Molecular Voids Formed from Effective Attraction in Submonolayer

Apr 28, 2009 - an attractive force driven molecular void formation for DNA chemically ... mechanisms for the void formation implies an effective attra...
0 downloads 0 Views 3MB Size
Download PDF
pubs.acs.org/Langmuir © 2009 American Chemical Society

Molecular Voids Formed from Effective Attraction in Submonolayer DNA Deposited on Au(111) Pengshun Luo,† Norman L. Bemelmans,‡ Michael S. Woody,§ and Thomas P. Pearl*,† †

Department of Physics, North Carolina State University, Raleigh, North Carolina 27695-7518, ‡Department of Chemistry, North Carolina State University, Raleigh, North Carolina 27695-8204, and §Department of Physics and Astronomy, University of North Carolina, Chapel Hill, North Carolina 27599-3255 Received February 6, 2009. Revised Manuscript Received April 3, 2009

The development of DNA-based biosensors requires a deep understanding of how DNA molecules adsorb and organize on solid state surfaces as well as the electronic properties of individual and aggregates of DNA molecules. Using scanning tunneling microscopy (STM) and atomic force microscopy (AFM), we have successfully characterized an attractive force driven molecular void formation for DNA chemically adsorbed on Au(111) as a function of strand length and deposition conditions. Here we report the observation of these void structures formed on the Au(111) surface by adsorption of both 45 and 90 base pair long, thiolated double-stranded DNA. We found that the average void diameter decreases when increasing the number of base pairs exposed to the surface. The critical determinant in the molecular void formation is the total charge delivered to the surface via the adsorption of the DNA strands and the related counterions, which can ultimately be quantified by the number of base pairs in each adsorbed DNA molecule. Complementary measurements involving STM and AFM suggest that an intact Au(111) surface area is preserved inside the void and is surrounded by a submonolayer of DNA molecules adsorbed on the surface. The discussion of the possible mechanisms for the void formation implies an effective attraction between the DNA molecules.

Introduction Besides its genetic function, the DNA molecule is well suited for biosensing because of the specific recognition between its complementary base pairs. Most DNA-based biosensors consist of a probe layer of single-stranded DNA (ssDNA) immobilized on a solid state surface.1-3 The biosensing events occur when complementary target ssDNA strands hybridize with the probe DNA while an optical or electrical signal is detected depending on the transducer mechanism. The efficiency and accuracy of the DNAbased biosensors depend greatly on how the probe ssDNA adsorb and organize on the solid state surface. A model system involving DNA molecules chemically bonded to Au surfaces has been widely investigated for the purpose of extracting molecular details associated with the integration and interaction of DNA with metallic contacts. The hybridization efficiency has been found to depend on the probe DNA density and alignment.4-7 An optimal probe DNA coverage and alignment are critical to produce better accessibility for the target DNA. The density of the probe DNA can be manipulated by controlling the exposure time,4,5,7-9 DNA solution concentration,5,8 and salt concentrations in buffer4,7,9,10 *To whom correspondence should be addressed. (1) Wang, J. Nucleic Acids Res. 2000, 28, 3011–3016. (2) Drummond, T. G.; Hill, M. G.; Barton, J. K. Nat. Biotechnol. 2003, 21, 1192–1199. (3) Odenthal, K. J.; Gooding, J. J. Analyst 2007, 132, 603–610. (4) Herne, T. M.; Tarlov, M. J. J. Am. Chem. Soc. 1997, 119, 8916–8920. (5) Huang, E.; Satjapipat, M.; Han, S. B.; Zhou, F. M. Langmuir 2001, 17, 1215–1224. (6) Satjapipat, M.; Sanedrin, R.; Zhou, F. M. Langmuir 2001, 17, 7637–7644. (7) Peterson, A. W.; Heaton, R. J.; Georgiadis, R. M. Nucleic Acids Res. 2001, 29, 5163–5168. (8) Steel, A. B.; Levicky, R. L.; Herne, T. M.; Tarlov, M. J. Biophys. J. 2000, 79, 975–981. (9) Petrovykh, D. Y.; Kimura-Suda, H.; Whitman, L. J.; Tarlov, M. J. J. Am. Chem. Soc. 2003, 125, 5219–5226. (10) Sakao, Y.; Ueno, N.; Nakamura, F.; Ito, E.; Hayasi, J.; Hara, M. Mol. Cryst. Liq. Cryst. 2003, 407, 537–542. (11) Arinaga, K.; Rant, U.; Knezevic, J.; Pvoidsheim, E.; Tornow, M.; Fujita, S.; Abstreiter, G.; Yokoyama, N. Biosens. Bioelectron. 2007, 23, 326–331.

Langmuir 2009, 25(14), 7995–8000

and also by controllable desorption of a densely packed layer of DNA.11 An alkanethiol self-assembled monolayer has also been used to remove the nonspecific adsorption of DNA and at the same time form the spacers between DNA to enhance the accessibility for the target DNA.4-8 The immobilization of DNA on a surface is a complicated process which involves a wide variety of cooperative and competing interactions, including the specific DNA surface chemistry12-15 as well as the intercommunication between DNA strands and between DNA and buffer counterions in solution. The negatively charged backbones of DNA molecules interact electrostatically with each other as well as with ions existing in the solution. For example, the DNA coverage on a Au surface can be increased by increasing the salt concentration in solution which effectively reduces the electrostatic repulsion between DNA molecules by counterions.4,7,9,10 The alignment of DNA molecules has been adjusted by applying a negative potential on the Au substrate, as a result producing a well-ordered DNA self-assembled monolayer.15-17 On the other hand, it has been widely observed that high molecular weight DNA can condense to a compact, usually highly ordered toroidal structure in vitro in the presence of multivalent cations.18,19 The experimental evidence shows that DNA condensation occurs when about 90% of its charge is neutralized by counterions. An effective attractive (12) Mourougou-Candoni, N.; Naud, C.; Thibaudau, F. Langmuir 2003, 19, 682–686. (13) Casero, E.; Darder, M.; Diaz, D. J.; Pariente, F.; Martin-Gago, J. A.; Abruna, H.; Lorenzo, E. Langmuir 2003, 19, 6230–6235. (14) Wolf, L. K.; Gao, Y.; Georgiadis, R. M. Langmuir 2004, 20, 3357–3361. (15) Wackerbarth, H.; Grubb, M.; Zhang, J. D.; Hansen, A. G.; Ulstrup, J. Langmuir 2004, 20, 1647–1655. (16) Wackerbarth, H.; Grubb, M.; Zhang, J. D.; Hansen, A. G.; Ulstrup, J. Angew. Chem., Int. Ed. 2004, 43, 198–203. (17) Grubb, M.; Wackerbarth, H.; Ulstrup, J. J. Am. Chem. Soc. 2006, 128, 7734–7735. (18) Bloomfield, V. A. Biopolymers 1997, 44, 269–282. (19) Lyubchenko, Y. L. Cell Biochem. Biophys. 2004, 41, 75–98.

Published on Web 04/28/2009

DOI: 10.1021/la900470h

7995

Article

Luo et al.

force is required to favor the DNA condensation, which may arise from the fluctuating counterion electrostatics and hydration force.18 In this paper, we have studied the adsorption and organization of double-stranded DNA (dsDNA) anchored on the Au(111) surface by a chemical Au-S bond, a model system for both fundamental research and application development. We chose to use a short deposition time and low DNA deposition concentrations to prepare a submonolayer of DNA lying down on the Au(111) surface. The deposition time is selected to be 5 min where chemical adsorption through Au-S interaction dominates over any nonspecific adsorption in solution. In this specific deposition condition window, we have observed a novel structure, the molecular void, with both STM and AFM techniques. Our measurement and sample preparation approach, including judicious choice of solution compositions from which the DNA is deposited on the surface, address the mechanism and possible interactions that give rise to the observed molecular voids.

Experimental Details All DNA strands were purchased from Integrated DNA technologies (IDT) with HPLC purification. The dsDNA was obtained by hybridizing the two complementary single strands purchased from IDT together. The 50 thiol modified ssDNA were received in a disulfide form (IDT modification code: /5ThioMC6-D/), protected by a mercaptohexanol group. Both the thiol modified ssDNA and their complementary strands were suspended in either PBS buffer or 1 TE buffer. The PBS buffer consists of 10 mM phosphate buffer (Sigma-Aldrich, pH 8.0), 300 mM NaCl, and 1 mM ethylenediaminetetraacetic acid (EDTA, titrated to pH 8.0 with NaOH, Sigma-Aldrich). The 1 TE buffer consists of 10 mM Tris (tris(hydroxymethyl)aminomethane, Fischer Scientific) and 1 mM EDTA. The 50 thiol modified dsDNA were prepared by heating the mixed solution of these two strands to 95 °C for 5 min and then cooling down slowly to room temperature. The disulfide protection was cleaved by treating the dsDNA with dithiothreitol (DTT, Pierce) in a mole ratio of 1:500 (DNA:DTT). After a 1 h treatment at room temperature, the thiolated dsDNA, abbreviated as HS-dsDNA, was isolated and collected by using a size exclusion chromatography PD-10 column (GE Healthcare). The collected HS-dsDNA was then stored at -20 °C for future use. Two lengths of DNA, 45 and 90 bp, were used for the DNA strand length dependence study.20 The Au(111) substrate was prepared by melting a high-purity Au wire (0.02 in. diameter, 99.999% purity, Alfa Aesar) into a gold bead termination and annealing with a hydrogen flame. The other end of the Au wire was welded to a platinum foil (0.004 in. thick, 99.99% purity, Alfa Aesar), which was in turn clamped to a molybdenum sample holder for placement in different microscope setups. Large terraces up to 1 μm wide of Au(111) facets on the gold bead can be routinely achieved by this method as measured by STM and AFM. The large terraces provide a facile platform to observe the dispersed molecular void structures on the surface. The gold bead can be reannealed and used after cleaning in piranha solution (1:3 mixture of H2O2 and H2SO4) and then ultrasonic cleaning in separate acetone and methanol baths. The dsDNA molecules were deposited onto the surface by placing the gold bead in contact with a small droplet (∼20 μL) of HS-dsDNA solution of varying concentration in a 1.5 mL (20) Integrated DNA Technologies-45 bp: 50 -CGT GAG GCT GCT ACC GCT TTT TTT TTT TTT TTT TAC TAT CTT GTG-30 , 50 -/5ThioMC6-D/CAC AAG ATA GTA AAA AAA AAA AAA AAA AGC GGT AGC AGC CTC ACG-30 ; 90 bp: 50 -ACT GCT GTT CGA CAA TGC TCT TAG GCG AAA AAA AAA AAA ACT GAG AAA TAC CGT AGC TAA CGC TCA TAA AAA AAA AAA AAG ACT CTC TCG-30 , 50 -/5ThioMC6-D/CGA GAG AGT CTT TTT TTT TTT TTA TGA GCG TTA GCT ACG GTA TTT CTC AGT TTT TTT TTT TTT CGC CTA AGA GCA TTG TCG AAC AGC AGT-30 .

7996 DOI: 10.1021/la900470h

centrifuge tube for 5 min. The sample was then rinsed thoroughly with 18 MΩ deionized water and dried with dry nitrogen gas. STM measurements were performed using a custom built, high gap impedance, ambient microscope setup21 with typical imaging conditions of sample biased with respect to the tip, Vb = 750 mV, and a tunneling current, It = 30 pA. AFM measurements were recorded with either a ThermoMicroscopes Autoprobe CP Research System or an Asylum Research MFP-3D-SA microscope setup. The AFM images were taken in tapping mode with a microMasch NSC14 cantilever (typical resonant frequency: ∼160 kHz; force constant: ∼5.0 N/m). All images were processed with a free image processing software package.22

Results and Discussion The deposition of HS-dsDNA was performed in a broad DNA concentration range from several nanomolar to the highest concentration of the stock solution (∼286 nM). The HS-dsDNA used in this study was suspended primarily in PBS buffer except where noted. The variation of the DNA concentration was obtained by diluting the stock solution with PBS buffer. Figure 1 presents a set of AFM measurements of 45 bp HS-dsDNA taken in a specific concentration window (60-160 nM) where we can see the development of DNA adsorption on the surface. At low concentrations (60 nM or lower), we are able to resolve the individual DNA strands on the surface. As the deposition concentration is increased, the adsorbed DNA molecules pack more and more densely as expected.5,8 At a certain stage, the distance between the DNA molecules becomes too small to be resolved with an AFM tip. Besides the expected concentration dependence of the DNA adsorption, a novel structure, which will be denoted as a molecular void, has been observed. The large-scale image (Figure 1e) shows that the molecular voids disperse sparsely on the surface. We have not observed such molecular void structures in the deposition of PBS buffer alone, which confirms that the molecular void is not formed from the residue of PBS buffer solution (see Figure S1a). We can see the precursor of the molecular void at low concentration of 60 nM where the DNA surface density is too low to form a complete circular boundary of the void. The surface area inside of the void is not occupied by DNA molecules. With increasing DNA surface coverage (as a result of increasing concentration of DNA in solution), we note that the void diameter shrinks. The statistical analysis of the voids shows that a broad distribution of void diameter size exists with a standard deviation as large as 10 nm for each concentration, while the average void diameter decreases with increasing DNA concentration (see Figure 2). The saturation of the average void diameter at low concentration is due to the limited resolution of the AFM tip. We have performed a similar set of experiments on the 90 bp long HSdsDNA. The results from 90 bp HS-dsDNA data follow a similar dependence of the void diameter on the concentration when we scale the HS-dsDNA molar concentration to the molar concentration of base pairs (see Figure 2b). Since the DNA backbone carries approximately one fundamental charge for each base in solution, this result implies that the void diameter is related to the amount of base pairs as well as the charge density on the surface. The molecular void structure is also found to vary when we dilute the stock HS-dsDNA with deionized water as opposed to maintaining a constant buffer concentration for varying molar concentrations of DNA. As the PBS buffer is diluted when (21) Santagata, N. M.; Luo, P. S.; Lakhani, A. M.; DeWitt, D. J.; Day, B. S.; Norton, M. L.; Pearl, T. P. IEEE Sens. J. 2008, 8, 758–766. (22) Horcas, I.; Fernandez, R.; Gomez-Rodriguez, J. M.; Colchero, J.; GomezHerrero, J.; Baro, A. M. Rev. Sci. Instrum. 2007, 78, 013705.

Langmuir 2009, 25(14), 7995–8000

Luo et al.

Article

Figure 1. AFM images of 45 bp HS-dsDNA deposited on Au(111) from a PBS buffered solution with varying DNA concentrations: (a-d) 60, 80, 120, and 160 nM, scan size 800 nm  800 nm; (e) large-scale image of (b), scan size 3 μm  3 μm. One of the void structures is indicated by the arrow in each image.

diluting the stock HS-dsDNA with deionized water, the variation implies a salt or ion concentration dependence of the void structures. The void structures prepared from the HS-dsDNA solution of the same concentration, but diluted with deionized water from the stock solution, have more pronounced protruding edges as shown in Figure 3 and Figure S2a. As can be seen in the line profile in Figure 3b, the void is ∼58 nm in diameter. Its edge is 0.7 nm higher than the surface outside of the void. The void is 0.5 nm deeper than the rest of the surface as are most of the voids prepared in pure PBS buffer (see an example of a void profile from Figure 1d as shown in Figure 3b). The difference in height between the inside and the outside of the void is comparable to the height of lambda DNA (BstE II digested bacteria phage Lambda DNA, Sigma-Aldrich) deposited on mica measured with the same AFM instrument and the same type of cantilever. It should be noted here that gold vacancy islands have been observed in alkanethiol self-assembled monolayers on the Au(111) surface15-17,23-25 ascribed to the Au-S abstraction interaction. Those vacancy islands have a depth identical to one Au(111) atomic step and are distinctly different from the void structures observed here. The lateral size of such a gold vacancy is only several nanometers in the ordered thiolated ssDNA monolayer.15-17 The much larger diameter of the void also excludes the possible explanation of the structures as gold vacancies due to Au-S interaction. The comparable depth of the void to the height of lambda DNA suggests that the void is an intact gold surface area surrounded by a submonolayer of DNA molecules lying down on the substrate. The AFM image for DNA deposited at a low concentration of 60 nM (Figure 1a) also provides a more direct evidence for this supposition. Further analysis by comparing the AFM topography (Figure 4a) and phase (Figure 4b) images also supports this scenario. The tapping mode AFM phase image is a measure of energy dissipation due to the AFM tip interaction with the surface.26-28 The phase image is a map of the difference in (23) Haussling, L.; Michel, B.; Voidsdorf, H.; Rohrer, H. Angew. Chem., Int. Ed. Engl. 1991, 30, 569–572. (24) Poirier, G. E. Langmuir 1997, 13, 2019–2026. (25) Poirier, G. E. Chem. Rev. 1997, 97, 1117–1127. (26) Tamayo, J.; Garcia, R. Langmuir 1996, 12, 4430–4435. (27) Garcia, R.; Tamayo, J.; San Paulo, A. Surf. Interface Anal. 1998, 27, 312–316. (28) Garcia, R.; Gomez, C. J.; Martinez, N. F.; Patil, S.; Dietz, C.; Magerle, R. Phys. Rev. Lett. 2006, 97, 016103.

Langmuir 2009, 25(14), 7995–8000

phase between the resonant cantilever and its excitation source. In general, the amplitude of the phase shift increases with increasing surface viscosity or with decreasing surface stiffness.26-28 The AFM phase image in Figure 4b shows that the phase contrast only appears between the inside and the outside of the void, while there is no phase difference between Au (111) terraces covered with DNA. Smaller (darker) phase values in the void compared to the outside of the void imply that a stiffer material (Au) inside of the void is interacting with the AFM tip, which is consistent with our conclusions. To further assess the chemical composition of the void , we chose to image the same features with STM. In contrast to the AFM images, the void structures have a different appearance in the constant current STM image (see Figure 5). The void appears flatter and ∼1 A˚ higher than the outside of the void (see Figure 5c,d). The constant current STM image depends not only on the surface topographic structure but also on the electronic properties of the surface and the electron tunneling path. The DNA molecules adsorbed on the Au(111) surface have been reported to appear as a gentle depression, which can be explained in terms of resonant tunneling through virtual states between STM tip and the DNA molecule.21,29,30 The depression outside of the void supports our suggestion that the Au(111) surface is encircled with DNA molecules outside of the void. The flat surface inside of the void also tells us that a clean Au(111) surface area is preserved. The scanning tunneling spectroscopy measurement of the void structures also supports our conclusion (see Figure S3). Both the differential conductance (dI/dV) map and dI/dZ map show a difference only between the inside and the outside of the void on the surface. The void has a higher conductivity and work function than outside of the void. While the above experimental evidence strongly supports the supposition that the void structure is an intact Au(111) surface surrounded by a submonolayer of DNA molecules, the mechanism of void formation is less straightforward. Since we have been able to determine that electrostatic forces play an important role in the formation of the voids, due to the dependence of the void statistics on DNA base pair concentration, let us discuss what (29) Shapir, E.; Yi, J. Y.; Cohen, H.; Kotlyar, A. B.; Cuniberti, G.; Porath, D. J. Phys. Chem. B 2005, 109, 14270–14274. (30) Kim, D. H.; Shapir, E.; Jeong, H.; Porath, D.; Yi, J. Phys. Rev. B 2006, 73, 235416.

DOI: 10.1021/la900470h

7997

Article

Figure 2. (a) Statistical analysis of the void diameters as a function of the HS-dsDNA solution concentration for 45 bp dsDNA. Each set of data is offset for clarity. The data are extracted from more than 80 voids from at least three different surface locations for each concentration. (b) Dependence of the mean value of the void diameter on HS-dsDNA solution concentration, in terms of base pairs, for both 45 and 90 bp HS-dsDNA. The error bars represent the standard deviation.

attractive or repulsive interactions may be involved in the void formation. On one hand, the DNA backbones in solution are negatively charged with a linear charge density of approximately 1 e/1.7 A˚.31 On the other hand, it has been well established that the negative charges on the DNA are compensated by the condensation of counterions in solution.18,19,31-33 The counterion condensation on the DNA strands happens in a delocalized and dynamic way with higher local concentration of the small mobile cations aggregating around the DNA backbones. In terms of the Manning counterion condensation theory,31 the effective charge per unit length of DNA is only a fraction 1/Zξ (Z is the counterion valency) of its bare value. The Manning parameter, ξ, is defined as the ratio of lB/b, where lB = e2/εkBT is the Bjerrum length and 1/b is the linear density of charge on DNA. The counterions are considered to be bound to the DNA molecules defined by a (31) Manning, G. S. Q. Rev. Biophys. 1978, 11, 179–246. (32) Wilson, R. W.; Bloomfield, V. A. Biochemistry 1979, 18, 2192–2196. (33) Anderson, C. F.; Record, M. T. Annu. Rev. Biophys. Biophys. Chem. 1990, 19, 423–465.

7998 DOI: 10.1021/la900470h

Luo et al.

Figure 3. (a) AFM image of 45 bp HS-dsDNA deposited on Au (111) from a 60 nM HS-dsDNA solution showing void structures on the multiple terraces of the surface. For this example, the 60 nM HS-dsDNA solution was prepared by diluting the stock HSdsDNA suspended in PBS buffer with deionized water. Scan size: 3 μm  3 μm. (b) Line profiles across the voids along the lines marked in the inset of (a) (empty circle) and in Figure 1d (solid circle). In Figure 1d, the line crosses a single gold atom step which is measured to be ∼0.27 nm.

Figure 4. AFM topography (a) and phase (b) images of the void recorded simultaneously. Scan size: 600 nm  600 nm. A 70 nM 45 bp HS-dsDNA solution diluted with deionized water from the stock solution was used for deposition on to the Au(111) surface. The void structure is indicated by the arrows.

volume of region, Vp. When using univalent Na+ as counterions, the effective charge density is ∼0.24 of the bare value for native DNA and independent of Na+ concentration up to 0.1 M.31 Langmuir 2009, 25(14), 7995–8000

Luo et al.

Figure 5. AFM image (a) and STM image (b) recorded from the same Au(111) surface exposed to 70 nM 45 bp dsDNA suspended in PBS. Scan sizes: 1 μm  1 μm. One of the void structures is indicated by the arrow in each image. (c) Small scale STM image of the void structure recorded from Au(111) surface exposed to 45 nM 90 bp dsDNA suspended in PBS. Scan size: 100 nm  100 nm. (d) Line profile along the line as marked in (c).

The thickness of cylindrical shells of Vp is 7 A˚. With further increase of the Na+ concentration, this value slightly decreases to 0.17 at 1 M of Na+. The compensation of the negative charge on the DNA greatly reduces the electrostatic repulsion between DNA strands. This is part of the reason why the graft density of DNA increases with an increase of the salt concentration.4,7,9 Furthermore, in the presence of multivalent cations, the interaction between DNA molecules can become effectively attractive instead of repulsive due to fluctuating counterion electrostatics and hydration force which makes the DNA condensation happen.18,19,32 The effective attractive force can arise from induced dipole interactions between the fluctuating ion atmospheres surrounding DNA34-36 or correlated ionic fluctuations37 and has been demonstrated by simulations using Monte Carlo methods38-40 and Brownian dynamics simulations.41 In the framework of counterion condensation theory, Ray and Manning have also calculated the potential of the mean force between a pair of parallel oriented like-charged polyions in a solution of monovalent ions.42 They found the interaction becomes attractive when the distance falls in an intermediate range in which the distance is not small but less than the Debye length, and the polyions have mutually penetrated each other’s Debye atmosphere and share a common population of condensed counterions. Away from this distance range, either closer or farther, the force becomes repulsive. The hydration force, which is due to reconfiguration of water between the surface (34) Oosawa, F. Biopolymers 1968, 6, 1633–1647. (35) Marquet, R.; Houssier, C. J. Biomol. Struct. Dyn. 1991, 9, 159–167. (36) Ha, B. Y.; Liu, A. J. Phys. Rev. Lett. 1997, 79, 1289–1292. (37) Rouzina, I.; Bloomfield, V. A. J. Phys. Chem. 1996, 100, 9977–9989. (38) Guldbrand, L.; Nilsson, L. G.; Nordenskiold, L. J. Chem. Phys. 1986, 85, 6686–6698. (39) Hribar, B.; Vlachy, V. J. Phys. Chem. B 1997, 101, 3457–3459. (40) Khan, M. O.; Jonsson, B. Biopolymers 1999, 49, 121–125. (41) GronbechJensen, N.; Mashl, R. J.; Bruinsma, R. F.; Gelbart, W. M. Phys. Rev. Lett. 1997, 78, 2477–2480. (42) Ray, J.; Manning, G. S. Langmuir 1994, 10, 2450–2461.

Langmuir 2009, 25(14), 7995–8000

Article

of macromolecules, may be another source of an attractive force.18,43 The hydration force can be either attractive or repulsive depending on the mutual arrangement of the water on the surface of the DNA. Leikin et al. stated that attractions result from a complementary ordering, while repulsion is due to symmetrical structuring.44 On the other hand, steric interactions may also keep the DNA collapsing to each other. The effective attraction favors 2D molecular void structure formation rather than the repulsive interaction. If one considers only the electrostatic repulsion, the void structures may form only if negative charges accumulate at a particular surface location, and the local charges would furthermore need to be large enough to expel the DNA molecules out of that specific area and thus form a molecular void. If we consider the charge neutralization of DNA backbones by the counterions, even larger local charges are needed to create the same repulsive force compared to the bare DNA backbones case. Considering NaCl in the buffer solution, the chlorine anions (Cl-) may adsorb on the Au(111) surface and then serve as a source of the local negative charges. The counterparts, sodium cations (Na+), condense on the DNA molecules to compensate the negative charges on the DNA strands.31,33 However, such separation of cations and anions would be unlikely to happen as it enhances the total free energy. However, an effective attraction makes DNA molecules lying on the surface aggregate together in a monolayer fashion when the DNA is confined to the surface. At very low coverage, the DNA molecules are separated from each other and the interaction is too weak to overcome the Au-S bonding or the energy barrier of moving Au-S complexes, so that the DNA molecules disperse on the surface randomly. When the surface coverage of DNA is increased, the DNA molecules attract each other and aggregate to form a denser monolayer (DNA should still lie down on the surface in our deposition concentration window) and leave some empty space unoccupied with DNA. From an energetic standpoint, the empty spaces prefer a circular boundary which has the highest 2D symmetry. This may explain the mechanism of the molecular void structure formation. With further increase of the surface coverage of DNA, the void diameter decreases and finally the whole surface is covered with DNA molecules. Here we should note, based on this model, the void diameter depends on base pair coverage on the surface, which is consistent with our experimental observation. In the absence of counterions, the electrostatic repulsion is large enough to distribute the DNA molecules on the surface in order to minimize the total free energy and thus no void structures should be observed. In order to further clarify this issue, we have performed control experiments by suspending the HS-dsDNA either in 1 TE buffer or in 1 TE buffer with 300 mM NaCl (see Figures S1b and S1c). We have not observed any molecular voids without NaCl added to the solution. This further confirms the importance of the counterion condensation and ion fluctuating electrostatic effects, which enables the production of an effective attractive interaction between DNA molecules. The effective attraction is supposed to depend on the salt concentration, as the variation of counterion concentration changes the Debye screening length, the ion fluctuation electrostatics, and also the water configuration on the DNA molecular surface. It is not surprising that the void structures depend on the dilution methods and thus the salt concentration in Figure 3, but further investigation will be needed for a proper explanation of the protruding edges. (43) Leikin, S.; Parsegian, V. A.; Rau, D. C.; Rand, R. P. Annu. Rev. Phys. Chem. 1993, 44, 369–395. (44) Leikin, S.; Rau, D. C.; Parsegian, V. A. Phys. Rev. A 1991, 44, 5272–5278.

DOI: 10.1021/la900470h

7999

Article

Luo et al.

Conclusions We have studied the HS-dsDNA organization on the Au(111) surface in a specific molecular coverage window. A novel structure, a molecular void, has been observed and identified to be an intact Au(111) surface area surrounded with submonolayer DNA. The discussion of the mechanism of the void formation implies that interaction between the DNA molecules is effectively attractive instead of repulsive in our study. The purely attractive interaction is a superposition effect of both the counterion condensation and other attractive interactions, such as fluctuating counterion electrostatic and hydration forces. Acknowledgment. The authors thank Michael L. Norton and Brian S. Day for help regarding the thiolated DNA sample

8000 DOI: 10.1021/la900470h

preparation and Nancy M. Santagata, Shawn M. Huston, Satyaveda C. Bharath, and Jae-Ryang Hahn for fruitful discussions. We also thank Xianhua Kong and Robert J. Nemanich for help with AFM measurements. We acknowledge funding and support from the Army Research Office and Defense Threat Reduction Agency for this work. Supporting Information Available: Included in the supporting information are AFM images recorded from the control experiments using different buffer solutions, additional comparison of results with/without diluted PBS buffer solution, and experimental details and results of STM/STS measurements. This material is available free of charge via the Internet at http://pubs.acs.org.

Langmuir 2009, 25(14), 7995–8000