Letter pubs.acs.org/JPCL
Next-Generation Epigenetic Detection Technique: Identifying Methylated Cytosine Using Graphene Nanopore Towfiq Ahmed,*,† Jason T. Haraldsen,‡ Jian-Xin Zhu,†,§ and Alexander V. Balatsky*,¶,∥ †
Theoretical Division, §Center for Integrated Nanotechnologies, and ∥Institute for Materials Science, Los Alamos National Laboratory, Los Alamos, New Mexico 87545, United States ‡ Department of Physics and Astronomy, James Madison University, Harrisonburg, Virginia 22807, United States ¶ Nordic Institute for Theoretical Physics, KTH Royal Institute of Technology and Stockholm University, Roslagstullsbacken 23, 106 91 Stockholm, Sweden S Supporting Information *
ABSTRACT: DNA methylation plays a pivotal role in the genetic evolution of both embryonic and adult cells. For adult somatic cells, the location and dynamics of methylation have been very precisely pinned down with the 5-cytosine markers on cytosine-phosphate-guanine (CpG) units. Unusual methylation on CpG islands is identified as one of the prime causes for silencing the tumor suppressant genes. Early detection of methylation changes can diagnose the potentially harmful oncogenic evolution of cells and provide promising guideline for cancer prevention. With this motivation, we propose a cytosine methylation detection technique. Our hypothesis is that electronic signatures of DNA acquired as a molecule translocates through a nanopore would be significantly different for methylated and nonmethylated bases. This difference in electronic fingerprints would allow for reliable real-time differentiation of methylated DNA. We calculate transport currents through a punctured graphene membrane while the cytosine and methylated cytosine translocate through the nanopore. We also calculate the transport properties for uracil and cyanocytosine for comparison. Our calculations of transmission, current, and tunneling conductance show distinct signatures in their spectrum for each molecular type. Thus, in this work, we provide a theoretical analysis that points to a viability of our hypothesis. SECTION: Biophysical Chemistry and Biomolecules
W
methylation between normal and cancerous DNA strands. Although methylation was recently found to be on other DNA bases such as adenine,7,8 it is predominantly on cytosine bases in mammalian cells. Therefore, in this Letter, we mostly focus on methylation on cytosine. Fast and high-fidelity identification of methylated regions is an important step for the efficient diagnosis of cancer. With the development of next-generation sequencing using solid-state techniques, one can expect commensurate progress in identification of methylated regions using next-generation sequencing. One question that arises in this context is whether new generation sequencing approaches like solid-state electronics-based sequencing provide a significantly better methylation detection. While the ultimate answer to this question will come from experiments, in this Letter, we model the cytosine methylation detection scheme based on the graphene nanopore.
ith the recent progress on the Human Epigenome Project (HEP), the need for a better understanding of DNA methylation patterns and its consequences is felt more than ever.1,2 There is a significant scientific interest in developing techniques for accurate and fast detection of methylated DNA as it is recognized to be pivotal for epigenetic therapy and DNA-based drug discovery.3 In a world with an increasing rate of malignant diseases, like cancers, these new techniques will allow for better detection, prevention, and ultimately treatments of the malignancies. Since the early triumph of Rollin Hotchkiss in 1948 for discovering the presence of a methyl group in fifth atom of the six-atom ring of cytosine,4 methylation was found to be a genewide phenomenon and is often necessary for important genetic transcription and an evolution mechanism. The attachment of a methyl group at the five-position of cytosine followed by a guanine (CpG) has a special role in reducing gene expression. Such a methylation pattern can have serious consequences in developing a large number of human malignancies by suppressing the oncogene, which primarily regulates cell division.5,6 In Figure 1a and b, we schematically show the molecular form of CpG islands and the difference in © 2014 American Chemical Society
Received: May 29, 2014 Accepted: July 11, 2014 Published: July 11, 2014 2601
dx.doi.org/10.1021/jz501085e | J. Phys. Chem. Lett. 2014, 5, 2601−2607
The Journal of Physical Chemistry Letters
Letter
particularly for identifying single nucleotides using transverse conductance.19−21 It has been established that experimental methods are capable of achieving single-base resolution, which has prompted investigations into the local electrical properties of single DNA bases.11,22 Concurrently, the theoretical underpinnings of this approach are being continuously developed.9,10,17,19,20 The single-molecule sensitivity of nanopore sequencing has been recently demonstrated by Tsutsui et al.23 and Chang et al.24 The sequence of DNA/RNA oligomers and microRNA by tunneling has also been demonstrated.25 Recently, DNA detection using transverse current through a graphene nanopore was experimentally realized by Traversi et al.26 High-throughput sequencing methods for methylcytosine detection on plant DNA has recently been demonstrated by Huang et al.27 In the search for a suitable substrate, the experimental realization of a single-layer graphene-based nanopore device is made possible by combining several stateof-the-art techniques, for example, mechanical exfoliation from graphite on a SiO2 substrate. Ionic current measurements, as the double-stranded (ds)DNA translocates through a monolayer graphene nanopore, were previously reported by Schneider et al.28 AFM studies22 and theoretical simulations of scanning tunneling spectroscopy (STS)17 support the identification of electronic features with varying spatial extent and intensity near the HOMO−LUMO band. In an effort to detect methylated cytosine, we propose the measurement of the transverse current in a graphene nanopore. Our first-principles calculations based on the nonequilibrium Green’s function (NEGF-DFT) formalism29 show distinctive features in the current spectrum of methylated and nonmethylated cytosine. We also present our calculations for cyanocytosine and uracil, which can be obtained by
Figure 1. (a) Schematic representation of the cytosine-phosphateguanine (CpG) segment of DNA. CpGs are represented with square symbols, while the methyl group is shown as circles. (b) Schematic presentation of DNA methylation. Excessive methylation in the promoter gene is responsible for gene silencing of the antioncogenic part, which is often associated with different types of cancer. (c) The experimental setup is shown schematically using a graphene nanopore. The transport currents can be measured for the translocating nucleotides with or without methylation.
Previous theoretical9,10 and experimental11−14 studies of the interactions between DNA bases and nanopore devices have revealed the local electronic structure of single bases. Nanopore-based biomolecular fingerprinting15,16 and serial methods9,17 provide promising alternatives to the wellestablished Sanger method18 or other conventional methods,
Figure 2. (a) Molecular representation of cytosine, methylcytosine, uracil, and cyanocytosine. White circles stand for hydrogen, gray ones for carbon, red for oxygen, and blue for nitrogen. (b) Schematic drawing of contact and electrodes regions in the proposed experimental setup. The bias voltage of left and right electrodes are −0.35 and +0.35 V, respectively. (c) Five different orientation angles of cytosine while translocating through the pore in the presence of a ±0.35 V bias. This calculation is performed for these angles for all other molecules as shown in (a). 2602
dx.doi.org/10.1021/jz501085e | J. Phys. Chem. Lett. 2014, 5, 2601−2607
The Journal of Physical Chemistry Letters
Letter
Figure 3. (a) Transmission in the presence of a ±0.35 V voltage bias for cytosine (red), methylcytosine (blue), uracil (green), and cyanocytosine (gray) for a 0, 30, 45, 60, and 90° angular orientation with the nanopore plane. Shaded regions show the prominent peaks within the chemical potentials of left and right electrodes (vertical black dashed lines at ±0.35 eV). Vertical dashed lines at ±0.35 eV are at the chemical potential of the left/right electrodes, and the middle dashed line at 0 eV is at the Fermi energy EF. (b) The dominant peak positions and intensities near EF are shown for each molecular type and their angular orientation. Peak intensities are qualitatively proportional to the size of the marker.
deamination of cytosine.30 Such functional transformations of cytosine into uracil30 or modification by additional groups (e.g., CN−, NH2+) using chemical substitution are proposed with the motivation of achieving enhanced and characteristically distinct signals in nanopore experiments. The proposed experimental set up for a typical nanopore arrangement is schematically shown in Figures 1c and 2b. To be specific, we focus on graphene as the porous material because it is atomically thick and exhibits extraordinary thermal and electronic properties. Besides these geometric advantages and good conductivity, graphene also possesses high tensile strength and can endure a high transmembrane pressure environment.31 Consequently, graphene has been proposed as an effective substrate and conducting medium for nanopore sequencing by numerous groups.11,28,32−35 To take advantage of the metallic behavior of graphene nanoribbons (GNRs), we have considered the zigzag configuration. Previously, the preferable zigzag configuration of GNRs for DNA detection was also demonstrated by other authors using NEGF-DFT36 and molecular dynamics (MD) simulation.37 We first discuss our first-principles calculations of transmittance for individual methylcytosine, cytosine, and its chemical variations inside of the graphene nanopore for different incident angles, as presented in Figures 2c and 3. Then in Figure 4, we show the translational averaged transmission signals for each type of molecule for a fixed voltage bias. Finally, in Figure 5, we present the voltagedependent current and conductance spectrum, which distinguishes methylcytosine from the rest of the other types of molecules. In our first-principles approach, for each molecular type, we have taken five random orientation angles with the graphene membrane, while calculating the transmittance between the two
electrodes with 0.7 V voltage bias. For each angular orientation, the dominant peaks near the Fermi energy, within the chemical potential between left and right electrodes, are shown with the shaded region for each molecular type in Figure 3a. The blueshaded regions stand for methylcytosine, which is distinct from the rest of the chemical species for any given angular configuration. The transmission spectra, as shown in Figure 3a, are correlated with the local density of states (LDOS) of the system. To be precise, the LDOS can be written as a sum of contributions from the individual eigen channels of the transmission matrix.38 While the HOMO and LUMO states in molecular obitals of the independent bases are not affected by angular orientation, the strong chemical interaction and hybridization between the orbitals of the graphene substrate and its nearest atoms in the DNA base can contribute to the transient LDOS and, therefore, the peaks in the transmission spectra as the bases translocate through the pore. Thus, the transient HOMO−LUMO states of the system, which are proportional to the dominant (shaded) transmission peaks in Figure 3a, strongly depend on the angular orientation of the DNA bases with respect to the graphene sheet. In Figure 3, we do see a dependence of the transmission peak intensity on the molecular orientation due to the complex and nontrivial interactions of the local base orbitals and the graphene as it passes through the nanopore. Figure 3b clearly shows the chemical distinction between them. Comparing with cytosine, uracil, and cyanocytosine, methylcytosine has distinct signatures in both the positive and negative voltage bias regions. The size of the markers in Figure 3b are qualitatively proportional to the peak intensity. The vertical dashed lines are at −0.35 and +0.35 eV, which are the chemical potentials of the left and right electrodes, respectively. 2603
dx.doi.org/10.1021/jz501085e | J. Phys. Chem. Lett. 2014, 5, 2601−2607
The Journal of Physical Chemistry Letters
Letter
Figure 4. (a) Single-molecule translocation through the nanopore with time. Five snapshots with a space separation of Δz are shown for a single molecule (here, cytosine) as it passes through the graphene nanopore. Only the 60° orientation is considered, where the angle between the cytosine ring plane and graphene plane is 60° at various displacements. (b) The translational averaged transmission (T(E)) for five different positions for cytosine (red), methylcytosine (blue), cyanocytosine (black), and uracil (green). Vertical dashed lines are the chemical potentials (±0.35 eV) of the left and right electrodes. Only the 60° configuration is considered for each molecular type.
For a systematic study of the difference between the transmittance among the four molecular types, we also plotted the translational averaged transmittance in Figure 4. The angular conformation of single-stranded DNA (sDNA) as it translocates through the graphene nanopore was demonstrated previosuly39 using MD simulation. Suggested by this earlier work, we have considered the 60° angle as one of the most probable angular orientations of the cytosine base with respect to the graphene plane. As shown in Figure 4a, five different locations at a displacement of Δz are considered as the molecule approaches and passes the nanopore. The averaged transmissions are shown in Figure 4b. The left and right electrode chemical potentials are displayed with vertical dashed lines. To study the effect of distance on the characteristic transmission signals for different chemical variations of cytosine in the vicinity of the nanopore, our calculations are done using only the 60° molecular orientation, as shown in Figure 4a. The voltage bias is fixed at 0.7 V. The sharp peak near the Fermi energy for methylcytosine is distinct from other chemical variations (solid blue curve in Figure 4b). This indicates that the experimental setup with single-base resolution ability using a graphene nanopore can, in principle, identify transmission signals coming from methylated cytosine. Figures 3 and 4 shows distinct signatures for methylated cytosine for all angular and spatial cases, comparing not only the peak positions but also the peak intensities in the positive and(or) negative energy regions. It is also important to note that transmission spectra in general are a theoretical quantity and are often not directly measurable in experiments. However, in today’s data-centric and modeling-guided experimental research, our proof of principle calculations show that an accurate and state-of-the-art NEGF-DFT technique can predict transmission spectra with
distinct signatures of methylated cytosine for many angular orientations, and at the same time, such data can help guide the experiments to detect DNA methylation from the current and conductance measurements. We emphasize that in these NEGF-DFT calculations, although our device system (graphene + nanopore) is much smaller and simpler compared to real experimental situations, the results indicate the potential significance of computational data-guided experimental effort, which may open up possibilities for the next-generation DNA methylation detection techniques. The transmission curves are qualitatively proportional to the density of states for small voltage bias. For a given potential difference between the electrodes, the total current can be measured and calculated. In our proposed experimental setup, one can vary the voltage bias and obtain substantially more information including the voltage-dependent current and conductance. In Figure 5, our first-principles calculations of current I(V) and tunneling conductance dI(V)/dV are shown for cytosine, methylcytosine, cyanocytosine, and uracil. Here, we again considered the 60° angular orientation of the DNA base plane with the graphene plane. Taking advantage of the fact that this is one of the most probable angular orientations,39 our NEGF-DFT-based proof of principle calculation shows potentially identifiable signatures coming from methylated cytosine compared with the nonmethylated one. Both methylated- and cyanocytosine have enhanced peaks in the current (Figure 5a) and conductance (Figure 5b) spectrum compared to cytosine and uracil. Addition of functional groups (CH3 and CN) to cytosine base is responsible for this characteristic difference and thus provides a potential and promising technique for methylation detection. The steep rise in the solid blue curve in Figure 5 indicates enhanced electron transport in the presence of the 2604
dx.doi.org/10.1021/jz501085e | J. Phys. Chem. Lett. 2014, 5, 2601−2607
The Journal of Physical Chemistry Letters
Letter
Figure 5. (a) Bias voltage dependence of the current I(V) for cytosine (red), methylcytosine (blue), cyanocytosine (black), and uracil (green) as the molecules translocate through the graphene nanopore. Only the 60° angular orientation between the biomolecular ring plane and graphene plane is considered. (b) Bias voltage dependence of the tunneling conductance dI/dV is shown for each molecular type.
(2) Second, we varied the voltage between the two electrodes and subsequently calculated the voltage-dependent current and conductance using calculated T(E). Calculations of transmission were performed by taking each type of molecule inside of the nanopore with five different orientation angles (Figure 3a) and five different positions (Figure 4a) and using the Landauer−Buttikker40 formalism implemented in the ab initio software ATK.29,41 We emphasize that our approach does not rely or requires a geometry optimization of molecules in the pores. The translocation is a dynamical process with significant variations of configuration found for molecules inside of a pore. Thus, the same molecule can arrive in different orientation angles at each pore, a process contributing to the configuration noise sources that we addressed in an early paper.21 Therefore, we did not optimize the configurations and instead used the set of various configurations as the set. In these calculations, we have taken a GNR with 208 carbon atoms in the conduction region, where the nanopore is constructed by removing central carbon atoms and capping the inner wall with hydrogen atoms because hydrogenated edges were found35 to enhance the average experimental conductivity. The voltage bias between the left and right electrodes was fixed at +0.35 and −0.35 V as in Figures 3 and 4 and was varied to obtain tunneling conductance, as shown in Figure 5. In this work, we assume the nanopore diameter at about 5 Å, which is smaller than that modeled by other groups.34,39,42 This nanopore size is comparable with the single cytosine base. For our proof of principles calculations, we primarily focused on the characteristic differences in electronic transport between methylated and nonmethylated cytosine and thus restricted ourselves with a smaller nanopore dimension. In our calculation, we have taken a punctured GNR suspended between two electrodes. The inner wall of the nanopore was hydrogen-passivated in order to avoid any
methyl group due the large number of molecular states near the Fermi energy of the graphene−bimolecular hybrid. This can also be seen in our calculated transmission spectra in Figure 4b in the presence of a 0.7 V bias voltage. In this Letter, we have considered a nanopore size comparable to the size of DNA bases. It was assumed that the nanopore size can significantly change the peak intensity and some of the peak locations without altering the key conclusions of this work. This is evident from our calculated transmission peaks for various chemical variants of cytosine, which indicates that the added groups (e.g., −CH3, −CN) play key roles in the local interaction with the graphene and therefore significantly change the dominant peaks in the transmission spectra. Nanopore size effects and other relevant consideration (e.g., nanopore fictionalization) will be addressed in a future publication. In calculating transmission coefficients, we ignore the background contribution from the large phosphate backbone, which is typically present in DNA or in CpG islands. This simplification is based on the assumption that by identifying and subtracting the background noise coming from the heavy and rigid backbone structure, one can isolate the relevant signal from the individual molecules. Also, addition of the backbone structure will require a larger nanopore and therefore a significantly large GNR. However, we focus on our proof of principle calculations for next-generation methylation detection with single-base resolution. A more extensive and realistic simulation of the backbone effect will be considered in the future. Here, we have built on earlier work10,17,19,35 to model the transverse conductance containing a molecule in two steps: (1) First, we carried out ab initio calculations of the transmission (T(E)) and current (I) as a single DNA base translocates through the nanopore of a graphene monolayer. 2605
dx.doi.org/10.1021/jz501085e | J. Phys. Chem. Lett. 2014, 5, 2601−2607
The Journal of Physical Chemistry Letters
Letter
(6) Momparler, R. L.; Bovenzi, V. J. Cell Physiol 2000, 183, 145−154. (7) Karrer, K. M.; VanNuland, T. A. Nucleic Acids Res. 2002, 30, 1364−1470. (8) Ratel, D.; Ravanat, J.-L.; Berger, F.; Wion, D. BioEssays 2006, 28, 309−315. (9) Kilina, S.; Tretiak, S.; Yarotski, D. A.; Zhu, J.-X.; Modine, N.; Taylor, A.; Balatsky, A. V. J. Phys. Chem. C 2007, 111, 14541−14551. (10) Kilina, S.; Yarotski, D. A.; Talin, A. A.; Tretiak, S.; Taylor, A. J.; Balatsky, A. V. J. Drug Delivery 2011, 2011, 415621. (11) Tanaka, H.; Kawai, T. Nat. Nanotechnol. 2009, 4, 518−522. (12) Wanunu, M.; Cohen-Karni, D.; Johnson, R. R.; Fields, L.; Benner, J.; Peterman, N.; Zheng, Y.; Klein, M. L.; Drndic, M. J. Am. Chem. Soc. 2011, 133, 486−492. (13) Garaj, S.; Hubbard, W.; Reina, A.; Kong, J.; Branton, D.; Golovchenko, J. A. Nature 2010, 467, 190−193. (14) Shim, J.; Humphreys, G. I.; Venkatesan, B. M.; Munz, J. M.; Zou, X.; Sathe, C.; Schulten, K.; Kosari, F.; Nardulli, A. M.; Vasmatzis, G.; Bashir, R. Sci. Rep. 2013, 3, 1389. (15) Zwolak, M.; Di Ventra, M. Rev. Mod. Phys. 2008, 80, 141−165. (16) Branton, D.; et al. Nat. Biotechnol. 2008, 26, 1146−1153. (17) Ahmed, T.; Kilina, S.; Das, T.; Haraldsen, J. T.; Rehr, J. J.; Balatsky, A. V. Nano Lett. 2012, 12, 927−931. (18) Sanger, F.; Nicklen, S.; Coulson, A. R. Proc. Natl. Acad. Sci. U.S.A. 1977, 74, 5463−5467. (19) Zwolak, M.; Di Ventra, M. Nano Lett. 2005, 5, 421−424. (20) Lagerqvist, J.; Zwolak, M.; Di Ventra, M. Nano Lett. 2006, 6, 779−782. (21) Ahmed, T.; Haraldsen, J. T.; Rehr, J. J.; Ventra, M. D.; Schuller, I.; Balatsky, A. V. Nanotechnology 2014, 25, 125705. (22) Yarotski, D. A.; Kilina, S. V.; Talin, A. A.; Tretiak, S.; Prezhdo, O. V.; Balatsky, A. V.; Taylor, A. J. Nano Lett. 2009, 9, 12−17. (23) Tsutsui, M.; Taniguchi, M.; Yokota, K.; Kawai, T. Nat. Nanotechnol. 2010, 5, 286−290. (24) Chang, S.; Huang, S.; He, J.; Liang, F.; Zhang, P.; Li, S.; Chen, X.; Sankey, O.; Lindsay, S. Nano Lett. 2010, 10, 1070−1075. (25) Ohshiro, T.; Matsubara, K.; Tsutsui, M.; Furuhashi, M.; Taniguchi, M.; Kawai, T. Sci. Rep. 2012, 2, 1070−1075. (26) Traversi, F.; Raillon, C.; Benameur, S. M.; Liu, K.; Khlybov, S.; Tosun, M.; Krasnozhon, D.; Kis, A.; Radenovic, A. Nat. Nanotechnol. 2013, 8, 939−945. (27) Huang, X.; Lu, H.; Wang, J.-W.; Xu, L.; Liu, S.; Sun, J.; Gao, F. BMC Genet. 2013, 14, 1−9. (28) Schneider, G. F.; Kowalczyk, S. W.; Calado, V. E.; Pandraud, G.; Zandbergen, H. W.; Vandersypen, L. M. K.; Dekker, C. Nano Lett. 2010, 10, 3163−3167. (29) Brandbyge, M.; Mozos, J.-L.; Ordejón, P.; Taylor, J.; Stokbro, K. Phys. Rev. B 2002, 65, 165401. (30) Chahwan, R.; Wontakal, S. N.; Roa, S. Trends Genet. 2010, 26, 443−448. (31) Lee, C.; Wei, X.; Kysar, J. W.; Hone, J. Science 2008, 321, 385− 388. (32) Branton, D.; et al. Nat. Biotechnol. 2008, 26, 1146−1153. (33) Merchant, C. A.; Healy, K.; Wanunu, M.; Ray, V.; Peterman, N.; Bartel, J.; Fischbein, M. D.; Venta, K.; Luo, Z.; Johnson, A. T. C.; Drndic, M. Nano Lett. 2010, 10, 2915−2921. (34) Nelson, T.; Zhang, B.; Prezhdo, O. V. Nano Lett. 2010, 10, 3237−3242. (35) He, Y.; Scheicher, R. H.; Grigoriev, A.; Ahuja, R.; Long, S.; Huo, Z.; Liu, M. Adv. Funct. Mater. 2011, 21, 2602−2602. (36) Saha, K. K.; Drndić, M.; Nikolić, B. K. Nano Lett. 2012, 12, 50− 55. (37) Avdoshenko, S. M.; Nozaki, D.; Gomes de Rocha, C.; González, J. W.; Lee, M. H.; Gutierrez, R.; Cuniberti, G. Nano Lett. 2013, 13, 1969−1976. (38) Brandbyge, M.; Tsukada, M. Phys. Rev. B 1998, 57, R15088− R15091. (39) Wells, D. B.; Belkin, M.; Comer, J.; Aksimentiev, A. Nano Lett. 2012, 12, 4117−4123.
transient bond formation between the translocating molecules and carbon atoms. For the device geometry, we treated a twoprobe system with the electrodes and the contact region. For the results shown in Figure 5, the left electrode voltage was kept fixed, and right electrode voltage was varied with a step size of 0.1 V to obtain the voltage-dependent transmission T(E), current I(V), and conductance dI(V)/dV. The details of our first-principles NEGF method can be found in earlier publications.21,29,43,44 A brief description is also given in our Supporting Information. In conclusion, we propose graphene-nanopore-based transverse current measurements to identify methylated DNA components. We have performed first-principles calculation of transmittance, currents, and tunneling conductance for a systematic study of the identification of methylated cytosine from other biomolecules translocating through graphene nanopores. We find that methylated cytosine possesses unique fingerprints that potentially could enable methylation detection based on only electronic conductance. To illustrate this approach, we have adopted a first-principles nonequilibrium Green’s function approach for simulating the transport properties under various geometric configurations of biomolecules. Our calculations have shown that different functional variations of cytosine have significant characteristic differences in various transport spectra. Both the electronic structure in graphene and its robust mechanical properties serve to make it an excellent material of choice for designing a potentially superior methylation detection device. Thus, our ab initio calculations provide guidelines for the development of next-generation epigenetic detection of DNA methylation.
■
ASSOCIATED CONTENT
S Supporting Information *
Single-particle scattering theory and calculation details. This material is available free of charge via the Internet at http:// pubs.acs.org.
■
AUTHOR INFORMATION
Corresponding Authors
*E-mail: atowfi
[email protected] (T.A.). *E-mail:
[email protected] (A.V.B.). Notes
The authors declare no competing financial interest.
■
ACKNOWLEDGMENTS We are grateful to Los Alamos National Laboratory, an affirmative action equal opportunity employer, operated by Los Alamos National Security, LLC, for the National Nuclear Security Administration of the U.S. Department of Energy under Contract DE-AC52-06NA25396. This work was supported by the U.S. DOE Office of Basic Energy Sciences, by VR 621-2012-2983 and ERC 321031-DM, and, in part, by the Center for Integrated Nanotechnologies, a U.S. DOE BES user facility.
■
REFERENCES
(1) Jones, P. A.; Baylin, S. B. Nat. Rev. Genet 2002, 3, 415−428. (2) Ehrlich, M. Oncogene 2002, 21, 5400−5413. (3) Schulz, W.; Seifert, H.-H. DNA Methylation and Cancer Therapy; Medical Intelligence Unit; Springer: New York, 2005; pp 42−58. (4) Hotchkiss, R. D. J. Biol. Chem. 1948, 175, 315−332. (5) Villa, R.; Santis, F. D.; Gutierrez, A.; Minucci, S.; Pelicci, P.; Croce, L. D. Biochem. Pharmacol. 2004, 68, 1247−1254. 2606
dx.doi.org/10.1021/jz501085e | J. Phys. Chem. Lett. 2014, 5, 2601−2607
The Journal of Physical Chemistry Letters
Letter
(40) Du, X.; Skachko, I.; Barker, A.; Andrei, E. Y. Nat. Nanotechnol. 2008, 3, 491−495. (41) Atomistic Toolkit, version 11.8. http://www.quantumwise.com (2012). (42) Postma, H. W. C. Nano Lett. 2010, 10, 420−425. (43) Di Ventra, M. Electrical Transport in Nanoscale Systems; Cambridge University Press: New York, 2008. (44) Brandbyge, M.; Sørensen, M. R.; Jacobsen, K. W. Phys. Rev. B 1997, 56, 14956−14959.
2607
dx.doi.org/10.1021/jz501085e | J. Phys. Chem. Lett. 2014, 5, 2601−2607
