NANO LETTERS
Sequence-Independent Helical Wrapping of Single-Walled Carbon Nanotubes by Long Genomic DNA
2006 Vol. 6, No. 2 159-164
Brittany Gigliotti, Brenda Sakizzie, Donald S. Bethune, Robert M. Shelby, and Jennifer N. Cha* IBM Almaden Research Center, 650 Harry Road, San Jose, California 95120 Received September 20, 2005; Revised Manuscript Received November 9, 2005
ABSTRACT Because of their nanometer sizes and molecular recognition capabilities, biological systems have garnered much attention as vehicles for the directed assembly of nanoscale materials.1-6 One of the greatest challenges of this research has been to successfully interface biological systems with electronic materials, such as semiconductors and metals. As a means to address some of these issues, Sarikaya, Belcher, and others have used a combinatorial technique called phage display7-9 to discover new families of peptides that showed binding affinities to various substrates. More recently, Zheng and co-workers used combinatorial DNA libraries to isolate short DNA oligomers (30−90 bases) that could disperse single-walled carbon nanotubes (SWCNT) in water.10 Through a systematic analysis, they found that short oligonucleotides having repeating sequences of gunanines and thymines (dGdT)n could wrap in a helical manner around a CNT with periodic pitch.11 Although helix formation around SWCNTs having regular pitches is an effective method for dispersing and separating CNTs, the need for specific repeating sequences limits use to nonnatural DNA that must be synthesized with optimal lengths of less than 150 bases. In contrast, we demonstrate here that long genomic single-stranded DNA (.100 bases) of a completely random sequence of bases can be used to disperse CNTs efficiently through the single-stranded DNA’s (ssDNA) ability to form tight helices around the CNTs with distinct periodic pitches. Although this process occurs irrespective of the DNA sequence, we show that this process is highly dependent on the removal of complementary strands. We also demonstrate that although the helix pitch-to-pitch distances remain constant down the length of a single CNT, the distances are variable from one DNA-CNT to another. Finally, we report initial work that shows that methods developed to align long dsDNA can be applied in a similar fashion to produce highly dense arrays of aligned ssDNA-CNT hybrids.
In the interest of using DNA as a vehicle to disperse and pattern CNTs for microelectronics, ideal DNA lengths should be on the order of micrometers, comparable to the CNT dimensions being used currently for fabricating field effect transistors.12 Although the final channel lengths are less than a few hundred nanometers, the patterning of the source drain electrodes are done on much longer CNTs. Because DNA can only be synthesized to approximately 150 nucleotides, corresponding to a double stranded DNA length of ∼50 nm, genomic DNA was targeted because of its diverse pool of lengths. Lambda DNA was chosen because its entire 48 502 base pair sequence is known and its restriction enzyme map has been fully characterized. The restriction enzyme NdeI was chosen because of the number of bases yielded in the DNA fragments and the digested DNA piece corresponding to 3796 base pairs (bp) was targeted initially and gel purified. Sonicating this denatured λ DNA fragment with bulk HiPco SWCNTs (Carbon Nanotechnologies) did not disperse the CNTs visibly into solution, and few if any carbon nanotubes could be detected by either spectroscopy or * To whom correspondence should be addressed. E-mail: chaj@ us.ibm.com; tel: 408-927-1585; fax: 408-927-3310. 10.1021/nl0518775 CCC: $33.50 Published on Web 01/11/2006
© 2006 American Chemical Society
electron and atomic force microscopy (AFM). As demonstrated in Figure 1, thermal denaturation of the λ DNA fragment shown in Figure 1a followed by sonication with bulk SWCNTs produced no obvious carbon nanotubes in solution that could be detected by AFM. In fact, what was observed by AFM looked very similar in structure to what is often detected when imaging coils of single-stranded DNA on mica surfaces,13 indicating that only ssDNA remained in solution after sonication with the CNTs. Varying the temperatures and sonication times did not improve the dispersion of the bulk CNTs markedly. Using the other fragments from the NdeI digested λ DNA (27630 bp and 2433 bp) gave similarly negative results. Although ssDNA appears to be an obvious biopolymer to interface with CNT surfaces because of its aromatic bases, genomic DNA was clearly unable to disperse the CNTs effectively. One reason for this could be that the ssDNAs used were composed of random base sequences, preventing helix formation around the CNTs. Another plausible explanation is that the solutions of ssDNAs contained equimolar ratios of complementary DNA sequences because the ssDNAs were formed by simple heat denaturation of genomic
Figure 1. Lambda DNA fragment and products detected after sonicating its ssDNA with bulk SWCNTs. (a) Height AFM image of the gel-purified 3796 bp dsDNA fragment obtained after digestion of λ DNA with restriction enzyme NdeI. Arrows point to strands of dsDNA. (b) Height AFM image of the supernatant after sonication of denatured samples of the dsDNA shown in a with bulk HiPco SWCNTs. Arrows point to coils of single-stranded DNA.
dsDNA. The competitive hybridization between complementary sequence domains among the long base pair chains could afford many opportunities for preventing a significant portion of the DNA chains from interacting effectively with the CNT surfaces. To investigate the competitive binding of ssDNA to its complementary sequence versus a CNT, a method to separate genomic ssDNA from its complementary strand was developed. Because two complementary genomic strands of DNA have an equal number of bases and similar molecular weights, separating them cleanly is not trivial because it cannot be achieved by either gel electrophoresis or sizeexclusion filtration or chromatography. As is shown in Scheme 1, a technique was developed to produce amplified segments of λ DNA where one of the two strands of the dsDNA fragment was modified at its 5′ end with a thiol group while its complementary strand remained unmodified. The thiolated genomic dsDNA was then conjugated to gold nanoparticles, and the umodified ssDNA was removed from its nanoparticle conjugated complementary strand by centrifugation (Scheme 1). Lambda DNA was used as the DNA template for DNA polymerase chain reaction (PCR), and appropriate primers were designed so as to amplify a 3796 base-pair region of the lambda DNA sequence. One of the 50-base primers was synthesized modified at its 5′ end with a thiol group; the other primer was left unmodified. Using a molar excess of lambda DNA (1 µg/100 µL PCR reaction) over primers provided the best amplification results with little primer dimer formation (Figure 2a). The amplified thiol-modified dsDNA solutions were next mixed with 15-nm phosphinecapped gold nanoparticles14 using 0.5 to 1 molar ratios of 160
DNA to gold. After reactions for 1 h at room temperature, the gold nanoparticles were spun down at high centrifugation speeds to separate the dsDNA bound to the gold nanoparticles from free dsDNA. Both the resuspended gold pellet and the supernatant were run on 0.7% ethidium bromide (EtBr) agaraose gels. As demonstrated in Figure 2b, the gold band detected under white illumination is at exactly the same location within the gel as the approximately 4000 base pair dsDNA as detected by UV light, indicating that the thiolmodified genomic DNA did bind to the surface of the gold nanoparticles. However, as shown in Figure 2c, some of the thiol-modified dsDNA remains unbound to the gold nanoparticles because dsDNA is still detected in the supernatant; these results demonstrate the need to spin down the DNAconjugated gold nanoparticles away from unbound DNA left in solution. Because it has been shown by Li et al. that ssDNA has substantial affinities for gold surfaces,15 prior to thermal denaturation of the dsDNA bound to the gold nanoparticles (Au-dsDNA), the Au-dsDNA conjugates were diluted substantially to limit the interactions between the nanoparticles and any ssDNA not covalently tethered to a gold particle. Because only one of the two sequences of the initial genomic dsDNA was thiol modified, only one of the two complementary sequences should remain covalently bound to the gold nanoparticles after heating (Scheme 1). After denaturation of the dsDNA, the complementary unbound strands could therefore be separated away by centrifugation, producing a solution of ssDNA (ssDNA-1seq) that cannot self-hybridize through complementary base pairing. The Au-dsDNA conjugates were heated at 98 °C for 5 min and quenched in an ice-water bath and the mixture of tethered ssDNA-Au Nano Lett., Vol. 6, No. 2, 2006
Scheme 1. Schematic Diagram of the Method Developed to Produce High Yields of ssDNA with Their Complementary Sequences Removed
conjugates (Au-ssDNA) and the freely dispersed ssDNA (ssDNA-1seq) was centrifuged for 30 min at high speeds; the supernatant was separated from the gold pellet to obtain samples of unbound ssDNA. All of the ssDNA-1seq samples were lyophilized to dryness and reconstituted in deionized water to final concentrations of 10 µg/mL. The ssDNA-1seq solutions were next sonicated in a 4 °C ice-water bath with bulk samples of HiPco SWCNTs for a total of 10 to 20 min. As control experiments, solutions of ssDNA that had been produced by denaturing samples of the original amplified 3796 base pair dsDNA (Figure 2a) were also sonicated with the HiPco SWCNTs; these ssDNA solutions contained equimolar ratios of both complementary strands (ssDNA-2seq). All of the CNTs that did not disperse in solution were then centrifuged away at mild spin speeds and the supernatants were dried on both mica and (aminopropyl)triethoxysilane (APTES)-terminated silicon surfaces for AFM imaging. One of the immediate differences seen between using the ssDNA-1seq and ssDNA-2seq was that although very few if any of the CNTs dispersed in the ssDNA-2seq solutions, a gray color was detected in just a few minutes in the ssDNA1seq solutions (Figure 3a), indicating that the CNTs dispersed easily in the ssDNA samples having no complementary strands present (Figure 3a). AFM analyses revealed that the dispersed nanotubes remained on average greater than 700 nm in length (Figure 3b) and could reach in a few instances close to 2 micrometers. This most likely is due to both the lengths of DNA being used (3796 bases, calculated extended length of 1.4 microns) as well as the relatively short sonication times needed (10-20 min) to disperse the CNTs. Nano Lett., Vol. 6, No. 2, 2006
The correlation between the lengths of the DNA used and the CNTs dispersed in solution is currently being investigated. UV-visible absorption measurements of the bulk DNA CNT samples showed that the CNTs were relatively well dispersed in solution (Figure 3c), with spectra similar to those obtained using sodium dodecyl sulfate as the dispersing agent prior to centrifugation.16 Because mild sonication methods and times (20 min) were used in the work presented here, the presence of small nanotube bundles even after centrifugation is not surprising. The effective dispersion of the SWCNTs using single sequences of ssDNA (ssDNA-1seq) was revealed, by both height and phase AFM analyses, to be due to the ssDNAs’ ability to wrap as tight helices around the CNTs (Figure 3df) having very uniform pitches. In contrast to work published earlier,11 these results revealed that the DNA’s tight helix formation around CNTs might not necessarily be sequencedependent because the sequence of the ssDNA used here was genomic and completely random. Furthermore, the need to remove the DNA’s complementary strand from solution reveals that inhibiting self-recognition is a key element of allowing the DNA strand to interface with the carbon nanotube’s aromatic surface. The random sequence of bases of the DNA used also precludes the notion that any recognition of specific types of CNTs is occurring. This is supported by the observation that although only one sequence of DNA was used, almost 90% of the CNTs added initially dispersed in solution. Because of the known polydispersity of diameters, lengths, and chiralities of commercially produced CNTs, the ssDNA used here clearly could not be binding only one type of CNT and still demonstrating such 161
Figure 2. Gel electrophoresis analyses of PCR reactions and DNA conjugation to gold nanoparticles. (a) Results of PCR reaction to produce amplified amounts of thiolated dsDNA. MW: molecular weight markers; λ DNA: λ DNA (template DNA); control: products of control PCR reaction (no template DNA, only primers); PCR: results of PCR reaction (primers and template DNA). (b) Results of the attachment of the thiolated dsDNA to 15-nm gold particles. Under white light illumination, the gold nanoparticles are detected, whereas under UV light the ethidium bromide (EtBr) stained dsDNA is seen. A gold band and the DNA band are detected at the same location within the gel, indicating binding of DNA to gold. (c) Results of attaching thiolated dsDNA to gold followed by centrifugation of the gold nanoparticle reactions. Lane 1: molecular weight markers; lane 2: dsDNA products of PCR reaction (2a); lane 3: unconjugated 15-nm gold nanoparticles; lane 4: Resuspended Au pellet after centrifugation of the reactions of thiolated dsDNA and gold nanoparticles; lane 5: Supernatant obtained after centrifugation. As is observed in lanes 4 and 5, although some of the thiolated dsDNA did indeed bind to the gold nanoparticles, a portion of the DNA did not bind the gold nanoparticles.
dispersion yields. Preliminary AFM studies also show that although only one genomic DNA sequence is used, helices with a range of pitches can be detected for different carbon nanotubes (Figure 3e). Similar to the work by Zheng et al.,11 these early observations give rise to the possibility of separating these ssDNA-CNT materials by charge density, albeit in this case with much longer CNTs. Systematic analyses are currently underway to understand the relationship between CNT diameters and chiralities with the helical wrapping of genomic ssDNA around these nanotubes. Because the ssDNA forms a helix around each CNT with a distinct periodic pitch and with the aromatic bases presumably facing the CNT surface, a regular set of negative charges is thought to evenly distribute down the length of the DNA-CNTs because of the phosphate groups of the DNA backbone. Because the DNA-CNT can be considered to be 162
a structure similar to that of genomic dsDNA, albeit with half the negative charge and with an increase in the helix pitch, it was thought that one could use the techniques developed for combing DNA on surfaces to produce aligned arrays of the DNA-CNTs.17-18 As shown in Figure 4a and b, high-density arrays of aligned ssDNA-CNT hybrids could be produced at receding edges of air-drying droplets on (aminopropyl)triethoxysilane (APTES)-terminated silicon substrates. Extending genomic DNA physics to these DNACNT hybrid structures provides fascinating new avenues of research that can have a possible impact on nanoelectronics. It has been demonstrated here that in the absence of complementary strands, genomic ssDNA having lengths far greater than 100 bases and composed of an arbitrary sequence of bases not only absorb onto SWCNTs but can also disperse the CNTs effectively in water by forming tight helices around Nano Lett., Vol. 6, No. 2, 2006
Figure 3. Results of reacting ssDNA with bulk HiPco SWCNTs. (a) Visual comparison of the ssDNA-CNT solutions obtained after sonication of CNTs with either a ssDNA solution containing complementary sequences (ssDNA-2seq, left) or a ssDNA solution obtained after removal of complementary strand (ssDNA-1seq, right). (b) Large area height AFM scan of DNA bound to CNTs on mica. Inset: close-up of large area scan. (c) UV-visible absorption spectrum of ssDNA dispersed HiPco CNTs. (d) Height AFM image of one ssDNAwrapped SWCNT and section analysis of the DNA-CNT structure indicated by the blue arrow. Section analysis gives a large pitch of ∼60 nm for that particular DNA-CNT. The difference in pitch should be noted in the DNA-CNT structure shown in the center of the image and in the zoomed image. (e) Height AFM image of a single ssDNA-wrapped CNT. (f) Phase AFM image of several ssDNA-wrapped CNTs.
individual CNTs. Recent studies of the elastic forces of long ssDNAs have demonstrated that under base-pairing conditions (neutral pH, mid ionic strengths), the force needed to extend long ssDNA goes to a constant value of ∼2 piconewtons (pN), which is much smaller than the 15 pNs needed to open up a 10 base pair DNA helix-loop.19 It has furthermore been demonstrated that such self-complementary Nano Lett., Vol. 6, No. 2, 2006
regions that require forces of 15 pN in order to open do not exist in long genomic ssDNA.19 These studies provide a plausible explanation for the dramatic differences observed in dispersing CNTs when using solutions of genomic ssDNA containing both complementary sequences as opposed to those having only one of the two sequences. Interestingly, we found that short 50-base oligomers having random base 163
Figure 4. Aligned array of ssDNA-CNT hybrids at edge of air-dried droplet on APTES functionalized Si wafers. (a) Height AFM image, 5 × 5 µm2 scan. (b) Height AFM image, 2 × 2 µm2 scan.
sequences were significantly less effective in dispersing CNTs. Although the short DNA strands did initially appear to partially disperse the CNTs, after centrifugation very few CNTs could be detected visually or by spectroscopy. This might in part be due to the different folding characterstics long genomic ssDNA has compared to short ssDNA strands that have random base sequences and are less than 100 bases in length. The work presented here demonstrates that long genomic ssDNA having random base sequences can efficiently form complexes with SWNTs by wrapping around the nanotubes as tight helices, provided that the complementary ssDNA strands are first removed. It was furthermore observed that the ssDNA helices exhibited different pitches for different nanotubes. This novel type of hybrid material, combined with the biological recognition and binding capabilities of DNA, may make possible the development of new methods for separating different types of CNTs, and for aligning and assembling them into larger scale structures. Further work to characterize the effects of CNT diameter and chirality on ssDNA-CNT helix formation is needed to evaluate these possibilities. Acknowledgment. We thank Ho-Cheol Kim and Willi Wolksen for providing the APTES modified substrates. We would also like to thank Jane Frommer and Kerem Unal for assistance with the AFM measurements. We would also like to acknowledge Robert Miller, Geraud Dubois, James Hedrick, and Jia Chen for additional helpful discussions and suggestions. Brittany Gigliotti acknowledges the INROADS program for sponsoring her summer research employmernt
164
at the IBM Almaden Research Center, and Brenda Sakizzie acknowledges NSF GOALI Grant CHE9625628 for funding her summer internship at IBM. References (1) Mirkin, C. A.; Letsinger, R. L.; Mucic, R. C.; Storhoff, J. J. Nature 1996, 382, 607. (2) Fu, A.; Micheel, C. M.; Cha, J.; Chang, H.; Yang, H.; Alivisatos, A. P. J. Am. Chem. Soc. 2004, 126, 10832. (3) Alivisatos, A. P.; Johnsson, K. P.; Peng, X.; Wilson, T. E.; Loweth, C. J.; Bruchez, M. P.; Schultz, P. G. Nature 1996, 382, 609. (4) Shenton, W.; Davis, S. A.; Mann, S. AdV. Mater. 1999, 11, 449. (5) Lee, S.-W.; Mao, C.; Flynn, C. E.; Belcher, A. M. Science 2002, 296, 892. (6) Hartgerink, J. D.; Beniash, E.; Stupp, S. I. Science 2001, 294, 1684. (7) Sarikaya, M.; Tamerler, C.; Jen, A. K.-Y.; Schulten, K.; Baneyx, F. Nat. Mater. 2003, 2, 577. (8) Oren, E. E.; Tamerler, C.; Sarikaya, M. Nano Lett. 2005, 5, 415. (9) Whaley, S. R.; English, D. S.; Hu, E. L.; Barbara, P. F.; Belcher, A. M. Nature 2000, 405, 665. (10) Zheng, M.; Jagota, A.; Semke, E. D.; Diner, B. A.; Mclean, R. S.; Lustig, S. R.; Richardson, R. E.; Tassi, N. G. Nat. Mater. 2003, 2, 338. (11) Zheng, M. et al. Science 2003, 302, 1545. (12) Wind, S. J.; Appenzeller, J.; Martel, R.; Derycke, V.; Avouris, P. J. Vac. Sci. Technol., B. 2002, 20, 2798. (13) Hansma, H. G.; Sinsheimer, R. L.; Li, M. Q.; Hansma, P. K. Acids Res. 1992, 20, 3585. (14) Loweth, C. J.; Caldwell, W. B.; Peng, X.; Alivisatos, A. P.; Schultz, P. G. Angew. Chem., Int. Ed. 1999, 38, 1808. (15) Li, H.; Rothberg, L. Proc. Natl. Acad. Sci. 2004, 101, 14036. (16) O’Connell, M. J. et al. Science 2002, 297, 593. (17) Bensimon, A.; Simon, A.; Chiffaudel, A.; Croquette, V.; Heslot, F.; Bensimon, D. Science 1994, 265, 2096. (18) Michalet, X. et al. Science 1997, 277, 1518. (19) Cocco, S.; Marko, J. F.; Monasson, R.; Sarkar, A.; Yan, J. Eur. Phys. J. E. 2003, 10, 249.
NL0518775
Nano Lett., Vol. 6, No. 2, 2006
