- Email: [email protected]
Structural study on gold nanoparticle functionalized with DNA and its non-cross-linking aggregation
Journal of Colloid and Interface Science 368 (2012) 629–635
Contents lists available at SciVerse ScienceDirect
Journal of Colloid and Interface Science www.elsevier.com/locate/jcis
Structural study on gold nanoparticle functionalized with DNA and its non-cross-linking aggregation Masahiro Fujita a,⇑, Yoshizumi Katafuchi b, Kazuki Ito c, Naoki Kanayama a, Tohru Takarada a, Mizuo Maeda a a b c
Bioengineering Laboratory, RIKEN, Hirosawa 2-1, Wako, Saitama 351-0198, Japan Faculty of Industrial Science and Technology, Tokyo University of Science, Yamazaki 2641, Noda, Chiba 278-8510, Japan Structural Materials Science Laboratory, RIKEN SPring-8 Center, Kouto 1-1-1, Sayo, Hyogo 679-5148, Japan
a r t i c l e
i n f o
Article history: Received 8 September 2011 Accepted 9 November 2011 Available online 18 November 2011 Keywords: Bioconjugate DNA Gold nanoparticle Soft interface Synchrotron radiation SAXS
a b s t r a c t Hybridization of DNA tethered on colloidal nanoparticles with fully matched complementary one induces the aggregation of the particles in a non-cross-linking configuration. Here, we performed a structural study on DNA-functionalized gold nanoparticle and its non-cross-linking aggregation mainly using synchrotron radiation small-angle X-ray scattering. To understand the non-cross-linking aggregation, the nanoparticles with various DNA lengths and core sizes were used. In the aggregation, the surface distance between the gold nanoparticles increased with the length of DNA duplex, although the increment of the distance per base pair was not constant and showed the tendency to become small with increasing DNA length, meaning the interdigitation of DNA layers. The aggregation was also found to occur between the identical cores, without being affected by tethered DNA. Furthermore, it was proved that the relative increase in DNA length to core size leads to the increase in colloidal stability. Even the nanoparticles with full-matched DNA duplex were dispersed stably. These facts suggested that van der Waals interaction between core particles rather than end-to-end stacking between DNA duplexes is a dominant attractive interaction. The steric repulsion force arising from entropic loss of thermal fluctuation of DNA molecules might be a key factor to characterize the non-cross-linking aggregation. Ó 2011 Elsevier Inc. All rights reserved.
1. Introduction Colloidal materials are fundamental for new functional materials and widely used in several fields. The change in optical property of the colloidal system, resulting from, e.g., colloidal dispersion/aggregation, is useful for application as a sensor. Biomolecule-functionalized particle system is a representative example of the application to colorimetric diagnostics and prognostics. The use of gold nanoparticle as a core particle is advantageous for those sensors because the surface plasmon shift by dispersion/aggregation is visible. For example, DNA/gold nanoparticle materials have been drawn much attention toward a colorimetric detection method of single-base mutation such as single-nucleotide polymorphisms. Mirkin and coworkers first reported a DNA sensing system using DNA/gold nanoparticles [1,2]. Two sets of single-stranded DNA (ssDNA)-functionalized gold nanoparticles assemble in the presence of a complementary target DNA, which hybridizes to both tethered ssDNA molecules on the gold nanoparticles, acting as a cross-linker. Using this system, the discrimination between fully ⇑ Corresponding author. Fax: +81 48 462 4658. E-mail address: [email protected] (M. Fujita). 0021-9797/$ - see front matter Ó 2011 Elsevier Inc. All rights reserved. doi:10.1016/j.jcis.2011.11.021
complementary and mismatched targets is visible at an appropriate temperature because the dissociation temperatures of the duplexes are different. The cross-linking aggregation of DNA/gold nanoparticles assisted by hybridization has been widely applied to other detection methods such as metal ions and small molecules [3–5]. This technique has been extended to the methodology of programmable nanoassemblies directed by DNA [6,7]. Meanwhile, it was discovered that DNA-functionalized nanoparticles assemble by hybridization in a non-cross-linking configuration. When ssDNA molecules grafted on the nanoparticles hybridize to the fully complementary ones, the resulting nanoparticles, which are covered with double-stranded DNA (dsDNA), immediately assemble without molecular cross-linking at high salt concentrations, while the nanoparticles with ssDNA molecules remain stable [8–10]. This phenomenon called non-cross-linking aggregation occurs very rapidly. More remarkable point is that the dispersion/aggregation of the particles is very sensitive to the terminal base pair of DNA duplex. When the tethered ssDNA hybridizes with a single-base mismatched DNA, the nanoparticles still disperse stably at the high salt concentrations. Accordingly, the discrimination between fully complementary and mismatched targets is easy in short time without temperature control. The non-cross-linking
630
M. Fujita et al. / Journal of Colloid and Interface Science 368 (2012) 629–635
aggregation was observed in poly(N-isopropylacrylamide)-DNA [8], gold nanoparticle-DNA [9], and polystyrene nanoparticle-DNA [10]. Therefore, the non-cross-linking aggregation system is also potentially useful for single-base mutation detection. Furthermore, this system has been applied to detections of metal ions, and small molecules, using the nanoparticles covered with functional nucleic acids such as DNA aptamer [11,12]. However, the nature of the non-cross-linking aggregation interaction has not been understood yet. The hybridization of ssDNA with its complementary one might bring about the change in electrostatic, steric repulsion potential, or both. On the other hand, the possibility of a specific interaction between the terminal base pairs of DNA duplexes, end-to-end stacking [13–17], has been supposed in order to explain the feature of the non-cross-linking aggregation interaction. Such interparticle interactions must be reflected in the configuration of the assembling particles. No structural information on DNA-functionalized nanoparticle and its assemble has been, however, provided so far. In this study, the non-cross-linking aggregation was explored from the structural point of view, using mainly small-angle X-ray scattering (SAXS). The structural characterization of the gold nanoparticles with ss- or dsDNA was performed, and the effects of DNA length, core size, and temperature on the non-cross-linking aggregation/dispersion were also investigated. 2. Experimental 2.1. Materials Gold nanoparticles with diameters of ca. 5, 15, and 40 nm (Au5, Au15, and Au40, respectively, hereafter) were purchased from British-Biocell, UK. The gold nanoparticles were functionalized with ssDNA according to the method reported in Ref. [9]. In this study, the DNAs with seven different lengths (15, 20, 25, 30, 35, 40, and 45 mer) (Operon Biotechnologies, Japan) were used. The ssDNA has a C6 thiol linker at the 50 -end. The sequences of probe DNAs are summarized in Table 1. The functionalized gold nanoparticles were suspended in 10 mM phosphate buffer (PB) (pH 7.0) containing 0.1 M NaCl. The graft number of probe DNA on gold nanoparticle was evaluated according to the procedure reported in Ref. [9]. Unless otherwise noted, the concentration of the gold nanoparticle was prepared at around 1 OD at 520 nm (5.0 1010 particles/lL for Au5, 1.4 109 particles/lL for Au15 and 9.0 107 particles/lL for Au40). For hybridization of DNA, the complementary DNA (Operon Biotechnologies, Japan) was added to the suspension (final concentration of 0.5 lM) and, for aggregation experiments, the NaCl was further added to it (final concentration of 1 M). 2.2. Characterization The dynamic light scattering (DLS) measurements were conducted on a Zetasizer Nano ZS ZEN3600 with a He–Ne laser (633 nm) (Malvern Instrument Limited, UK). According to cumulants analysis, z-averaged size and variance (hydrodynamic Table 1 Sequences of probe DNA used in this study. Length (bases)
Sequence (50 –30 )
15 20 25 30 35 40 45
TAC GCC ACC AGC TCC TAC TCT ACG CCA CCA GCT CC TAC TCC TTA TTA CGC CAC CAG CTC C TAC TCC TTA TTC TTT TAC GCC ACC AGC TCC TAC TTT TCT CCT TAT TCT TTT ACG CCA CCA GCT CC TAC TTT TCT TTT CTC CTT ATT CTT TTA CGC CAC CAG CTC C TAC TTT TCT TTT TTC TAC TCC TTA TTC TTT TAC GCC ACC AGC TCC
radius (Rh), and polydispersity index (PDI)) of the nanoparticles dispersed in 10 mM PB (pH 7.0) containing 0.1 M NaClat 25 °C were evaluated. UV–vis spectra were measured with a Cary 50 Bio UV–Visible Spectrometer (VARIAN, Inc., CA, USA) in order to confirm the dispersion/aggregation of the nanoparticles at 25 °C. The UV–vis spectra at various temperatures were taken using a UV2550 UV–Vis spectrophotometer (Shimadzu Corporation, Japan). Circular dichroism (CD) spectra were measured with a JASCO J-720WI Spectropolarimeter (JASCO Corporation, Japan) at 25 °C. In CD, the spectra of hybridized DNA (5 lM) without gold nanoparticles (10 mM PB containing 0.1 M or 1.0 M NaCl) were also observed. For gold nanoparticles with dsDNA, the samples concentrated to more than 8 times (10 mM PB containing 0.1 M or 1.0 M NaCl, target DNA: 4–5 lM) were used for CD measurements.
2.3. Solution small-angle X-ray scattering Solution SAXS measurements were carried out at the BL45XU RIKEN Structural Biology Beamline I (wavelength, k = 0.09 nm) of the SPring-8, Harima, Japan [18]. The camera length was set to be about 2.5 m and calibrated using a silver behenate. Two-dimensional (2D) SAXS images were recorded with a CCD camera (C4880-1014A, Hamamatsu Photonics, Japan) coupled with an X-ray image intensifier (V5445P MOD, Hamamatsu Photonics, Japan). The pixel size of CCD camera was about 0.15 mm 0.15 mm. The 2-D SAXS images were converted into one-dimensional profiles by circular averaging. The sample solution (ca. 50 lL) was placed in a sample cell and thermostatted at 25 °C except for temperature-controlled experiments. For aggregation experiments, the SAXS images were acquired about 3 min after the complementary DNA and then NaCl was added to the suspension.
3. Results and discussion In this study, the gold nanoparticles with the diameters of 15 and 40 nm were mainly used as core particles, and their DNA-functionalized gold particles were first characterized. The values of DNA graft density C evaluated for the samples subjected to SAXS experiments are summarized in Table 2. The graft density was dependent on the core size as reported previously [19]and evaluated at about 0.25 strands/nm2 for Au15 and 0.15 strands/nm2 for Au 40, respectively, although the values somewhat varied with the length of DNA. The hydrodynamic radii of the DNA-functionalized gold particles measured by DLS are given in Table 3. The values of Rh strongly depended on DNA length. There was little difference of Rh between ss- and dsDNA-functionalized particles. This is likely because Rh results from the slow motion of particle diffusion and the difference in rapid molecular motion of DNA between single- and double-stranded structures might little affect the overall Rh. Although some particles showed somewhat high
Table 2 Graft density of probe DNA. Length (bases)
15 20 25 30 35 40 45
C (strands/nm2) Au15
Au40
0.25 0.17 0.21 0.20 0.27 0.17 0.13
0.14 0.10 0.18 0.11 0.17 0.12 0.054
631
M. Fujita et al. / Journal of Colloid and Interface Science 368 (2012) 629–635 Table 3 The radii of gold nanoparticles without and with ss- and dsDNA. Sample type
a
Au 15 hRi (nm)
No DNA 15 mer 20 mer 25 mer 30 mer 35 mer 40 mer 45 mer a b c
ss ds ss ds ss ds ss ds ss ds ss ds ss ds
7.46 7.70 7.72 7.51 7.51 7.48 7.53 7.59 7.66 7.45 7.52 7.28 7.35 7.44 7.64
Au 40 b
r (nm) 0.60 0.45 0.47 0.49 0.60 0.49 0.60 0.60 0.50 0.65 0.60 0.64 0.48 0.70 0.49
b
Rh (nm)
c
9.15 15.0 14.9 14.1 14.9 14.9 15.3 16.7 17.2 16.7 18.0 16.3 17.1 20.3 20.5
PDI
c
hRi (nm)
0.02 0.16 0.13 0.09 0.07 0.06 0.07 0.18 0.14 0.05 0.06 0.24 0.23 0.15 0.09
18.0 18.0 18.1 18.1 18.1 18.2 18.0 17.7 17.7 18.2 18.1 17.6 17.8 18.1 18.3
b
r (nm) b
Rh (nm)
2.0 2.0 2.0 2.0 2.0 2.0 2.0 2.0 2.0 2.0 2.0 2.0 2.0 2.0 2.0
19.5 24.8 25.0 26.7 26.3 26.7 26.2 25.5 26.8 28.2 30.0 34.1 35.5 30.4 31.5
c
PDI
c
0.11 0.23 0.23 0.14 0.14 0.11 0.12 0.14 0.13 0.10 0.09 0.20 0.20 0.12 0.09
ss and ds indicate single- and double-stranded DNA, respectively. hRi and r are the average radius and the standard deviation, evaluated by theoretical fit of SAXS data. Rh and PDI are the hydrodynamic radius and the polydisperse index by DLS.
PDIs, non-cross-linking aggregation occurred even for such particles, as shown later. Solution SAXS measurements for dispersed colloidal solutions of bare gold, gold nanoparticles with ssDNA, and dsDNA were performed. Fig. 1 shows the circularly averaged experimental data from the bare Au15, Au15 functionalized with 15 mer ssDNA, and Au15 with 15 mer dsDNA (Au15-ss15 and Au15-ds15), represented with the circular symbols. The DNA-functionalized gold nanoparticles suspended in 10 mM PB (pH 7.0) containing 0.1 M NaCl. From the SAXS profiles, the structural parameters of such colloidal nanoparticles were derived. Generally, the scattering intensity I(q) for a polydisperse solution of particle is given by:
IðqÞ ¼ n
Z
DðRÞPðq; RÞdRSðqÞ
ð1Þ
where q (=4p sin h/k, 2h: scattering angle) is the scattering vector, nis the number density of particle, R is the radius of particle, D(R) is the size distribution function, P(q, R) is the form factor of particle, and S(q) is the structure factor [20]. If the particle is a hard sphere of radius R, the form factor is expressed by:
Fig. 1. SAXS profiles from bare Au15, Au15-ss15, and Au15-ds15at 25 °C. The gold nanoparticles with DNA dispersed in 10 mM PB (pH 7.0) containing 0.1 M NaCl. The solid lines are the theoretical fitting curves.
2
2
Pðq; RÞ ¼ m ðDqÞ
" #2 3ðsinðqRÞ qR cosðqRÞÞ ðqRÞ3
ð2Þ
where v is the volume of particle, and Dq is the difference in electron density between the particle and the solvent. In this study, the size distribution was assumed to be a Gaussian distribution expressed by:
( ) 1 ðR hRiÞ2 DðRÞ ¼ pffiffiffiffiffiffiffi exp 2r2 2pr
ð3Þ
where r is the standard deviation, hRi is the mean radius of particle. If the interparticle interference is neglected, S(q) can be regarded as unity. The hydrodynamic radius Rh of the nanoparticle increased with increasing DNA length as shown in Table 3 so that a core–shell model was expected as a structural model. However, all the obtained data for DNA-functionalized particles were well fitted according to the above-mentioned hard sphere model as shown in Fig. 1 (the solid lines). This is because the observed intensities are dominated by the core scattering, due to higher scattering power of gold. Actually, from this theoretical curve fitting, the mean radii of Au15, Au15-ss15, and Au15-ds15 were evaluated at 7.46 ± 0.60, 7.70 ± 0.45, 7.72 ± 0.47 nm, respectively. Similarly, the structural parameters of other nanoparticles were obtained by the curve fitting. The values of hRi and r are summarized in Table 3. The colloidal nanoparticles covered with full-matched dsDNA are less stable against salt [9]. The scattering intensities from Au15-ds15, -ds30, and -ds45 in the presence of 1 M NaCl are shown in Fig. 2. Intense peaks due to interparticle interference effect were clearly observed in the profiles. The structure factor S(q) was derived by dividing the observed scattering data from the assemblies (1 M NaCl) by that from dispersion state (0.1 M NaCl). The obtained structure factors for the particles (Au15 and Au40) with 15 mer, 30 mer, and 45 mer dsDNA are shown in Fig. 3. The interference peaks are recognized at qx/ q1 1, 1.7, etc., which are characteristic of a crystalline structural order [21,22]. But, the peaks are broad because of less long-range ordering. This feature may be attributable to the structure factor of less ordered face-centered cubic (FCC) or possibly random hexagonal structure [7]. For such less ordered colloid crystal, a structural analysis by paracrystalline theory have been proposed [23,24]. In this study, this method was applied to analyze the structure factor, assumed a FCC configuration with a disorder.
632
M. Fujita et al. / Journal of Colloid and Interface Science 368 (2012) 629–635
1 qaðsin h sin / þ cos hÞ 2 1 q a2 ¼ qað sin h cos / þ cos hÞ 2 1 q a3 ¼ qað sin h cos / þ sin h sin /Þ 2 q a1 ¼
Fig. 2. SAXS profiles from Au15-ds15, -ds30, and -ds45 in 10 mM PB (pH 7.0) containing 1 M NaCl at 25 °C.
Fig. 3. Structure factors: (a) Au15-ds15, -ds30, and -ds45, (b) Au40-ds15, -ds30, and -ds45. The solid lines indicate the theoretical structural factors.
ð8Þ ð9Þ ð10Þ
The thermal oscillation of the lattice points and the size effect of the paracrystal were also considered [23,24]. The solid lines in Fig. 3 are the fitting curves calculated by the paracrystal model. For Au15-ds15, this theoretical fitting yielded the lattice dimension p a of 34.7 nm, so that the nearest neighbor distance (a/ 2 for FCC) between the particles was calculated at 24.5 nm. Accordingly, the distance between the surfaces of core gold particles, the intersurface distance, for Au15-ds15 was evaluated as 9.6 nm. For other assemblies with different DNAs and core size, the values were obtained similarly. The inter-surface distances against the number of base (Nbp) are plotted in Fig. 4a. As shown in this figure, the intersurface distance was found to increase with increasing length of tethered DNA. To consider the relation between the length of DNA and the inter-surface distance, it is necessary to obtain structural information of DNA molecules on the gold nanoparticles. The SAXS analysis provided only the core size, while DLS analysis yielded the hydrodynamic radius. Unfortunately, the structural information of DNA layer, especially in the aggregate, could not be obtained directly. Here, we examined the conformation of dsDNA on gold nanoparticle using CD [25]. Fig. 5 demonstrates the CD spectra for free 45 mer dsDNA (without gold nanoparticle) and for tethered 45 mer dsDNA on the surface of gold nanoparticle. The dashed and solid lines represent the spectra taken at 0.1 M and 1 M NaCl, respectively. Similar to the case of the free duplex, the spectrum of the tethered dsDNA in the dispersion state showed a negative band at around 250 nm and a positive band at around 280 nm. Even at 1 M NaCl, the similar spectrum was observed although the signal was weak probably due to the precipitation followed by aggregation. The conformation of dsDNA tethered on gold nanoparticle is most likely in B-form (0.34 nm/bp). Considering the length of C6 linker of DNA (1.0 nm) [26], the total thickness of DNA layer on the surface of gold nanoparticle was thus estimated. When the tethered dsDNA orients perpendicularly to the surface of gold particle, the layer thickness is given by (Nbp 1) 0.34 + 1.0 nm. In the case of end-to-end stacking between the terminal base pairs of duplexes, the inter-surface distance is expected to be twice
According to Refs. [23,24], for the ideal perfect lattice of a threedimensional cubic lattice system, three fundamental vectors ak (k = 1, 2, 3), which represent the directions of the nearest neighbor points, are defined. If the lattice distortions from the ideal lattice points are independent each other and isotropic, expressed by a Gaussian function with a standard deviation Da, the lattice factor Z(q) is given by:
ZðqÞ ¼ P3k¼1 Z k ðqÞ
ð4Þ 1 j F k j
Z k ðqÞ ¼
2
1 2 j F k j cosðak qÞþ j F k j2 o 1 Da2 n j F k ðqÞ j¼ exp ða1 qÞ2 þ ða2 qÞ2 þ ða3 qÞ2 2 2 a
ð5Þ ð6Þ
where a is a lattice dimension. For randomly oriented paracrystals, the lattice factor Z(q), which corresponds to S(q) of our case, is calculated by:
ZðqÞ ¼
1 4p
Z 2p 0
d/
Z p
dhZ 1 ðq; h; /ÞZ 2 ðq; h; /ÞZ 3 ðq; h; /Þ sin h
ð7Þ
0
Considering a FCC configuration with a lattice dimension of a, the scalar products ak q in Eqs. (5) and (6) are given by:
Fig. 4. The inter-surface distances against the number of base (a) and their differential values (b) for Au15 (s, d) and Au40 (h, j).
M. Fujita et al. / Journal of Colloid and Interface Science 368 (2012) 629–635
Fig. 5. CD spectra of free (a) and tethered 45 mer dsDNA (b) in 10 mM PB (pH 7.0) containing 0.1 M (dashed) or 1 M NaCl (solid) at 25 °C.
the layer thickness. But any inter-surface distances showing in Fig. 4a were shorter. For the cross-linking configuration of the DNA-functionalized gold nanoparticles, for example, the rise of inter-surface distance per one base pair of DNA linker was found to be constant at 0.25 nm/bp, which is shorter than 0.34 nm/bp, for oligo-DNA up to 72 bp [27]. In our case, actual thickness of DNA layer might be also smaller because of tilting of tethered DNA molecules [26] and the others. Thus, the increment per one base pair was evaluated. The differential values with respect to Nbp are plotted in Fig. 4b. Note that Nbp is the number of base of one tethered DNA (e.g., the rise per one base pair should be 0.68 (=0.34 2) nm/bp in the case of end-to-end stacking for duplexes oriented perpendicularly to the core surface.) As can be seen in this figure, the rise per base pair was not constant and rather showed the tendency to become small with increasing Nbp. It appears that the degree of overlap depends on the graft density of DNA. This is probably the result of overlap between DNA layers, because oligo-DNA duplex used here is considered to be a rigid rod. To further clarify the nature of non-cross-linking aggregation, we investigated the aggregation of binary mixtures of DNA-functionalized gold nanoparticles with two different particle sizes. First, we examined the binary mixture of the two different cores functionalized with identical DNA sequence, Au15-ds15, and Au40-ds15, which overall radii are 13.5 and 23.9 nm, respectively. Fig. 6a shows the structure factor derived from the SAXS data of this mixed system. Two interference peaks at q = 0.18 and 0.32 nm1 were clearly observed in this system. For comparison, the structure factors obtained for the individual components are also given in this figure. The structure factor of the mixed system superposes those of the single components. If the non-cross-linking aggregation is induced by a specific interaction between the DNA duplexes, end-to-end stacking, the nanoparticles would assemble irrespective of core size. In practice, however, the aggregation of mixed nanoparticles was found to occur separately between the particles with the same core. For a longer DNA, two sets of Au15-ds45 and Au40-ds45, the particles also assembled separately (data not shown). Next, another two sets of the same core size with different dsDNAs (Au15-ds15 and Au15-ds45) were examined. As similar to the above mixed system, the overall radii of the nanoparticles are widely different from each other (13.5 nm for Au15-ds15 and 23.6 nm for Au15-ds45). In sharp contrast with Fig. 6a, a major interference peak appeared at around q = 0.26 nm1 between those peaks from the single components (q = 0.20 nm1 for Au15-ds45
633
Fig. 6. Structure factors of binary mixtures: (a) Au15-ds15 and Au40-ds15, and (b) Au15-ds15 and Au15-ds45. In each part, the structure factors for individual components are also shown.
and 0.32 nm1 for Au15-ds15) as shown in Fig. 6b. This indicates that the two components can assemble miscible with each other, regardless of DNA length, if the core particles are identical. The inter-surface distance derived from the curve fitting by paracrystalline model was calculated at 15.1 nm, which was intermediate between the inter-surface distances in each single system. For a bigger core, the mixed system of Au40-ds15 and Au40-ds45, the two components were also found to be miscible with each other in the aggregation (data not shown). The above experiments of binary mixtures clarified that the assembling between the DNA-functionalized particles having identical cores is thermodynamically preferable. It is considered that van der Waals potential between core particles is responsible for non-cross-linking aggregation as an attractive interaction. On the other hand, the end-to-end stacking interaction between DNA duplexes seems not to be involved. As reported previously, the endto-end stacking interaction of DNA duplexes might not occur in the presence of monovalent cation such as sodium ion [15,16]. It is rather reasonable to consider the repulsive interactions due to electrostatic potential and steric stabilization ascribed to the characteristics of DNA molecules in order to understand the non-crosslinking aggregation. The electrostatic potential arising from the negatively charged phosphate groups of DNA contributes strongly to the colloidal stability of the particle. This repulsion force reduces by adding a salt because of the suppression of electric double layer. Compared to the nanoparticles covered with dsDNA, the nanoparticles with ssDNA disperse rather stable against salt although the number of phosphate groups on the particle is smaller than the case of dsDNA. In practice, it is considered that the effective charge densities of ssand dsDNA are almost the same each other due to compensation by counter-cation condensation. According to the counterion condensation theory by Manning, the numbers of associated countercations per phosphate group for ssDNA and dsDNA are estimated as 0.44 and 0.76, respectively [28]. It is, of course, no wonder that there is little difference in the negative charge densities between fully matched and one base mismatched duplexes. Accordingly, it is considered that there is little difference in the electrostatic potential energy among the DNA structures. At the high salt concentration of 1 M NaCl, where the aggregation experiments of the nanoparticles with dsDNA were performed, the electric double layer is considerably suppressed (Debye length j1 0.3 nm at 25 °C) so that the contribution of electrostatic potential can be neglected.
634
M. Fujita et al. / Journal of Colloid and Interface Science 368 (2012) 629–635
Fig. 8. (a) UV–vis spectra of Au5-ss15 (dashed) and -ds15 (solid) in 10 mM PB (pH 7.0) containing 1 M NaCl. The spectra were obtained every 5 °C on cooling (25 °C to 5 °C). For Au5-ds15, heating process (5 °C to 25 °C) was also carried out. (b) Structure factors obtained from SAXS data of Au5-ds15 in 10 mM PB (pH 7.0) containing 1 M NaCl. The SAXS data were obtained every 5 °C on cooling.
Fig. 7. Plasmon absorbance peak of DNA-functionalized gold nanoparticles as a function of temperature for three core particles which are covered with ds15 (a), ds30 (b), and ds45 (c). The UV–vis spectra from the particles in 10 mM PB (pH 7.0) containing 1 M NaCl were obtained every 5 °C on heating.
The effect of steric stabilization is another important factor to determine the difference in colloidal stability between fullmatched dsDNA and the others (ssDNA and terminal mismatched duplex). Aggregation experiments using a smaller gold nanoparticle would serve in order to make the effect of the steric stabilization clearer, because the van der Waals potential between the core particles should reduce. Here, DNA-functionalized gold nanoparticles with a smaller core diameter of ca. 5 nm (Au5) were also examined. Fig. 7 shows the change in absorbance peak of UV–vis spectra with core size and temperature for ds15, -30, and -45 in the presence of 1 M NaCl. DNA dehybridization upon heating was expected to lead to the redispersion of the assembled DNA-functionalized nanoparticles. For example, the peak shift to shorter wavelengths observed at 55–60 °C for Au40-ds15 is the result of DNA dehybridization because the melting temperature of 15 mer DNA duplex was evaluated at 50 °C by UV–vis. However, for smaller core particles (Au5 and Au15) with the same DNA, the peak shift was found to occur at lower temperatures, meaning that the particles are redispersed without DNA melting. As shown in Fig. 8a,the nanoparticles of Au5-ds15 dispersed stably even at 25 °C in the presence of 1 M NaCl. This was also recognized by the experimental evidence that the corresponding structure factor S(q) in Fig. 8b shows no interference peak. Below 25 °C, the surface plasmon shift to longer wavelength in UV–vis spectra and the interference peak in S(q) (q = 0.47 nm1)for Au5ds15 were gradually observed, indicating the aggregation of the nanoparticles. The inter-surface distance at 5 °C was evaluated at 10.1 nm, which was close to those for Au15- and Au40-ds15. It should be noted that the nanoparticles of Au5-ss15 still disperse stably at the temperatures in the presence of 1 M NaCl. Fig. 7 also shows that even for nanoparticles composed of bigger cores, the nanoparticles covered with longer dsDNA can disassemble without DNA melting. As demonstrated in the case of Au40-ds45, the disassembly occurred at a lower temperature than its expected melting point of 45 mer dsDNA. This behavior may be caused by the increase in steric repulsion with increasing DNA length.
The steric overlap repulsion arising from the avoidance of crowding between neighboring molecules [12,29] can be estimated for unhybridized DNA [30], using the theoretical approach by Milner [31,32]. This overlap repulsion energy depends on the height of brush (layer thickness), graft density, diameter, and stiffness of DNA molecule [30–32]. Applied to the case of hybridization, which corresponds to increase in effective diameter and persistence length of DNA molecule, this computation predicted that the gold nanoparticle with dsDNA is more stable. However, this is contrary to the fact that the particle with fully matched dsDNA is less stable than that with ssDNA. Other entropic factors arising from the thermal fluctuation of DNA molecules or layer might be more important [12,29].When the particles approach each other, the entropic loss due to the limitation of molecular motion of DNA likely occurs [29]. The flexible ssDNA might be more mobile than the stiff dsDNA. The mobility of distal end of single-base mismatched dsDNA is expected to be higher than that of full-matched dsDNA. The hybridization of DNA tethered on the particle with full-matched DNA might result in a decrease in such thermal fluctuation, leading to reduction in colloidal stability of DNA-functionalized nanoparticle. In order to further understand of the mechanism, additional detailed studies will be performed. 4. Concluding remarks In this study, we explored the DNA-functionalized gold nanoparticles and the non-cross-linking aggregation mainly using synchrotron radiation SAXS and succeeded in obtaining the structural information. A series of SAXS measurements were performed to investigate the effects of DNA length and core size on non-crosslinking aggregation. The aggregate of the nanoparticles has a crystalline structural order. On the basis of paracrystalline theory, the surface distance between the neighboring particles was evaluated. The surface distance was found to increase with increasing dsDNA length. From the experiments using the binary mixtures, it was unveiled that the DNA-functionalized particles with the same core size can assemble together without being affected by tethered DNA. Furthermore, it was demonstrated that for smaller core sizes or longer DNA lengths, their nanoparticles covered with full-matched dsDNA can disperse stably. This may be caused by steric stabilization of DNA molecules. It was thus considered that the attractive interaction in the non-cross-linking aggregation is
M. Fujita et al. / Journal of Colloid and Interface Science 368 (2012) 629–635
attributable to van der Waals potential between core particles, and the contribution of the end-to-end stacking attraction between dsDNA might be little. The feature of non-cross-linking aggregation interaction is likely characterized by steric repulsion stabilization due to entropic loss in conformation and mobility of DNA molecules. Acknowledgments This work has been supported by Grant-in-Aid for Scientific Research on Innovative Areas, ‘‘Molecular Soft-Interface Science’’, from the Ministry of Education, Culture, Sports, Science and Technology of Japan. We thank N. Kobayashi for assistance in the CD measurements. References [1] C.A. Mirkin, R.L. Letsinger, R.C. Mucic, J.J. Storhoff, Nature 382 (1996) 607. [2] R. Elghanian, J.J. Storhoff, R.C. Mucic, R.L. Letsinger, C.A. Mirkin, Science 277 (1997) 1078. [3] J. Liu, Y. Lu, J. Am. Chem. Soc. 127 (2005) 12677. [4] J.S. Lee, M.S. Han, C.A. Mirkin, Angew. Chem. Int. Ed. 46 (2007) 4093. [5] J. Liu, Y. Lu, Angew. Chem. Int. Ed. 45 (2006) 90. [6] D. Nykypanchuk, M.M. Maye, D. van der Lelie, O. Gang, Nature 451 (2008) 549. [7] S.Y. Park, A.K.R. Lytton-Jean, B. Lee, S. Weigand, G.C. Schatz, C.A. Mirkin, Nature 451 (2008) 553. [8] T. Mori, M. Maeda, Polym J. 34 (2002) 624. [9] K. Sato, K. Hosokawa, M. Maeda, J. Am. Chem. Soc. 125 (2003) 8102. [10] K. Sato, M. Sawayanagi, K. Hosokawa, M. Maeda, Anal. Sci. 20 (2004) 893. [11] D. Miyamoto, Z. Tang, T. Takarada, M. Maeda, Chem. Commun. (2007) 4743.
635
[12] W. Zhao, W. Chiuman, J.C.F. Lam, S.A. McManus, W. Chen, Y. Cui, R. Pelton, M.A. Brook, Y. Li, J. Am. Chem. Soc. 130 (2008) 3610. [13] H. Yan, T.H. LaBean, L. Feng, J.H. Reif, Proc. Natl. Acad. Sci. USA 100 (2003) 8103. [14] M. Nakata, G. Zanchetta, B.D. Chapman, C.D. Jones, J.O. Cross, R. Pindak, T. Bellini, N.A. Clark, Science 318 (2007) 1276. [15] X. Qui, K. Andresen, L.W. Kwok, J.S. Lamb, H.Y. Park, L. Pollack, Phys. Rev. Lett. 99 (2007) 038104. [16] L. Li, S.A. Pabit, J.S. Lamb, H.Y. Park, L. Pollack, Appl. Phys. Lett. 92 (2008) 223901. [17] Y. Sato, K. Hosokawa, M. Maeda, Colloids Surf. B62 (2008) 71. [18] T. Fujisawa, K. Inoue, T. Oka, H. Iwamoto, T. Uruga, T. Kumasaka, Y. Inoko, N. Yagi, M. Yamamoto, T. Ueki, J. Appl. Crystallogr. 33 (2000) 797. [19] H.D. Hill, J.E. Millstone, M.J. Banholzer, C.A. Mirkin, ACS Nano 3 (2009) 418. [20] L.A. Feigin, D.I. Svergun, Structure Analysis by Small-Angle X-Ray and Neutron Scattering, Plenum Press, New York, 1987. [21] B. Abécassis, F. Testard, O. Spalla, Phys. Rev. Lett. 100 (2008) 115504. [22] H. Guo, T. Narayanan, M. Sztuchi, P. Schall, G.H. Wedgdam, Phys Rev. Lett. 100 (2008) 188303. [23] H. Matsuoka, H. Tanaka, T. Hashimoto, N. Ise, Phys. Rev. B36 (1987) 1754. [24] H. Matsuoka, H. Tanaka, N. Iizuka, T. Hashimoto, N. Ise, Phys. Rev. B41 (1990) 3854. [25] R. Jin, G. Wu, Z. Li, C.A. Mirkin, G.C. Schatz, J. Am. Chem. Soc. 125 (2003) 1643. [26] R. Levicky, T.M. Herene, M.J. Tarlov, S.K. Satija, J. Am. Chem. Soc. 120 (1998) 9787. [27] S.J. Park, A.A. Lazarides, J.J. Storhoff, L. Pesce, C.A. Mirkin, J. Phys. Chem. B108 (2004) 12375. [28] G.S. Manning, Q. Rev. Biophys. 11 (1978) 179. [29] J.N. Israelachivili, Intermolecular and Surface Forces, second ed., Academic Press Ltd., London, 1992. [30] D. Nykypanchuk, M.M. Maye, D. van der Lelie, O. Gang, Langmuir 23 (2007) 6305. [31] S.T. Milner, Europhys. Lett. 7 (1988) 695. [32] S.T. Milner, Science 251 (1991) 905.
Contents lists available at SciVerse ScienceDirect
Journal of Colloid and Interface Science www.elsevier.com/locate/jcis
Structural study on gold nanoparticle functionalized with DNA and its non-cross-linking aggregation Masahiro Fujita a,⇑, Yoshizumi Katafuchi b, Kazuki Ito c, Naoki Kanayama a, Tohru Takarada a, Mizuo Maeda a a b c
Bioengineering Laboratory, RIKEN, Hirosawa 2-1, Wako, Saitama 351-0198, Japan Faculty of Industrial Science and Technology, Tokyo University of Science, Yamazaki 2641, Noda, Chiba 278-8510, Japan Structural Materials Science Laboratory, RIKEN SPring-8 Center, Kouto 1-1-1, Sayo, Hyogo 679-5148, Japan
a r t i c l e
i n f o
Article history: Received 8 September 2011 Accepted 9 November 2011 Available online 18 November 2011 Keywords: Bioconjugate DNA Gold nanoparticle Soft interface Synchrotron radiation SAXS
a b s t r a c t Hybridization of DNA tethered on colloidal nanoparticles with fully matched complementary one induces the aggregation of the particles in a non-cross-linking configuration. Here, we performed a structural study on DNA-functionalized gold nanoparticle and its non-cross-linking aggregation mainly using synchrotron radiation small-angle X-ray scattering. To understand the non-cross-linking aggregation, the nanoparticles with various DNA lengths and core sizes were used. In the aggregation, the surface distance between the gold nanoparticles increased with the length of DNA duplex, although the increment of the distance per base pair was not constant and showed the tendency to become small with increasing DNA length, meaning the interdigitation of DNA layers. The aggregation was also found to occur between the identical cores, without being affected by tethered DNA. Furthermore, it was proved that the relative increase in DNA length to core size leads to the increase in colloidal stability. Even the nanoparticles with full-matched DNA duplex were dispersed stably. These facts suggested that van der Waals interaction between core particles rather than end-to-end stacking between DNA duplexes is a dominant attractive interaction. The steric repulsion force arising from entropic loss of thermal fluctuation of DNA molecules might be a key factor to characterize the non-cross-linking aggregation. Ó 2011 Elsevier Inc. All rights reserved.
1. Introduction Colloidal materials are fundamental for new functional materials and widely used in several fields. The change in optical property of the colloidal system, resulting from, e.g., colloidal dispersion/aggregation, is useful for application as a sensor. Biomolecule-functionalized particle system is a representative example of the application to colorimetric diagnostics and prognostics. The use of gold nanoparticle as a core particle is advantageous for those sensors because the surface plasmon shift by dispersion/aggregation is visible. For example, DNA/gold nanoparticle materials have been drawn much attention toward a colorimetric detection method of single-base mutation such as single-nucleotide polymorphisms. Mirkin and coworkers first reported a DNA sensing system using DNA/gold nanoparticles [1,2]. Two sets of single-stranded DNA (ssDNA)-functionalized gold nanoparticles assemble in the presence of a complementary target DNA, which hybridizes to both tethered ssDNA molecules on the gold nanoparticles, acting as a cross-linker. Using this system, the discrimination between fully ⇑ Corresponding author. Fax: +81 48 462 4658. E-mail address: [email protected] (M. Fujita). 0021-9797/$ - see front matter Ó 2011 Elsevier Inc. All rights reserved. doi:10.1016/j.jcis.2011.11.021
complementary and mismatched targets is visible at an appropriate temperature because the dissociation temperatures of the duplexes are different. The cross-linking aggregation of DNA/gold nanoparticles assisted by hybridization has been widely applied to other detection methods such as metal ions and small molecules [3–5]. This technique has been extended to the methodology of programmable nanoassemblies directed by DNA [6,7]. Meanwhile, it was discovered that DNA-functionalized nanoparticles assemble by hybridization in a non-cross-linking configuration. When ssDNA molecules grafted on the nanoparticles hybridize to the fully complementary ones, the resulting nanoparticles, which are covered with double-stranded DNA (dsDNA), immediately assemble without molecular cross-linking at high salt concentrations, while the nanoparticles with ssDNA molecules remain stable [8–10]. This phenomenon called non-cross-linking aggregation occurs very rapidly. More remarkable point is that the dispersion/aggregation of the particles is very sensitive to the terminal base pair of DNA duplex. When the tethered ssDNA hybridizes with a single-base mismatched DNA, the nanoparticles still disperse stably at the high salt concentrations. Accordingly, the discrimination between fully complementary and mismatched targets is easy in short time without temperature control. The non-cross-linking
630
M. Fujita et al. / Journal of Colloid and Interface Science 368 (2012) 629–635
aggregation was observed in poly(N-isopropylacrylamide)-DNA [8], gold nanoparticle-DNA [9], and polystyrene nanoparticle-DNA [10]. Therefore, the non-cross-linking aggregation system is also potentially useful for single-base mutation detection. Furthermore, this system has been applied to detections of metal ions, and small molecules, using the nanoparticles covered with functional nucleic acids such as DNA aptamer [11,12]. However, the nature of the non-cross-linking aggregation interaction has not been understood yet. The hybridization of ssDNA with its complementary one might bring about the change in electrostatic, steric repulsion potential, or both. On the other hand, the possibility of a specific interaction between the terminal base pairs of DNA duplexes, end-to-end stacking [13–17], has been supposed in order to explain the feature of the non-cross-linking aggregation interaction. Such interparticle interactions must be reflected in the configuration of the assembling particles. No structural information on DNA-functionalized nanoparticle and its assemble has been, however, provided so far. In this study, the non-cross-linking aggregation was explored from the structural point of view, using mainly small-angle X-ray scattering (SAXS). The structural characterization of the gold nanoparticles with ss- or dsDNA was performed, and the effects of DNA length, core size, and temperature on the non-cross-linking aggregation/dispersion were also investigated. 2. Experimental 2.1. Materials Gold nanoparticles with diameters of ca. 5, 15, and 40 nm (Au5, Au15, and Au40, respectively, hereafter) were purchased from British-Biocell, UK. The gold nanoparticles were functionalized with ssDNA according to the method reported in Ref. [9]. In this study, the DNAs with seven different lengths (15, 20, 25, 30, 35, 40, and 45 mer) (Operon Biotechnologies, Japan) were used. The ssDNA has a C6 thiol linker at the 50 -end. The sequences of probe DNAs are summarized in Table 1. The functionalized gold nanoparticles were suspended in 10 mM phosphate buffer (PB) (pH 7.0) containing 0.1 M NaCl. The graft number of probe DNA on gold nanoparticle was evaluated according to the procedure reported in Ref. [9]. Unless otherwise noted, the concentration of the gold nanoparticle was prepared at around 1 OD at 520 nm (5.0 1010 particles/lL for Au5, 1.4 109 particles/lL for Au15 and 9.0 107 particles/lL for Au40). For hybridization of DNA, the complementary DNA (Operon Biotechnologies, Japan) was added to the suspension (final concentration of 0.5 lM) and, for aggregation experiments, the NaCl was further added to it (final concentration of 1 M). 2.2. Characterization The dynamic light scattering (DLS) measurements were conducted on a Zetasizer Nano ZS ZEN3600 with a He–Ne laser (633 nm) (Malvern Instrument Limited, UK). According to cumulants analysis, z-averaged size and variance (hydrodynamic Table 1 Sequences of probe DNA used in this study. Length (bases)
Sequence (50 –30 )
15 20 25 30 35 40 45
TAC GCC ACC AGC TCC TAC TCT ACG CCA CCA GCT CC TAC TCC TTA TTA CGC CAC CAG CTC C TAC TCC TTA TTC TTT TAC GCC ACC AGC TCC TAC TTT TCT CCT TAT TCT TTT ACG CCA CCA GCT CC TAC TTT TCT TTT CTC CTT ATT CTT TTA CGC CAC CAG CTC C TAC TTT TCT TTT TTC TAC TCC TTA TTC TTT TAC GCC ACC AGC TCC
radius (Rh), and polydispersity index (PDI)) of the nanoparticles dispersed in 10 mM PB (pH 7.0) containing 0.1 M NaClat 25 °C were evaluated. UV–vis spectra were measured with a Cary 50 Bio UV–Visible Spectrometer (VARIAN, Inc., CA, USA) in order to confirm the dispersion/aggregation of the nanoparticles at 25 °C. The UV–vis spectra at various temperatures were taken using a UV2550 UV–Vis spectrophotometer (Shimadzu Corporation, Japan). Circular dichroism (CD) spectra were measured with a JASCO J-720WI Spectropolarimeter (JASCO Corporation, Japan) at 25 °C. In CD, the spectra of hybridized DNA (5 lM) without gold nanoparticles (10 mM PB containing 0.1 M or 1.0 M NaCl) were also observed. For gold nanoparticles with dsDNA, the samples concentrated to more than 8 times (10 mM PB containing 0.1 M or 1.0 M NaCl, target DNA: 4–5 lM) were used for CD measurements.
2.3. Solution small-angle X-ray scattering Solution SAXS measurements were carried out at the BL45XU RIKEN Structural Biology Beamline I (wavelength, k = 0.09 nm) of the SPring-8, Harima, Japan [18]. The camera length was set to be about 2.5 m and calibrated using a silver behenate. Two-dimensional (2D) SAXS images were recorded with a CCD camera (C4880-1014A, Hamamatsu Photonics, Japan) coupled with an X-ray image intensifier (V5445P MOD, Hamamatsu Photonics, Japan). The pixel size of CCD camera was about 0.15 mm 0.15 mm. The 2-D SAXS images were converted into one-dimensional profiles by circular averaging. The sample solution (ca. 50 lL) was placed in a sample cell and thermostatted at 25 °C except for temperature-controlled experiments. For aggregation experiments, the SAXS images were acquired about 3 min after the complementary DNA and then NaCl was added to the suspension.
3. Results and discussion In this study, the gold nanoparticles with the diameters of 15 and 40 nm were mainly used as core particles, and their DNA-functionalized gold particles were first characterized. The values of DNA graft density C evaluated for the samples subjected to SAXS experiments are summarized in Table 2. The graft density was dependent on the core size as reported previously [19]and evaluated at about 0.25 strands/nm2 for Au15 and 0.15 strands/nm2 for Au 40, respectively, although the values somewhat varied with the length of DNA. The hydrodynamic radii of the DNA-functionalized gold particles measured by DLS are given in Table 3. The values of Rh strongly depended on DNA length. There was little difference of Rh between ss- and dsDNA-functionalized particles. This is likely because Rh results from the slow motion of particle diffusion and the difference in rapid molecular motion of DNA between single- and double-stranded structures might little affect the overall Rh. Although some particles showed somewhat high
Table 2 Graft density of probe DNA. Length (bases)
15 20 25 30 35 40 45
C (strands/nm2) Au15
Au40
0.25 0.17 0.21 0.20 0.27 0.17 0.13
0.14 0.10 0.18 0.11 0.17 0.12 0.054
631
M. Fujita et al. / Journal of Colloid and Interface Science 368 (2012) 629–635 Table 3 The radii of gold nanoparticles without and with ss- and dsDNA. Sample type
a
Au 15 hRi (nm)
No DNA 15 mer 20 mer 25 mer 30 mer 35 mer 40 mer 45 mer a b c
ss ds ss ds ss ds ss ds ss ds ss ds ss ds
7.46 7.70 7.72 7.51 7.51 7.48 7.53 7.59 7.66 7.45 7.52 7.28 7.35 7.44 7.64
Au 40 b
r (nm) 0.60 0.45 0.47 0.49 0.60 0.49 0.60 0.60 0.50 0.65 0.60 0.64 0.48 0.70 0.49
b
Rh (nm)
c
9.15 15.0 14.9 14.1 14.9 14.9 15.3 16.7 17.2 16.7 18.0 16.3 17.1 20.3 20.5
PDI
c
hRi (nm)
0.02 0.16 0.13 0.09 0.07 0.06 0.07 0.18 0.14 0.05 0.06 0.24 0.23 0.15 0.09
18.0 18.0 18.1 18.1 18.1 18.2 18.0 17.7 17.7 18.2 18.1 17.6 17.8 18.1 18.3
b
r (nm) b
Rh (nm)
2.0 2.0 2.0 2.0 2.0 2.0 2.0 2.0 2.0 2.0 2.0 2.0 2.0 2.0 2.0
19.5 24.8 25.0 26.7 26.3 26.7 26.2 25.5 26.8 28.2 30.0 34.1 35.5 30.4 31.5
c
PDI
c
0.11 0.23 0.23 0.14 0.14 0.11 0.12 0.14 0.13 0.10 0.09 0.20 0.20 0.12 0.09
ss and ds indicate single- and double-stranded DNA, respectively. hRi and r are the average radius and the standard deviation, evaluated by theoretical fit of SAXS data. Rh and PDI are the hydrodynamic radius and the polydisperse index by DLS.
PDIs, non-cross-linking aggregation occurred even for such particles, as shown later. Solution SAXS measurements for dispersed colloidal solutions of bare gold, gold nanoparticles with ssDNA, and dsDNA were performed. Fig. 1 shows the circularly averaged experimental data from the bare Au15, Au15 functionalized with 15 mer ssDNA, and Au15 with 15 mer dsDNA (Au15-ss15 and Au15-ds15), represented with the circular symbols. The DNA-functionalized gold nanoparticles suspended in 10 mM PB (pH 7.0) containing 0.1 M NaCl. From the SAXS profiles, the structural parameters of such colloidal nanoparticles were derived. Generally, the scattering intensity I(q) for a polydisperse solution of particle is given by:
IðqÞ ¼ n
Z
DðRÞPðq; RÞdRSðqÞ
ð1Þ
where q (=4p sin h/k, 2h: scattering angle) is the scattering vector, nis the number density of particle, R is the radius of particle, D(R) is the size distribution function, P(q, R) is the form factor of particle, and S(q) is the structure factor [20]. If the particle is a hard sphere of radius R, the form factor is expressed by:
Fig. 1. SAXS profiles from bare Au15, Au15-ss15, and Au15-ds15at 25 °C. The gold nanoparticles with DNA dispersed in 10 mM PB (pH 7.0) containing 0.1 M NaCl. The solid lines are the theoretical fitting curves.
2
2
Pðq; RÞ ¼ m ðDqÞ
" #2 3ðsinðqRÞ qR cosðqRÞÞ ðqRÞ3
ð2Þ
where v is the volume of particle, and Dq is the difference in electron density between the particle and the solvent. In this study, the size distribution was assumed to be a Gaussian distribution expressed by:
( ) 1 ðR hRiÞ2 DðRÞ ¼ pffiffiffiffiffiffiffi exp 2r2 2pr
ð3Þ
where r is the standard deviation, hRi is the mean radius of particle. If the interparticle interference is neglected, S(q) can be regarded as unity. The hydrodynamic radius Rh of the nanoparticle increased with increasing DNA length as shown in Table 3 so that a core–shell model was expected as a structural model. However, all the obtained data for DNA-functionalized particles were well fitted according to the above-mentioned hard sphere model as shown in Fig. 1 (the solid lines). This is because the observed intensities are dominated by the core scattering, due to higher scattering power of gold. Actually, from this theoretical curve fitting, the mean radii of Au15, Au15-ss15, and Au15-ds15 were evaluated at 7.46 ± 0.60, 7.70 ± 0.45, 7.72 ± 0.47 nm, respectively. Similarly, the structural parameters of other nanoparticles were obtained by the curve fitting. The values of hRi and r are summarized in Table 3. The colloidal nanoparticles covered with full-matched dsDNA are less stable against salt [9]. The scattering intensities from Au15-ds15, -ds30, and -ds45 in the presence of 1 M NaCl are shown in Fig. 2. Intense peaks due to interparticle interference effect were clearly observed in the profiles. The structure factor S(q) was derived by dividing the observed scattering data from the assemblies (1 M NaCl) by that from dispersion state (0.1 M NaCl). The obtained structure factors for the particles (Au15 and Au40) with 15 mer, 30 mer, and 45 mer dsDNA are shown in Fig. 3. The interference peaks are recognized at qx/ q1 1, 1.7, etc., which are characteristic of a crystalline structural order [21,22]. But, the peaks are broad because of less long-range ordering. This feature may be attributable to the structure factor of less ordered face-centered cubic (FCC) or possibly random hexagonal structure [7]. For such less ordered colloid crystal, a structural analysis by paracrystalline theory have been proposed [23,24]. In this study, this method was applied to analyze the structure factor, assumed a FCC configuration with a disorder.
632
M. Fujita et al. / Journal of Colloid and Interface Science 368 (2012) 629–635
1 qaðsin h sin / þ cos hÞ 2 1 q a2 ¼ qað sin h cos / þ cos hÞ 2 1 q a3 ¼ qað sin h cos / þ sin h sin /Þ 2 q a1 ¼
Fig. 2. SAXS profiles from Au15-ds15, -ds30, and -ds45 in 10 mM PB (pH 7.0) containing 1 M NaCl at 25 °C.
Fig. 3. Structure factors: (a) Au15-ds15, -ds30, and -ds45, (b) Au40-ds15, -ds30, and -ds45. The solid lines indicate the theoretical structural factors.
ð8Þ ð9Þ ð10Þ
The thermal oscillation of the lattice points and the size effect of the paracrystal were also considered [23,24]. The solid lines in Fig. 3 are the fitting curves calculated by the paracrystal model. For Au15-ds15, this theoretical fitting yielded the lattice dimension p a of 34.7 nm, so that the nearest neighbor distance (a/ 2 for FCC) between the particles was calculated at 24.5 nm. Accordingly, the distance between the surfaces of core gold particles, the intersurface distance, for Au15-ds15 was evaluated as 9.6 nm. For other assemblies with different DNAs and core size, the values were obtained similarly. The inter-surface distances against the number of base (Nbp) are plotted in Fig. 4a. As shown in this figure, the intersurface distance was found to increase with increasing length of tethered DNA. To consider the relation between the length of DNA and the inter-surface distance, it is necessary to obtain structural information of DNA molecules on the gold nanoparticles. The SAXS analysis provided only the core size, while DLS analysis yielded the hydrodynamic radius. Unfortunately, the structural information of DNA layer, especially in the aggregate, could not be obtained directly. Here, we examined the conformation of dsDNA on gold nanoparticle using CD [25]. Fig. 5 demonstrates the CD spectra for free 45 mer dsDNA (without gold nanoparticle) and for tethered 45 mer dsDNA on the surface of gold nanoparticle. The dashed and solid lines represent the spectra taken at 0.1 M and 1 M NaCl, respectively. Similar to the case of the free duplex, the spectrum of the tethered dsDNA in the dispersion state showed a negative band at around 250 nm and a positive band at around 280 nm. Even at 1 M NaCl, the similar spectrum was observed although the signal was weak probably due to the precipitation followed by aggregation. The conformation of dsDNA tethered on gold nanoparticle is most likely in B-form (0.34 nm/bp). Considering the length of C6 linker of DNA (1.0 nm) [26], the total thickness of DNA layer on the surface of gold nanoparticle was thus estimated. When the tethered dsDNA orients perpendicularly to the surface of gold particle, the layer thickness is given by (Nbp 1) 0.34 + 1.0 nm. In the case of end-to-end stacking between the terminal base pairs of duplexes, the inter-surface distance is expected to be twice
According to Refs. [23,24], for the ideal perfect lattice of a threedimensional cubic lattice system, three fundamental vectors ak (k = 1, 2, 3), which represent the directions of the nearest neighbor points, are defined. If the lattice distortions from the ideal lattice points are independent each other and isotropic, expressed by a Gaussian function with a standard deviation Da, the lattice factor Z(q) is given by:
ZðqÞ ¼ P3k¼1 Z k ðqÞ
ð4Þ 1 j F k j
Z k ðqÞ ¼
2
1 2 j F k j cosðak qÞþ j F k j2 o 1 Da2 n j F k ðqÞ j¼ exp ða1 qÞ2 þ ða2 qÞ2 þ ða3 qÞ2 2 2 a
ð5Þ ð6Þ
where a is a lattice dimension. For randomly oriented paracrystals, the lattice factor Z(q), which corresponds to S(q) of our case, is calculated by:
ZðqÞ ¼
1 4p
Z 2p 0
d/
Z p
dhZ 1 ðq; h; /ÞZ 2 ðq; h; /ÞZ 3 ðq; h; /Þ sin h
ð7Þ
0
Considering a FCC configuration with a lattice dimension of a, the scalar products ak q in Eqs. (5) and (6) are given by:
Fig. 4. The inter-surface distances against the number of base (a) and their differential values (b) for Au15 (s, d) and Au40 (h, j).
M. Fujita et al. / Journal of Colloid and Interface Science 368 (2012) 629–635
Fig. 5. CD spectra of free (a) and tethered 45 mer dsDNA (b) in 10 mM PB (pH 7.0) containing 0.1 M (dashed) or 1 M NaCl (solid) at 25 °C.
the layer thickness. But any inter-surface distances showing in Fig. 4a were shorter. For the cross-linking configuration of the DNA-functionalized gold nanoparticles, for example, the rise of inter-surface distance per one base pair of DNA linker was found to be constant at 0.25 nm/bp, which is shorter than 0.34 nm/bp, for oligo-DNA up to 72 bp [27]. In our case, actual thickness of DNA layer might be also smaller because of tilting of tethered DNA molecules [26] and the others. Thus, the increment per one base pair was evaluated. The differential values with respect to Nbp are plotted in Fig. 4b. Note that Nbp is the number of base of one tethered DNA (e.g., the rise per one base pair should be 0.68 (=0.34 2) nm/bp in the case of end-to-end stacking for duplexes oriented perpendicularly to the core surface.) As can be seen in this figure, the rise per base pair was not constant and rather showed the tendency to become small with increasing Nbp. It appears that the degree of overlap depends on the graft density of DNA. This is probably the result of overlap between DNA layers, because oligo-DNA duplex used here is considered to be a rigid rod. To further clarify the nature of non-cross-linking aggregation, we investigated the aggregation of binary mixtures of DNA-functionalized gold nanoparticles with two different particle sizes. First, we examined the binary mixture of the two different cores functionalized with identical DNA sequence, Au15-ds15, and Au40-ds15, which overall radii are 13.5 and 23.9 nm, respectively. Fig. 6a shows the structure factor derived from the SAXS data of this mixed system. Two interference peaks at q = 0.18 and 0.32 nm1 were clearly observed in this system. For comparison, the structure factors obtained for the individual components are also given in this figure. The structure factor of the mixed system superposes those of the single components. If the non-cross-linking aggregation is induced by a specific interaction between the DNA duplexes, end-to-end stacking, the nanoparticles would assemble irrespective of core size. In practice, however, the aggregation of mixed nanoparticles was found to occur separately between the particles with the same core. For a longer DNA, two sets of Au15-ds45 and Au40-ds45, the particles also assembled separately (data not shown). Next, another two sets of the same core size with different dsDNAs (Au15-ds15 and Au15-ds45) were examined. As similar to the above mixed system, the overall radii of the nanoparticles are widely different from each other (13.5 nm for Au15-ds15 and 23.6 nm for Au15-ds45). In sharp contrast with Fig. 6a, a major interference peak appeared at around q = 0.26 nm1 between those peaks from the single components (q = 0.20 nm1 for Au15-ds45
633
Fig. 6. Structure factors of binary mixtures: (a) Au15-ds15 and Au40-ds15, and (b) Au15-ds15 and Au15-ds45. In each part, the structure factors for individual components are also shown.
and 0.32 nm1 for Au15-ds15) as shown in Fig. 6b. This indicates that the two components can assemble miscible with each other, regardless of DNA length, if the core particles are identical. The inter-surface distance derived from the curve fitting by paracrystalline model was calculated at 15.1 nm, which was intermediate between the inter-surface distances in each single system. For a bigger core, the mixed system of Au40-ds15 and Au40-ds45, the two components were also found to be miscible with each other in the aggregation (data not shown). The above experiments of binary mixtures clarified that the assembling between the DNA-functionalized particles having identical cores is thermodynamically preferable. It is considered that van der Waals potential between core particles is responsible for non-cross-linking aggregation as an attractive interaction. On the other hand, the end-to-end stacking interaction between DNA duplexes seems not to be involved. As reported previously, the endto-end stacking interaction of DNA duplexes might not occur in the presence of monovalent cation such as sodium ion [15,16]. It is rather reasonable to consider the repulsive interactions due to electrostatic potential and steric stabilization ascribed to the characteristics of DNA molecules in order to understand the non-crosslinking aggregation. The electrostatic potential arising from the negatively charged phosphate groups of DNA contributes strongly to the colloidal stability of the particle. This repulsion force reduces by adding a salt because of the suppression of electric double layer. Compared to the nanoparticles covered with dsDNA, the nanoparticles with ssDNA disperse rather stable against salt although the number of phosphate groups on the particle is smaller than the case of dsDNA. In practice, it is considered that the effective charge densities of ssand dsDNA are almost the same each other due to compensation by counter-cation condensation. According to the counterion condensation theory by Manning, the numbers of associated countercations per phosphate group for ssDNA and dsDNA are estimated as 0.44 and 0.76, respectively [28]. It is, of course, no wonder that there is little difference in the negative charge densities between fully matched and one base mismatched duplexes. Accordingly, it is considered that there is little difference in the electrostatic potential energy among the DNA structures. At the high salt concentration of 1 M NaCl, where the aggregation experiments of the nanoparticles with dsDNA were performed, the electric double layer is considerably suppressed (Debye length j1 0.3 nm at 25 °C) so that the contribution of electrostatic potential can be neglected.
634
M. Fujita et al. / Journal of Colloid and Interface Science 368 (2012) 629–635
Fig. 8. (a) UV–vis spectra of Au5-ss15 (dashed) and -ds15 (solid) in 10 mM PB (pH 7.0) containing 1 M NaCl. The spectra were obtained every 5 °C on cooling (25 °C to 5 °C). For Au5-ds15, heating process (5 °C to 25 °C) was also carried out. (b) Structure factors obtained from SAXS data of Au5-ds15 in 10 mM PB (pH 7.0) containing 1 M NaCl. The SAXS data were obtained every 5 °C on cooling.
Fig. 7. Plasmon absorbance peak of DNA-functionalized gold nanoparticles as a function of temperature for three core particles which are covered with ds15 (a), ds30 (b), and ds45 (c). The UV–vis spectra from the particles in 10 mM PB (pH 7.0) containing 1 M NaCl were obtained every 5 °C on heating.
The effect of steric stabilization is another important factor to determine the difference in colloidal stability between fullmatched dsDNA and the others (ssDNA and terminal mismatched duplex). Aggregation experiments using a smaller gold nanoparticle would serve in order to make the effect of the steric stabilization clearer, because the van der Waals potential between the core particles should reduce. Here, DNA-functionalized gold nanoparticles with a smaller core diameter of ca. 5 nm (Au5) were also examined. Fig. 7 shows the change in absorbance peak of UV–vis spectra with core size and temperature for ds15, -30, and -45 in the presence of 1 M NaCl. DNA dehybridization upon heating was expected to lead to the redispersion of the assembled DNA-functionalized nanoparticles. For example, the peak shift to shorter wavelengths observed at 55–60 °C for Au40-ds15 is the result of DNA dehybridization because the melting temperature of 15 mer DNA duplex was evaluated at 50 °C by UV–vis. However, for smaller core particles (Au5 and Au15) with the same DNA, the peak shift was found to occur at lower temperatures, meaning that the particles are redispersed without DNA melting. As shown in Fig. 8a,the nanoparticles of Au5-ds15 dispersed stably even at 25 °C in the presence of 1 M NaCl. This was also recognized by the experimental evidence that the corresponding structure factor S(q) in Fig. 8b shows no interference peak. Below 25 °C, the surface plasmon shift to longer wavelength in UV–vis spectra and the interference peak in S(q) (q = 0.47 nm1)for Au5ds15 were gradually observed, indicating the aggregation of the nanoparticles. The inter-surface distance at 5 °C was evaluated at 10.1 nm, which was close to those for Au15- and Au40-ds15. It should be noted that the nanoparticles of Au5-ss15 still disperse stably at the temperatures in the presence of 1 M NaCl. Fig. 7 also shows that even for nanoparticles composed of bigger cores, the nanoparticles covered with longer dsDNA can disassemble without DNA melting. As demonstrated in the case of Au40-ds45, the disassembly occurred at a lower temperature than its expected melting point of 45 mer dsDNA. This behavior may be caused by the increase in steric repulsion with increasing DNA length.
The steric overlap repulsion arising from the avoidance of crowding between neighboring molecules [12,29] can be estimated for unhybridized DNA [30], using the theoretical approach by Milner [31,32]. This overlap repulsion energy depends on the height of brush (layer thickness), graft density, diameter, and stiffness of DNA molecule [30–32]. Applied to the case of hybridization, which corresponds to increase in effective diameter and persistence length of DNA molecule, this computation predicted that the gold nanoparticle with dsDNA is more stable. However, this is contrary to the fact that the particle with fully matched dsDNA is less stable than that with ssDNA. Other entropic factors arising from the thermal fluctuation of DNA molecules or layer might be more important [12,29].When the particles approach each other, the entropic loss due to the limitation of molecular motion of DNA likely occurs [29]. The flexible ssDNA might be more mobile than the stiff dsDNA. The mobility of distal end of single-base mismatched dsDNA is expected to be higher than that of full-matched dsDNA. The hybridization of DNA tethered on the particle with full-matched DNA might result in a decrease in such thermal fluctuation, leading to reduction in colloidal stability of DNA-functionalized nanoparticle. In order to further understand of the mechanism, additional detailed studies will be performed. 4. Concluding remarks In this study, we explored the DNA-functionalized gold nanoparticles and the non-cross-linking aggregation mainly using synchrotron radiation SAXS and succeeded in obtaining the structural information. A series of SAXS measurements were performed to investigate the effects of DNA length and core size on non-crosslinking aggregation. The aggregate of the nanoparticles has a crystalline structural order. On the basis of paracrystalline theory, the surface distance between the neighboring particles was evaluated. The surface distance was found to increase with increasing dsDNA length. From the experiments using the binary mixtures, it was unveiled that the DNA-functionalized particles with the same core size can assemble together without being affected by tethered DNA. Furthermore, it was demonstrated that for smaller core sizes or longer DNA lengths, their nanoparticles covered with full-matched dsDNA can disperse stably. This may be caused by steric stabilization of DNA molecules. It was thus considered that the attractive interaction in the non-cross-linking aggregation is
M. Fujita et al. / Journal of Colloid and Interface Science 368 (2012) 629–635
attributable to van der Waals potential between core particles, and the contribution of the end-to-end stacking attraction between dsDNA might be little. The feature of non-cross-linking aggregation interaction is likely characterized by steric repulsion stabilization due to entropic loss in conformation and mobility of DNA molecules. Acknowledgments This work has been supported by Grant-in-Aid for Scientific Research on Innovative Areas, ‘‘Molecular Soft-Interface Science’’, from the Ministry of Education, Culture, Sports, Science and Technology of Japan. We thank N. Kobayashi for assistance in the CD measurements. References [1] C.A. Mirkin, R.L. Letsinger, R.C. Mucic, J.J. Storhoff, Nature 382 (1996) 607. [2] R. Elghanian, J.J. Storhoff, R.C. Mucic, R.L. Letsinger, C.A. Mirkin, Science 277 (1997) 1078. [3] J. Liu, Y. Lu, J. Am. Chem. Soc. 127 (2005) 12677. [4] J.S. Lee, M.S. Han, C.A. Mirkin, Angew. Chem. Int. Ed. 46 (2007) 4093. [5] J. Liu, Y. Lu, Angew. Chem. Int. Ed. 45 (2006) 90. [6] D. Nykypanchuk, M.M. Maye, D. van der Lelie, O. Gang, Nature 451 (2008) 549. [7] S.Y. Park, A.K.R. Lytton-Jean, B. Lee, S. Weigand, G.C. Schatz, C.A. Mirkin, Nature 451 (2008) 553. [8] T. Mori, M. Maeda, Polym J. 34 (2002) 624. [9] K. Sato, K. Hosokawa, M. Maeda, J. Am. Chem. Soc. 125 (2003) 8102. [10] K. Sato, M. Sawayanagi, K. Hosokawa, M. Maeda, Anal. Sci. 20 (2004) 893. [11] D. Miyamoto, Z. Tang, T. Takarada, M. Maeda, Chem. Commun. (2007) 4743.
635
[12] W. Zhao, W. Chiuman, J.C.F. Lam, S.A. McManus, W. Chen, Y. Cui, R. Pelton, M.A. Brook, Y. Li, J. Am. Chem. Soc. 130 (2008) 3610. [13] H. Yan, T.H. LaBean, L. Feng, J.H. Reif, Proc. Natl. Acad. Sci. USA 100 (2003) 8103. [14] M. Nakata, G. Zanchetta, B.D. Chapman, C.D. Jones, J.O. Cross, R. Pindak, T. Bellini, N.A. Clark, Science 318 (2007) 1276. [15] X. Qui, K. Andresen, L.W. Kwok, J.S. Lamb, H.Y. Park, L. Pollack, Phys. Rev. Lett. 99 (2007) 038104. [16] L. Li, S.A. Pabit, J.S. Lamb, H.Y. Park, L. Pollack, Appl. Phys. Lett. 92 (2008) 223901. [17] Y. Sato, K. Hosokawa, M. Maeda, Colloids Surf. B62 (2008) 71. [18] T. Fujisawa, K. Inoue, T. Oka, H. Iwamoto, T. Uruga, T. Kumasaka, Y. Inoko, N. Yagi, M. Yamamoto, T. Ueki, J. Appl. Crystallogr. 33 (2000) 797. [19] H.D. Hill, J.E. Millstone, M.J. Banholzer, C.A. Mirkin, ACS Nano 3 (2009) 418. [20] L.A. Feigin, D.I. Svergun, Structure Analysis by Small-Angle X-Ray and Neutron Scattering, Plenum Press, New York, 1987. [21] B. Abécassis, F. Testard, O. Spalla, Phys. Rev. Lett. 100 (2008) 115504. [22] H. Guo, T. Narayanan, M. Sztuchi, P. Schall, G.H. Wedgdam, Phys Rev. Lett. 100 (2008) 188303. [23] H. Matsuoka, H. Tanaka, T. Hashimoto, N. Ise, Phys. Rev. B36 (1987) 1754. [24] H. Matsuoka, H. Tanaka, N. Iizuka, T. Hashimoto, N. Ise, Phys. Rev. B41 (1990) 3854. [25] R. Jin, G. Wu, Z. Li, C.A. Mirkin, G.C. Schatz, J. Am. Chem. Soc. 125 (2003) 1643. [26] R. Levicky, T.M. Herene, M.J. Tarlov, S.K. Satija, J. Am. Chem. Soc. 120 (1998) 9787. [27] S.J. Park, A.A. Lazarides, J.J. Storhoff, L. Pesce, C.A. Mirkin, J. Phys. Chem. B108 (2004) 12375. [28] G.S. Manning, Q. Rev. Biophys. 11 (1978) 179. [29] J.N. Israelachivili, Intermolecular and Surface Forces, second ed., Academic Press Ltd., London, 1992. [30] D. Nykypanchuk, M.M. Maye, D. van der Lelie, O. Gang, Langmuir 23 (2007) 6305. [31] S.T. Milner, Europhys. Lett. 7 (1988) 695. [32] S.T. Milner, Science 251 (1991) 905.