Designing a C84 fullerene as a specific voltage-gated sodium channel blocker
© Hilder and Chung; licensee Springer. 2013
Received: 11 June 2013
Accepted: 9 July 2013
Published: 16 July 2013
Fullerene derivatives demonstrate considerable potential for numerous biological applications, such as the effective inhibition of HIV protease. Recently, they were identified for their ability to indiscriminately block biological ion channels. A fullerene derivative which specifically blocks a particular ion channel could lead to a new set of drug leads for the treatment of various ion channel-related diseases. Here, we demonstrate their extraordinary potential by designing a fullerene which mimics some of the functions of μ-conotoxin, a peptide derived from cone snail venom which potently binds to the bacterial voltage-gated sodium channel (NavAb). We show, using molecular dynamics simulations, that the C84 fullerene with six lysine derivatives uniformly attached to its surface is selective to NavAb over a voltage-gated potassium channel (Kv1.3). The side chain of one of the lysine residues protrudes into the selectivity filter of the channel, while the methionine residues located just outside of the channel form hydrophobic contacts with the carbon atoms of the fullerene. The modified C84 fullerene strongly binds to the NavAb channel with an affinity of 46 nM but binds weakly to Kv1.3 with an affinity of 3 mM. This potent blocker of NavAb may serve as a structural template from which potent compounds can be designed for the targeting of mammalian Nav channels. There is a genuine need to target mammalian Nav channels as a form of treatment of various diseases which have been linked to their malfunction, such as epilepsy and chronic pain.
KeywordsC84 fullerene derivative Bacterial voltage-gated sodium channel NavAb Voltage-gated potassium channel Kv1.3 Binding Molecular dynamics 87.85.Qr 87.85.Rs 87.10.Tf
Typically, toxins from venomous species such as cone snails, spiders, and snakes are investigated as possible drug leads for ion channel blockers. Converting these toxins to drugs represents a considerable challenge . For example, disulfide bridges in these peptides, abundant in all toxins, are vulnerable to scrambling and reduction in certain extracellular environments and therefore must be replaced [1–4]. Nanomaterials designed to mimic the main features of these complex toxin structures present exciting opportunities to specifically target a particular ion channel subtype and may alleviate some of the challenges of these peptides.
Increasing attention is being given to fullerenes for biological applications including antiviral and antibacterial agents, antioxidants, vectors for drug/gene delivery, photodynamic therapy, enzyme inhibitors, and diagnostics (e.g., magnetic resonance imaging) [5, 6]. For example, fullerene derivatives have been shown to bind to and inhibit the activity of HIV protease . Fullerenes consist of a hollow carbon cage structure formed by 20 to as many as 300 carbon atoms [8, 9]. The most abundantly produced are those with 60 and 70 carbon atoms. Fullerenes are insoluble in aqueous solution and aggregate easily. Therefore, there has been significant work into making these structures soluble so that they can be utilized for their potential biomedical applications. One method which increases their solubility is chemical functionalization with moieties such as amino acids and carboxylic acid . Fullerene chemistry has been intensely developed, and the main efforts are now devoted to broaden their application .
In 2003, Park et al.  identified non-functionalized carbon nanotubes and C60 fullerenes as a novel class of ion channel blockers. Their experiments on various biological ion channels demonstrated that these nanostructures indiscriminately interfere with the activity of potassium channels depending on their geometric structure and size. Similarly, experiments by Chhowalla et al.  and Xu et al.  demonstrated that functionalized multi-walled and single-walled carbon nanotubes affected the activity of ion channels in Chinese hamster ovary cells (CHO) and undifferentiated pheochromocytoma (PC12) cells, respectively. Park et al.  also examined the binding of their fullerenes and nanotubes to KcsA using docking simulations and proposed that the molecules block the entrance to the pore. In contrast, Kraszewski et al.  showed using molecular dynamics simulations that C60 fullerenes do not bind to the selectivity filter. Instead, they demonstrated that C60 fullerenes bind strongly to the hydrophobic residues of the extracellular loops in the three potassium channels they examined, namely KcsA, MthK, and Kv1.2, and suggest that these fullerenes may hinder the function of potassium channels . Similarly, Monticelli et al. examined the interaction of a C70 fullerene with the Kv1.2 potassium channel using molecular dynamics and found that they made contact with hydrophobic residues in the extracellular or intracellular loops, but not the selectivity filter . They also examined C70 fullerenes fully coated in gallic acid to stabilize the fullerenes in solution. These gallic acid coated fullerenes were also shown to make contact with the extracellular or intracellular loops, but not the selectivity filter . Monticelli and co-workers [14, 15] have also shown using molecular dynamics that non-functionalized fullerenes agglomerate within the hydrophobic layer of lipid bilayers.
Although C60 and C70 fullerenes are the most abundantly produced in carbon soot, higher fullerenes such as C76, C78, and C84 have also been isolated [24, 25] and are among the most abundant higher fullerenes . We generate an initial C84 fullerene molecule using the fullerene library available in the Nanotube Modeler 1.7.3 software . The C84 fullerene has six favorable isomers , and of these, the D2 and D2d have the lowest energy . We choose the structure with D2d symmetry (structure number 23 in Nanotube Modeler) as this has also been reported as the most commonly observed in experiments . The C84 fullerene has an approximate diameter of 8 Å.
Ideally, an ion channel blocker design would have flexible side chains which can bind to the channel and block the entrance to the pore. The D2d isomer of C84 has been shown to have the most localized π bonding of the fullerenes that have been isolated and has therefore been suggested as being the most reactive toward addition reactions . Researchers [30–32] have also shown that it is possible to attach various chemical species to the outside of fullerene molecules. For example, phenylalanine and lysine amino acid derivatives have been attached to the C60 fullerene [30, 31]. Therefore, we import the C84 fullerene structure into ArgusLab 4.0.1 and attach six lysine derivatives to its outside surface . A similar water-soluble amino-fullerene derivative with five cysteine moieties attached to the surface of C60 fullerene has previously been synthesized and characterized by Hu et al. . They demonstrated the ability of this fullerene derivative to prevent oxidative-induced cell death without evident toxicity . We choose positively charged residues with the aim of mimicking the function of μ-conotoxin to NavAb. The distance between nitrogen atoms on opposing lysine chains is approximately 21 Å.
The modified fullerene (C84(C4H8NH3+)6 structure is optimized in ArgusLab  and is shown in Figure 1. The geometry optimizations were performed using default parameters, the Broyden-Fletcher-Goldfarb-Shanno algorithm and the universal force field. Restricted Hartree-Fock method was used, where the molecule is a closed shell system with all orbitals doubly occupied. All optimization processes are performed until the Hartree-Fock self-consistent field converged to 10−10 kcal/mol and the gradient converged to 10−1 kcal/mol/Å. Throughout this paper, this modified C84 fullerene is referred to as [Lys]-fullerene.
The coordinates of NavAb are obtained from the protein database [PDB:3RVY] . We obtain a homology model of Kv1.3 using the refined structure of the Kv1.2 channel (PDB:SLUT) as a template . The generation of the homology model for Kv1.3 is described in detail in Chen et al. , where it was shown to reproduce experimental conduction properties and binding affinities to charybdotoxin and ShK. We chose to examine the binding of the [Lys]-fullerene to Kv1.3, giving us the opportunity to directly compare our results with the binding of polypeptide toxins [37, 38].
Molecular dynamics (MD) simulations are used to determine the bound configuration of the [Lys]-fullerene and calculate the potential of mean force (PMF) of the [Lys]-fullerene binding to the channel. All MD simulations are performed using NAMD 2.8 and visualized using VMD 1.9 [39, 40]. Throughout, we use the CHARMM36 force field [41, 42] and TIP3P water, with a time step of 2 fs, at constant pressure (1 atm), and temperature (300 K). The channel and fullerene complex are embedded in a POPC lipid bilayer, solvated in approximately a 100 × 100 × 100 Å3 box of water. Potassium/sodium (for Kv1.3/NavAb) and chloride ions are added to both neutralize the system and simulate a 250-mM ionic concentration. The protein is initially held fixed to allow the water and ions to equilibrate during the simulation period of 0.1 ns, and in subsequent simulations, the protein and lipid bilayer center of mass is held by a harmonic constraint of 0.2 kcal/mol/Å2. A similar methodology has been used to investigate the binding of toxins to ion channels [16, 37, 43].
The [Lys]-fullerene is initially placed near the entrance of the selectivity filter (at z = 22 Å) and the system is allowed to equilibrate for 1 to 3 ns with the fullerene unconstrained. The PMF for the binding of the [Lys]-fullerene to the NavAb and Kv1.3 channels is determined using umbrella sampling with this equilibrated structure. Umbrella sampling windows are generated using steered MD simulations with a force of 30 kcal/mol/Å applied to pull the fullerene out of the binding site. During the steered MD simulations the backbone atoms of the protein are held fixed and the atoms of the fullerene are held by a harmonic constraint of 0.2 kcal/mol/Å2 to maintain the root-mean-square deviation, with reference to a starting configuration below 0.25 Å so that no significant distortion takes place. The channel central axis (z-axis) is used as the reaction coordinate. Pulling generates a continuous number of configurations along the permeation pathway so that umbrella sampling windows can be constructed every 0.5 Å.
During umbrella sampling the center of mass of the backbone atoms of the fullerene is confined to be within a cylinder of 8 and 13 Å centered on the channel axis for Kv1.3 and NavAb, respectively, and beyond this, a harmonic potential of 20 kcal/mol/Å2 is applied. These values are shown to provide adequate sampling. Moreover, a force constant of 30 kcal/mol/Å2 is applied in the z direction to constrain the center of mass of fullerene to the sampling window. The center of mass coordinates of the backbone atoms of the fullerene is saved every 0.5 ps. The PMF is then constructed along the z direction using the weighted histogram analysis method  and the center of mass coordinates. Each sampling window is run for 5 ns. The PMF is shown to converge as the depth changes by less than 0.5 kT when simulations are run for a further 1 ns.
where W(z) is the 1D PMF with the zero point located at the bulk, 1,000 NA is used to convert from cubic meter to liter per mole, kB and T are Boltzmann's constant and temperature, respectively, z1 is in the binding pocket, and z2 is in the bulk . Although Equation 1 was originally derived for the binding of an ion to a channel , it has also been successfully applied to toxin binding [16, 37, 43]. Note that the windows at 38.5 and 42.0 Å for NavAb and Kv1.3, respectively, are assumed to be bulk, and the PMF is therefore set to zero at this z position. The fullerene is docked to NavAb and Kv1.3 at z = 20.5 Å and z = 23.0 Å, respectively, and the center of mass is located at z = 0 Å. A hydrogen bond is assumed to be formed if the donor-acceptor distance is within 3.0 Å and the donor-hydrogen-acceptor angle is ≥150. A salt bridge is formed between the fullerene and ion channel if the distance between any of the nitrogen atoms on the fullerene side chains and the oxygen atoms of an acidic residue on the ion channel is <4 Å.
Results and discussion
Using Equation 1, we obtain dissociation constants, Kd, of 46 nM and 3 mM for the NavAb and Kv1.3 channels, respectively. In comparison, in MD simulations, Chen and Chung  found that μ-conotoxin binds to NavAb with a well depth of approximately −25 kT and a binding affinity of 0.1 nM. French and colleagues  have recently confirmed this result experimentally and obtained a binding affinity of 0.005 nM for the NaChBac, a bacterial channel closely related to NavAb. This [Lys]-fullerene mimic of μ-conotoxin is specific to NavAb over Kv1.3 and presents exciting opportunities for future drug development research.
Sequence alignment between Kv1.3, Na v Ab, and Nav1.8
Bacterial and mammalian channels differ significantly in both sequence and structure. In an attempt to understand how the [Lys]-fullerene might bind to a mammalian Nav channel, we align the sequence of NavAb to Nav1.8. Although μ-conotoxin is sensitive to Nav1 channels, Nav1.8 is both tetrodotoxin and μ-conotoxin insensitive [19, 49]. The Nav1.8 sequence has recently been studied for gain-of-function mutations which have been linked to painful peripheral neuropathy . A few selective blockers of Nav1.8 have been identified, such as A-803467 and μO-conotoxin, and have been shown to suppress chronic pain behavior [19, 20]. Therefore, it is interesting to consider the sensitivity of Nav1.8 to [Lys]-fullerene. Amino acid sequences of the NavAb and Nav1.8 ion channels were obtained from the NCBI protein database (NCBI:3RVY_A, NCBI:NP_006505.2, respectively) [35, 50], and the sequences were aligned using MUSCLE . A comparison of the two sequences, shown in Table 1, demonstrates that Glu177 in NavAb aligns with the Asp-Glu-Lys-Ala (DEKA) residues of the selectivity filter of Nav1.8. As mentioned, the four methionine residues at position 181 form hydrophobic bonds with the fullerene molecule ‘coordinating’ it to the pore of NavAb. In Nav1.8, there are four hydrophobic residues in a similar position to Met181 and in particular Leu-Met-Iso-Leu (LMIL). It may be possible that a similar hydrophobic bond could form between the fullerene and this mammalian Nav channel. However, in Kv1.3, the methionine residue does not contribute to the binding of [Lys]-fullerene and instead stabilizes the channel. A similar mechanism could occur in Nav1.8. Unfortunately, no crystal structure of Nav1.8 or any other mammalian Nav channel is currently available. Therefore, to confirm such a hypothesis requires significant future work such as building a Nav1.8 homology model and conducting molecular dynamics simulations to ascertain the binding affinity of the [Lys]-fullerene.
Carbon-based nanomaterials offer the possibility of some exciting applications, such as drug delivery, selective cell destructive agents, biosensors, and diagnostics. Further development is needed regarding the toxicity of these materials in both biological and environmental environments, in the short and long terms, for these applications to be brought into widespread use. We refer the reader to recent reviews on the use of carbon nanotubes and fullerenes in biology and medicine [5, 6, 51]. Typically, non-functionalized carbon-based nanomaterials are considered to be toxic, but significant work has been done to make these structures soluble and biocompatible. For example, it has been demonstrated that C60 fullerene with five cysteine residues attached to its surface is water soluble and does not cause cellular toxicity .
As with any drug lead, to move from an idea to a marketable drug can take between 10 to 15 years. Therefore, significant research effort is required to develop this theoretical [Lys]-fullerene design into a drug for therapeutic use. Future simulations are required to determine whether these compounds are potent blockers of mammalian Nav channels and if they are specific to a particular channel sub-type. Following this, experiments would need to be performed to confirm theoretical findings and determine toxicity profiles. Polypeptide toxins from venomous animals have evolved over millions of years, aimed at rapidly immobilizing and capturing prey. Since they act on a broad spectrum of ion channel families and are rapidly degraded in vivo, converting these toxins to drugs represents a considerable challenge, and attempts are being made to synthesize smaller and more durable mimetic structures [1–4]. The use of nanomaterials, which replace the rigid backbone of the naturally occurring toxins, may prove to be a fruitful approach for such an endeavor. In the past, fullerenes suffered from high production costs which generated an obstacle to the development of fullerene-based applications, but the cost has rapidly declined .
Voltage-gated sodium channels are present throughout muscle and neuronal cells in mammals. Their dysfunction has long been linked to disorders such as epilepsy and chronic pain. Toxins from venomous species such as cone snails and scorpions have demonstrated activity against sodium channels. One example is the polypeptide toxin μ-conotoxin (PIIIA), extracted from the cone snail, which has been shown to potently block both bacterial and mammalian Nav channels [16, 17, 52]. Unfortunately, converting toxins to drugs represents a considerable challenge [1–4]. We attempt to mimic the structure of μ-conotoxin by (1) replacing its bulky core with a C84 fullerene and (2) chemically attaching positively charged groups to the fullerene surface. Although fullerenes have previously been identified as possible ion channel blockers [10–15], no studies have demonstrated the potential of designing fullerenes through chemical modification to target specific ion channels. We demonstrate that [Lys]-fullerene blocks the ion-conducting pathway of the bacterial voltage-gated sodium channel, NavAb with a strong binding affinity, similar in magnitude to μ-conotoxin , and with selectivity over potassium voltage-gated channels. This fullerene design may serve as a structural template from which a new set of potent compounds can be designed for the treatment of various diseases linked to sodium channel dysfunction.
We thank A. Robinson and R. Chen for their scientific advice. This work was supported by the NCI National Facility at the Australian National University. We gratefully acknowledge the support from the Australian Research Council through a Discovery Early Career Researcher Award and the National Health and Medical Council.
- Norton RS: μ-Conotoxins as leads in the development of new analgesics. Molecules 2010, 15: 2825–2844. 10.3390/molecules15042825View ArticleGoogle Scholar
- Clark RJ, Akcan M, Kaas Q, Daly NL, Craik DJ: Cyclization of conotoxins to improve their biopharmaceutical properties. Toxicon 2012, 59: 446–455. 10.1016/j.toxicon.2010.12.003View ArticleGoogle Scholar
- Dekan Z, Vetter I, Daly NL, Craik DJ, Lewis RJ, Alewood PF: α-Conotoxin Iml incorporating stable cystathionine bridges maintains full potency and identical three-dimensional structure. J Amer Chem Soc 2011, 133: 15866–15869. 10.1021/ja206408qView ArticleGoogle Scholar
- Khoo KK, Wilson MJ, Smith BJ, Zhang MM, Gulyas J, Yoshikami D, Rivier JE, Bulaj G, Norton RS: Lactam-stabilized helical analogues of the analgesic μ-conotoxin KIIIA. J Med Chem 2011, 54: 7558–7566. 10.1021/jm200839aView ArticleGoogle Scholar
- Bakry R, Vallant RM, Najam-ul-Haq M, Rainer M, Szabo Z, Huck CW, Bonn GK: Medicinal applications of fullerenes. Int J Nanomed 2007, 2: 639–649.Google Scholar
- Da Ros T: Twenty years of promises: fullerene in medicinal chemistry. In Medicinal Chemistry and Pharmacological Potential of Fullerenes and Carbon Nanotubes. Edited by: Cataldo F, Da Ros T. Netherlands: Springer Science; 2008:1–21.View ArticleGoogle Scholar
- Friedman SH, DeCamp DL, Sijbesma RP, Srdanov G, Wudl F, Kenyon GL: Inhibition of the HIV-1 protease by fullerene derivatives: model building studies and experimental verification. J Am Chem Soc 1993, 115: 6506–6509. 10.1021/ja00068a005View ArticleGoogle Scholar
- Prinzbach H, Weiler A, Landenberger P, Wahl F, Wörth J, Scott LT, Gelmont M, Olevano D, Issendorff B: Gas-phase production and photoelectron spectroscopy of the smallest fullerene, C20. Nature 2000, 407: 60–63. 10.1038/35024037View ArticleGoogle Scholar
- Shinohara H, Sato H, Saito Y, Takayama M, Izuoka A, Sugawara T: Formation and extraction of very large all-carbon fullerenes. J Phys Chem 1991, 95: 8449–8451.View ArticleGoogle Scholar
- Park KH, Chhowalla M, Iqbal Z, Sesti F: Single-walled carbon nanotubes are a new class of ion channel blockers. J Biol Chem 2003, 278: 50212–50216. 10.1074/jbc.M310216200View ArticleGoogle Scholar
- Chhowalla M, Unalan HE, Wang Y, Iqbal Z, Park K, Sesti F: Irreversible blocking of ion channels using functionalized single-walled carbon nanotubes. Nanotechnology 2005, 16: 2982–2986. 10.1088/0957-4484/16/12/042View ArticleGoogle Scholar
- Xu H, Bai J, Meng J, Hao W, Xu H, Cao JM: Multi-walled carbon nanotubes suppress potassium channel activities in PC12 cells. Nanotechnology 2009, 20: 285102. 10.1088/0957-4484/20/28/285102View ArticleGoogle Scholar
- Kraszewski S, Tarek M, Treptow W, Ramseyer C: Affinity of C60 neat fullerenes with membrane proteins: a computational study on potassium channels. ACS Nano 2010, 4: 4158–4164. 10.1021/nn100723rView ArticleGoogle Scholar
- Monticelli L, Barnoud J, Orlowski A, Vattulainen I: Interaction of C70 with the Kv1.2 potassium channel. Phys Chem Chem Phys 2012, 14: 12526–12533. 10.1039/c2cp41117bView ArticleGoogle Scholar
- Wong-Ekkabut J, Baoukina S, Triampo W, Tang IM, Tieleman DP, Monticelli L: Computer simulation study of fullerene translocation through lipid membranes. Nature Nanotech 2008, 3: 363–368.View ArticleGoogle Scholar
- Chen R, Chung SH: Binding modes of μ-conotoxin to the bacterial sodium channel (NavAb). Biophys J 2012, 102: 483–488. 10.1016/j.bpj.2011.12.041View ArticleGoogle Scholar
- Finol-Urdaneta RK, Glavica R, McArthur JR, French RJ: Polymodal, high affinity actions of μ-conotoxins on a bacterial voltage-gated sodium channel [abstract]. Biophys J 2013, 104: 136a-137a.View ArticleGoogle Scholar
- Stevens M, Peigneur S, Tytgat J: Neurotoxins and their binding areas on voltage-gated sodium channels. Front Pharmacol 2011, 2: 1–13.View ArticleGoogle Scholar
- Eijkelkamp N, Linley JE, Baker MD, Minett MS, Cregg R, Werdehausen R, Rugiero F, Wood JN: Neurological perspectives on voltage-gated sodium channels. Brain 2012, 135: 2585–2612. 10.1093/brain/aws225View ArticleGoogle Scholar
- Ekberg J, Jayamanne A, Vaughan CW, Aslan S, Thomas L, Mould J, Drinkwater R, Baker MD, Abrahamsen B, Wood JN, Adams DJ, Christie MJ, Lewis RJ: μO-Conotoxin MrVIB selectively blocks Nav1.8 sensory neuron specific sodium channels and chronic pain behavior without motor deficits. Proc Natl Acad Sci USA 2006, 103: 17030–17035. 10.1073/pnas.0601819103View ArticleGoogle Scholar
- Koishi R, Xu H, Ren D, Navarro B, Spiller BW, Shi Q, Clapham DE: A superfamily of voltage-gated sodium channels in bacteria. J Biol Chem 2004, 279: 9532–9538. 10.1074/jbc.M313100200View ArticleGoogle Scholar
- Macnab RM: The bacterial flagellum: reversible rotary propeller and type III export apparatus. J Bacteriol 1999, 181: 7149–7153.Google Scholar
- Wadhams GH, Armitage JP: Making sense of it all: bacterial chemotaxis. Nature Rev Mol Cell Biol 2004, 5: 1024–1037. 10.1038/nrm1524View ArticleGoogle Scholar
- Diederich F, Ettl R, Rubin Y, Whetten RL, Beck R, Alvarez M, Anz S, Sensharma D, Wudl F, Khemani KC, Koch A: The higher fullerenes: isolation and characterization of C76, C84, C90, C94, and C70O, an oxide of D5h-C70. Science 1991, 252: 548–551. 10.1126/science.252.5005.548View ArticleGoogle Scholar
- Liu X, Schmalz TG, Klein DJ: Favorable structures for higher fullerenes. Chem Phys Lett 1992, 188: 550–554. 10.1016/0009-2614(92)80864-8View ArticleGoogle Scholar
- Diederich F, Whetten RL: Beyond C60: the higher fullerenes. Acc Chem Res 1992, 25: 119–126. 10.1021/ar00015a004View ArticleGoogle Scholar
- JCrystalSoft: Nanotube modeler. Version 1.7.3. Copyright JCrystalSoft, 2005–2012. [http://www.jcrystal.com] 
- Balch AL, Ginwalla AS, Lee JW, Noll BC, Olmstead MM: Partial separation and structural characterization of C84 isomers by crystallization of (η2-C84)Ir(CO)Cl(P(C6H5)3)2. J Am Chem Soc 1994, 116: 2227–2228. 10.1021/ja00084a106View ArticleGoogle Scholar
- Bakowies D, Kolb M, Thiel W, Richard S, Ahlrichs R, Kappes MM: Quantum-chemistry study of C84 fullerene isomers. Chem Phys Lett 1992, 200: 411–417. 10.1016/0009-2614(92)87013-FView ArticleGoogle Scholar
- Strom TA, Barron AR: A simple quick route to fullerene amino acid derivatives. Chem Commun 2010, 46: 4764–4766. 10.1039/c003019hView ArticleGoogle Scholar
- Yang J, Alemany LB, Driver J, Hartgerink JD, Barron AR: Fullerene-derivatized amino acids: synthesis, characterization, antioxidant properties, and solid-phase peptide synthesis. Chem Eur J 2007, 13: 2530–2545. 10.1002/chem.200601186View ArticleGoogle Scholar
- Taylor R: Lecture Notes on Fullerene Chemistry. London: Imperial College Press; 1999.View ArticleGoogle Scholar
- ArgusLab: Mark Thompson and Planaria Software LLC. Version 4.0.1. Copyright 1997–2004. [http://www.arguslab.com] 
- Hu Z, Guan W, Wang W, Huang L, Xing H, Zhu Z: Protective effect of a novel cystine C60 derivative on hydrogen peroxide-induced apoptosis in rat pheochromocytoma PC12 cells. Chem Biol Interact 2007, 167: 135–144. 10.1016/j.cbi.2007.02.009View ArticleGoogle Scholar
- Payandeh J, Scheuer T, Zheng N, Catterall WA: The crystal structure of a voltage-gated sodium channel. Nature 2011, 475: 353–359. 10.1038/nature10238View ArticleGoogle Scholar
- Chen X, Wang Q, Ni F, Ma J: Structure of the full-length Shaker potassium channel Kv1.2 by normal-mode-based X-ray crystallographic refinement. Proc Natl Acad Sci USA 2010, 107: 11352–11357. 10.1073/pnas.1000142107View ArticleGoogle Scholar
- Chen R, Robinson A, Gordon D, Chung SH: Modeling the binding of three toxins to the voltage-gated potassium channel (Kv1.3). Biophys J 2011, 101: 2652–2660. 10.1016/j.bpj.2011.10.029View ArticleGoogle Scholar
- Chen R, Chung SH: Engineering a potent and specific blocker of voltage-gated potassium channel Kv1.3, a target for autoimmune diseases. Biochem 2012, 51: 1976–1982. 10.1021/bi201811jView ArticleGoogle Scholar
- Phillips JC, Braun R, Wang W, Gumbart J, Tajkhorshid E, Villa E, Chipot C, Skeel RD, Kale L, Schulten K: Scalable molecular dynamics with NAMD. J Comput Chem 2005, 26: 1781–1802. 10.1002/jcc.20289View ArticleGoogle Scholar
- Humphrey W, Dalke A, Schulten K: VMD: visual molecular dynamics. J Mol Graphics 1996, 14: 33–38. 10.1016/0263-7855(96)00018-5View ArticleGoogle Scholar
- MacKerell AD Jr, Feig M, Brooks CL III: Extending the treatment of backbone energetics in protein force fields: limitations of gas-phase quantum mechanics in reproducing protein conformational distributions in molecular dynamics simulations. J Comput Chem 2004, 25: 1400–1415. 10.1002/jcc.20065View ArticleGoogle Scholar
- MacKerell AD Jr, Bashford D, Bellott M, Dunbrack RL Jr, Evanseck JD, Field MJ, Fischer S, Gao J, Guo H, Ha S, Joseph-McCarthy D, Kuchnir L, Kuczera K, Lau FTK, Mattos C, Michnick S, Ngo T, Nguyen DT, Prodhom B, Reiher WE III, Roux B, Schlenkrich M, Smith JC, Stote R, Straub J, Watanabe M, Wiórkiewicz-Kuczera J, Yin D, Karplus M: All-atom empirical potential for molecular modeling and dynamic studies of proteins. J Phys Chem B 1998, 102: 3586–3616. 10.1021/jp973084fView ArticleGoogle Scholar
- Hilder TA, Chung SH: Conduction and block of inward rectifier K+ channels: predicted structure of a potent blocker of Kir2.1. Biochem 2013, 52: 967–974. 10.1021/bi301498xView ArticleGoogle Scholar
- Kumar S, Rosenberg JM, Bouzida D, Swendsen RH, Kollman PA: Multidimensional free-energy calculations using the weighted histogram analysis method. J Comput Chem 1995, 16: 1339–1350. 10.1002/jcc.540161104View ArticleGoogle Scholar
- Allen TW, Andersen OS, Roux B: Energetics of ion conduction through the gramicidin channel. Proc Natl Acad Sci USA 2004, 101: 117–122. 10.1073/pnas.2635314100View ArticleGoogle Scholar
- Gordon D, Chen R, Chung SH: Computational methods of studying the binding of toxins from venomous animals to biological ion channels: theory and application. Physiol Rev 2013, 93: 767–802. 10.1152/physrev.00035.2012View ArticleGoogle Scholar
- De Leon A, Jalbout AF, Basiuk VA: Fullerene-amino acid interactions. A theoretical study. Chem Phys Lett 2008, 452: 306–314. 10.1016/j.cplett.2007.12.065View ArticleGoogle Scholar
- EMBL-EBI: MUSCLE—multiple sequence comparison by log-expectation. Copyright © EMBL-EBI 2013. [http://www.ebi.ac.uk/Tools/msa/muscle/] 
- Zhang MM, Wilson MJ, Gajewiak J, Rivier JE, Bulaj G, Olivera BM, Yoshikami D: Pharmacological fractionation of tetrodotoxin-sensitive sodium currents in rat dorsal root ganglion neurons by μ-conotoxins. British J Pharmacology 2013, 169: 102–114. 10.1111/bph.12119View ArticleGoogle Scholar
- Faber CG, Lauria G, Merkies ISJ, Cheng X, Han C, Ahn HS, Persson AK, Hoeijmakers JGJ, Gerrits MM, Pierro T, Lombardi R, Kapetis D, Dib-Hajj SD Waxman SG: Gain-of-function Nav1.8 mutations in painful neuropathy. Proc Natl Acad Sci USA 2012, 109: 19444–19449. 10.1073/pnas.1216080109View ArticleGoogle Scholar
- Heister E, Brunner EW, Dieckmann GR, Jurewicz I, Dalton AB: Are carbon nanotubes a natural solution? Applications in biology and medicine. ACS Appl Mater Interfaces 2013, 5: 1870–1891. 10.1021/am302902dView ArticleGoogle Scholar
- Safo P, Rosenbaum T, Shcherbatko A, Choi DY, Han E, Toledo-Aral JJ, Olivera BM, Brehm P, Mendel G: Distinction among neuronal subtypes of voltage-activated sodium channels by μ-conotoxin PIIIA. J Neurosci 2000, 20: 76–80.Google Scholar