Exploring the Dynamic Behaviors and Transport Properties of Gas

Nov 18, 2013 - The dynamic behaviors and transport properties of O2, CO2, and NH3 molecules through a transmembrane cyclic peptide nanotube (CPNT) of ...
2 downloads 14 Views 2MB Size
Download PDF
Article pubs.acs.org/JPCB

Exploring the Dynamic Behaviors and Transport Properties of Gas Molecules in a Transmembrane Cyclic Peptide Nanotube Rui Li, Jianfen Fan,* Hui Li, Xiliang Yan, and Yi Yu College of Chemistry, Chemical Engineering and Materials Science, Soochow University, Suzhou 215123, People’s Republic of China ABSTRACT: The dynamic behaviors and transport properties of O2, CO2, and NH3 molecules through a transmembrane cyclic peptide nanotube (CPNT) of 8×cyclo-(WL)4/POPE have been investigated by steered molecular dynamics (SMD) simulations and adaptive biasing force (ABF) samplings. Different external forces are needed for three gas molecules to enter the channel. The periodic change of the pulling force curve for a gas traveling through the channel mainly arises from the regular and periodic arrangement of the composed CP subunits of the CPNT. Radial distribution functions (RDFs) between gas and water disclose the density decrease of channel water, which strongly aggravates the discontinuity of H-bond formation between a gas molecule and the neighboring water. Compared to hardly any H-bond formation between CO2 (or O2) and the framework of the CPNT, NH3 can form abundant H-bonds with the carbonyl/amide groups of the CPNT, leading to a fierce competition to NH3−water H-bonded interactions. In addition to direct H-bonded interactions, all three gases can form water bridges with the tube. The potential profile of mean force coincides with the occurring probability of a gas molecule along the tube axis. The energy barriers at two mouths of the CPNT elucidate the phenomenon that CO2 and O2 are thoroughly confined in the narrow lumen while NH3 can easily go outside the tube. Intermolecular interactions of each gas with channel water and the CPNT framework and the formation of H-bonds and water bridges illuminate the different gas translocation behaviors. The results uncover interesting and comprehensive mechanisms underlying the permeation characteristics of three gas molecules traveling through a transmembrane CPNT.

■

found that CO2 molecules have natural tendencies to fill carbon nanotubes (CNTs),19 and the transport diffusivities of CO2 in CNTs with the diameters ranging from ∼1 to 5 nm are roughly independent of pressure.20 A modified CNT with a shrunk neck in the middle exhibits high transport resistance to N2 while allowing O2 to pass.21,22 The high-frequency oscillation regime for Ne traveling inside a boron-nitride nanotube (BNNT) was found to be possible for a large range of temperatures.23 The size, curvature, and chirality of a silicon carbide nanotube (SiCNT) have been found to have strong influences on the pressure dependences of the adsorptions and self-diffusions of N2, H2, CO2, CH4, etc., in the tube.24 As to biological channels, aquaporin (AQP) families applied for gas permeation have been a long-standing hot-discussed issue. Significant AQP1-mediated CO2 permeation was found to arise only in a membrane with a low intrinsic permeability.25 Simulations on the permeabilities of O2 and CO2 molecules in AQP1 showed that either gas molecule can easily pass through the central pore of AQP1. However, the water pores show very low permeability to O2 due to the more hydrophobic nature of

INTRODUCTION Cyclic peptide nanotubes (CPNTs) are a class of artificial nanochannels which were first synthesized by Ghadiri’s group in 1993.1 CPNTs have open-ended hollow tubular structures based on the flat, ring-shaped cyclic peptide (CP) subunits formed by α- or β-, etc., amino acid residues.2−6 CP subunits usually stack by antiparallel self-assembly and can contiguously furnish H-bonded β-sheet-like tubular ensembles in an appropriate environment, providing a wide range of structural and functional capabilities of biological relevancy. The hydrophilic/hydrophobic property of the outer surface and the internal diameter of a CPNT can be adjusted simply by the choice of the side chain functionalities of composed amino acid residues and the ring size of the peptide subunit employed. CP subunits are generally separated by ∼4.8 Å.7,8 In all classifications, a CPNT made up of eight CP subunits with suitable external hydrophobicity can be easily embedded into a lipid bilayer membrane and acts as a transmembrane channel.9,10 For decades, the potential and useful properties of transmembrane CPNTs have been widely studied and mainly focused on conducting small species such as water,6,11,12 ions,13−15 glucose,16 and drug molecules.17,18 The dynamics and transport properties of gas molecules in abiological nanochannels have been widely studied. It has been © 2013 American Chemical Society

Received: September 2, 2013 Revised: November 15, 2013 Published: November 18, 2013 14916

dx.doi.org/10.1021/jp408769u | J. Phys. Chem. B 2013, 117, 14916−14927

The Journal of Physical Chemistry B

Article

O2 compared with CO2 molecule.26 It has been demonstrated that AQP4 presents low energy barriers against NO and O2 permeation.27 Moreover, it has been clearly determined that the ubiquitous ammonium transport protein AmtB is an ammoniaconducting channel rather than an ammonium ion transporter.28 A picture of the transport process of ammonia gas conducting in the AmtB channel has been painted.29 The NH3 transport in plasma membrane aquaporin SoPIP2;1 from spinach indicates that the ar/R region is the major energetic barrier for NH3 conduction.30 Similar results have recently been obtained from the simulations on human and bacterial homologues.26 Exploring the dynamics and transport properties of gas molecules in nanochannels, especially in biological channels, is significant and indispensable for understanding the physiological functions of animals and plants. Throughout the literature, it can be found that the micro permeation process of a gas molecule is extremely associated with channel structure. Though CPNTs have outstanding performances in mimicking biological channels, an initial picture that explains the dynamics and transport properties of a gas molecule permeating through a CPNT has not been painted so far. Taking into account the universal distributions of O2, CO2, and NH3 in living organisms, CO2 and O2 as typical linear gas molecules and NH3 as a typical polar one are chosen as research targets. This study aims at exploring the permeation and transport properties of CO2, NH3, and O2 through a transmembrane octa-CPNT of 8×(WL)4/POPE. Steered molecular dynamics (SMD) schemes and adaptive biasing force (ABF) sampling techniques have been applied to investigate the transport processes and potentials of mean force (PMFs) of three gas systems. The results uncover interesting and comprehensive mechanisms underlying the permeation characteristics of these gas molecules in an octa-CPNT.

Figure 1. Side view of an octa-CPNT of 8×cyclo-(WL)4 embedded in a fully hydrated POPE lipid bilayer. The N and P atoms of lipid units are represented in vdW spheres. The framework of the octa-CPNT is described in a stick style. As an example, a CO2 molecule is inserted at the position of z = −25.00 Å along the tube axis and represented in vdW spheres. Bulk and channel water are drawn in surf and line representations, respectively.

speed, non-bonded interactions were only searched within 14 Å of atomic spacing. The Nosé−Hoover Langevin piston method38 and Langevin dynamics39 were applied to maintain the system pressure and temperature at 1 bar and 310 K, respectively. The distances between hydrogen and heavy atoms were constrained to the equilibrium values using the SHAKE algorithm.40 A time step of 1.0 fs was used, and data were collected every 1.0 ps. Steered Molecular Dynamics (SMD) Simulation. SMD simulation has been proved to be an effective computational approach to simulate the transport process of a guest molecule through a nanochannel. In a conventional SMD simulation, a guest molecule of interest is steered by an imaginary atomic force microscopy (AFM) tip, and thus, an external force, which may be constant or variable, is exerted on the guest molecule to facilitate its transport through a channel. In the present study, a constant-velocity SMD (cv-SMD) simulation was performed to explore how a gas molecule transports through a CPNT channel and to provide reasonable starting structures for later ABF samplings. In each system, a gas molecule was first introduced at the position of z = −25.0 Å along the tube axis. After the whole system undergoing 5000-step energy minimization and 1 ns NVT ensemble equilibration, a harmonic constraint was attached to the center of mass (COM) of the gas molecule, which was pulled to travel through the lumen of the CPNT from its initial point (z = −25.0 Å) to the destination at z = +25.0 Å. To find an appropriate pulling velocity, three SMD simulations were performed using different pulling velocities (2.5 × 10−3, 5.0 × 10−3, and 0.01 Å ps−1). Our testing results showed that such three SMD simulations produced similar force profiles. Therefore, a pulling rate of 2.5 × 10−3 Å ps−1 was chosen. To prevent the CPNT channel to drift badly, a harmonic restraint potential of 20 kcal mol−1 Å−2 was applied to each Cα atom of the CPNT backbone by learning from Hwang’s SMD simulation practice.15 The simulation time of the cv-SMD scheme for each gas system was 20 ns, and trajectories were collected for data analysis. Potential of Mean Force (PMF). The PMF of each gas molecule along the CPNT axis (z) has been obtained to explore the quantitative characteristic of free energy related to

■

MATERIALS AND METHODS System Preparations. The equilibrium structure of 8×(WL)4/POPE was taken from our previous result10 based on a 20 ns MD simulation. It was composed of one POPE (palmitoyloleylphosphatedylethanolamine) bilayer containing 104 lipid units (52 per layer), one octa-CPNT embedded in the POPE membrane, and 5700 water molecules on both sides of the bilayer. A single O2, CO2, or NH3 was separately inserted at the position of z = −25.00 Å along the tube axis, and a certain number of water molecules were deleted because of overlapping with the introduced gas molecule. Three newly constructed systems, namely, the CO2 system, NH3 system, and O2 system shown in Figure 1, were then applied in the following SMD simulations and ABF samplings. The parameters of CO2 were directly applied from the CHARMM27 biomolecular force field,31 and those of NH3 were taken from Chen’s work.32 The parameters of heme oxygen were referenced for O2 by setting the oxygen charge to zero.33 All the simulations were performed using the program NAMD 2.934 with the CHARMM27 force field31 and TIP3P water model35 for the CPNT, POPE membrane, and water, respectively. Analysis and visualization were made using the molecular graphics program VMD 1.9.1.36 Assuming periodic boundary conditions, the particle mesh Ewald (PME) method37 was employed for the computation of full electrostatic interactions. The cutoff radii for long-range electrostatic and van der Waals interactions (vdW) were set to be 12 Å, with smoothing functions applied from 10 Å. To improve computing 14917

dx.doi.org/10.1021/jp408769u | J. Phys. Chem. B 2013, 117, 14916−14927

The Journal of Physical Chemistry B

Article

Such periodic alternations of the force curves mainly arise from the regular and periodic arrangement of the composed CP subunits of the CPNT. Stereohindrance caused by the relatively narrow space in an α-plane region and more frequent collisions between a gas molecule with the neighboring water both make the gas molecule need more “power” to pass through the peptide ring, resulting in the formation of a local maximum of force in an α-plane region. A similar phenomenon has also been reported in the transport of 5-fluorouracil through an octaCPNT.17,18 In addition, compared with CO2 and O2, NH3 molecule needs less pulling force to cross over the last CP subunit. H-Bonded Interactions. In order to uncover the contribution of the interaction between a gas molecule and its surrounding to the molecular transferring process, we explore the H-bond formation of a gas molecule with water and the backbone carbonyl/amide groups of the octa-CPNT by analyzing the trajectory of the SMD simulation. There are two criteria determining whether a H-bond exists, namely, the distance l between the donor and acceptor and the angle φ (or the supplementary angle θ) formed by the donor, hydrogen, and acceptor. When l is longer than 3.1 Å and θ is bigger than 35°, the H-bond can be considered as a broken bond.44 In our practice, we use 3.0 Å of l and 20° of θ as the criteria of Hbonded interactions, which are the default in the “hydrogen bonds” plugin implanted in VMD 1.9.1. (1). H-Bonded Interactions in the CO2 System. Water molecules can form at least one H-bond with a CO2 molecule in the neutral hydration of carbon dioxide.45 Undoubtedly, in such “H−O−H···O−C−O” H-bonded configuration, a water molecule acts as a H donor and CO2 an acceptor. Here, the number of H-bonds between a CO2 molecule and water during the 20 ns pulling process of the SMD simulation is depicted in Figure 3 (a1). It can be easily found that no more than two Hbonds between a CO2 molecule and water occur. The double H-bonds formed between CO2 and water just sporadically occur four times. What is more, the hydration pattern of CO2 is discontinuous during the translocation. Such discontinuity seems more obvious and serious when CO2 is traveling through the CPNT. To throw light on the H-bonded interaction between a CO2 molecule and water, the radial distribution functions (RDFs) between CO2 and water through the whole SMD process have been calculated, shown in Figure 4a. It can be found that the height of the first peak of each RDF curve in bulk is ∼3 times the one in the lumen environment. Evidently, the density of the first water shell surrounding CO2 molecule is dramatically reduced when CO2 is traveling through the nanotube. For this reason, the number of H-bonds decreases, characterized by the obvious discontinuity in the tube region in Figure 3 (a1). Besides, the first peak position of each RDF curve in the tube is found to shift left, indicating that the water shell is slightly closer to CO2 due to the confinement of the lumen. The number of H-bonds between the CO2 molecule and the backbone of the CPNT during the SMD simulation is depicted in Figure 3 (a2). It is noteworthy that there are scarcely such Hbond events during CO2 translocation through the lumen. Instead, when going in or out of the tube, CO2 can form a Hbond with either mouth of the tube. Careful investigation shows that such a H-bond is achieved by a bare amide group stretching outward over there. In all, it is the channel water mainly participating in the intermittent H-bonded interaction when CO2 is traveling in the lumen. When accessing an

the transport of the gas molecule through the CPNT. Here, the adaptive biasing force (ABF) method41,42 was applied in its NAMD formulation and implementation.43 In detail, the absolute position of a gas molecule along the channel axis was defined as the reaction coordinate. In order to enhance the efficiency of the ABF algorithm, the CPNT channel, from −20 to 20 Å along the tube axis (z), was divided into eight equal spaced windows, each with a width of 5 Å. A separate ABF simulation of 10 ns was performed for each window. On the basis of the trajectory from the SMD simulation for a gas molecule moving through the whole CPNT, a PDB file for the gas molecule in each window was created as the starting point for the ABF simulation for each window. In order to generate a smooth PMF profile, each window was further divided into 50 bins, each with a width of 0.1 Å. A threshold of 1000 force samples should be set when applying the ABF sampling to obtain a reasonable estimate of average force. After accomplishing the sampling, we constructed the whole PMF by combining the output from eight separate windows. In all, a total of 80 ns simulations in NVT ensembles were performed to generate the whole PMF profile for each gas system. Considering the symmetry of the nanotube,16 the final PMF curve was reconstructed by averaging the data on both sides of the channel.

■

RESULTS AND DISCUSSION Pulling Force. The variations of the pulling forces exerting on O2, CO2, and NH3 molecules moving through the tube are collectively depicted in Figure 2. First, all three gas molecules

Figure 2. Variations of the pulling forces exerting on O2, CO2, and NH3 molecules traveling along the tube axis (z) during a 20 ns SMD simulation, respectively. The gray bars represent the positions of the entrance, dividual rings, and the exit of the octa-CPNT.

encounter obstacles at the entrance of the tube. External forces of ∼210, 190, and 100 kcal mol−1 Å−1 are needed for O2, CO2, and NH3 molecules to enter the lumen, respectively. Obviously, the process into the tube is of the most difficulty for O2 molecule, and NH3 is opposite. Then, it can be found that each profile displays a similar variation trend; namely, different local peaks and valleys appear successively during the transfer of a gas molecule through the CPNT. When a gas molecule reaches a midplane region, the force becomes a local minimum and increases later when the molecule leaves for an α-plane zone. 14918

dx.doi.org/10.1021/jp408769u | J. Phys. Chem. B 2013, 117, 14916−14927

The Journal of Physical Chemistry B

Article

Figure 3. Time evolutions of the number of H-bonds formed between a CO2 (a) or NH3 (b), O2 (c) molecule and water (red lines) or the tube (black lines) during the 20 ns pulling process of a SMD simulation. The nanotube zone is shown by the light gray background.

entrance of the CPNT, CO2 can form H-bonds either with water or an amide group at a tube mouth. However, it is of extraordinary difficulty for CO2 to form H-bonds with the inner backbone of the CPNT. (2). H-Bonded Interactions in the NH3 System. A fair amount of experimental and theoretical research has been devoted to a detailed understanding of H-bonded interactions between NH3 and water. Some of them draw a conclusion that NH3 interacts with water by forming H-bonded clusters.46−49 Here, the number of H-bonds formed between NH3 and the surrounding water during the SMD simulation has been counted and shown in Figure 3 (b1). A NH3 molecule can form one, two, three, and even four H-bonds with water. The probability of each kind of H-bond has been further calculated. The result shows that the occurrences of one, two, three, and four H-bonds account for ∼81.85, 11.32, 0.82, and 0.01%, respectively. A NH3 molecule can simultaneously form four Hbonds with the surrounding water. However, such an event occurs only once near one mouth of the tube during a total of 20 ns of the SMD simulation. Compared with CO2, it is more stable for NH3 molecule existing in a water environment with much more H-bonds. The density of water molecules around NH3 can be found from the RDF profiles shown in Figure 4b. The evident decay of the water density around NH3 molecule in the lumen intensifies the intermittence of H-bonds when NH3 traveling through the CPNT. It should be noted that there are two peaks in the RDFs between the N atom of NH3 and the H atom of water, representing the densities of Hw1 and Hw2, depicted in Figure 5a, where NH3 acts as a H acceptor and a donor,

Figure 4. RDF profiles between the C atom of CO2 (or N of NH3, O of O2) and the H (or O) atom of water molecules in bulk (black and red lines) and in the tube (green and blue lines) during a SMD simulation, respectively. For the sake of clarity, the RDF curves for NH3 molecule in the tube in partial image (b) are shown with dashed lines.

respectively. When NH3 is traveling in the lumen, the height of the second peak decreases heavily, causing the ratio of two peak areas to increase from 1.60/28.15 to 1.13/7.80. Therefore, the channel environment may significantly decrease the probability of the H-bond formation between H of NH3 and O of water. Figure 3 (b2) shows that a NH3 molecule can form one or two H-bonds with the CPNT framework when moving through the lumen. Counting the number of each H-bonded pattern 14919

dx.doi.org/10.1021/jp408769u | J. Phys. Chem. B 2013, 117, 14916−14927

The Journal of Physical Chemistry B

Article

calculating the RDFs, and the results are shown in Figure 4c. Comparison with the RDFs calculated for CO2 and NH3 systems shows that the much lesser density of water around O2 molecule in the lumen dramatically intensifies the discontinuity of the H-bond between O2 and water when O2 is traveling through the CPNT. Figure 3 (c2) indicates that there is scarcely no H-bond formation between the O2 molecule and the backbone of the CPNT except at two ends of the tube, where bare outwardstretching amide groups can exert H-bonded interaction with O2 molecule. Because there is no competition between the Hbond formations for O2 with channel water or channel wall during the O2 transport process, the discontinuity in the O2 system mainly results from the density decrease of channel water around the gas molecule. Water Bridges. A water bridge was once proposed and verified by Liu and Vijayaraj’s groups, exploring the transport mechanism of antitumor drug 5-fluorouracil through an octaCPNT.17,18 It is defined as a water-mediated H-bond between a transporting molecule and a backbone carbonyl/amide group of a CPNT, a type of indirect interaction bridged by water(s) in the molecular translocation process. (1). Water Bridges in the CO2 System. Water bridges in the CO2 system are depicted in Figure 7a. Although no direct H-

Figure 5. NH3 acting as a H acceptor or donor in the formation of Hbonds with the surrounding water (a) and the framework of the CPNT (b). Hw1 and Hw2 represent H atoms of two representative surrounding water molecules, corresponding to the two peaks of the RDF profiles between the N atom of NH3 and the H of water in Figure 4. The H-bonds are shown with dashed lines.

shows that the occurrence of one H-bond accounts for ∼99.4%, and that of two H-bonds ∼0.6%. This result means that it is the one H-bonded pattern which occupies the overwhelming majorities of all H-bonded patterns. It is noteworthy that our SMD simulation shows that NH3 can also form two types of Hbonds with the tube, depicted in Figure 5b, where NH3 acts as a H donor or an acceptor. To demonstrate it, we carefully investigated the trajectories and captured the typical snapshots, depicted in Figure 6. Partial image a contributes to type A,

Figure 6. Snapshots of two H-bonds formed between the NH3 molecule and the framework of the CPNT during a SMD simulation. Partial images (a and b) contribute to the H-bonds of types A and B shown in Figure 5b, respectively. The H-bonds are shown with dashed lines.

Figure 7. Number of the water bridges formed between a CO2 (a) or NH3 (b), O2 (c) molecule, and the tube during a 20 ns SMD simulation. The nanotube zone is shown by the light gray background.

where NH3 molecule acts as a H donor. Analysis of the trajectory shows that H atoms of NH3 may bond with one carboxylate oxygen with an absolutely high probability of ∼99.3%. The H-bond of type B only exhibits as partial image b, with an extremely low probability of ∼0.7%. In a word, nearly all the H-bonds between NH3 and the tube are made up of Hbonds of type A. (3). H-Bonded Interactions in the O2 System. It is the common sense that oxygen as a non-polar molecule is insoluble and unstable in water. Here, the number of H-bonds formed between O2 molecule and water during the SMD simulation is shown in Figure 3 (c1). An O2 molecule can form only one Hbond with the surrounding water, discontinuously and infrequently. The quantity of H-bonds is much less than those in CO2 and NH3 systems. The discontinuity of H-bonds becomes more evident when O2 is traveling through the lumen, compared with the cases in CO2 and NH3 systems. The density of water molecules around O2 has been investigated by

bonded interaction between CO2 and the CPNT framework was observed, water bridges really exist during the CO2 translocation in the lumen. Figure 7a shows that there is just a single-water bridge, occurring repeatedly but not frequently. To further explore such pattern of indirect H-bonded interaction, we traced the trajectories and searched for various water bridges composed of a different number of mediated water molecules. In total, four kinds of water bridges occur, mediated by one, two, three, and four water molecules, respectively, shown in Figure 8. A water bridge mediated by more than four water molecules in the CO2 system has not been found, possibly due to the confined environment in the tube and the moderate ability of CO2 forming H-bonds with water. The probability of each kind of water bridge has been calculated. The result shows that the water bridges with one, two, three, and four mediated water molecules account for ∼59.5, 32.4, 2.7, and 5.4%, respectively, indicating that the 14920

dx.doi.org/10.1021/jp408769u | J. Phys. Chem. B 2013, 117, 14916−14927

The Journal of Physical Chemistry B

Article

92.4%. Therefore, single-water bridges in the NH3 system are mainly composed of the N atom of NH3. Second, a NH3 molecule can form double-water bridges at the same time, which never happens in the systems of CO2 and O2. Partial image c of Figure 9 represents that two water H atoms participate in forming two water bridges between the N atom of NH3 and the tube, while one water H atom and one water O atom participate in partial image d. Due to needing more water molecules and occupying more spaces, such double-water bridges occur few, only accounting for ∼1.3%. In addition, a NH3 molecule obtains more opportunities forming water bridges when approaching both entrances of the tube, compared with CO2 and O2 molecules. (3). Water Bridges in the O2 System. The number of water bridges formed between O2 molecule and the CPNT framework during the SMD simulation is displayed in Figure 7c. The trajectory shows that O2 molecule can form a singlewater bridge with an amide group of the CPNT. However, such a water bridge occurs exceedingly less than those in CO2 and NH3 systems, with only five times when O2 is traveling in the tube. Therefore, an O2 molecule is not easy to form water bridges as well as H-bonds with the tube when transporting in the lumen. VdW and Electrostatic Interactions. The vdW and electrostatic interactions of each gas molecule with the octaCPNT and water molecules have been studied. To do this, a gas molecule was placed at the positions of z = −30.0, −26.4, −24.0, −21.6, ..., +21.6, +24.0, +26.4, +30.0 Å along the tube axis (z), respectively. There are a total of 25 sampling sites for each gas molecule, and the interval of adjacent sites except for two entrances is 2.4 Å. A simulation of 2.0 ns was performed for each gas molecule positioned above each specified site, and the last 1.0 ns trajectories were collected for data analysis. (1). VdW and Electrostatic Interactions in the CO2 System. The vdW and electrostatic interactions of a CO2 molecule with water and the CPNT are depicted in Figure 10a. The vdW interaction energies with water and the CPNT reach global maxima of ∼ −4.0 and −6.0 kcal mol−1 in bulk and in the tube, respectively. Entering deeply into the nanotube, the vdW interaction energy with water decreases to ∼ −0.5 kcal mol−1. However, that with the channel increases to ∼ −6.0 kcal mol−1, which mainly makes up the vdW interactions between the CO2 molecule and the surroundings. The electrostatic interaction between CO2 and water is significantly fluctuant, especially when CO2 is accessing a CPNT entrance. Moreover, the electrostatic interaction with water in the lumen cherishes the characteristic of periodic alternation, reaching a maximum in an α-plane zone and decaying to a minimum in a midplane region, which may result from relatively more water locating in the neighboring midplane region when CO2 is in an α-plane zone. In contrast to the significant fluctuation of the electrostatic interaction with water, that with the tube changes slightly, especially when CO2 is in the tube. The profile fluctuates obviously only in two entrance regions, caused by the electrostatic interaction between CO2 and a bare amide group at the mouth of the tube. On the basis of the detailed analysis of all non-bonded interactions, it is the vdW interaction between CO2 and the tube that mainly makes up the entire vdW interactions, and the electrostatic interaction between CO2 and water to the entire electrostatic interactions. Among all the interactions, the vdW interaction between CO2 and the tube is the greatest contributor to the whole non-bonded interactions of CO2 with the surroundings.

Figure 8. Four kinds of water bridges composed of one (a), two (b), three (c), and four (d) mediated water molecules in the CO2 system during a SMD simulation. The H-bonds are shown with dashed lines.

water bridges with one or two medicated water molecules seem to more likely occur. (2). Water Bridges in the NH3 System. The number of water bridges formed between NH3 and the CPNT framework during the SMD simulation is depicted in Figure 7b. NH3 molecule can form more abundant water bridges compared with CO2 and O2 molecules. The number of mediated water molecules in the water bridges in the NH3 system can be more than four but no more than seven. There are two points worth mentioning. First, both N and H atoms of NH3 molecule can participate in forming water bridges, shown in Figure 9. Partial image a associated with a H atom of NH3 only accounts for 7.6%, while partial image b related with the N atom of NH3 accounts for

Figure 9. Single- and double-water bridges formed between the NH3 molecule and the framework of the CPNT. The H-bonds are shown with dashed lines. 14921

dx.doi.org/10.1021/jp408769u | J. Phys. Chem. B 2013, 117, 14916−14927

The Journal of Physical Chemistry B

Article

major contributor to the entire non-bonded interactions of NH3 with the surroundings, the powerful electrostatic interaction with water makes the NH3 molecule possess the lowest energy when transporting through the CPNT. (3). VdW and Electrostatic Interactions in the O2 System. The vdW and electrostatic interactions of an O2 molecule with water and the CPNT are depicted in Figure 10c. The electrostatic interactions with the tube and water are both 0.0 kcal mol−1 due to the electrical neutrality of the O atom in the O2 molecule. The vdW interactions of O2 with water and the tube show similar trends to those in the CO2 system. However, the maxima of two interactions, about −3.0 and −4.0 kcal mol−1, respectively, are both lower than those in the CO2 system. In addition, the curve of vdW interaction with the tube shows a slight fluctuation when O2 is traveling through the lumen. Wave peaks and valleys alternately occur in midplane regions and α-plane zones, which may result from the periodic structure of the CPNT. Potential of Mean Force (PMF). The PMF profiles derived from the ABF samplings for CO2, NH3, and O2 molecules moving through the octa-CPNT are collectively depicted in Figure 11. Several features in the PMF profiles are

Figure 11. PMF profiles of O2 (red), CO2 (green), and NH3 (blue) molecules moving through the CPNT, respectively. The vertical dashed lines mean the positions of dividual CP subunits.

Figure 10. VdW and electrostatic interactions of CO2 (a) or NH3 (b), O2 (c) with water and the framework of the CPNT. The vdW and electrostatic interactions with water are represented by red square and green diamond lines, respectively. Those with the tube are described by black square and blue triangle lines, respectively. Vertical dashed lines represent the positions of dividual CP subunits.

worth addressing. First, all three PMF profiles display similar trends of periodical variation in the transmembrane region, namely, a local minimum occurs in a midplane region and a maximum in an α-plane zone, coinciding well with the equally spaced ring arrangement of the CPNT. The oscillations of the PMF profiles are similar to those of the pulling force curves shown in Figure 2. Second, local maxima in individual α-plane regions do not go beyond 2.0 kcal mol−1, which is low enough for three gas molecules to effortlessly overcome by the means of thermal motion, especially for NH3 and CO2 molecules. This finding holds good agreement with the fact that these three gas molecules can freely shuttle in the CPNT channel. Third, the energy barriers for three gas molecules at two entrances of the CPNT are different. The energy barriers, ∼4.4 and ∼4.9 kcal mol−1, for CO2 and O2 molecules, respectively, indicate that these two molecules can be confined in the lumen, not easily going out. On the contrary, the energy barriers for the NH3 molecule at two entrances are even slightly lower than the local maxima inside the channel. It can be speculated that NH3 molecule can go out of the tube easily.

(2). VdW and Electrostatic Interactions in the NH3 System. The vdW and electrostatic interactions of a NH3 molecule with water and the CPNT are depicted in Figure 10b. When NH3 is entering the CPNT lumen, the vdW interaction with the tube rises by ∼3.0 kcal mol−1. The trend of the profile is similar to that of the CO2 system, but the amount of increase is lower. The electrostatic interaction with the tube does not change too much except for the increases at two entrances of the tube. The vdW interaction with water reaches a maximum of ∼4.5 kcal mol−1 at two entrances of the tube. However, it doesn’t fall to zero in the lumen compared with the CO2 system. In addition, both the vdW and electrostatic interactions with water become local maxima in α-plane zones and minima in midplane regions. It is worth mentioning that the electrostatic interaction energy with water is much higher than that in the CO2 system. As a 14922

dx.doi.org/10.1021/jp408769u | J. Phys. Chem. B 2013, 117, 14916−14927

The Journal of Physical Chemistry B

Article

Figure 12. Probability distributions of CO2 (a) and O2 (b) molecules along the tube axis (z), obtained from molecular random walks through the CPNT, respectively.

Molecular Occurring Probability along the Tube Axis. In order to further clarify the motion of each gas molecule synchronously varied with the PMF, the random walk of a gas molecule in the tube under no artificial external force has been investigated. To do this, a gas molecule was first restricted at the center of the CPNT lumen. After equilibrating the whole system, the restriction was then canceled and the gas molecule was allowed to walk randomly along with channel water in the tube for 20 ns NVT ensemble equilibration. The trajectories were analyzed, and the probabilities of the gas molecule appearing at specified locations in the channel were recorded. The occurring probabilities of CO2 and O2 molecules along the tube axis (z) are shown in Figure 12. What are the similarities is that CO2 and O2 molecules are both confined in the narrow lumen of the CPNT throughout 20 ns because there are no distributions beyond the tube. This finding holds good agreement with the global maxima of energy barriers at two entrances of the tube in the PMF profiles of CO2 and O2 molecules, shown in Figure 11. Furthermore, the two profiles both show peaks in midplane regions and valleys in α-plane zones, meaning that two gas molecules both prefer to stay at midplane regions due to the potential wells locating there, shown in the PMF profiles. However, some differences obviously exist between the two profiles. The profile of CO2 displays much narrower peaks, while that of O2 seems relatively coarse with more peaks spanning more spacing. Such different characteristics fully show that O2 molecule is more active and frequently oscillating in the lumen of the CPNT compared with CO2. As for the NH3 molecule, things become interesting. The NH3 molecule was found to run out of the tube in the −z direction after 3.4 ns. In order to verify this phenomenon, another two parallel tests for the NH3 molecule were performed, resulting in similar phenomena: the NH3 molecule runs out in the +z direction after 2.5 ns and the −z direction after 1.7 ns, respectively. Such phenomena coincide very well with the PMF profile of the NH3 molecule, shown in Figure 11. The low energy barriers at two entrances of the tube let the NH3 molecule step over the barriers effortlessly. Molecular Orientation. When traveling through a nanochannel, a molecule may rotate or roll itself to move forward with the best posture. Herein, the orientation of a gas molecule, defined as the angle (θ) between the molecular bond axis (or

dipole) and the tube axis, has been investigated. The results are depicted in Figure 13. Distinctly, the diatomic O2 molecule seems to be very flexible, mostly orientating itself with a wide range of angle (θ) with the tube axis. It infers that O2 molecule keeps strong rocking motions when traveling in the lumen. The angle (θ) of 100° shows the maximum probability, meaning that the O2 molecule mostly moves forward in a way nearly perpendicular to the tube axis. The linear triatomic CO2 also takes a wide range of angle (θ) with the tube axis. However, the highest probability of the angle (θ) of ∼22 or 158° shows that the molecule mostly travels with the orientation parallel or reverse to the tube axis. As a polar molecule, NH3 prefers to take an angle of ∼25, 125, 135, or 160° with the tube axis when moving in the lumen, also traveling with the orientation parallel or reverse to the tube axis. However, compared with the angle profiles of O2 and CO2, the NH3 molecule orientates itself with much less probability of θ = 90° and the peaks of the NH3 profile are much narrower and sharper than those of CO2. In order to further explore the orientation changes of gas molecules traveling through the lumen, the distributions of the orientation angles (θ) of O2, CO2, and NH3 in dividual gaps of the CPNT have been analyzed, shown in Figure 14. All the curves in partial image a display almost the same variation trends no matter which gap the O2 molecule occurs in, showing similar movement of O2 molecule in each gap. In partial image b, all the curves show similar trends except those in gap 3 and gap 5, where the CO2 molecule rolls flexibly, orientating itself with a wide range of angle (θ) with the tube axis. When the CO2 molecule is traveling in the regions close to an entrance (i.e., gaps 1, 2, 6, and 7) or in the middle zone (i.e., gap 4) of the CPNT, it mostly takes an angle (θ) of ∼25 or 175° with the tube axis, traveling with the orientation parallel or reverse to the tube axis. It can be speculated that the CO2 molecule may make an instantaneous flip in these gaps. However, in partial image c, nearly each curve exhibits just one sharp peak but is associated with a different angle (θ). Figure 15 illustrates the variation of the dipole orientation of NH3 molecule traveling from gap 1 to gap 7. Evidently, the dipole of NH3 just reverses when traveling through the whole channel, pointing to both entrances when close to them. It was speculated that the bare carbonyl groups at the tube mouths strongly affect the dipole orientation of the polar NH3 molecule. Such similarity also occurs in the dipole orientations of the water chain at the two end gaps of a 14923

dx.doi.org/10.1021/jp408769u | J. Phys. Chem. B 2013, 117, 14916−14927

The Journal of Physical Chemistry B

Article

Figure 13. Distributions of the orientation angles (θ) of O2 (a), CO2 (b), and NH3 (c) molecules, obtained from molecular random walks through the CPNT, respectively. Here, the orientation of a gas molecule is defined as the angle (θ) between the molecular bond axis (of O2 and CO2) or dipole (of NH3) and the tube axis.

Figure 14. Distributions of the orientation angles (θ) of O2 (a), CO2 (b), and NH3 (c) molecules in dividual gaps, respectively. The definition of molecular orientation angle (θ) can be found from the caption of Figure 13.

CPNT.10 The relatively lower and broader peak in gap 4, shown in Figure 14c, infers that NH3 may take moderate rotating or rolling motion for angle adjustment in the middle of the CPNT. The dipole change of the NH3 molecule from gap 1 to gap 7 is gradual. No such busy rocking or flip motions as CO2 and O2 occurs. Figure 16 shows the flip frequencies of CO2, O2, and NH3 molecules in dividual gaps. Here, the flip frequency is defined as the number of molecular flips per nanosecond. We set a status of angle orientation for each gas molecule. “status + 1”, “status

Figure 15. Variation of the dipole orientation of NH3 molecule traveling from gap 1 to gap 7, obtained from its random walk through the CPNT.

14924

dx.doi.org/10.1021/jp408769u | J. Phys. Chem. B 2013, 117, 14916−14927

The Journal of Physical Chemistry B

Article

NH3 acts as a H donor, coordinating with a carboxylate oxygen group of the tube. CO2 and O2 molecules hardly form H-bonds with the tube except at an entrance, where they exert H-bonded interactions with bare outward-stretched amide groups. (3) RDFs between gas and water show the density decrease of channel water due to the confinement of the lumen environment, which significantly results in the strong decay of the continuity of H-bond formation between a gas molecule and water environment. (4) In addition to the direct H-bonded interactions, all three gases can form water bridges with the tube. A NH3 molecule can even form a few double-water bridges at the same time, which never happens in the systems of CO2 and O2. (5) As a major contributor to the interactions of NH3 with the surroundings, the powerful electrostatic interaction with water makes the NH3 molecule possess the lowest energy when transporting through the CPNT. The vdW interactions of CO2 and O2 molecules with the tube are both the greatest contributors to the whole non-bonded interactions with the surroundings. (6) The PMF profiles elucidate the alternation of energy maxima and minima in α-plane regions and midplane zones, coinciding with the molecular occurring probability along the tube axis. The energy barriers at two entrances reasonably explain the phenomena that NH3 can easily go outside the tube, while CO2 and O2 are confined in the narrow lumen. (7) The diatomic O2 molecule keeps strong rocking motions when traveling in the lumen and moves forward mostly in a way nearly perpendicular to the tube axis. The linear triatomic CO2 mostly travels with the orientation parallel or reverse to the tube axis. The dipole change of NH3 molecule from gap 1 to gap 7 is gradually varied, just reversing when traveling through the whole channel. No such busy rocking motions as CO2 and O2 occur. (8) The trajectories show that O2 is the most active, reaching ∼500 flips/ns in each gap. Oppositely, scare dipole reverse motions for the NH3 molecule were observed except in gap 4. CO2 has a flip frequency between those of O2 and NH3 molecules. The findings in this work shed light on the interesting and comprehensive mechanisms underlying the microprocesses of O2, CO2, and NH3 transporting in an octa-CPNT. They could provide relevant information for designing suitable and applicable artificial gas nanochannels.

Figure 16. Flip frequencies of O2, CO2, and NH3 molecules in dividual gaps, obtained from the SMD simulation processes, respectively.

− 1”, and “status 0” correspond to the status in which the molecular orientation angle is larger than, less than, and equal to 90°, respectively. For every frame of a 20 ns MD trajectory, the orientation angle of a gas molecule was calculated, and marked with the corresponding status, further compared with that in the last frame. When the status changes from +1 to −1 or vice versa, the gas molecule was thought to have completed a flip event. We individually recorded the times (t1, t2, t3, ..., tn) for the dividual flipping events (event 1, 2, 3, ..., n) and further calculated the average time needed for completing one flipping event. The flip frequency was computed as the reciprocal of the average time. It can be clearly found that the O2 molecule is the most active and has a similar flip frequency in each gap, reaching ∼500 times/ns. Oppositely, scarce molecular flip of the NH3 molecule occurs except for in gap 4. The findings hold good agreement with the strong dipole orientation of NH3 in each gap except in the middle of the channel, shown in Figure 14c. The partial profile for the CO2 molecule shows that its flip frequencies in middle regions (i.e., gaps 3, 4, and 5) are relatively higher than those in the regions close to an entrance, which uncovers the similar result displayed in Figure 14b. In addition, CO2 has the flip frequency between those of O2 and NH3 molecules, showing moderate rolling motion when traveling through the tube.

■

■

CONCLUSION This is the first time the dynamic behaviors and transport properties of O2, CO2, and NH3 molecules through a transmembrane octa-CPNT have been systematically investigated. The following points have been obtained: (1) SMD simulations show that external forces of ∼210, 190, and 100 kcal mol−1 Å−1 are needed for O2, CO2, and NH3 molecules to enter the CPNT channel, respectively. The periodic change of the pulling force curve for a gas molecule traveling through the channel mainly arises from the regular and periodic arrangement of the composed CP subunits of the CPNT. (2) Analysis of intermolecular interactions shows that the transport of a gas molecule is mediated by direct and indirect (i.e., water-bridged) H-bonded interactions with the surroundings. The NH3 molecule possesses the strongest abilities forming H-bonds, acting as a H donor or an acceptor, both with water and the tube. The channel environment significantly decreases the probability of the H-bond formation between H of NH3 and O of water. In nearly all the H-bonds between NH3 and the tube,

AUTHOR INFORMATION

Corresponding Author

*Phone: 0086-0512-65883271. E-mail: jff[email protected] and jff[email protected]. Notes

The authors declare no competing financial interest.

■

ACKNOWLEDGMENTS

This work has been supported by the National Natural Science Foundation of China (Grant No. 21173154) and the Priority Academic Program Development of Jiangsu Higher Education Institutions. The authors are grateful to Mr. Jan Liu, Yi Wang, Bert L. de Groot, and Emad Tajkhorshid for their insightful suggestions and are indebted to School of Computer Science & Technology in Soochow University for providing us generous amounts of computer facilities assignment on the high-powered computers. 14925

dx.doi.org/10.1021/jp408769u | J. Phys. Chem. B 2013, 117, 14916−14927

The Journal of Physical Chemistry B

■

Article

(22) Arora, G.; Sandler, S. I. Nanoporous Carbon Membranes for Separation of Nitrogen and Oxygen: Insight from Molecular Simulations. Fluid Phase Equilib. 2007, 259, 3−8. (23) Garcez, K. M.; Moreira, E.; Azevedo, D. L.; Galvao, D. S. Neon Atoms Oscillating Inside Carbon and Boron-Nitride Nanotubes: A Fully Atomistic Molecular Dynamics Investigation. Mol. Simul. 2010, 36, 639−643. (24) Malek, K.; Sahimi, M. Molecular Dynamics Simulations of Adsorption and Diffusion of Gases in Silicon-Carbide Nanotubes. J. Chem. Phys. 2010, 132, 014310−014319. (25) Hub, J. S.; de Groot, B. L. Mechanism of Selectivity in Aquaporins and Aquaglyceroporins. Proc. Natl. Acad. Sci. U.S.A. 2008, 105, 1198−1203. (26) Wang, Y.; Cohen, J.; Boron, W. F.; Schulten, K.; Tajkhorshid, E. Exploring Gas Permeability of Cellular Membranes and Membrane Channels with Molecular Dynamics. J. Struct. Biol. 2007, 157, 534− 544. (27) Wang, Y.; Tajkhorshid, E. Nitric Oxide Conduction by the Brain Aquaporin AQP4. Proteins: Struct., Funct., Bioinf. 2010, 78, 661−670. (28) Winkler, F. K. Amt/MEP/Rh Proteins Conduct Ammonia. Pflügers Arch. 2006, 451, 701−707. (29) Khademi, S.; Stroud, R. M. The Amt/MEP/Rh Family: Structure of AmtB and the Mechanism of Ammonia Gas Conduction. Physiology 2006, 21, 419−429. (30) Dynowski, M.; Mayer, M.; Moran, O.; Ludewig, U. Molecular Determinants of Ammonia and Urea Conductance in Plant Aquaporin Homologs. FEBS Lett. 2008, 582, 2458−2462. (31) MacKerell, A. D., Jr.; Bashford, D.; Bellott, M.; Dunbrack, R. L.; Evanseck, J. D.; Field, M. J.; Fischer, S.; Gao, J.; Guo, H.; Ha, S.; Joseph-McCarthy, D.; Kuchnir, L.; Kuczera, K.; Lau, F. T. K.; Mattos, C.; Michnick, S.; Ngo, T.; Nguyen, D. T.; Prodhom, B.; Reiher, W. E., III; Roux, B.; Schlenkrich, M.; Smith, J. C.; Stote, R.; Straub, J.; Watanabe, M.; Wiorkiewicz-Kuczera, J.; Yin, D.; Karplus, M. All-Atom Empirical Potential for Molecular Modeling and Dynamics Studies of Proteins. J. Phys. Chem. B 1998, 102, 3586−3616. (32) Chen, I. J.; Yin, D.; MacKerell, A. D. Combined Ab initio/ Empirical Approach for Optimization of Lennard-Jones Parameters for Polar-Neutral Compounds. J. Comput. Chem. 2002, 23, 199−213. (33) Cohen, J.; Arkhipov, A.; Braun, R.; Schulten, K. Imaging the Migration Pathways for O2, CO, NO, and Xe Inside Myoglobin. Biophys. J. 2006, 91, 1844−1857. (34) 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. (35) Jorgensen, W. L.; Chandrasekhar, J.; Madura, J. D.; Impey, R. W.; Klein, M. L. Comparison of Simple Potential Functions for Simulating Liquid Water. J. Chem. Phys. 1983, 79, 926−935. (36) Humphrey, W.; Dalke, A.; Schulten, K. VMD: Visual Molecular Dynamics. J. Mol. Graphics 1996, 14, 33−38. (37) 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. (38) Martyna, G. J.; Tobias, D. J.; Klein, M. L. Constant Pressure Molecular Dynamics Algorithms. J. Chem. Phys. 1994, 101, 4177− 4189. (39) Pastor, R. W.; Brooks, B. R.; Szabo, A. An Analysis of the Accuracy of Langevin and Molecular Dynamics Algorithms. Mol. Phys. 1988, 65, 1409−1419. (40) 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. (41) Darve, E.; Pohorille, A. Calculating Free Energies Using Average Force. J. Chem. Phys. 2001, 115, 9169−9183. (42) Darve, E.; Rodríguez-Gómez, D.; Pohorille, A. Adaptive Biasing Force Method for Scalar and Vector Free Energy Calculations. J. Chem. Phys. 2008, 128, 144120−144132.

REFERENCES

(1) Ghadiri, M. R.; Granja, J. R.; Milligan, R. A.; McRee, D. E.; Khazanovich, N. Self-Assembling Organic Nanotubes Based on a Cyclic Peptide Architecture. Nature 1993, 366, 324−327. (2) Hartgerink, J. D.; Granja, J. R.; Milligan, R. A.; Ghadiri, M. R. Self-Assembling Peptide Nanotubes. J. Am. Chem. Soc. 1996, 118, 43− 50. (3) Gauthier, D.; Baillargeon, P.; Drouin, M.; Dory, Y. L. SelfAssembly of Cyclic Peptides into Nanotubes and Then into Highly Anisotropic Crystalline Materials. Angew. Chem., Int. Ed. 2001, 40, 4635−4638. (4) Amorín, M.; Castedo, L.; Granja, J. R. New Cyclic Peptide Assemblies with Hydrophobic Cavities: the Structural and Thermodynamic Basis of a New Class of Peptide Nanotubes. J. Am. Chem. Soc. 2003, 125, 2844−2845. (5) Horne, W. S.; Stout, C. D.; Ghadiri, M. R. A Heterocyclic Peptide Nanotube. J. Am. Chem. Soc. 2003, 125, 9372−9376. (6) Raghavender, U. S.; Aravinda, S.; Shamala, N.; Balaram, P. Hydrophobic Peptide Channels and Encapsulated Water Wires. J. Am. Chem. Soc. 2009, 132, 1075−1086. (7) Ghadiri, M. R. Self-Assembled Nanoscale Tubular Ensembles. Adv. Mater. 1995, 7, 675−677. (8) Khazanovich, N.; Granja, J. R.; McRee, D. E.; Milligan, R. A.; Ghadiri, M. R. Nanoscale Tubular Ensembles with Specified Internal Diameters. Design of a Self-Assembled Nanotube with a 13-Å Pore. J. Am. Chem. Soc. 1994, 116, 6011−6012. (9) Gong, X.; Li, J.; Zhang, H.; Wan, R.; Lu, H.; Wang, S.; Fang, H. Enhancement of Water Permeation Across a Nanochannel by the Structure outside the Channel. Phys. Rev. Lett. 2008, 101, 257801. (10) Liu, J.; Fan, J.; Tang, M.; Cen, M.; Yan, J.; Liu, Z.; Zhou, W. Water Diffusion Behaviors and Transportation Properties in Transmembrane Cyclic Hexa-, Octa-and Decapeptide Nanotubes. J. Phys. Chem. B 2010, 114, 12183−12192. (11) Zhu, J.; Cheng, J.; Liao, Z.; Lai, Z.; Liu, B. Investigation of Structures and Properties of Cyclic Peptide Nanotubes by Experiment and Molecular Dynamics. J. Comput.-Aided Mol. Des. 2008, 22, 773− 781. (12) Liu, J.; Fan, J.; Cen, M.; Song, X.; Liu, D.; Zhou, W.; Liu, Z.; Yan, J. Dependences of Water Permeation through Cyclic Octapeptide Nanotubes on Channel Length and Membrane Thickness. J. Chem. Inf. Model. 2012, 52, 2132−2138. (13) Dehez, F.; Tarek, M.; Chipot, C. Energetics of Ion Transport in a Peptide Nanotube. J. Phys. Chem. B 2007, 111, 10633−10635. (14) Choi, K. M.; Kwon, C. H.; Kim, H. L.; Hwang, H. Potential of Mean Force Calculations for Ion Selectivity in a Cyclic Peptide Nanotube. Bull. Korean Chem. Soc. 2012, 33, 911−916. (15) Hwang, H.; Schatz, G. C.; Ratner, M. A. Steered Molecular Dynamics Studies of the Potential of Mean Force of a Na+ or K+ Ion in a Cyclic Peptide Nanotube. J. Phys. Chem. B 2006, 110, 26448−26460. (16) Granja, J. R.; Ghadiri, M. R. Channel-Mediated Transport of Glucose across Lipid Bilayers. J. Am. Chem. Soc. 1994, 116, 10785− 10786. (17) Liu, H.; Chen, J.; Shen, Q.; Fu, W.; Wu, W. Molecular Insights on the Cyclic Peptide Nanotube-Mediated Transportation of Antitumor Drug 5-Fluorouracil. Mol. Pharmaceutics 2010, 7, 1985− 1994. (18) Vijayaraj, R.; Van Damme, S.; Bultinck, P.; Subramanian, V. Theoretical Studies on the Transport Mechanism of 5-Fluorouracil through Cyclic Peptide Based Nanotubes. Phys. Chem. Chem. Phys. 2013, 15, 1260−1270. (19) Alexiadis, A.; Kassinos, S. Molecular Dynamic Simulations of Carbon Nanotubes in CO2 Atmosphere. Chem. Phys. Lett. 2008, 460, 512−516. (20) Skoulidas, A. I.; Sholl, D. S.; Johnson, J. K. Adsorption and Diffusion of Carbon Dioxide and Nitrogen through Single-Walled Carbon Nanotube Membranes. J. Chem. Phys. 2006, 124, 054708− 054714. (21) Arora, G.; Sandler, S. I. Molecular Sieving Using Single Wall Carbon Nanotubes. Nano Lett. 2007, 7, 565−569. 14926

dx.doi.org/10.1021/jp408769u | J. Phys. Chem. B 2013, 117, 14916−14927

The Journal of Physical Chemistry B

Article

(43) Hénin, J.; Chipot, C. Overcoming Free Energy Barriers Using Unconstrained Molecular Dynamics Simulations. J. Chem. Phys. 2004, 121, 2904−2914. (44) Khan, A. A liquid water model: density variation from supercooled to superheated states, prediction of H-bonds, and temperature limits. J. Phys. Chem. B 2000, 104, 11268−11274. (45) Nguyen, M. T.; Raspoet, G.; Vanquickenborne, L. G.; Van Duijnen, P. T. How Many Water Molecules Are Actively Involved in the Neutral Hydration of Carbon Dioxide? J. Phys. Chem. A 1997, 101, 7379−7388. (46) Stockman, P. A.; Bumgarner, R. E.; Suzuki, S.; Blake, G. A. Microwave and Tunable Far-Infrared Laser Spectroscopy of the Ammonia-Water Dimer. J. Chem. Phys. 1992, 96, 2496−2510. (47) Fraser, G. T.; Suenram, R. D. Perturbations in the Infrared Spectrum of the NH Umbrella Mode of HOH-NH. J. Chem. Phys. 1992, 96, 7287−7297. (48) Sadlej, J.; Moszynski, R.; Dobrowolski, J. C.; Mazurek, A. P. Structure and Energetics of the Weakly Bound NH3···H2O Complex. J. Phys. Chem. A 1999, 103, 8528−8536. (49) Donaldson, D. J. Adsorption of Atmospheric Gases at the Air− Water Interface. 1. NH3. J. Phys. Chem. A 1999, 103, 62−70.

14927

dx.doi.org/10.1021/jp408769u | J. Phys. Chem. B 2013, 117, 14916−14927