be involved in the line broadening of the double CP. To decide whether or not the C−OH group of the choline is also involved in the interaction with the silica surface, we have performed 13C−29Si CP-REDOR experiments for 13C−29Si distance measurements. In the 13C−CP-MAS NMR spectra, three signals show up for choline salts like choline chloride or bromide due to the three chemically different carbon atoms of the choline molecule (see Table 1). For the choline−silica xerogels, the signal assignment Table 1. Chemical Shifts of the Different Carbon Species in Choline Salts 1 HO carbon species CH2 CH2 CH3

Figure 1. (Top) Single-pulse 1H-MAS NMR spectrum of 1−13C−CSX exhibiting signals at 4.9 ppm for the proton of the alcohol and 4.2, 3.6, and 3.3 ppm for the C1-methylene group adjacent to the alcohol, the C2-methylene group, and the methyl group bound to the nitrogen, respectively. (Bottom) 1H−29Si−1H double CP spectrum of 1−13C− CSX together with the signal decomposition for 5250 μs contact time. The peaks of δ(OH + H2O) (4.9 ppm) and δ(CH3) (3.4 ppm) indicate close contact of the strongly adsorbed water and/or alcohol protons and the protons of the methyl group with the silica surface.

CH2

2 position 1 2 3

3

CH2

N(Me)3 chemical shift/ppm 56.6 68.6 55.0

was performed using singly isotopically labeled samples, which are commercially available. Two different samples with 13Cenriched carbons at position 1 (C−OH) and position 3 (methyl groups) were used. The xerogels obtained from these samples are denoted as 1−13C−CSX and Me−13C−CSX, respectively. Interestingly, two different signals are observed for the 1−13C position in 1−13C−CSX (see Figure 2). The increased chemical

is given in Figure 1 (top). Note that the signal at 4.9 ppm cannot be solely due to the OH protons of choline because its intensity is by far too high compared with the CH3 protons. Therefore, it is concluded that this signal also contains a significant contribution from strongly adsorbed H2O molecules even in this dried state (6 days at 37 °C). 1H−29Si−1H double CP experiments were performed in order to detect spatial proximities between species at silica/organic interfaces.26 The corresponding spectra mainly exhibit the two signals from CH3 at the quaternary nitrogen and OH (C−OH and H2O). In contrast to CH3 and OH, the weak signals of the N- and Obound methylene groups are hardly separated from the dominating signals (see also Supporting Information, Figure S1). The intensities of the two observed signals, CH3 and OH, were used to measure the buildup curve for the second CP step (29Si → 1H; see Supporting Information, Figure S2). The buildup time constant TCP was determined for short contact times (450 μs ≤ tcontact ≤ 1450 μs) by linear approximation (see Supporting Information). For the OH protons, the buildup time TSI(OH) = 2.15 ms is slightly shorter than the value of TSI(CH3) = 2.75 ms for the protons corresponding to the three methyl groups. These observations indicate that both OH and methyl protons at the positively charged quaternary nitrogen are in close contact with silica. With respect to the C−OH group, however, the 1H−29Si−1H double CP does not allow an

Figure 2. Rotor synchronized 1H−13C CP Hahn echo spectra of methyl−13C−choline bromide silica xerogel (methyl−13C−CSX) and 1−13C−choline chloride silica xerogel (1−13C−CSX). Number of scans: 400 and 1024, respectively; recycle delay: 3 s; line broadening: 15 Hz. Fitted curve (using dmfit program “dm2011vs/rel/release #20111221”59) for the adsorbed species displayed in black. C


Article

Langmuir shift of the weaker signal at 58.6 ppm compared to the value for the intense signal at 56.6 ppm indicates a stronger adsorption than for the former signal. This assumption is based on earlier observations for carboxyl groups of amino acids where 13C NMR chemical shift changes of a comparable order of magnitude were found due to hydrogen-bond formation.70,71 13 C−29Si CP-REDOR experiments were performed for 1−13C− CSX (see Figure 3). The highest REDOR fraction occurs for

Table 2. Mean Values with Standard Deviations of the Smallest C1−Si and N−Si Distances and the Total Time of Occurence of Hydrogen Bonds for Different States with Different Ionization and Residual Water dried state IA (residual water, 9% ionization)a

dried state IB (residual water, 0% ionization)a

distances (Å) C1−Si N−Si hydrogen bonds donor acceptor choline silica silica choline choline water water choline a

4.34 ± 0.28 4.63 ± 0.17

4.53 ± 0.41 4.86 ± 0.22

time (%) 74.91 0.34

time (%) 20.24 11.16 36.15 0.02

0.30

Area density of water molecules ≤0.54 H2O/nm2

The above-described 1H−29Si−1H double CP experiments indicated also a close contact between the methyl groups and the silica surface. To further substantiate this observation, we have also performed 1H−13C−29Si CP-REDOR experiments on Me−13C−CSX. Measuring REDOR fractions up to a dephasing time of 7.5 ms, a 13C−29Si second moment of 9.1 × 103 Hz2 was determined. Note that the 13C signal is caused by three methyl groups, which cannot be all in contact with the silica surface. They may also perform rotations around the C3-axis. Therefore, a distance determination is questionable. In that context, a 1H−15N−29Si CP-REDOR experiment would deliver, in principle, more precise information. However, the corresponding experiments did not show any detectable dephasing (see Supporting Information, Figure S5). Computational Studies of Choline−Silica Surface Interactions. Atomistic molecular dynamics (MD) simulations are increasingly important for understanding and interpreting the interactions and formation mechanisms governing silica biomineralization.74 One main advantage of classical MD is that such simulations can describe the adsorption of biomolecules on silica surfaces at the molecular level, i.e., with atomistic resolution. The interaction of choline with silica in the presence and absence of water and in dependence on the silica surface charge is modeled within this section and compared with the experimental data shown above. Because the NMR measurements were carried out on airdried samples, we have performed simulations for a dried state I that still contains residual water (see above) assuming a residual amount of water corresponding to an area density ≤0.54 H2O/ nm2 as previously proposed.45 Hydrated silica surfaces are partially ionized because the surface SiOH groups dissociate depending on the pH. At pH 7, we anticipate ∼9% dissocation for surface SiOH.54 This is also assumed here for the simulations and will be denoted as dried state IA. However, the air-dried sample with only residual amounts of water may exhibit a lower degree of surface SiOH dissociation because completely water-free surfaces are expected to exhibit 0% SiOH dissociation. To evaluate this influence, we have also performed simulations for the dried state in the presence of residual water and 0% surface SiOH dissociation (dried state IB). Moreover, the fully hydrated state (55 mol/L H2O, 9% SiOH dissociation) and the fully dried state (water-free, 0% SiOH dissociation) were simulated for comparison (see Supporting Information). Let us first consider the situation with 9% ionization and a residual amount of water. As follows from the inset in Figure

Figure 3. REDOR fractions (ΔS /S0) of 1−13C−CSX for the signals at 58.6 ppm (tightly bond state) and 56.6 ppm (loosely bound state).

the weaker signal at ∼58.6 ppm, corresponding to a second moment of 9.1 × 104 Hz2. In a two-spin-approximation,72 this value would correspond to a 13C−29Si distance of 3.8 Å. However, because it is likely that the 13C nucleus interacts with more than one 29Si, this value only constitutes a lower bound for the 13C−29Si distance. In any case, the observation of a dipolar contact is in agreement with the assumption that hydrogen bonding is likely to occur, causing the low field shift of the tightly adsorbed species of the carbon at the C−OH (see above). A more elaborate analysis is provided below, based on the results of our computational studies. The intense signal at 56.6 ppm exhibits a smaller REDOR fraction corresponding to a second moment of 2.0 × 104 Hz2 and a 13C−29Si distance of 4.9 Å. This observation supports the hypothesis that the carbon signal at ∼58.6 ppm is caused by 1−13C-carbons of tightly silica-attached choline species. Nevertheless, tightly and loosely bound choline species are both in detectable contact with the silica surface. The second moment observed for the loosely bound choline species could also be influenced by its higher mobility. The observation of the two different adsorption states may be caused by the degree of surface ionization and the choline silica ratio of the solution before the formation of the gel. The degree of surface SiOH dissociation in the relevant pH regime is the subject of various publications.54,73 In general, it is concluded that only a limited number is dissociated (9% is given in ref 54). The number of ionized silanol groups is thus probably much lower than the number of choline molecules present. The MD simulations predict that the interaction of choline with ionized SiO− moieties is stronger than with nonionized SiOH (see Table 2). Therefore, it is tempting to speculate that the signal at 58.6 ppm might be due to choline on SiO− moieties. D


Article

Langmuir

Figure 4. (A) Probability distributions of smallest C1−Si and N−Si distances for an amorphous Q3 silica surface with 9% ionization and residual amount of water. (B) Normalized 2D projection of the smallest choline (center of mass)−silicon distances onto the silica surface. (C) Partial charges of atoms for amorphous Q3 silica with 9% ionization are illustrated. Atom labeling: red, silanol, siloxide, and bulk oxygen atoms; gray, hydrogen; blue, silanol, siloxide, and bulk silicon atoms.

Figure 5. Hydrogen bonding between choline and (A) a negatively charged deprotonated Si−O− moiety and (B) an interfacial water molecule in the presence of a residual amount of water. Only nearby water molecules are shown. The area density of water molecules in both cases is ≤0.54 H2O/ nm2.

4A, the relative orientation of choline at the surface is hardly changing during the simulation period. Only minor variations of the C1−Si and N−Si internuclear distances occur, indicating a well-defined adsorption state. The mean values for the distances are 4.34 ± 0.28 Å for C1−Si and 4.63 ± 0.17 Å for N−Si (see Table 2). In addition, both distributions are more symmetric compared to the hydrated state (see Supporting Information) and overlap with each other. This indicates an adsorbed state where the choline is oriented almost parallel with respect to the silica surface. The distribution of the partial charges at the amorphous silica surface for the state with a residual amount of water and 9% ionization is shown in Figure 4C. Various simulations with different initial configurations of choline relative to the silica surface always lead to an adsorbed state of choline in the vicinity of SiO−, i.e., an ionized silanol group as shown in Figure 4C (more-intense red color). Silanols are obviously the preferred interaction sites. Both hydrogen bonding and electrostatic attraction by the same charged silanol Si−O−

moiety contribute to surface binding of choline. Compared with the fully hydrated state, the influence of electrostatic interactions increases (cf. Supporting Information, Figure S6 and Table S1). This can be explained by the absence of a hydration shell around the quaternary nitrogen. Therefore, the methylated amino group is able to come closer to the silica surface than in the hydrated state. An analysis of the presence of hydrogen bonds for the different considered states is also provided in Table 2. For the dried state IA with residual water and 9% surface ionization, hydrogen bonding between the C− OH group of choline and siloxide/silanol groups of amorphous silica is observed for ∼75% of the total simulation time, that is, hydrogen bonding between choline and a negatively charged deprotonated Si−O− species is characteristic for this state (see Figure 5A). A significant change occurs, however, for nonionized surfaces. The choline C−OH group then interacts with the residual water by forming hydrogen bonds for >36% of the time (see Table 2 and Figure 5B). Moreover, the C1−Si and N−Si mean distances and their standard deviations are larger E


Article

Langmuir

example, found no interaction of the choline headgroup with the silica surface, which might be due to the above-mentioned differences between DPPC and choline. Our experimental observations are in good agreement with the MD simulations, which also reveal that especially the hydrogen-bond formation via the C−OH groups significantly depends on the hydration state as well as on the degree of ionization of the surface. In the dried state with partial ionization (dried state IA), the choline molecule is tightly bound and shows strong interactions with the surface siloxide/ silanol groups via hydrogen bonding. In the absence of dissociated silanol groups, interactions with the lone-pair electrons of the oxygen atoms of the undissociated Si−OH dominate. The C−OH group of choline is then less important for silica binding.

for 0% surface SiOH ionization, as can be seen in Table 2. This indicates that the choline molecule is less tightly bound to the surface. Nevertheless, choline remains in close contact with the surface and prefers the silanols due to the interaction between the positively charged quaternary ammonium and the lone pair electrons of the oxygen atoms of Si−OH groups. Complete water removal and 0% ionization practically led to complete disappearance of hydrogen bonding (see Supporting Information, Table S1). The resulting changes in the predicted C1−Si and N−Si distances are relatively small, however. For the two dried states IA and IB, which are relevant with respect to the samples under study, we have calculated the second moments M2 for 13C−29Si interactions. In these calculations, we have included the four next-nearest 29Si as predicted by the described MD simulations (cf. Figures S7, S8, and S9). It turned out that the distances predicted for dried state IA result in an M2 value of 11.0 × 104 Hz2. This is close to the value of 9.1 × 104 Hz2 measured for strongly silica-attached choline in the REDOR experiment. The distances obtained for dried state IB result in a second moment of 8.7 × 104 Hz2. Therefore, we conclude that the experimentally observed tightly silica-attached choline molecules giving rise to the 13C1 signal at 58.6 ppm correspond to choline situated in a state similar to the predictions of the MD simulations for dried states IA and IB.

■

ASSOCIATED CONTENT

S Supporting Information *

The Supporting Information is available free of charge on the ACS Publications website at DOI: 10.1021/acs.langmuir.6b03311. Double CP, HETCOR, and REDOR spectra and additional details on MD simulations (PDF)

■

■

CONCLUSION The interactions at the silica/organic interface of silica/choline nanocomposites were studied by an approach combining MD simulations with solid-state NMR spectroscopy. The MD simulations have revealed the following: (i) Both electrostatic interactions and hydrogen bonding are important for the adsorption of choline at silica surfaces. Silanol groups are the preferred interaction sites. (ii) Hydrogen bonding between the C−OH group and SiO− sites is predominating, especially for strongly adsorbed choline. (iii) Water removal results in an increasing influence of electrostatic interactions. (iv) The adsorption state depends on the silanol dissociation, i.e., surface ionization. Weaker adsorption is observed for nonionized surfaces. The predictions (i) and (ii) obtained from the MD simulations are verified by solid-state NMR spectroscopy on air-dried samples. The 1H−29Si−1H double-CP and 1H−13C− {29Si} CP-REDOR experiments reveal a contact of the methyl groups with the surface. This observation shows the relevance of electrostatic interactions in agreement with the simulations. In terms of energy, the interaction between silica and the positively charged, methylated nitrogen is the leading interaction (see Supporting Information, Figure S10). The positive charge is capable of interacting with partially negatively charged sites of the bulk silica, thus directing the choline molecule toward preferred interaction sites, ionized silanol groups. Interactions between the C−OH group and the silica are also important despite the lower contribution to the total interaction energies. The 1H−13C−{29Si} CP-REDOR experiments show that the C−OH group is in close contact with the silica surface. Compared to other publications30,32 where phosphatidylcholine was investigated, our study used the simpler choline cation as a model. It shows a different sterical situation, a lower number of possible hydrogen bonds, and another partial charge by using an alcohol as the hydrogenbonding functional group in contrast to a phosphate ester of dipalmitoylphosphatidylcholine (DPPC). Folliet et al.32, for

AUTHOR INFORMATION

Corresponding Authors

*E-mail: [email protected]. Phone: +49 (0)351 463 32631. Fax: +49 (0)351 463 37188. *E-mail: [email protected]. Phone: +49 (0)351 463 32631. Fax: +49 (0)351 463 37188. Notes

The authors declare no competing financial interest.

■

ACKNOWLEDGMENTS The authors thank the Deutsche Forschungsgemeinschaft (FOR 2038: Nanopatterned Organic Matrices in Biological Silica Mineralization) for financial support and the Center for Information Services and High Performance Computing (ZIH) at TU Dresden for computational resources. The authors also thank Prof. C. C. Perry (Nottingham, U.K.) for providing the INTERFACE force field.

■

REFERENCES

(1) Frankel, R. B.; Blakemore, R.; Wolfe, R. S. Magnetite in Freshwater Magnetotactic Bacteria. Science 1979, 203, 1355−1356. (2) Bazylinski, D. A.; Frankel, R. B. Magnetosome formation in prokaryotes. Nat. Rev. Microbiol. 2004, 2, 217−230. (3) Kröger, N.; Bergsdorf, C.; Sumper, M. Frustulins: Domain Conservation in a Protein Family Associated with Diatom Cell Walls. Eur. J. Biochem. 1996, 239, 259−264. (4) Grö g er, C.; Lutz, K.; Brunner, E. NMR studies of biomineralisation. Prog. Nucl. Magn. Reson. Spectrosc. 2009, 54, 54−68. (5) Kröger, N.; Deutzmann, R.; Bergsdorf, C.; Sumper, M. Speciesspecific polyamines from diatoms control silica morphology. Proc. Natl. Acad. Sci. U. S. A. 2000, 97, 14133−14138. (6) Kröger, N.; Lorenz, S.; Brunner, E.; Sumper, M. Self-Assembly of Highly Phosphorylated Silaffins and Their Function in Biosilica Morphogenesis. Science 2002, 298, 584−586. (7) Richthammer, P.; Börmel, M.; Brunner, E.; van Pée, K.-H. Biomineralization in Diatoms: The Role of Silacidins. ChemBioChem 2011, 12, 1362−1366. (8) Scheffel, A.; Poulsen, N.; Shian, S.; Kröger, N. Nanopatterned protein microrings from a diatom that direct silica morphogenesis. Proc. Natl. Acad. Sci. U. S. A. 2011, 108, 3175−3180. F


Article

Langmuir

Dominant Role of Polyethylene Glycol-Induced Hydrogen Bonding. J. Am. Chem. Soc. 2014, 136, 4236−4244. (29) Chunbo, Y.; Daqing, Z.; Aizhuo, L.; Jiazuan, N. A NMR Study of the Interaction of Silica with Dipalmitoylphosphatidylcholine Liposomes. J. Colloid Interface Sci. 1995, 172, 536−538. (30) Murray, D. K.; Harrison, J. C.; Wallace, W. E. A 13C CP/MAS and 31P NMR study of the interactions of dipalmitoylphosphatidylcholine with respirable silica and kaolin. J. Colloid Interface Sci. 2005, 288, 166−170. (31) Snyder, J. A.; Madura, J. D. Interaction of the Phospholipid Head Group with Representative Quartz and Aluminosilicate Structures: An Ab initio Study. J. Phys. Chem. B 2008, 112, 7095− 7103. (32) Folliet, N.; Roiland, C.; Bégu, S.; Aubert, A.; Mineva, T.; Goursot, A.; Selvaraj, K.; Duma, L.; Tielens, F.; Mauri, F.; Laurent, G.; Bonhomme, C.; Gervais, C.; Babonneau, F.; Azaïs, T. Investigation of the Interface in Silica-Encapsulated Liposomes by Combining Solid State NMR and First Principles Calculations. J. Am. Chem. Soc. 2011, 133, 16815−16827. (33) Roehrich, A.; Drobny, G. Solid-State NMR Studies of Biomineralization Peptides and Proteins. Acc. Chem. Res. 2013, 46, 2136−2144. (34) Shaw, W. J. Solid-state NMR studies of proteins immobilized on inorganic surfaces. Solid State Nucl. Magn. Reson. 2015, 70, 1−14. (35) Ravera, E.; Martelli, T.; Geiger, Y.; Fragai, M.; Goobes, G.; Luchinat, C. Biosilica and bioinspired silica studied by solid-state NMR. Coord. Chem. Rev. 2016, DOI: 10.1016/j.ccr.2016.06.003. (36) Gullion, T.; Schaefer, J. Rotational - echo double - resonance NMR. J. Magn. Reson. (1969-1992) 1989, 81, 196−200. (37) Jaroniec, C. P.; Tounge, B. A.; Rienstra, C. M.; Herzfeld, J.; Griffin, R. G. Measurement of 13C-15N Distances in Uniformly 13C Labeled Biomolecules: J-Decoupled REDOR. J. Am. Chem. Soc. 1999, 121, 10237−10238. (38) Jaroniec, C. P.; Tounge, B. A.; Herzfeld, J.; Griffin, R. G. Frequency Selective Heteronuclear Dipolar Recoupling in Rotating Solids: Accurate 13C−15N Distance Measurements in Uniformly 13 15 C, N-labeled Peptides. J. Am. Chem. Soc. 2001, 123, 3507−3519. (39) Mehta, A. K.; Schaefer, J. Rotational - echo double resonance of uniformly labeled 13C clusters. J. Magn. Reson. 2003, 163, 188−191. (40) Beusen, D. D.; McDowell, L. M.; Slomczynska, U.; Schaefer, J. Solid-State Nuclear Magnetic Resonance Analysis of the Conformation of an Inhibitor Bound to Thermolysin. J. Med. Chem. 1995, 38, 2742− 2747. (41) Ndao, M.; Ash, J. T.; Breen, N. F.; Goobes, G.; Stayton, P. S.; Drobny, G. P. A 13C{31P} REDOR NMR Investigation of the Role of Glutamic Acid Residues in Statherin- Hydroxyapatite Recognition. Langmuir 2009, 25, 12136−12143. (42) Masica, D. L.; Ash, J. T.; Ndao, M.; Drobny, G. P.; Gray, J. J. Toward a Structure Determination Method for Biomineral-Associated Protein Using Combined Solid- State NMR and Computational Structure Prediction. Structure 2010, 18, 1678−1687. (43) Ben Shir, I.; Kababya, S.; Amitay-Rosen, T.; Balazs, Y. S.; Schmidt, A. Molecular Level Characterization of the Inorganic Bioorganic Interface by Solid State NMR: Alanine on a Silica Surface, a Case Study. J. Phys. Chem. B 2010, 114, 5989−5996. (44) Ben Shir, I.; Kababya, S.; Schmidt, A. Binding Specificity of Amino Acids to Amorphous Silica Surfaces: Solid-State NMR of Glycine on SBA-15. J. Phys. Chem. C 2012, 116, 9691−9702. (45) Ben Shir, I.; Kababya, S.; Schmidt, A. Molecular Details of Amorphous Silica Surfaces Determine Binding Specificity to Small Amino Acids. J. Phys. Chem. C 2014, 118, 7901−7909. (46) Jayanthi, S.; Kababya, S.; Schmidt, A.; Vega, S. Deuterium MAS NMR and Local Molecular Dynamic Model to Study AdsorptionDesorption Kinetics of a Dipeptide at the Inner Surfaces of SBA-15. J. Phys. Chem. C 2016, 120, 2797−2806. (47) Christiansen, S. C.; Hedin, N.; Epping, J. D.; Janicke, M. T.; del Amo, Y.; Demarest, M.; Brzezinski, M.; Chmelka, B. F. Sensitivity considerations in polarization transfer and filtering using dipole-dipole couplings: Implications for biomineral systems. Solid State Nucl. Magn.

(9) Kröger, N.; Wetherbee, R. Pleuralins are Involved in Theca Differentiation in the Diatom Cylindrotheca f usiformis. Protist 2000, 151, 263−273. (10) Hedrich, R.; Machill, S.; Brunner, E. Biomineralization in diatoms-phosphorylated saccharides are part of Stephanopyxis turris biosilica. Carbohydr. Res. 2013, 365, 52−60. (11) Brunner, E.; Richthammer, P.; Ehrlich, H.; Paasch, S.; Simon, P.; Ueberlein, S.; van Pée, K.-H. Chitin-Based Organic Networks: An Integral Part of Cell Wall Biosilica in the Diatom Thalassiosira pseudonana. Angew. Chem., Int. Ed. 2009, 48, 9724−9727. (12) Sumper, M.; Lorenz, S.; Brunner, E. Biomimetic Control of Size in the Polyamine-Directed Formation of Silica Nanospheres. Angew. Chem., Int. Ed. 2003, 42, 5192−5195. (13) Brunner, E.; Lutz, K.; Sumper, M. Biomimetic synthesis of silica nanospheres depends on the aggregation and phase separation of polyamines in aqueous solution. Phys. Chem. Chem. Phys. 2004, 6, 854−857. (14) Jantschke, A.; Spinde, K.; Brunner, E. Electrostatic interplay: The interaction triangle of polyamines, silicic acid, and phosphate studied through turbidity measurements, silicomolybdic acid test, and 29 Si NMR spectroscopy. Beilstein J. Nanotechnol. 2014, 5, 2026−2035. (15) Bridoux, M. C.; Annenkov, V. V.; Keil, R. G.; Ingalls, A. E. Widespread distribution and molecular diversity of diatom frustule bound aliphatic long chain polyamines (LCPAs) in marine sediments. Org. Geochem. 2012, 48, 9−20. (16) Bridoux, M. C.; Keil, R. G.; Ingalls, A. E. Analysis of natural diatom communities reveals novel insights into the diversity of long chain polyamine (LCPA) structures involved in silica precipitation. Org. Geochem. 2012, 47, 9−21. (17) Belton, D.; Patwardhan, S. V.; Perry, C. C. Putrescine homologues control silica morphogenesis by electrostatic interactions and the hydrophobic effect. Chem. Commun. 2005, 3475−3477. (18) Belton, D. J.; Patwardhan, S. V.; Perry, C. C. Spermine, spermidine and their analogues generate tailored silicas. J. Mater. Chem. 2005, 15, 4629−4638. (19) Belton, D. J.; Patwardhan, S. V.; Annenkov, V. V.; Danilovtseva, E. N.; Perry, C. C. From biosilicification to tailored materials: Optimizing hydrophobic domains and resistance to protonation of polyamines. Proc. Natl. Acad. Sci. U. S. A. 2008, 105, 5963−5968. (20) Wieneke, R.; Bernecker, A.; Riedel, R.; Sumper, M.; Steinem, C.; Geyer, A. Silica precipitation with synthetic silaffin peptides. Org. Biomol. Chem. 2011, 9, 5482−5486. (21) Wenzl, S.; Deutzmann, R.; Hett, R.; Hochmuth, E.; Sumper, M. Quaternary Ammonium Groups in Silica-Associated Proteins. Angew. Chem., Int. Ed. 2004, 43, 5933−5936. (22) Sumper, M.; Brunner, E.; Lehmann, G. Biomineralization in diatoms: Characterization of novel polyamines associated with silica. FEBS Lett. 2005, 579, 3765−3769. (23) Kotzsch, A.; Pawolski, D.; Milentyev, A.; Shevchenko, A.; Scheffel, A.; Poulsen, N.; Shevchenko, A.; Kröger, N. Biochemical Composition and Assembly of Biosilica-associated Insoluble Organic Matrices from the Diatom Thalassiosira pseudonana. J. Biol. Chem. 2016, 291, 4982−4997. (24) Behrens, S. H.; Grier, D. G. The charge of glass and silica surfaces. J. Chem. Phys. 2001, 115, 6716−6721. (25) Goloub, T. P.; Koopal, L. K.; Bijsterbosch, B. H.; Sidorova, M. P. Adsorption of Cationic Surfactants on Silica. Surface Charge Effects. Langmuir 1996, 12, 3188−3194. (26) Baccile, N.; Laurent, G.; Bonhomme, C.; Innocenzi, P.; Babonneau, F. Solid-State NMR Characterization of the SurfactantSilica Interface in Templated Silicas: Acidic versus Basic Conditions. Chem. Mater. 2007, 19, 1343−1354. (27) Jantschke, A.; Koers, E.; Mance, D.; Weingarth, M.; Brunner, E.; Baldus, M. Insight into the Supramolecular Architecture of Intact Diatom Biosilica from DNP-Supported Solid-State NMR Spectroscopy. Angew. Chem., Int. Ed. 2015, 54, 15069−15073. (28) Preari, M.; Spinde, K.; Lazic, J.; Brunner, E.; Demadis, K. D. Bioinspired Insights into Silicic Acid Stabilization Mechanisms: The G


Article

Langmuir Reson. 2006, 29, 170−182. Special Issue honoring Prof. Alexander Pines on his 60th birthday. (48) Rawal, A.; Smith, B. J.; Athens, G. L.; Edwards, C. L.; Roberts, L.; Gupta, V.; Chmelka, B. F. Molecular Silicate and Aluminate Species in Anhydrous and Hydrated Cements. J. Am. Chem. Soc. 2010, 132, 7321−7337. (49) Sangodkar, R. P.; Smith, B. J.; Gajan, D.; Rossini, A. J.; Roberts, L. R.; Funkhouser, G. P.; Lesage, A.; Emsley, L.; Chmelka, B. F. Influences of Dilute Organic Adsorbates on the Hydration of LowSurface-Area Silicates. J. Am. Chem. Soc. 2015, 137, 8096−8112. (50) Wisser, D.; Brückner, S. I.; Wisser, F. M.; Althoff-Ospelt, G.; Getzschmann, J.; Kaskel, S.; Brunner, E. 1H-13C-29Si triple resonance and REDOR solid-state NMR - A tool to study interactions between biosilica and organic molecules in diatom cell walls. Solid State Nucl. Magn. Reson. 2015, 66−67, 33−39. (51) Ndao, M.; Goobes, G.; Emani, P. S.; Drobny, G. P. A REDOR ssNMR Investigation of the Role of an N-Terminus Lysine in R5 Silica Recognition. Langmuir 2016, DOI: 10.1021/acs.langmuir.5b04114. (52) Tielens, F.; Gervais, C.; Lambert, J. F.; Mauri, F.; Costa, D. Ab Initio Study of the Hydroxylated Surface of Amorphous Silica: A Representative Model. Chem. Mater. 2008, 20, 3336−3344. (53) Rimola, A.; Costa, D.; Sodupe, M.; Lambert, J.-F.; Ugliengo, P. Silica Surface Features and Their Role in the Adsorption of Biomolecules: Computational Modeling and Experiments. Chem. Rev. 2013, 113, 4216−4313. (54) Emami, F. S.; Puddu, V.; Berry, R. J.; Varshney, V.; Patwardhan, S. V.; Perry, C. C.; Heinz, H. Force Field and a Surface Model Database for Silica to Simulate Interfacial Properties in Atomic Resolution. Chem. Mater. 2014, 26, 2647−2658. (55) Emami, F. S.; Puddu, V.; Berry, R. J.; Varshney, V.; Patwardhan, S. V.; Perry, C. C.; Heinz, H. Prediction of Specific Biomolecule Adsorption on Silica Surfaces as a Function of pH and Particle Size. Chem. Mater. 2014, 26, 5725−5734. (56) Schwarz, R.; Menner, E. Zur Kenntnis der Kieselsäuren (I.). Ber. Dtsch. Chem. Ges. B 1924, 57, 1477−1481. (57) Fung, B.; Khitrin, A.; Ermolaev, K. An Improved Broadband Decoupling Sequence for Liquid Crystals and Solids. J. Magn. Reson. 2000, 142, 97−101. (58) Bräuniger, T.; Wormald, P.; Hodgkinson, P. Improved Proton Decoupling in NMR Spectroscopy of Crystalline Solids Using the SPINAL-64 Sequence. Monatsh. Chem. 2002, 133, 1549−1554. (59) Massiot, D.; Fayon, F.; Capron, M.; King, I.; Le Calvé, S.; Alonso, B.; Durand, J.-O.; Bujoli, B.; Gan, Z.; Hoatson, G. Modelling one- and two-dimensional solid-state NMR spectra. Magn. Reson. Chem. 2002, 40, 70−76. (60) Phillips, J. C.; Braun, R.; Wang, W.; Gumbart, J.; Tajkhorshid, E.; Villa, E.; Chipot, C.; Skeel, R. D.; Kalé, L.; Schulten, K. Scalable molecular dynamics with NAMD. J. Comput. Chem. 2005, 26, 1781− 1802. (61) Wang, J.; Cieplak, P.; Kollman, P. A. How well does a restrained electrostatic potential (RESP) model perform in calculating conformational energies of organic and biological molecules? J. Comput. Chem. 2000, 21, 1049−1074. (62) Hornak, V.; Abel, R.; Okur, A.; Strockbine, B.; Roitberg, A.; Simmerling, C. Comparison of multiple Amber force fields and development of improved protein backbone parameters. Proteins: Struct., Funct., Genet. 2006, 65, 712−725. (63) Vanquelef, E.; Simon, S.; Marquant, G.; Garcia, E.; Klimerak, G.; Delepine, J. C.; Cieplak, P.; Dupradeau, F.-Y. R.E.D. Server: a web service for deriving RESP and ESP charges and building force field libraries for new molecules and molecular fragments. Nucleic Acids Res. 2011, 39, W511−W517. (64) Wang, F.; Becker, J.; Cieplak, P.; Dupradeau, F.-Y. R.E.D. Python: Object oriented programming for Amber force fields. 247th American Chemical Society national meeting 2014, 296. (65) Bhandarkar, M. et al. NAMD User’s Guide; Theoretical and Computational Biophysics Group, Beckman Institute, University of Illinois: Urbana, IL, 2015.

(66) 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. (67) Ryckaert, J.-P.; Ciccotti, G.; Berendsen, H. J. Numerical integration of the cartesian equations of motion of a system with constraints: molecular dynamics of n-alkanes. J. Comput. Phys. 1977, 23, 327−341. (68) Humphrey, W.; Dalke, A.; Schulten, K. VMD − Visual Molecular Dynamics. J. Mol. Graphics 1996, 14, 33−38. (69) Stone, J. An Efficient Library for Parallel Ray Tracing and Animation. M.Sc. Thesis, Computer Science Department, University of MissouriRolla, 1998. (70) Asakawa, N.; Kuroki, S.; Kurosu, H.; Ando, I.; Shoji, A.; Ozaki, T. Hydrogen-bonding effect on carbon-13 NMR chemical shifts of Lalanine residue carbonyl carbons of peptides in the solid state. J. Am. Chem. Soc. 1992, 114, 3261−3265. (71) Gu, Z.; Zambrano, R.; McDermott, A. Hydrogen Bonding of Carboxyl Groups in Solid-State Amino Acids and Peptides: Comparison of Carbon Chemical Shielding, Infrared Frequencies, and Structures. J. Am. Chem. Soc. 1994, 116, 6368−6372. (72) Bertmer, M.; Eckert, H. Dephasing of spin echoes by multiple heteronuclear dipolar interactions in rotational echo double resonance NMR experiments. Solid State Nucl. Magn. Reson. 1999, 15, 139−152. (73) Pfeiffer-Laplaud, M.; Costa, D.; Tielens, F.; Gaigeot, M.-P.; Sulpizi, M. Bimodal Acidity at the Amorphous Silica/Water Interface. J. Phys. Chem. C 2015, 119, 27354−27362. (74) Belton, D. J.; Deschaume, O.; Perry, C. C. An overview of the fundamentals of the chemistry of silica with relevance to biosilicification and technological advances. FEBS J. 2012, 279, 1710−1720.

H


Probing SilicaâBiomolecule Interactions by Solid-State NMR and

Recommend Documents

Probing SilicaâBiomolecule Interactions by Solid-State NMR and