Available online at www.sciencedirect.com
ScienceDirect Procedia Materials Science 11 (2015) 423 – 427
5th International Biennial Conference on Ultrafine Grained and Nanostructured Materials, UFGNSM15
Multi-Scale Simulation of Carbon Nanotubes Interactions with Cell Membrane: DFT Calculations and Molecular Dynamic Simulation S. H. Tabaria, Y. Jamalib,c, R. Poursalehia,* a
Nanomaterials group, Department of Materials Engineering, Tarbiat Modares University, Tehran 14115-143, Iran Department of Applied Mathematics, School of Mathematical Sciences, Tarbiat Modares University, 14115-134, Iran c Computational Physics Research Laboratory, School of Nano-Science, Institute for research in Fundamental Sciences (IPM), Tehran, Iran b
Abstract Due to unique physiochemical properties of carbon nanotubes (CNTs), they are utilized in biomedicine applications including biomedical engineering, tissue engineering, drug delivery, gene therapy and biosensor engineering. Studies have been shown that carbon nanotube have a large propensity to interact with biomembranes; therefore, the toxicity of these nanomaterials and understanding the way CNTs interact with cell become more important. Recently, enormous efforts have been made to understand the molecular mechanism of CNTs and lipid bilayer interaction. Because of the relative large scale of lipid bilayer, quantum ab-initio methods are not common for such systems; consequently, both all atom and coarse grained molecular dynamic simulation (MD) are widely performed. For simulation of such system it is important to apply the correct physical and structural properties of carbon nanotube in the simulation, for instance with density functional theory (DFT) modeling it has been clear that, for open-ended CNTs not only is there unsaturated bonds at the end of tube, but also has a significant dipole moment which should not neglected and their quantity correspond to nanotube chirality and diameter. The aim of this work is to discuss the partial charges distribution along the carbon nanotubes and describe how they affect in CNTs -lipid bilayer interaction by free energy calculation. Our Simulation consist of two stage, first partial charges for pristine and functional carbon nanotube were determined by DFT calculation then results were applied to molecular dynamic force field. The DFT simulation computes partial charges quantity at the ends of nanotube and its results indicate that partial charges for functional nanotubes are higher than pristine nanotubes. The MD simulation determines that nanotubes interaction energy with biomembrane is highly strong, about 120 kcal/mol. © 2015 2015Published The Authors. Published by Elsevier Ltd. by Elsevier Ltd. This is an open access article under the CC BY-NC-ND license Peer-review under responsibility of the organizing committee of UFGNSM15. (http://creativecommons.org/licenses/by-nc-nd/4.0/). Peer-review under responsibility of the organizing committee of UFGNSM15 Keywords: molecular dynamic; partial charge; free energy; carbon nanotube; cell membrane; density functional theory.
* Corresponding author. Tel : +98-218-288-3997 Fax : +98-218-288-3997 . E-mail address:
[email protected]
2211-8128 © 2015 Published by Elsevier Ltd. This is an open access article under the CC BY-NC-ND license (http://creativecommons.org/licenses/by-nc-nd/4.0/). Peer-review under responsibility of the organizing committee of UFGNSM15 doi:10.1016/j.mspro.2015.11.025
424
S.H. Tabari et al. / Procedia Materials Science 11 (2015) 423 – 427
1. Introduction Due to extraordinary physiochemical properties of single wall carbon nanotubes (SWCNTs), Baughman et al. (2002), they not only are used in numerous technological applications, but also attracted great interest in biological applications, Kostarelos et al. (2009). Examples of their usage in biomedical applications are drug delivery, cellular imaging, and artificial ion channels, Holt et al. (2006). Studies have been shown that SWCNTs have a large propensity to interact with biomembranes, Kraszewski et al. (2012) ; therefore, the toxicity of this nanomaterials become more important, Brown et al. (2007). In the past 20 years, the number of studies that assess SWCNTs toxicity have been significantly increased; For instance it was reported that SWCNTs cause inflammation and carcinogenesis, Poland et al. (2008). Generally, nanoparticles interact through four mechanisms with bio-membrane, including adsorption onto the membrane surface, pore formation, direct penetration and endocytosis, He et al. (2015). Although many studies have been conducted to elucidate the carbon nanotubes (CNTs) and bio-membrane interactions, the uptake mechanism of CNTs into cells has long been a topic of controversy. Experimental study has been shown that CNTs insert in lipid bilayer perpendicularly and form a nanopore, which is named CNT porins and have an excellent transport properties as well as biological membrane channels, Geng et al. (2014) . Recently, enormous efforts have been made to understand the molecular mechanism of SWCNTs and lipid bilayer interaction. Coarse grained molecular dynamic (CG-MD) simulation was used with the Martini force field for different size and functional group of carbon nanotube to calculate free energy insertion into a bio-membrane, Baoukina et al. (2013). Uptake mechanism of amino functionalized and pristine SWCNTs into lipid bilayer were computationally studied with using CG-MD and implementation of free energy calculation. Their result suggests that SWCNTs passively diffuse into lipid bilayer which free energy difference for the mechanism equals -21 kcal/mol, Kraszewski et al. (2012). Molecular dynamic simulation could make not accurate prediction as a result of force fields limitations. Main challenge for simulation of such system is carbon nanotube force field, but with density functional theory (DFT) simulation can respond to the challenge, Dandpat and Sarkar (2015). For simulation of such system it is important to apply the correct physical and structural properties of carbon nanotube in the simulation, for instance with DFT modeling it has been clear that, for open-ended CNTs not only there is unsaturated bonds at the end of tube, but also has a significant dipole moment which should not neglected and their quantity correspond to nanotube chirality and diameter, Won et al. (2006). This paper discusses the partial charge distribution along the carbon nanotube and shows how they affect in SWCNTs - lipid bilayer interaction by free energy calculation. We determine partial charge for pristine and functional carbon nanotube by DFT calculation and then apply the result to molecular dynamic and modify force field parameterization. Because of strong dipole moment of functionalized carbon nanotube, they strongly interact with bio-membrane. Our results shed light on the effect of partial charges on uptake mechanism that could help to facilitate drug designing. 2. Simulation setups The DFT method with the B3LYP model and the 6–31G basis** were implemented to calculate partial charge. GAUSSIAN package was used to optimize the geometries and to calculate the partial charges, Frisch et al. (2004). We considered a (10,0) pristine CNT, Won et al., (2006) and a (10,0) functionalized CNT with hydroxyl functional group with length of 1.3 nm as shown in figure 1. Initial C-C, C-O and O-H length were fixed at 1.42 Å, 1.36 Å and 0.93 Å and also initial COH bond angle was chosen 109°. The partial charges on the SWCNT were acquired by fitting the electrostatic potential to fixed charge at the carbon atoms using the Charges from Electrostatic Potentials using a Grid based method (CHELPG). The CHELPG charges are applied to MD simulations because they capture higher order effects arising from dipoles, Breneman et al. (1990). Molecular dynamics simulations were performed with NAMD package, Phillips et al. (2005) and visualized in VMD, Humphrey et al. (1996). The simulated system consisted of 121 1,2-dioleoyl-sn-glycerol-3-phosphocholine (DOPC) , 8932 water molecules and a CNT molecule with 4nm length, which simulation box was 6×6×12 nm 3 .The Charmm36 force field, MacKerell et al. (2004) and TIP3 water model were used for DOPC and water molecule, and
S.H. Tabari et al. / Procedia Materials Science 11 (2015) 423 – 427
also nanotubes. All simulations are carried out using 2 fs time steps and under a constant pressure of 1 atm using a Langevin piston and a constant temperature of 298 K using Langevin dynamics. Each simulation employs a periodic boundary condition in all three dimensions, and the particle mesh Ewald method is used to determine electrostatic interactions. A switch distance of 10 Å and a cutoơ distance of 12 Å are used for electrostatic and van der Waals interactions. The free energy profile was estimated using GPU accelerated NAMD software with the adaptive biasing force (ABF), Darve et al. (2008) in the Collective Variables module and under the same conditions as described for the MD simulations. In order to calculate insertion free energy, distance between center of mass (COM) of the CNT and the COM of lipid bilayer along z axis taken as reaction coordinates which vary from 0 to 6nm and in all of simulation, CNT was constrained in x-y plane to prevent CNT rotation and also calculate free energy in CNT perpendicular state with respect to the membrane plane.
Fig. 1. Visualization of a single-walled carbon nanotube with hydroxyl functional group.
3. Results and discussion Figure 2 illustrates the CHELPG partial charges for the (10,0) zigzag SWCNTs with and without functional group along the axis of the tube. The partial charges were averaged for atoms with the same axial coordinate. The amount of the electrostatic charges at the tube ends is much greater compared to the charges in the middle region of the tube. The partial charges quickly decline about 2 Å from the ends of the tube. As result of end effects, partial charges distribute unequally just over nanotube ends, therefore, there is no correlation between end effects and middle region of the tube. With applying partial charges distribution to nanotubes, it is possible to increase the middle region length in order to simulate longer nanotubes.
Fig. 2. Partial charge distribution on a SWCNT along axial direction. Blue line: partial charges for pristine CNT. Red line: partial charges for functionalized CNT.
425
426
S.H. Tabari et al. / Procedia Materials Science 11 (2015) 423 – 427
The effect of functional group on the nanotube electrostatics can be understood by comparing blue and red line in Fig. 2. The magnitude of the average partial charge at the ends of the functionalized SWCNT is greater than the pristine one. This reveals that the electrostatic potential at the ends of the nanotubes depends greatly on end effects and functional groups. As result of strong tendency of oxygen to attract electrons, the electrons are more localized near the oxygen, which increase the magnitude of the oxygen partial charges.
Fig. 3. The free energy profiles obtained using the ABF approach of nanotubes inserting into a DOPC bilayer. Blue line: free energy for pristine CNT with partial charge modification. Red line: free energy for functionalized CNT.
Using ABF method, we calculated free energy for pristine (with partial charge modification) and functionalized nanotube insert into lipid bilayer after applying DFT results. The free energy difference between water and the bilayer center in perpendicular orientation for pristine and functionalized nanotube equal 117 kcal/mol and 125 kcal/mol, which these magnitudes are expected, Baoukina et al. (2013). DFT calculation suggests that a functionalized nanotube is more polar than the pristine nanotube; hence, functionalized nanotube has greater tendency for insertion into lipid bilayer. The profiles shown in Figure 3 exhibit an attractive potentials because geometrical structure of nanotubes are analogous to lipid bilayer, they have polar heads with hydrophobic bodies. Free energy curve for pristine nanotube has a local minimum in 5nm corresponding to interaction between lipid upper head groups and CNT lower polar part. Derivative of free energy curve in Fig. 3 corresponds to applied force of CNT. The average thickness of the lipid in our system is 5nm and also length of CNTs are 4 nm, when difference in COMs is 0.5 nm, polar region of CNTs come into contact with lipid head groups; therefore, variation of free energy slope occurs in 0.5nm. 4. Conclusion Pristine and functionalized carbon nanotube with (10,0) chirality have different electrostatic behaviors for their different functional groups. We have shown that the difference in electrostatic behaviors of nanotubes exert substantial effect on their interaction with lipid bilayer. With variation of functional groups, attractive potential of nanotubes can be tuned to favored purposes. Calculating free energy reveals that attractive potential of functionalized nanotube for insertion is stronger than pristine one; however, pristine nanotube has metastable states at earlier stages. Different forces, which correspond to slope of free energy curve, apply to nanotubes for the duration of insertion and variations originate from interaction between polar region of lipid bilayer and partially charged parts of nanotubes.
S.H. Tabari et al. / Procedia Materials Science 11 (2015) 423 – 427
References Baoukina, S., Monticcelli, L.,Tieleman, D. P., 2013. Interaction of pristine and functionalized carbon nanotubes with lipid membranes. The Journal of Physical Chemistry B 117, 12113-12123. Baughman, R. H., Zakhidov, A. A., Deheer, W. A., 2002. Carbon nanotubes the route toward applications. Science 297, 787-792. Breneman, C. M., Wiberg, K. B., 1990. Determining atom centered monopoles from molecular electrostatic potentials. The need for high sampling density in formamide conformational analysis. Journal of Computational Chemistry 11, 361-373. Brown, D. M., Kinloch, I. A., Bangert, U., Windle, A., Walter D., Walker G., Scotchford C., Donaldson K., Stone V., 2007. An in vitro study of the potential of carbon nanotubes and nanofibres to induce inflammatory mediators and frustrated phagocytosis. Carbon 45, 1743-1756. Dandpat, S. S., Sarkar, M., 2015. Investigating the molecular and aggregated states of a drug molecule rutaecarpine using spectroscopy, microscopy, crystallography and computational studies. Physical Chemistry Chemical Physics 17, 13992-14002. Darve, E., Rodríguez-gómez, D., Pohorille, A., 2008. Adaptive biasing force method for scalar and vector free energy calculations. The Journal of Chemical Physics 128, 144120. Frisch, M., Trucks, G., Schlegel, H., Scuseria, G., Robb, M., Cheeseman, J., Montgomery, Jr J., Vreven, T., Kudin, K. , Burant, J., 2004. Gaussian 03, revision D. 01. Gaussian Inc. Wallingford, CT, 26. Geng, J., Kim, K., Zhang, J., Escalada, A., Tunuguntla, R., Comolli, L. R., Allen, F. I., Shnyrova, A. V., Cho, K. R., Munoz, D., 2014. Stochastic transport through carbon nanotubes in lipid bilayers and live cell membranes. Nature 514, 612-615. He, X. C., Lin, M., Li, F., Sha, B. Y., Xu, F., Qu, Z. G., Wang, L. 2015. Advances in studies of nanoparticle-biomembrane interactions. Nanomedicine 10, 121-141. Holt, J. K., Park, H. G., Wang, Y., Stadermann, M., Artyukhin, A. B., Grigoropoulos, C. P., Noy, A., Bakajin, O. 2006. Fast mass transport through sub-2-nanometer carbon nanotubes. Science 312, 1034-1037. Humphrey, W., Dalke, A., Schulten, K., 1996. VMD: visual molecular dynamics. Journal of Molecular Graphics 14, 33-38. Kostarelos, K., Bianco, A., Prato, M., 2009. Promises, facts and challenges for carbon nanotubes in imaging and therapeutics. Nature Nanotechnology 4, 627-633. Kraszewski, S., Bianco, A., Tarek, M., Ramseyer, C., 2012. Insertion of short amino-functionalized single-walled carbon nanotubes into phospholipid bilayer occurs by passive diffusion. PLoS One 7, 40703. Mackerell, A. D., Feig, M., Brooks, C. L., 2004. Extending the treatment of backbone energetics in protein force fields: Limitations of gasϋ phase quantum mechanics in reproducing protein conformational distributions in molecular dynamics simulations. Journal of Computational Chemistry 25, 1400-1415. Phillips, J. C., Braun, R., Wang, W., Gumbart, J., Tajkhorshid, E., Villa, E., Chipot, C., Skeel, R. D., Kale, L., Schulten, K., 2005. Scalable molecular dynamics with NAMD. Journal of Computational Chemistry 26, 1781-1802. Poland, C. A., Duffin, R., Kinloch, I., Maynard, A., Wallace, W. A., Seaton, A., Stone, V., Brown, S., Macnee, W., Donaldson, K., 2008. Carbon nanotubes introduced into the abdominal cavity of mice show asbestos-like pathogenicity in a pilot study. Nature Nanotechnology 3, 423428. Won, C. Y., Joseph, S., Aluru, N., 2006. Effect of quantum partial charges on the structure and dynamics of water in single-walled carbon nanotubes. The Journal of Chemical Physics 125, 114701.
427
