Biomaterials 35 (2014) 1771e1778
Contents lists available at ScienceDirect
Biomaterials journal homepage: www.elsevier.com/locate/biomaterials
Peptide encapsulation regulated by the geometry of carbon nanotubes Zhi-Sen Zhang 1, Yu Kang 1, Li-Jun Liang, Ying-Chun Liu*, Tao Wu, Qi Wang* Soft Matter Research Center and Department of Chemistry, Zhejiang University, Hangzhou 310027, PR China
a r t i c l e i n f o
a b s t r a c t
Article history: Received 21 October 2013 Accepted 14 November 2013 Available online 27 November 2013
In this work the encapsulation of an a-helical peptide in single carbon nanotubes (CNTs) with similar diameter and length but different geometry (armchair and zigzag) was investigated through molecular dynamics simulations and free energy calculations. Our simulation results showed that in vacuo it makes no evident difference whether the investigated peptide is encapsulated in armchair or zigzag CNTs; however, in aqueous solution the armchair CNT encapsulates the peptide remarkably easier than the zigzag CNT does. A detailed analysis revealed that the equilibrium conformation of the water molecules inside the CNTs with varying geometry mediates the peptide encapsulation. It suggests that the water molecules play an important role in regulating behaviors of biomolecules in bio-systems. Then the impact of the CNT geometry on the conformational changes of the confined peptide was studied. Analyses of secondary structures showed the a-helix of the peptide could be better maintained in the zigzag CNT. Ó 2013 Elsevier Ltd. All rights reserved.
Keywords: Geometry of CNTs Peptide encapsulation Interfacial water mediation Umbrella sampling
1. Introduction The properties and applications of hybrid organiceinorganic nanomaterial have aroused great expectation of broad importance for future advancements in nanotechnology. In particular, the biomolecule-CNT complex that consists of single-walled CNTs (SWNTs) wrapped and/or filled with self-assembled biomacromolecules continues to attract scientific attention. It has shown a wide range of interesting features with potential applications in CNT solubilization [1], single-molecule manipulation [2], biosensors [3], and biomedical devices [4], in which a few work have been reported by our group [5e9]. Meanwhile, these applications of SWNTs require their efficient dispersions based on geometry and diameter within cellular environment; separating and dispersing SWNTs, as well as the further detection of the CNTs with different geometry and/or diameters are therefore of importance [10]. To realize these applications, a detailed understanding of the fundamental molecular properties and interactions of the bio-nano complex is needed, which determines the potential applications of nanomaterial in biological settings. When concerning the interactions between CNTs and biomacromolecules, one should also pay attention to the role of the solvent molecules that extensively * Corresponding authors. Tel.: þ86 571 87952424; fax: þ86 571 87951895. E-mail addresses: [email protected] (Y.-C. Liu), [email protected] (Q. Wang). 1 These authors contributed equally to this study and share first authorship. 0142-9612/$ e see front matter Ó 2013 Elsevier Ltd. All rights reserved. http://dx.doi.org/10.1016/j.biomaterials.2013.11.041
exist in the bio-nano systems. Currently the dynamic mechanism of water-CNT systems has been extensively investigated by both experiments [11,12] and simulations [13,14]. For example, Kolesnikov and co-workers [15] investigated the dynamic behaviors of water molecules in CNTs through both molecular dynamics (MD) simulation and neutron scattering techniques. Hummer and co-workers [16,17] reported a one-dimensionally ordered chain of water molecules confined in SWNTs. Alexiadis and co-workers [18,19] studied the density of water in CNTs with various diameters and different geometry. Huang and co-workers [20] revealed the distribution patterns of water inside and outside single-walled carbon nanotubes, suggesting that the water molecules tend to adsorb around the center of a hexagon formed by carbon atoms. Recently, our group reported the effects of geometry and diameter of CNTs on the water diffusion and conduction through MD simulations and quantum mechanics (QM) calculations [21,22]. In these works we have observed the significantly different potential energy surfaces of water molecules in armchair and zigzag CNTs, and confirmed that the minima of potential energy surfaces are at the center of a hexagon formed by carbon atoms and the maxima of potential energy are around the carbon atoms, which is consistent with the previous report [20]. More comprehensive understanding of the interactions between water and CNTs can be found in a review report [23]. Although the effects of geometry of CNTs on the dynamics of small molecules have been widely investigated experimentally and theoretically, the problem became more complicated when macromolecules were incorporated. Currently few works reported the
1772
Z.-S. Zhang et al. / Biomaterials 35 (2014) 1771e1778
impact of CNT topologies on the dynamics of macromolecules. AlHaik and co-workers [24] investigated the binding affinity between polyethylene and CNTs with different geometry through MD simulations. In their work the CNTs with smaller chiral angles were observed to achieve higher adhesion energy, and tend to have smaller diameter and greater length. Zheng and co-workers [25] selected several CNTs with different geometry but similar diameters and lengths to investigate the interfacial binding of polymer-CNT composites, and the armchair CNT was considered to be optimal. However, the influence of the CNT geometry on the biomacromolecules in aqueous solution has not been reported so far, and the mechanism behind it remains obscure. In this study an a-helical peptide was selected as a model molecule to investigate encapsulation behaviors of the biomacromolecules in SWNTs with similar diameters and lengths but different geometries through MD simulations and free energy calculations. We here try to focus on the different encapsulation processes of the peptide in the SWNT with two distinct structures, armchair and zigzag, to elucidate the effects of the tube geometry and the remarkable role of the solvent. The effects of tube length and diameter on the encapsulation of peptide molecule are not involved in this work, which have been investigated and reported by our group [6,7].
2. Modeling and simulation details An a-helical peptide was taken as the model peptide from the Protein Data Bank (PDB), entry code 2OVN [26], with a sequence of NYHLENEVARLKKLCGE. According to the size of the model molecule, two types of uncapped SWNTs with similar tube length and diameter were selected: an armchair (14, 14) CNT and a zigzag (24, 0) CNT, with the tube diameter of 1.88 nm and 1.90 nm, and the tube length of 4.76 nm and 4.54 nm, respectively. The peptide was placed close to the entrance of each type of the nanotube, and was aligned along the tube axis. Then each peptide-CNT complex was immersed in a rectangular periodic box of TIP3P [27] water molecules (or in a rectangular periodic box of vacuum), in which the shortest distances between the complex surface and the box walls are larger than 1.0 nm. In this work, all simulations were performed in the isothermale isobaric (NpT) ensemble with Gromacs 4.5.2 package [28,29] and visualizations were made using VMD [30]. All bonds that involve Hatoms were fixed. A time step of 2 fs was used with atom coordinates saved every 1 ps. A temperature of 310 K and a pressure of 101.3 kPa was maintained using Berendsen method [31]. The CHARMM27 all-atoms (AA) force field [32] was used for describing the peptide and the CNT parameters were supplemented [33]. The atoms of the CNT remain fixed. The particle mesh Ewald (PME) summation [34] was used to calculate the full-system periodic electrostatic interactions, with a cutoff of 1.2 nm for the separation of the direct and reciprocal space summation. Carbon atoms of the CNTs are modeled to be uncharged. The cutoff distance for van der Waals interaction was 1.2 nm, and the parameters of the LennardJones potential for the cross interactions between nonbonded atoms were obtained from the venerable LorentzeBerthelot combination rule [35]. The instantaneous interaction energy between the peptide and the CNT is defined similarly to our previous work [9] as
Eint ðtÞ ¼ EpeptideþCNT ðtÞ Epeptide ðtÞ ECNT ðtÞ
(1)
where Eint(t) stands for the interaction energy between the peptide and the CNT, and EpeptideþCNT(t) refers to the total potential energy of the peptideeCNT complex. Epeptide(t) and ECNT(t) are the potential
energy of the peptide and that of the CNT, respectively. Each term indicates the energy as a function of the simulation time. The free energy profile (represented by potential of mean force, PMF) was calculated by using the umbrella sampling [36e38]. The PMF was calculated along the reaction coordinate x, which was defined as the z-component difference between the center-of-mass (CoM) of the peptide and that of the CNT, describing the entering path of the peptide. The entering path was divided into 1 Å wide equidistant windows perpendicular to the tube axis with the center of each window representing an umbrella center. To enhance sampling, the biased simulations were performed in 41 sampling windows, in which the CoM of the peptide was harmonically restrained in z direction with a harmonic force constant of 400 kJ$mol1$nm2. After carefully analysis, the first 6 ns was removed in each umbrella simulation for thermodynamic equilibration, followed by a 14 ns of production run. The PMFs were then unbiased and combined via the weighted histogram analysis method (WHAM) [39]. To clarify the behavior of the water molecules affected by the CNT, the density profiles of water molecules inside the CNTs were calculated based on the simulations of only a CNT (zigzag or amchair) in aqueous solution, regardless of the peptide. The two CNTs mentioned above were immersed respectively in an aqueous solvent box, with the shortest distance from the carbon atoms in the CNT to the box wall of 1.0 nm in radial direction and 3.0 nm in axial direction. Then 30 ns MD simulations were performed for the two CNT-water systems under the same conditions of temperature and pressure used in the peptide confining case, producing trajectories with 25,000 frames. The first 5000 frames (namely 6 ns) were discarded for thermodynamic equilibration, and the following 20,000 frames were used to extract density profiles of water molecules. 3. Results and discussion 3.1. Spontaneous encapsulation of the peptide into CNTs Fig. 1 shows the normalized CoM distance, d/d0, between the peptide and the CNTs as a function of the simulation time. First the peptide was placed close to the entrance of each type of the nanotube respectively, and for each complex a 5-ns simulation in vacuo was performed. As shown in Fig. 1, the normalized CoM distances decreased dramatically to around 0.0 within 1 ns simulations in both armchair and zigzag systems. It means that in vacuo both CNTs can encapsulate the peptide fast when the macromolecule was placed close to the entrance of the tube, and the geometry of the CNT (armchair or zigzag) barely affected the encapsulation for a given tube length and diameter. This is because the main driving force of the encapsulation is the van der Waals interaction between the peptide and the CNT [8], and the carbon atoms in CNTs with diverse geometry hardly affected their van der Waals attraction to a macromolecule, for example, a peptide. However, for small molecules such as water molecules, the effect of geometry of the CNTs on the confined guest molecules could not be negligible. In our previous work [22], the different potential energy surfaces of the water molecules inside the CNTs with different topologies have been observed. The potential energy surfaces reflected by the geometry of the CNT suggest obviously higher energy barrier of water molecules in a zigzag (24, 0) CNT compared to that in an armchair (14, 14) CNT in the axial direction. Based on these literatures, it is conceivable that the encapsulation of the confined macromolecules would be modulated or controlled through the equilibrium conformation of the solvent molecules inside the CNTs. We subsequently repeated the simulations in aqueous solution by immersing the peptide-CNT complex
Z.-S. Zhang et al. / Biomaterials 35 (2014) 1771e1778
1773
Fig. 1. Normalized CoM distance, d/d0, between the peptide and two types of CNTs in vacuo (a) and in aqueous solution (b) as a function of simulation time, where d0 is the initial CoM distance.
in water molecules. For both armchair and zigzag systems 150-ns simulations were performed, and the normalized CoM distance between the peptide and the CNTs as a function of simulation time was shown in Fig. 1(b). In contrast to the results in vacuo, an obviously different encapsulation processes were observed in armchair and zigzag systems. In the armchair system, a stepwise decrease of the d/d0 could be observed during the first 5 ns, before the normalized CoM distance reached around 0.0. It indicates that the spontaneous insertion of the peptide into the armchair CNT took place within 5 ns, and then the peptide kept fluctuating inside
the tube during the following 145 ns simulation. It also can be observed through the snapshots of the simulations as shown in Fig. 2, in which the water molecules were not displayed for clarity. In contrast, in the zigzag system, during the first 3 ns simulation, the d/d0 decreased as that in the armchair system, but afterwards fluctuated around 0.5 for 20 ns approximately. Then the normalized CoM distance continued to decrease sharply to around 0.0 during the following 5 ns simulation, which indicates that the peptide continued to move inside the zigzag CNT after an around 20-ns regulation. The difference demonstrates that in aqueous solution
Fig. 2. MD simulation snapshots of the peptide insertion into an armchair (14, 14) carbon nanotube (left) and a zigzag (24, 0) carbon nanotube (right), respectively. Water molecules are not displayed for clarity.
1774
Z.-S. Zhang et al. / Biomaterials 35 (2014) 1771e1778
the peptide would encounter greater resistance when entering a zigzag CNT compared with an armchair one, which could be attributed to the resistance from the water molecules in the system. As elucidated previously, the main driving force of the peptide encapsulation in CNTs is the van der Waals interaction between the peptide and the CNT, which is reflected by the decrease of the peptideeCNT van der Waals interaction energy during the simulation as shown in Fig. 3(a). In the armchair system, the peptidee CNT van der Waals interaction energy was decreased by 236.7 kcal$mol1 after 150 ns simulation; while in the zigzag system, it was decreased by 211.8 kcal$mol1. The more decreased peptideeCNT interaction energy (around 24.9 kcal$mol1) in the armchair (14, 14) CNT system could be attributed to the slightly smaller tube diameter and greater tube length, and to the different geometry of CNTs, which resulted in more carbon atoms in the armchair CNT within a certain distance of the peptide; however, this energy difference would not be adequate to cause the apparently different encapsulation process in aqueous solution. It can be also proven through the simulations in vacuo, where the difference in encapsulation for both CNTs is negligible. On the other hand, Fig. 3(b) shows the van der Waals interaction energy between water molecules and the CNTs as a function of simulation time. With the insertion of the peptide, the watereCNT interaction energy increased as the water molecules that were initially inside the CNTs were repelled out. In the armchair system, the watereCNT interaction energy was increased by 155.4 kcal$mol1 after 150 ns simulation (from 1376.1 kcal$mol1 to 1220.7 kcal$mol1); while in the zigzag system, it was increased by 131.3 kcal$mol1(from 1281.7 kcal$mol1 to 1150.4 kcal$mol1). The more increased water-CNT interaction energy (around 24.1 kcal$mol1) in the armchair (14, 14) CNT system could be compensated by the decrease of the peptideeCNT interaction energy, and thus in armchair and zigzag systems the interaction energy between the CNT and the guest molecules are almost equal. Therefore the attraction of the CNTs to the guest molecules would not cause the different peptide encapsulation process in the CNTs with diverse topologies. 3.2. Encapsulation free energy profiles of the peptide into CNTs It should be pointed out that the initial coordinate may have impact on the encapsulation process and thus need to be considered. In this work we slightly adjusted the initial coordinates and repeated the simulation 16 times for each system, and the difference in the encapsulation process can also be observed from these
Fig. 4. PMF as a function of the CoM distance between the peptide and the CNT in the axial direction x, together with snapshots taken at some key points.
parallel runs. Based on the sampling of the 16 trajectories for each system, the difference in CoM distance distribution is already observable, where the probability of the peptide located at the entrance of the CNT is apparently higher in zigzag case than that in armchair case (data not shown). Such difference in two systems was further investigated through free energy calculations to obtain a more statistic understanding. The free energy profiles (represented by the PMF) as a function of the CoM distance between the peptide and the CNT in the axial direction x were shown in Fig. 4. The peptide positions corresponding to x along the CNT at some key points were presented. The free energy profiles can be read from right to left corresponding to the spontaneous insertion processes of the peptide. First, in both armchair and zigzag systems the free energy decreased while the peptide moved from the outside to the inside of the CNTs, which confirmed that the encapsulation is a spontaneous process on certain condition. Differently, the free energy of the armchair system declined steadily during the encapsulation, whereas in the zigzag system, there was an obvious platform that occurred at the position corresponding to the d/d0 around 0.5 in Fig. 1(b), where half of the peptide stayed outside the zigzag CNT. This platform could only be found in the zigzag system but not in the armchair system. It hence suggested a higher distribution probability in the zigzag case for the peptide at the position where half of the molecule stayed outside of the tube, which
Fig. 3. (a) van der Waals interaction energy between the peptide and the CNTs as a function of simulation time; (b) van der Waals interaction energy between water molecules and the CNTs as a function of simulation time.
Z.-S. Zhang et al. / Biomaterials 35 (2014) 1771e1778
is in agreement with the phenomenon observed in MD simulations as shown in Fig. 1(b) and Fig. 2. As shown in the simulation results above, with similar tube diameter and length, the armchair (14, 14) CNT encapsulated the investigated peptide much faster than the zigzag (24, 0) CNT did in aqueous solution, which was shown in Fig. 1(b) and Fig. 2. However, the peptideeCNT interaction energy serving as the major driving force for the encapsulation was not considered to cause the different insertion processes. It suggests that the different encapsulation behaviors of the peptide into the CNTs with diverse geometry are not induced by the direct attraction to the peptide. A comparison with the results in vacuo clearly indicated that the water molecules play a remarkable role in modulating the confinement of the macromolecule. In our previous work [22], calculations of potential energy surfaces of water molecules inside armchair and zigzag CNTs revealed that there is a relatively large energy barrier in the axial direction for zigzag (24, 0) CNT, which is barely found for armchair (14, 14) system. Both for the armchair (14, 14) and zigzag (24, 0) CNTs, the minima of the potential energy surfaces are corresponding to the center of hexagons formed by carbon atoms, and the maxima are corresponding to location of the carbon atoms. In the system that a peptide molecule was added, as the macromolecule moved into the tube owing to the van der Waals attraction of the CNT, the water molecules that were initially inside the CNTs were repelled out to the bulk. The distribution of the confined water molecules was consequently changed. Based on the results of potential energy surfaces mentioned above, it would be easier for water molecules to move along the central axis of the armchair (14, 14) CNT; while in zigzag (24, 0) CNTs, the water molecules would encounter greater energy barrier along the tube axis. Under this circumstance, the density profiles of water molecules inside the CNTs were extracted. The results are shown in Fig. 5 and Fig. S1, and the average densities of water molecules inside of both CNTs match very well with the previous report [19]. There is a significant difference between the two density profiles of water molecules inside the two kinds of CNTs (Fig. 5 and Fig. S1). A highly organized distribution pattern of water molecules was observed inside of the zigzag (24, 0) CNT. When presented together with the zigzag (24, 0) CNT (Fig. 5(a)), the maxima in the density profile are found corresponding to the center of hexagons formed by carbon atoms, and the minima corresponding to the carbon atoms. This is consistent with the previous reports [20,22], in which the center of a hexagon formed by carbon atoms was found to be the binding site and the minima of potential energy surfaces. Unlike the result of zigzag (24, 0), the density profile of water molecules did not
1775
present explicitly organized structure inside of the armchair (14, 14) CNT (Fig. 5(b)). This significant discrepancy in density profiles of water molecules inside the two kinds of CNTs could be attributed to the different topology of carbon atoms. The geometric arrangement of carbon atoms in zigzag (24, 0) CNT is in radial direction, which is perpendicular to the encapsulating direction of the peptide (Fig. S2). It leads to a most energy favorable pathway for water molecules perpendicular to the axial direction. The peptide encapsulation would repel the water molecules out of the zigzag (24, 0) CNT, which would be naturally encumbered by the highly organized water molecules. In contrast, the geometric arrangement of carbon atoms in armchair (14, 14) CNT is in axial direction, which is parallel to the encapsulating direction of the peptide (Fig. S2). There would be little resistance caused by the water molecules inside the armchair (14, 14) CNT against the peptide encapsulation. Combined with the encapsulation of the peptide in vacuo, it can be concluded that the geometry arrangement of the carbon atoms in the CNTs could not directly affect the encapsulation behaviors of the biomolecules (in this case, a peptide). It is the confined water molecules with different distribution affected by the geometry of the CNTs that mediate the encapsulation of the peptide. Therefore, although the geometry of the CNT may not directly affect the encapsulation behavior of the macromolecules, it apparently affected the distributions of the inner water molecules and hence mediated the insertion of the macromolecules, indicating that the encapsulation behaviors of biomolecules could be modulated through the CNTs’ geometry, instead of changing the size of the nanomaterial. On the other hand, it also indicates the possibility of detection of the CNT geometry through monitoring the dynamic behaviors of the macromolecules. 3.3. Conformational changes of the peptide in CNTs During the encapsulation in the two types of CNTs, the peptide underwent different levels of conformational changes in response to the confined space. Fig. 6 shows the time dependence of the root mean square deviation (RMSD) of the peptide during the simulations, which indicates the process of stepwise conformational changes of the peptide in both types of the CNTs. In the armchair system, as shown in Fig. 6(a), the RMSD of the peptide changed dramatically within the first 5 ns simulation, corresponding to the insertion process. During the following 145 ns simulation, the conformational changes of the peptide under confinement barely appeared except a slight adjustment at around 55 ns. Similarly in the zigzag system, as shown in Fig. 6(b), the RMSD of the peptide
Fig. 5. Density profiles of water molecules inside the CNTs with different geometry: (a) zigzag (24, 0) and (b) armchair (14, 14).
1776
Z.-S. Zhang et al. / Biomaterials 35 (2014) 1771e1778
Fig. 6. Root mean square deviation (RMSD) of the peptide encapsulated in (a) CNT (14, 14) and (b) CNT (24, 0) as a function of simulation time.
changed obviously within the first 25 ns simulation which corresponds to the insertion process, but less dramatically than that in the armchair system. After completely confined inside the zigzag CNT, the peptide underwent stepwise conformational changes twice, which are more remarkable than those in the armchair system.
Fig. 7 illustrates the secondary structural change of the peptide during the encapsulation in two types of CNTs through the Define Secondary Structure of Proteins (DSSP) program [40]. Fig. 7(a) and (b) shows the details of each residue within the first 50 ns simulations in armchair and zigzag systems, where the color white, green, yellow, blue and gray means the secondary structure Coil,
Fig. 7. Plots of secondary structure change of the peptide during the encapsulation in (a) armchair (14, 14) CNT, and (b) zigzag (24, 0) CNT through DSSP. (c) Number of residues to form the a-helix in the peptide encapsulated in two types of CNTs.
Z.-S. Zhang et al. / Biomaterials 35 (2014) 1771e1778
Bend, Turn, a-helix, and 3/10-helix, respectively. During the encapsulation in the two types of CNTs, the peptide underwent different levels of conformational changes. In the armchair system, as shown in Fig. 7(a), the a-helix disappeared within the first 5 ns simulation. From 5 ns to 8 ns, the residues near the C-terminus transformed to Bend, and the main secondary structure of the peptide was converted to Turn. Then the secondary structure Turn and 3/10-helix alternated with each other as the main structure of the peptide, and the a-helix did not appear in the 145 ns simulation after the complete encapsulation. Differently in the zigzag system, as shown in Fig. 7(b), the peptide lost the a-helix from the N-terminus at around 2 ns, and then the middle of the peptide changed into Turn. The first half of the peptide lost the a-helix when this part entered the CNT, and the second half that stayed outside of the tube still kept the a-helix. After 25 ns simulation, the peptide was completely inside the CNT, and the 11the15th residues always kept as a-helix. The recovery of the a-helix could be observed in the first half of the peptide. During the whole encapsulation, the number of residues to form the a-helix in the peptide as a function of simulation time was shown in Fig. 7(c). In both systems, the decrease of the a-helix component was observed during the insertion. In the armchair system, corresponding to a fast insertion process that took place within 5 ns, the a-helix was destroyed completely and hardly recovered during the following 145-ns simulations. In contrast, although it took relatively long time to encapsulate the peptide in the zigzag CNT, we can observe frequent change in numbers of residues that form a-helix in the zigzag CNT as shown in Fig. 7(c), which was never observed in the armchair case within the total 150 ns simulation. Within the simulations of 150 ns, the difference of secondary structure stability was observe in different CNTs, indicating that the a-helix of the peptide could be better maintained in the zigzag (24, 0) CNT. 4. Conclusion In this study an a-helical peptide was selected as a model to investigate encapsulation of the biomacromolecules in armchair and zigzag SWNTs with similar diameter and length through MD simulations and free energy calculations. The results presented that there was no evident difference between the encapsulation in armchair and in zigzag CNTs in vacuo, whereas in aqueous solution the armchair CNT encapsulated the peptide much easier than the zigzag CNT did. It was found that the different density distribution of the water molecules affected by the geometry of the CNT mediated the different encapsulation processes of the peptide. This suggests that the water molecules play an important role in the bio-system. Although the geometry of the CNT may not directly affect the encapsulation processes of the macromolecules, it significantly affected the equilibrium conformation of the inner water molecules and hence mediated the insertion of the macromolecules. It also indicates that the dynamic behaviors of biomolecules could be modulated through the modification of the geometry of CNTs, instead of changing the size of the nanomaterial. On the other hand, it elucidates the possibility of detection of the CNT geometry through monitoring the dynamic behaviors of the macromolecules. The impact of the CNT geometry on the conformational changes of the peptide was also studied. Analyses of secondary structures showed the a-helix of the peptide could be better maintained in the zigzag CNT, although it spent relatively long time in encapsulating the peptide. Acknowledgments The authors acknowledge the National Natural Science Foundation of China (No. 21273200 and No. J1210042) for the financial supports of this work.
1777
Appendix A. Supplementary data Supplementary data related to this article can be found online at http://dx.doi.org/10.1016/j.biomaterials.2013.11.041. References [1] Zheng M, Jagota A, Semke ED, Diner BA, McLean RS, Lustig SR, et al. DNAassisted dispersion and separation of carbon nanotubes. Nat Mater 2003;2: 338e42. [2] Li Z, Yang W. Capture and manipulation of hybrid DNAs by carbon nanotube bundles. Nanotechnology 2010;21:195301. [3] Salimi A, Compton RG, Hallaj R. Glucose biosensor prepared by glucose oxidase encapsulated sol-gel and carbon-nanotube-modified basal plane pyrolytic graphite electrode. Anal Biochem 2004;333:49e56. [4] Bianco A, Kostarelos K, Partidos CD, Prato M. Biomedical applications of functionalised carbon nanotubes. Chem Commun 2005:571e7. [5] Shen JW, Wu T, Wang Q, Kang Y. Induced stepwise conformational change of human serum albumin on carbon nanotube surfaces. Biomaterials 2008;29: 3847e55. [6] Chen Q, Wang Q, Liu YC, Wu T, Kang Y, Moore JD, et al. Energetics investigation on encapsulation of protein/peptide drugs in carbon nanotubes. J Chem Phys 2009;131:015101. [7] Kang Y, Wang Q, Liu Y-C, Shen J-W, Wu T. Diameter selectivity of protein encapsulation in carbon nanotubes. J Chem Phys B 2010;114:2869e75. [8] Kang Y, Wang Q, Liu Y-C, Wu T, Chen Q, Guan W-J. Dynamic mechanism of collagen-like peptide encapsulated into carbon nanotubes. J Chem Phys B 2008;112:4801e7. [9] Kang Y, Liu Y-C, Wang Q, Shen J-W, Wu T, Guan W-J. On the spontaneous encapsulation of proteins in carbon nanotubes. Biomaterials 2009;30:2807e15. [10] Liu J, Hersam MC. Recent developments in carbon nanotube sorting and selective growth. MRS Bull 2010;35:315e21. [11] Guo D-Z, Zhang G-M, Zhang Z-X, Xue Z-Q, Gu Z-N. Visible-light-induced watersplitting in channels of carbon nanotubes. J Chem Phys B 2006;110:1571e5. [12] Rossi MP, Ye H, Gogotsi Y, Babu S, Ndungu P, Bradley J-C. Environmental scanning electron microscopy study of water in carbon nanopipes. Nano Lett 2004;4:989e93. [13] Zhou X, Li CQ, Iwamoto M. Equilibrium and kinetics: water confined in carbon nanotubes as one-dimensional lattice gas. J Chem Phys 2004;121:7996e8002. [14] Koga K, Gao GT, Tanaka H, Zeng XC. Formation of ordered ice nanotubes inside carbon nanotubes. Nature 2001;412:802e5. [15] Kolesnikov A, Zanotti J-M, Loong C-K, Thiyagarajan P, Moravsky A, Loutfy R, et al. Anomalously soft dynamics of water in a nanotube: a revelation of nanoscale confinement. Phys Rev Lett 2004;93:035503. [16] Hummer G, Rasaiah JC, Noworyta JP. Water conduction through the hydrophobic channel of a carbon nanotube. Nature 2001;414:188e90. [17] Sriraman S, Kevrekidis I, Hummer G. Coarse nonlinear dynamics and metastability of filling-emptying transitions: water in carbon nanotubes. Phys Rev Lett 2005;95:130603. [18] Alexiadis A, Kassinos S. Influence of water model and nanotube rigidity on the density of water in carbon nanotubes. Chem Eng Sci 2008;63:2793e7. [19] Alexiadis A, Kassinos S. The density of water in carbon nanotubes. Chem Eng Sci 2008;63:2047e56. [20] Huang B, Xia Y, Zhao M, Li F, Liu X, Ji Y, et al. Distribution patterns and controllable transport of water inside and outside charged single-walled carbon nanotubes. J Chem Phys 2005;122:084708. [21] Liu Y, Wang Q, Zhang L, Wu T. Dynamics and density profile of water in nanotubes as one-dimensional fluid. Langmuir 2005;21:12025e30. [22] Liu Y-C, Shen J-W, Gubbins KE, Moore JD, Wu T, Wang Q. Diffusion dynamics of water controlled by topology of potential energy surface inside carbon nanotubes. Phys Rev B 2008;77:125438. [23] Alexiadis A, Kassinos S. Molecular simulation of water in carbon nanotubes. Chem Rev 2008;108:5014e34. [24] Al-Haik M, Hussaini MY, Garmestani H. Adhesion energy in carbon nanotubepolyethylene composite: effect of chirality. J Appl Phys 2005;97:074306. [25] Zheng Q, Xue Q, Yan K, Gao X, Li Q, Hao L. Influence of chirality on the interfacial bonding characteristics of carbon nanotube polymer composites. J Appl Phys 2008;103:044302. [26] Steinmetz MO, Jelesarov I, Matousek WM, Honnappa S, Jahnke W, Missimer JH, et al. Molecular basis of coiled-coil formation. Proc Natl Acad Sci U S A 2007;104:7062e7. [27] Jorgensen WL, Chandrasekhar J, Madura JD, Impey RW, Klein ML. Comparison of simple potential functions for simulating liquid water. J Chem Phys 1983;79:926e35. [28] Van Der Spoel D, Lindahl E, Hess B, Groenhof G, Mark AE, Berendsen HJC. GROMACS: fast, flexible, and free. J Comput Chem 2005;26:1701e18. [29] van der Spoel D, Lindahl E, Hess B, van Buuren AR, Apol E, Meulenhoff PJ, et al. Gromacs user manual version 45; 2010. [30] Humphrey W, Dalke A, Schulten K. VMD: visual molecular dynamics. J Mol Graph 1996;14:33e8. [31] Berendsen HJC, Postma JPM, van Gunsteren WF, DiNola A, Haak JR. Molecular dynamics with coupling to an external bath. J Chem Phys 1984;81: 3684e90.
1778
Z.-S. Zhang et al. / Biomaterials 35 (2014) 1771e1778
[32] MacKerell AD, Bashford D, Bellott M, Dunbrack RL, Evanseck JD, Field MJ, et al. All-Atom empirical potential for molecular modeling and dynamics studies of proteins. J Chem Phys B 1998;102:3586e616. [33] Walther JH, Jaffe R, Halicioglu T, Koumoutsakos P. Carbon nanotubes in water: structural characteristics and energetics. J Chem Phys B 2001;105:9980e7. [34] Darden T, York D, Pedersen L. Particle mesh Ewald: an N $ log(N) method for Ewald sums in large systems. J Chem Phys 1993;98:10089e92. [35] Hirschfelder JO, Curtiss CF, Bird RB. Molecular theory of gases and liquids. New York: John Wiley and Sons; 1954. [36] Torrie GM, Valleau JP. Monte Carlo free energy estimates using nonBoltzmann sampling: application to the sub-critical Lennard-Jones fluid. Chem Phys Lett 1974;28:578e81.
[37] Roux B. The calculation of the potential of mean force using computer simulations. Comput Phys Commun 1995;91:275e82. [38] Souaille M, Roux Bt. Extension to the weighted histogram analysis method: combining umbrella sampling with free energy calculations. Comput Phys Commun 2001;135:40e57. [39] Kumar S, Rosenberg JM, Bouzida D, Swendsen RH, Kollman PA. The weighted histogram analysis method for free-energy calculations on biomolecules. I. The method. J Comput Chem 1992;13:1011e21. [40] Kabsch W, Sander C. Dictionary of protein secondary structure: pattern recognition of hydrogen-bonded and geometrical features. Biopolymers 1983;22:2577e637.
Contents lists available at ScienceDirect
Biomaterials journal homepage: www.elsevier.com/locate/biomaterials
Peptide encapsulation regulated by the geometry of carbon nanotubes Zhi-Sen Zhang 1, Yu Kang 1, Li-Jun Liang, Ying-Chun Liu*, Tao Wu, Qi Wang* Soft Matter Research Center and Department of Chemistry, Zhejiang University, Hangzhou 310027, PR China
a r t i c l e i n f o
a b s t r a c t
Article history: Received 21 October 2013 Accepted 14 November 2013 Available online 27 November 2013
In this work the encapsulation of an a-helical peptide in single carbon nanotubes (CNTs) with similar diameter and length but different geometry (armchair and zigzag) was investigated through molecular dynamics simulations and free energy calculations. Our simulation results showed that in vacuo it makes no evident difference whether the investigated peptide is encapsulated in armchair or zigzag CNTs; however, in aqueous solution the armchair CNT encapsulates the peptide remarkably easier than the zigzag CNT does. A detailed analysis revealed that the equilibrium conformation of the water molecules inside the CNTs with varying geometry mediates the peptide encapsulation. It suggests that the water molecules play an important role in regulating behaviors of biomolecules in bio-systems. Then the impact of the CNT geometry on the conformational changes of the confined peptide was studied. Analyses of secondary structures showed the a-helix of the peptide could be better maintained in the zigzag CNT. Ó 2013 Elsevier Ltd. All rights reserved.
Keywords: Geometry of CNTs Peptide encapsulation Interfacial water mediation Umbrella sampling
1. Introduction The properties and applications of hybrid organiceinorganic nanomaterial have aroused great expectation of broad importance for future advancements in nanotechnology. In particular, the biomolecule-CNT complex that consists of single-walled CNTs (SWNTs) wrapped and/or filled with self-assembled biomacromolecules continues to attract scientific attention. It has shown a wide range of interesting features with potential applications in CNT solubilization [1], single-molecule manipulation [2], biosensors [3], and biomedical devices [4], in which a few work have been reported by our group [5e9]. Meanwhile, these applications of SWNTs require their efficient dispersions based on geometry and diameter within cellular environment; separating and dispersing SWNTs, as well as the further detection of the CNTs with different geometry and/or diameters are therefore of importance [10]. To realize these applications, a detailed understanding of the fundamental molecular properties and interactions of the bio-nano complex is needed, which determines the potential applications of nanomaterial in biological settings. When concerning the interactions between CNTs and biomacromolecules, one should also pay attention to the role of the solvent molecules that extensively * Corresponding authors. Tel.: þ86 571 87952424; fax: þ86 571 87951895. E-mail addresses: [email protected] (Y.-C. Liu), [email protected] (Q. Wang). 1 These authors contributed equally to this study and share first authorship. 0142-9612/$ e see front matter Ó 2013 Elsevier Ltd. All rights reserved. http://dx.doi.org/10.1016/j.biomaterials.2013.11.041
exist in the bio-nano systems. Currently the dynamic mechanism of water-CNT systems has been extensively investigated by both experiments [11,12] and simulations [13,14]. For example, Kolesnikov and co-workers [15] investigated the dynamic behaviors of water molecules in CNTs through both molecular dynamics (MD) simulation and neutron scattering techniques. Hummer and co-workers [16,17] reported a one-dimensionally ordered chain of water molecules confined in SWNTs. Alexiadis and co-workers [18,19] studied the density of water in CNTs with various diameters and different geometry. Huang and co-workers [20] revealed the distribution patterns of water inside and outside single-walled carbon nanotubes, suggesting that the water molecules tend to adsorb around the center of a hexagon formed by carbon atoms. Recently, our group reported the effects of geometry and diameter of CNTs on the water diffusion and conduction through MD simulations and quantum mechanics (QM) calculations [21,22]. In these works we have observed the significantly different potential energy surfaces of water molecules in armchair and zigzag CNTs, and confirmed that the minima of potential energy surfaces are at the center of a hexagon formed by carbon atoms and the maxima of potential energy are around the carbon atoms, which is consistent with the previous report [20]. More comprehensive understanding of the interactions between water and CNTs can be found in a review report [23]. Although the effects of geometry of CNTs on the dynamics of small molecules have been widely investigated experimentally and theoretically, the problem became more complicated when macromolecules were incorporated. Currently few works reported the
1772
Z.-S. Zhang et al. / Biomaterials 35 (2014) 1771e1778
impact of CNT topologies on the dynamics of macromolecules. AlHaik and co-workers [24] investigated the binding affinity between polyethylene and CNTs with different geometry through MD simulations. In their work the CNTs with smaller chiral angles were observed to achieve higher adhesion energy, and tend to have smaller diameter and greater length. Zheng and co-workers [25] selected several CNTs with different geometry but similar diameters and lengths to investigate the interfacial binding of polymer-CNT composites, and the armchair CNT was considered to be optimal. However, the influence of the CNT geometry on the biomacromolecules in aqueous solution has not been reported so far, and the mechanism behind it remains obscure. In this study an a-helical peptide was selected as a model molecule to investigate encapsulation behaviors of the biomacromolecules in SWNTs with similar diameters and lengths but different geometries through MD simulations and free energy calculations. We here try to focus on the different encapsulation processes of the peptide in the SWNT with two distinct structures, armchair and zigzag, to elucidate the effects of the tube geometry and the remarkable role of the solvent. The effects of tube length and diameter on the encapsulation of peptide molecule are not involved in this work, which have been investigated and reported by our group [6,7].
2. Modeling and simulation details An a-helical peptide was taken as the model peptide from the Protein Data Bank (PDB), entry code 2OVN [26], with a sequence of NYHLENEVARLKKLCGE. According to the size of the model molecule, two types of uncapped SWNTs with similar tube length and diameter were selected: an armchair (14, 14) CNT and a zigzag (24, 0) CNT, with the tube diameter of 1.88 nm and 1.90 nm, and the tube length of 4.76 nm and 4.54 nm, respectively. The peptide was placed close to the entrance of each type of the nanotube, and was aligned along the tube axis. Then each peptide-CNT complex was immersed in a rectangular periodic box of TIP3P [27] water molecules (or in a rectangular periodic box of vacuum), in which the shortest distances between the complex surface and the box walls are larger than 1.0 nm. In this work, all simulations were performed in the isothermale isobaric (NpT) ensemble with Gromacs 4.5.2 package [28,29] and visualizations were made using VMD [30]. All bonds that involve Hatoms were fixed. A time step of 2 fs was used with atom coordinates saved every 1 ps. A temperature of 310 K and a pressure of 101.3 kPa was maintained using Berendsen method [31]. The CHARMM27 all-atoms (AA) force field [32] was used for describing the peptide and the CNT parameters were supplemented [33]. The atoms of the CNT remain fixed. The particle mesh Ewald (PME) summation [34] was used to calculate the full-system periodic electrostatic interactions, with a cutoff of 1.2 nm for the separation of the direct and reciprocal space summation. Carbon atoms of the CNTs are modeled to be uncharged. The cutoff distance for van der Waals interaction was 1.2 nm, and the parameters of the LennardJones potential for the cross interactions between nonbonded atoms were obtained from the venerable LorentzeBerthelot combination rule [35]. The instantaneous interaction energy between the peptide and the CNT is defined similarly to our previous work [9] as
Eint ðtÞ ¼ EpeptideþCNT ðtÞ Epeptide ðtÞ ECNT ðtÞ
(1)
where Eint(t) stands for the interaction energy between the peptide and the CNT, and EpeptideþCNT(t) refers to the total potential energy of the peptideeCNT complex. Epeptide(t) and ECNT(t) are the potential
energy of the peptide and that of the CNT, respectively. Each term indicates the energy as a function of the simulation time. The free energy profile (represented by potential of mean force, PMF) was calculated by using the umbrella sampling [36e38]. The PMF was calculated along the reaction coordinate x, which was defined as the z-component difference between the center-of-mass (CoM) of the peptide and that of the CNT, describing the entering path of the peptide. The entering path was divided into 1 Å wide equidistant windows perpendicular to the tube axis with the center of each window representing an umbrella center. To enhance sampling, the biased simulations were performed in 41 sampling windows, in which the CoM of the peptide was harmonically restrained in z direction with a harmonic force constant of 400 kJ$mol1$nm2. After carefully analysis, the first 6 ns was removed in each umbrella simulation for thermodynamic equilibration, followed by a 14 ns of production run. The PMFs were then unbiased and combined via the weighted histogram analysis method (WHAM) [39]. To clarify the behavior of the water molecules affected by the CNT, the density profiles of water molecules inside the CNTs were calculated based on the simulations of only a CNT (zigzag or amchair) in aqueous solution, regardless of the peptide. The two CNTs mentioned above were immersed respectively in an aqueous solvent box, with the shortest distance from the carbon atoms in the CNT to the box wall of 1.0 nm in radial direction and 3.0 nm in axial direction. Then 30 ns MD simulations were performed for the two CNT-water systems under the same conditions of temperature and pressure used in the peptide confining case, producing trajectories with 25,000 frames. The first 5000 frames (namely 6 ns) were discarded for thermodynamic equilibration, and the following 20,000 frames were used to extract density profiles of water molecules. 3. Results and discussion 3.1. Spontaneous encapsulation of the peptide into CNTs Fig. 1 shows the normalized CoM distance, d/d0, between the peptide and the CNTs as a function of the simulation time. First the peptide was placed close to the entrance of each type of the nanotube respectively, and for each complex a 5-ns simulation in vacuo was performed. As shown in Fig. 1, the normalized CoM distances decreased dramatically to around 0.0 within 1 ns simulations in both armchair and zigzag systems. It means that in vacuo both CNTs can encapsulate the peptide fast when the macromolecule was placed close to the entrance of the tube, and the geometry of the CNT (armchair or zigzag) barely affected the encapsulation for a given tube length and diameter. This is because the main driving force of the encapsulation is the van der Waals interaction between the peptide and the CNT [8], and the carbon atoms in CNTs with diverse geometry hardly affected their van der Waals attraction to a macromolecule, for example, a peptide. However, for small molecules such as water molecules, the effect of geometry of the CNTs on the confined guest molecules could not be negligible. In our previous work [22], the different potential energy surfaces of the water molecules inside the CNTs with different topologies have been observed. The potential energy surfaces reflected by the geometry of the CNT suggest obviously higher energy barrier of water molecules in a zigzag (24, 0) CNT compared to that in an armchair (14, 14) CNT in the axial direction. Based on these literatures, it is conceivable that the encapsulation of the confined macromolecules would be modulated or controlled through the equilibrium conformation of the solvent molecules inside the CNTs. We subsequently repeated the simulations in aqueous solution by immersing the peptide-CNT complex
Z.-S. Zhang et al. / Biomaterials 35 (2014) 1771e1778
1773
Fig. 1. Normalized CoM distance, d/d0, between the peptide and two types of CNTs in vacuo (a) and in aqueous solution (b) as a function of simulation time, where d0 is the initial CoM distance.
in water molecules. For both armchair and zigzag systems 150-ns simulations were performed, and the normalized CoM distance between the peptide and the CNTs as a function of simulation time was shown in Fig. 1(b). In contrast to the results in vacuo, an obviously different encapsulation processes were observed in armchair and zigzag systems. In the armchair system, a stepwise decrease of the d/d0 could be observed during the first 5 ns, before the normalized CoM distance reached around 0.0. It indicates that the spontaneous insertion of the peptide into the armchair CNT took place within 5 ns, and then the peptide kept fluctuating inside
the tube during the following 145 ns simulation. It also can be observed through the snapshots of the simulations as shown in Fig. 2, in which the water molecules were not displayed for clarity. In contrast, in the zigzag system, during the first 3 ns simulation, the d/d0 decreased as that in the armchair system, but afterwards fluctuated around 0.5 for 20 ns approximately. Then the normalized CoM distance continued to decrease sharply to around 0.0 during the following 5 ns simulation, which indicates that the peptide continued to move inside the zigzag CNT after an around 20-ns regulation. The difference demonstrates that in aqueous solution
Fig. 2. MD simulation snapshots of the peptide insertion into an armchair (14, 14) carbon nanotube (left) and a zigzag (24, 0) carbon nanotube (right), respectively. Water molecules are not displayed for clarity.
1774
Z.-S. Zhang et al. / Biomaterials 35 (2014) 1771e1778
the peptide would encounter greater resistance when entering a zigzag CNT compared with an armchair one, which could be attributed to the resistance from the water molecules in the system. As elucidated previously, the main driving force of the peptide encapsulation in CNTs is the van der Waals interaction between the peptide and the CNT, which is reflected by the decrease of the peptideeCNT van der Waals interaction energy during the simulation as shown in Fig. 3(a). In the armchair system, the peptidee CNT van der Waals interaction energy was decreased by 236.7 kcal$mol1 after 150 ns simulation; while in the zigzag system, it was decreased by 211.8 kcal$mol1. The more decreased peptideeCNT interaction energy (around 24.9 kcal$mol1) in the armchair (14, 14) CNT system could be attributed to the slightly smaller tube diameter and greater tube length, and to the different geometry of CNTs, which resulted in more carbon atoms in the armchair CNT within a certain distance of the peptide; however, this energy difference would not be adequate to cause the apparently different encapsulation process in aqueous solution. It can be also proven through the simulations in vacuo, where the difference in encapsulation for both CNTs is negligible. On the other hand, Fig. 3(b) shows the van der Waals interaction energy between water molecules and the CNTs as a function of simulation time. With the insertion of the peptide, the watereCNT interaction energy increased as the water molecules that were initially inside the CNTs were repelled out. In the armchair system, the watereCNT interaction energy was increased by 155.4 kcal$mol1 after 150 ns simulation (from 1376.1 kcal$mol1 to 1220.7 kcal$mol1); while in the zigzag system, it was increased by 131.3 kcal$mol1(from 1281.7 kcal$mol1 to 1150.4 kcal$mol1). The more increased water-CNT interaction energy (around 24.1 kcal$mol1) in the armchair (14, 14) CNT system could be compensated by the decrease of the peptideeCNT interaction energy, and thus in armchair and zigzag systems the interaction energy between the CNT and the guest molecules are almost equal. Therefore the attraction of the CNTs to the guest molecules would not cause the different peptide encapsulation process in the CNTs with diverse topologies. 3.2. Encapsulation free energy profiles of the peptide into CNTs It should be pointed out that the initial coordinate may have impact on the encapsulation process and thus need to be considered. In this work we slightly adjusted the initial coordinates and repeated the simulation 16 times for each system, and the difference in the encapsulation process can also be observed from these
Fig. 4. PMF as a function of the CoM distance between the peptide and the CNT in the axial direction x, together with snapshots taken at some key points.
parallel runs. Based on the sampling of the 16 trajectories for each system, the difference in CoM distance distribution is already observable, where the probability of the peptide located at the entrance of the CNT is apparently higher in zigzag case than that in armchair case (data not shown). Such difference in two systems was further investigated through free energy calculations to obtain a more statistic understanding. The free energy profiles (represented by the PMF) as a function of the CoM distance between the peptide and the CNT in the axial direction x were shown in Fig. 4. The peptide positions corresponding to x along the CNT at some key points were presented. The free energy profiles can be read from right to left corresponding to the spontaneous insertion processes of the peptide. First, in both armchair and zigzag systems the free energy decreased while the peptide moved from the outside to the inside of the CNTs, which confirmed that the encapsulation is a spontaneous process on certain condition. Differently, the free energy of the armchair system declined steadily during the encapsulation, whereas in the zigzag system, there was an obvious platform that occurred at the position corresponding to the d/d0 around 0.5 in Fig. 1(b), where half of the peptide stayed outside the zigzag CNT. This platform could only be found in the zigzag system but not in the armchair system. It hence suggested a higher distribution probability in the zigzag case for the peptide at the position where half of the molecule stayed outside of the tube, which
Fig. 3. (a) van der Waals interaction energy between the peptide and the CNTs as a function of simulation time; (b) van der Waals interaction energy between water molecules and the CNTs as a function of simulation time.
Z.-S. Zhang et al. / Biomaterials 35 (2014) 1771e1778
is in agreement with the phenomenon observed in MD simulations as shown in Fig. 1(b) and Fig. 2. As shown in the simulation results above, with similar tube diameter and length, the armchair (14, 14) CNT encapsulated the investigated peptide much faster than the zigzag (24, 0) CNT did in aqueous solution, which was shown in Fig. 1(b) and Fig. 2. However, the peptideeCNT interaction energy serving as the major driving force for the encapsulation was not considered to cause the different insertion processes. It suggests that the different encapsulation behaviors of the peptide into the CNTs with diverse geometry are not induced by the direct attraction to the peptide. A comparison with the results in vacuo clearly indicated that the water molecules play a remarkable role in modulating the confinement of the macromolecule. In our previous work [22], calculations of potential energy surfaces of water molecules inside armchair and zigzag CNTs revealed that there is a relatively large energy barrier in the axial direction for zigzag (24, 0) CNT, which is barely found for armchair (14, 14) system. Both for the armchair (14, 14) and zigzag (24, 0) CNTs, the minima of the potential energy surfaces are corresponding to the center of hexagons formed by carbon atoms, and the maxima are corresponding to location of the carbon atoms. In the system that a peptide molecule was added, as the macromolecule moved into the tube owing to the van der Waals attraction of the CNT, the water molecules that were initially inside the CNTs were repelled out to the bulk. The distribution of the confined water molecules was consequently changed. Based on the results of potential energy surfaces mentioned above, it would be easier for water molecules to move along the central axis of the armchair (14, 14) CNT; while in zigzag (24, 0) CNTs, the water molecules would encounter greater energy barrier along the tube axis. Under this circumstance, the density profiles of water molecules inside the CNTs were extracted. The results are shown in Fig. 5 and Fig. S1, and the average densities of water molecules inside of both CNTs match very well with the previous report [19]. There is a significant difference between the two density profiles of water molecules inside the two kinds of CNTs (Fig. 5 and Fig. S1). A highly organized distribution pattern of water molecules was observed inside of the zigzag (24, 0) CNT. When presented together with the zigzag (24, 0) CNT (Fig. 5(a)), the maxima in the density profile are found corresponding to the center of hexagons formed by carbon atoms, and the minima corresponding to the carbon atoms. This is consistent with the previous reports [20,22], in which the center of a hexagon formed by carbon atoms was found to be the binding site and the minima of potential energy surfaces. Unlike the result of zigzag (24, 0), the density profile of water molecules did not
1775
present explicitly organized structure inside of the armchair (14, 14) CNT (Fig. 5(b)). This significant discrepancy in density profiles of water molecules inside the two kinds of CNTs could be attributed to the different topology of carbon atoms. The geometric arrangement of carbon atoms in zigzag (24, 0) CNT is in radial direction, which is perpendicular to the encapsulating direction of the peptide (Fig. S2). It leads to a most energy favorable pathway for water molecules perpendicular to the axial direction. The peptide encapsulation would repel the water molecules out of the zigzag (24, 0) CNT, which would be naturally encumbered by the highly organized water molecules. In contrast, the geometric arrangement of carbon atoms in armchair (14, 14) CNT is in axial direction, which is parallel to the encapsulating direction of the peptide (Fig. S2). There would be little resistance caused by the water molecules inside the armchair (14, 14) CNT against the peptide encapsulation. Combined with the encapsulation of the peptide in vacuo, it can be concluded that the geometry arrangement of the carbon atoms in the CNTs could not directly affect the encapsulation behaviors of the biomolecules (in this case, a peptide). It is the confined water molecules with different distribution affected by the geometry of the CNTs that mediate the encapsulation of the peptide. Therefore, although the geometry of the CNT may not directly affect the encapsulation behavior of the macromolecules, it apparently affected the distributions of the inner water molecules and hence mediated the insertion of the macromolecules, indicating that the encapsulation behaviors of biomolecules could be modulated through the CNTs’ geometry, instead of changing the size of the nanomaterial. On the other hand, it also indicates the possibility of detection of the CNT geometry through monitoring the dynamic behaviors of the macromolecules. 3.3. Conformational changes of the peptide in CNTs During the encapsulation in the two types of CNTs, the peptide underwent different levels of conformational changes in response to the confined space. Fig. 6 shows the time dependence of the root mean square deviation (RMSD) of the peptide during the simulations, which indicates the process of stepwise conformational changes of the peptide in both types of the CNTs. In the armchair system, as shown in Fig. 6(a), the RMSD of the peptide changed dramatically within the first 5 ns simulation, corresponding to the insertion process. During the following 145 ns simulation, the conformational changes of the peptide under confinement barely appeared except a slight adjustment at around 55 ns. Similarly in the zigzag system, as shown in Fig. 6(b), the RMSD of the peptide
Fig. 5. Density profiles of water molecules inside the CNTs with different geometry: (a) zigzag (24, 0) and (b) armchair (14, 14).
1776
Z.-S. Zhang et al. / Biomaterials 35 (2014) 1771e1778
Fig. 6. Root mean square deviation (RMSD) of the peptide encapsulated in (a) CNT (14, 14) and (b) CNT (24, 0) as a function of simulation time.
changed obviously within the first 25 ns simulation which corresponds to the insertion process, but less dramatically than that in the armchair system. After completely confined inside the zigzag CNT, the peptide underwent stepwise conformational changes twice, which are more remarkable than those in the armchair system.
Fig. 7 illustrates the secondary structural change of the peptide during the encapsulation in two types of CNTs through the Define Secondary Structure of Proteins (DSSP) program [40]. Fig. 7(a) and (b) shows the details of each residue within the first 50 ns simulations in armchair and zigzag systems, where the color white, green, yellow, blue and gray means the secondary structure Coil,
Fig. 7. Plots of secondary structure change of the peptide during the encapsulation in (a) armchair (14, 14) CNT, and (b) zigzag (24, 0) CNT through DSSP. (c) Number of residues to form the a-helix in the peptide encapsulated in two types of CNTs.
Z.-S. Zhang et al. / Biomaterials 35 (2014) 1771e1778
Bend, Turn, a-helix, and 3/10-helix, respectively. During the encapsulation in the two types of CNTs, the peptide underwent different levels of conformational changes. In the armchair system, as shown in Fig. 7(a), the a-helix disappeared within the first 5 ns simulation. From 5 ns to 8 ns, the residues near the C-terminus transformed to Bend, and the main secondary structure of the peptide was converted to Turn. Then the secondary structure Turn and 3/10-helix alternated with each other as the main structure of the peptide, and the a-helix did not appear in the 145 ns simulation after the complete encapsulation. Differently in the zigzag system, as shown in Fig. 7(b), the peptide lost the a-helix from the N-terminus at around 2 ns, and then the middle of the peptide changed into Turn. The first half of the peptide lost the a-helix when this part entered the CNT, and the second half that stayed outside of the tube still kept the a-helix. After 25 ns simulation, the peptide was completely inside the CNT, and the 11the15th residues always kept as a-helix. The recovery of the a-helix could be observed in the first half of the peptide. During the whole encapsulation, the number of residues to form the a-helix in the peptide as a function of simulation time was shown in Fig. 7(c). In both systems, the decrease of the a-helix component was observed during the insertion. In the armchair system, corresponding to a fast insertion process that took place within 5 ns, the a-helix was destroyed completely and hardly recovered during the following 145-ns simulations. In contrast, although it took relatively long time to encapsulate the peptide in the zigzag CNT, we can observe frequent change in numbers of residues that form a-helix in the zigzag CNT as shown in Fig. 7(c), which was never observed in the armchair case within the total 150 ns simulation. Within the simulations of 150 ns, the difference of secondary structure stability was observe in different CNTs, indicating that the a-helix of the peptide could be better maintained in the zigzag (24, 0) CNT. 4. Conclusion In this study an a-helical peptide was selected as a model to investigate encapsulation of the biomacromolecules in armchair and zigzag SWNTs with similar diameter and length through MD simulations and free energy calculations. The results presented that there was no evident difference between the encapsulation in armchair and in zigzag CNTs in vacuo, whereas in aqueous solution the armchair CNT encapsulated the peptide much easier than the zigzag CNT did. It was found that the different density distribution of the water molecules affected by the geometry of the CNT mediated the different encapsulation processes of the peptide. This suggests that the water molecules play an important role in the bio-system. Although the geometry of the CNT may not directly affect the encapsulation processes of the macromolecules, it significantly affected the equilibrium conformation of the inner water molecules and hence mediated the insertion of the macromolecules. It also indicates that the dynamic behaviors of biomolecules could be modulated through the modification of the geometry of CNTs, instead of changing the size of the nanomaterial. On the other hand, it elucidates the possibility of detection of the CNT geometry through monitoring the dynamic behaviors of the macromolecules. The impact of the CNT geometry on the conformational changes of the peptide was also studied. Analyses of secondary structures showed the a-helix of the peptide could be better maintained in the zigzag CNT, although it spent relatively long time in encapsulating the peptide. Acknowledgments The authors acknowledge the National Natural Science Foundation of China (No. 21273200 and No. J1210042) for the financial supports of this work.
1777
Appendix A. Supplementary data Supplementary data related to this article can be found online at http://dx.doi.org/10.1016/j.biomaterials.2013.11.041. References [1] Zheng M, Jagota A, Semke ED, Diner BA, McLean RS, Lustig SR, et al. DNAassisted dispersion and separation of carbon nanotubes. Nat Mater 2003;2: 338e42. [2] Li Z, Yang W. Capture and manipulation of hybrid DNAs by carbon nanotube bundles. Nanotechnology 2010;21:195301. [3] Salimi A, Compton RG, Hallaj R. Glucose biosensor prepared by glucose oxidase encapsulated sol-gel and carbon-nanotube-modified basal plane pyrolytic graphite electrode. Anal Biochem 2004;333:49e56. [4] Bianco A, Kostarelos K, Partidos CD, Prato M. Biomedical applications of functionalised carbon nanotubes. Chem Commun 2005:571e7. [5] Shen JW, Wu T, Wang Q, Kang Y. Induced stepwise conformational change of human serum albumin on carbon nanotube surfaces. Biomaterials 2008;29: 3847e55. [6] Chen Q, Wang Q, Liu YC, Wu T, Kang Y, Moore JD, et al. Energetics investigation on encapsulation of protein/peptide drugs in carbon nanotubes. J Chem Phys 2009;131:015101. [7] Kang Y, Wang Q, Liu Y-C, Shen J-W, Wu T. Diameter selectivity of protein encapsulation in carbon nanotubes. J Chem Phys B 2010;114:2869e75. [8] Kang Y, Wang Q, Liu Y-C, Wu T, Chen Q, Guan W-J. Dynamic mechanism of collagen-like peptide encapsulated into carbon nanotubes. J Chem Phys B 2008;112:4801e7. [9] Kang Y, Liu Y-C, Wang Q, Shen J-W, Wu T, Guan W-J. On the spontaneous encapsulation of proteins in carbon nanotubes. Biomaterials 2009;30:2807e15. [10] Liu J, Hersam MC. Recent developments in carbon nanotube sorting and selective growth. MRS Bull 2010;35:315e21. [11] Guo D-Z, Zhang G-M, Zhang Z-X, Xue Z-Q, Gu Z-N. Visible-light-induced watersplitting in channels of carbon nanotubes. J Chem Phys B 2006;110:1571e5. [12] Rossi MP, Ye H, Gogotsi Y, Babu S, Ndungu P, Bradley J-C. Environmental scanning electron microscopy study of water in carbon nanopipes. Nano Lett 2004;4:989e93. [13] Zhou X, Li CQ, Iwamoto M. Equilibrium and kinetics: water confined in carbon nanotubes as one-dimensional lattice gas. J Chem Phys 2004;121:7996e8002. [14] Koga K, Gao GT, Tanaka H, Zeng XC. Formation of ordered ice nanotubes inside carbon nanotubes. Nature 2001;412:802e5. [15] Kolesnikov A, Zanotti J-M, Loong C-K, Thiyagarajan P, Moravsky A, Loutfy R, et al. Anomalously soft dynamics of water in a nanotube: a revelation of nanoscale confinement. Phys Rev Lett 2004;93:035503. [16] Hummer G, Rasaiah JC, Noworyta JP. Water conduction through the hydrophobic channel of a carbon nanotube. Nature 2001;414:188e90. [17] Sriraman S, Kevrekidis I, Hummer G. Coarse nonlinear dynamics and metastability of filling-emptying transitions: water in carbon nanotubes. Phys Rev Lett 2005;95:130603. [18] Alexiadis A, Kassinos S. Influence of water model and nanotube rigidity on the density of water in carbon nanotubes. Chem Eng Sci 2008;63:2793e7. [19] Alexiadis A, Kassinos S. The density of water in carbon nanotubes. Chem Eng Sci 2008;63:2047e56. [20] Huang B, Xia Y, Zhao M, Li F, Liu X, Ji Y, et al. Distribution patterns and controllable transport of water inside and outside charged single-walled carbon nanotubes. J Chem Phys 2005;122:084708. [21] Liu Y, Wang Q, Zhang L, Wu T. Dynamics and density profile of water in nanotubes as one-dimensional fluid. Langmuir 2005;21:12025e30. [22] Liu Y-C, Shen J-W, Gubbins KE, Moore JD, Wu T, Wang Q. Diffusion dynamics of water controlled by topology of potential energy surface inside carbon nanotubes. Phys Rev B 2008;77:125438. [23] Alexiadis A, Kassinos S. Molecular simulation of water in carbon nanotubes. Chem Rev 2008;108:5014e34. [24] Al-Haik M, Hussaini MY, Garmestani H. Adhesion energy in carbon nanotubepolyethylene composite: effect of chirality. J Appl Phys 2005;97:074306. [25] Zheng Q, Xue Q, Yan K, Gao X, Li Q, Hao L. Influence of chirality on the interfacial bonding characteristics of carbon nanotube polymer composites. J Appl Phys 2008;103:044302. [26] Steinmetz MO, Jelesarov I, Matousek WM, Honnappa S, Jahnke W, Missimer JH, et al. Molecular basis of coiled-coil formation. Proc Natl Acad Sci U S A 2007;104:7062e7. [27] Jorgensen WL, Chandrasekhar J, Madura JD, Impey RW, Klein ML. Comparison of simple potential functions for simulating liquid water. J Chem Phys 1983;79:926e35. [28] Van Der Spoel D, Lindahl E, Hess B, Groenhof G, Mark AE, Berendsen HJC. GROMACS: fast, flexible, and free. J Comput Chem 2005;26:1701e18. [29] van der Spoel D, Lindahl E, Hess B, van Buuren AR, Apol E, Meulenhoff PJ, et al. Gromacs user manual version 45; 2010. [30] Humphrey W, Dalke A, Schulten K. VMD: visual molecular dynamics. J Mol Graph 1996;14:33e8. [31] Berendsen HJC, Postma JPM, van Gunsteren WF, DiNola A, Haak JR. Molecular dynamics with coupling to an external bath. J Chem Phys 1984;81: 3684e90.
1778
Z.-S. Zhang et al. / Biomaterials 35 (2014) 1771e1778
[32] MacKerell AD, Bashford D, Bellott M, Dunbrack RL, Evanseck JD, Field MJ, et al. All-Atom empirical potential for molecular modeling and dynamics studies of proteins. J Chem Phys B 1998;102:3586e616. [33] Walther JH, Jaffe R, Halicioglu T, Koumoutsakos P. Carbon nanotubes in water: structural characteristics and energetics. J Chem Phys B 2001;105:9980e7. [34] Darden T, York D, Pedersen L. Particle mesh Ewald: an N $ log(N) method for Ewald sums in large systems. J Chem Phys 1993;98:10089e92. [35] Hirschfelder JO, Curtiss CF, Bird RB. Molecular theory of gases and liquids. New York: John Wiley and Sons; 1954. [36] Torrie GM, Valleau JP. Monte Carlo free energy estimates using nonBoltzmann sampling: application to the sub-critical Lennard-Jones fluid. Chem Phys Lett 1974;28:578e81.
[37] Roux B. The calculation of the potential of mean force using computer simulations. Comput Phys Commun 1995;91:275e82. [38] Souaille M, Roux Bt. Extension to the weighted histogram analysis method: combining umbrella sampling with free energy calculations. Comput Phys Commun 2001;135:40e57. [39] Kumar S, Rosenberg JM, Bouzida D, Swendsen RH, Kollman PA. The weighted histogram analysis method for free-energy calculations on biomolecules. I. The method. J Comput Chem 1992;13:1011e21. [40] Kabsch W, Sander C. Dictionary of protein secondary structure: pattern recognition of hydrogen-bonded and geometrical features. Biopolymers 1983;22:2577e637.
