The Reversible Phase Transition of DNA-Linked Colloidal Gold Assemblies

Abstract

We present direct evidence for a reversible phase transition of DNA-linked colloidal gold assemblies. Transmission electron microscopy and optical absorption spectroscopy are used to monitor the colloidal gold phase transition, whose behavior is dominated by DNA interactions. We use single-stranded DNA-capped colloidal gold that is linked by complementary linker DNA to form the assemblies. We found that, compared to free DNA, a sharp melting transition is observed for the DNA-linked colloidal gold assemblies. The structure of the assemblies is non-crystalline, much like a gel phase, consistent with theoretical predictions. Optical spectra and melting curves provide additional evidence of gelation of the colloidal system. The phase transition and separation are examples of percolation in a dilute solvent.

Full text

arXiv:physics/0504061v2 [physics.bio-ph] 21 Apr 2005
Physica A (2005), to appear.
The Reversible Phase Transition of
DNA-Linked Colloidal Gold Assemblies
Young Sun, Nolan C. Harris, and Ching-Hwa Kiang∗
Department of Physics and Astronomy
Rice University, Houston, TX 77005–1892
Abstract
We present direct evidence for a reversible phase transition of DNA-linked col-
loidal gold assemblies. Transmission electron microscopy and optical absorption
spectroscopy are used to monitor the colloidal gold phase transition, whose behav-
ior is dominated by DNA interactions. We use single-stranded DNA-capped colloidal
gold that is linked by complementary linker DNA to form the assemblies. We found
that, compared to free DNA, a sharp melting transition is observed for the DNA-
linked colloidal gold assemblies. The structure of the assemblies is non-crystalline,
much like a gel phase, consistent with theoretical predictions. Optical spectra and
melting curves provide additional evidence of gelation of the colloidal system. The
phase transition and separation are examples of percolation in a dilute solvent.
Key words: DNA phase transition, gold nanoparticle, DNA melting
PACS: 82.39.Pj, 87.15.By, 87.15.He, 87.68.+z, 87.15.-v
1 Introduction
Colloidal nanoparticles functionalized and linked with single-stranded DNA
exemplify a new class of complex fluids. Both equilibrium and nonequilibrium
phase transitions of complex particle systems are of great interest [1,2]. The
interaction between colloidal particles is controlled by the DNA intermolecular
potentials, which are dominated by hydrogen bonding. Owing to the complex
yet specific nature of DNA base pairing, the interaction between colloids in
such a system can be precisely controlled and “tailor made” to have a specific
potential. Both the number of components and the strength of the interaction
forces can be designed. A variety of different states of this system have been
investigated theoretically [3,4].
Preprint submitted to Elsevier Science 1 July 2018

DNA melting and hybridization, an important process in DNA replication
and translation, have been studied for decades [5,6,7,8,9,10,11,12]. Many of
the thermodynamic properties of free DNA are known, yet DNA interactions
in constrained spaces such as on surfaces, e.g. in microarrays, are still poorly
understood [13]. The sequence-specific hybridization properties of DNA have
been used for self-assembly of nanostructures [14,15]. The macroscopic prop-
erties of these novel systems can be easily detected and are a result of the
microscopic properties of DNA. Thus, studies of the phase behavior of these
self-assemblies provide valuable information on fundamental DNA properties
[3,16,17].
DNA-linked gold colloids were thought to self-assemble into crystalline struc-
tures [14]; however, it was proposed by Kiang [18] that the structure of the
assemblies is amorphous, much like a percolating cluster. Calculated optical
spectra are consistent with an amorphous structure [16]. The phase behav-
ior of this system has also been investigated theoretically, and it is believed
that the system undergoes a liquid-liquid phase separation, with the dense
liquid phase behaving as a solid (amorphous) gel [3]. More experimental stud-
ies of the structural phase transition are needed to unravel the true nature of
the phase transition, which is crucial in understanding the behavior of such
multi-component complex fluids.
Direct imaging is a powerful tool to study the structures arising during the
colloidal phase transition [19], and the results can be interpreted unambigu-
ously. In this paper, we present evidence of a liquid-gel phase transition of
DNA-linked gold colloid assemblies via direct imaging. The aggregation and
phase transition of the DNA-linked gold colloids were studied using transmis-
sion electron microscopy (TEM) [20] and optical absorption spectroscopy. The
results obtained suggest that gelation and phase separation occurs at room
temperature in the DNA-linked gold colloid system.
2 Results and Discussion
To prepare colloidal systems with different interactions, we start with a range
of sizes of gold nanoparticles (10 to 40 nm, <10% polydispersity) capped with
either 3’ or 5’ modified, single-stranded, 12-base DNA. The basic building
block is illustrated in Fig. 1. Details of the sample preparation are described
in Ref. [18]. We added linker DNA, which caused nanoparticles to form aggre-
gates, and we studied the melting transition with microscopy and spectroscopy.
We propose the growth mechanism of DNA-linked gold colloids (illustrated
in Fig. 2b) to be as follows. Particles initially are dissolved in the solution.
2

Fig. 1. The basic building block of DNA-linked gold particles.
With the addition of complementary linker DNA, hybridization occurs, and
the particles form a gel-like structure. The connectivity of the porous struc-
ture continues to increase past the percolation threshold, and eventually the
clusters become a dense amorphous structure. Phase separation occurs and
the gel-like aggregates eventually precipitate out of the solution.
Direct imaging using transmission electron microscopy (TEM) supports this
model. Fig. 2a displays a series of TEM images taken at different aggrega-
tion stages. Before adding linker DNA, the DNA-capped gold particles are
dispersed in the solution. Upon adding linker DNA, the colloidal gold forms
aggregates. The process is reversed by raising the temperature to above the
transition temperature.
UV-visible absorption spectroscopy is a powerful tool to study the aggregation
and phase transition of the DNA-linked gold colloids because DNA bases and
gold colloids have strong absorption in the UV region (∼260 nm) and the
visible light region (∼520 nm), respectively. The extinction coefficient as
well as the peak position are sensitive to the size of the aggregates of gold
colloids. The DNA double helix has a smaller extinction coefficient than does
Fig. 2. (a) Sequence of transmission electron microscopy images of 10 nm gold
colloids capped with thiol-modified DNA. The colloids are initially dispersed in so-
lution, upon adding linker DNA, gold colloids form gel-like porous and amorphous
aggregates. Phase separation occurs, and eventually the dense aggregates precipi-
tate out of the solution. (b) The growth mechanism of DNA-linked gold colloids.
The colloids are initially dispersed in solution. Upon adding linker DNA, gold col-
loids form gel-like porous and amorphous aggregates. Phase separation occurs, and
eventually the dense aggregates precipitate out of the solution.
3

30 40 50 60 70
Temperature (°C)
0
0.5
1
Normalized Extinction (@520 nm)
10 nm
20 nm
40 nm
30 40 50 60 70
0
0.5
1
Normalized Extinction (@260 nm)
10 nm
20 nm
40 nm
free DNA
Fig. 3. Melting curves for AB nanoparticle systems monitored at (a) 260 nm and
(b) 520 nm. The free DNA melting curve monitored at 260 nm is shown in (a).
single-stranded DNA due to hypocromism [21]. Thus, both the kinetics of
colloidal aggregation and the temperature-dependent melting transition can
be investigated using UV-visible spectroscopy. All spectra were taken on a
PerkinElmer Lambda 45 spectrophotometer.
Upon adding linker DNA, gold nanoparticles begin to aggregate, as indicated
by the change of the UV absorption. The aggregation rate at room temperature
is faster for systems with higher Tm. Solutions of DNA-linked gold colloids were
allowed to stand at room temperature for several days for the system to fully
aggregate before melting studies. We monitor the absorption intensity at 260
and 520 nm while slowly heating the DNA-linked gold colloids. The sample
was heated by a peltier temperature controller from 25 to 75 ◦C at a rate of 0.5
◦C/min. The 260 and 520 nm melting curves are very similar, indicating that
DNA and gold colloid melting are closely related. Fig. 3 shows the melting
curves of 10, 20, and 40 nm gold colloids with linker DNA. For comparison,
the melting transition of a free DNA is also shown. Apparently, the melting
transition of gold-attached DNA is much narrower than that of free DNA.
The melting temperature Tm, defined as the temperature at the midpoint of
the absorbency transition [22], is found to be a function of particle diame-
ter, D, and the data are fitted to Tm(◦C) = 68 −57/√D. The surface cover-
age of thiol-capped DNA bound to gold nanoparticles has been determined
via the fluorescence method to be approximately 160 DNA (12-mer) bound
for a 16 nm diameter particle [23]. Assuming the number of DNA bound to
gold particles scales with the particle surface area, D2, there are 50, 190, and
750 DNA on 10 nm, 20 nm, and 40 nm particles, respectively. An increasing
number of connections between particles as the gold particle size increases is
expected. Since the hydrogen bonding energy per DNA pair remain the same,
an increased number of connections effectively increases the enthalpy, ∆H,
between particles and, therefore, raises the melting temperature, Tm[16].
For short DNA (12–14 base pairs), melting and hybridization can be described
4

by a two-state model as an equilibrium between single- and double-stranded
DNA [22],
S+S⇀
↽D.
The melting curve is a slowly varying function of temperature and can be
described by the van’t Hoff relationship [21]. While short, free DNA does not
exhibit a phase transition, DNA bound to nanoparticles aggregate to form
networks that have a definite phase transition, and the melting curves cannot
be described with the van’t Hoff relationship.
The DNA-linked gold nanoparticle assemblies have several unusual features,
including a sharp melting transition compared to that of corresponding free
DNA, and the melting temperature, Tm, dependence on colloid size [18,24,25].
Recently, simulations based on the bond percolation model [16,26] agree qual-
itatively with the experimental optical spectra of DNA-capped gold colloids
at the melting transition. It has been suggested that at percolation, [1 −
p(Tc)]Ns/z = 1 −pc,where Ns, z, pc, Tcare number of single-stranded DNA on
each gold particle, number of nearest neighbors per gold particle, bond per-
colation threshold, and melting temperature, respectively [16]. The fraction
of single-stranded DNA that form links is p(T) and is temperature depen-
dent. The melting transition occurs when the fraction of links falls below the
percolation threshold. The calculated optical properties are found to change
dramatically when this threshold is passed. Moreover, the size dependence of
melting temperature can be explained in terms of the effect on link fraction.
The number of DNA on each particle, Ns, increases with increasing particle
size, D, with the relation Ns∝D2. The simulation confirmed that the melting
temperature dependence with particle size. Therefore, the percolation model
can qualitatively explain the experimental findings.
Fig. 4. Compositions of the AA versus AB systems. (a) In the AB system, each
particle is covered with with either DN A1 or D N A2 to create probes Aand B,
respectively. DN A1 and D N A2 are not complementary, and the linker DNA is
complementary to DN A1 and DN A2. As a result, probe Abinds exclusively with
probe Bupon the introduction of a target DNA sequence. (b) In the AA system,
each particle is covered with both DN A1 and D N A2 at a 1:1 ratio, resulting in
only one type of probe (A). Hence each particle can bind with any other particle in
the solution.
5

30 40 50 60 70
T (˚C)
0
0.5
1
Normalized Extinction
10 nm
20 nm
30 40 50 60 70
T (˚C)
0
0.5
1
Normalized Extinction
10 nm
20 nm
(a) (b)
Fig. 5. Melting curves for (a) AA and (b) AB nanoparticle systems monitored at
260 nm. In the AB (binary particle) system, particle Acan only bind with particle
Band vice versa, whereas in the AA system there is only one type of particle. The
melting transition width is different and the melting temperature scales differently
for the AA and AB systems.
The experimental results used for comparison, however, have two kinds of
particles, denoted as an AB system since each particle is either covered with
DN A1 (denoted as particle A) or DNA2 (denoted as particle B), but not
both, and particle Acan only be linked to particle B. Fig. 4a illustrates such
a binary system. In most simulations, the system is composed of only one kind
of particle, and each particle can be linked with any other particle. To this end,
we studied an AA system, where each particle is covered with both DN A1
and DN A2. This results in only one type of nanoparticle, and each particle
can be linked with any other particle in the system, as shown in Fig. 4b.
Melting curves for AA and AB systems are shown in Fig. 5. We noticed that
the width of the melting transition in the AA system is much narrower than
that of the AB system. The scaling of melting temperature with particle size
is also different for AA and AB systems. Unlike the AB system, there are
DN A1 and DN A2 on the same particle in the AA system. Since DN A1 only
connects with DN A2, the connections in the AA system is different from that
in the AB. More experimental data on the melting curves of AA systems of
different particle sizes are necessary to elucidate how the trend in melting
temperature changes from the AB to the AA systems.
As Lukatsky and Frenkel [3] recently proposed, the equilibrium phase be-
havior of DNA-linked colloidal assemblies is dominated by the temperature-
dependent binding free energy of a double-stranded DNA connecting a pair
of gold colloidal particles. Depending on the strength of the interaction, the
system can be in a homogeneous state, or separate into two coexisting phases.
Their calculation shows that there is a liquid-liquid phase separation in DNA-
linked gold nanoparticle assemblies. The origin of the sharp phase transition
is the entropic cooperativity of DNA-nanoparticle network. Upon cooling, the
6

Fig. 6. Proposed phase diagram of a DNA-linked colloidal gold system. The gelation
in the presence of solvent results in phase separation below the transition temper-
ature. The experiment follows the dashed line to the liquid/gel phase separation.
system undergoes liquid-liquid phase separation. The dense liquid phase is
strongly cross-linked and behaves as a solid gel.
In general, the percolation model is a crude representation of any gelation
processes [27]. In practice, the DNA-capped gold colloids are mixed with a
solvent, and this can be considered as “dilution effect,” i.e. gelation in dilute
solution. A gelation process in the presence of solvent always brings a trend
toward phase separation of the gelating species, and the phase diagram is
illustrated in Fig. 6. However, the critical exponents observed in our system
may be of the percolation type, as predicted by de Gennes [27], and supported
by simulations [16]. For our systems, we did observe phase separation, as pre-
dicted by de Gennes [27]. TEM images give direct evidence of the formation of
the gel phase, and support the expectation of phase transition and separation.
As shown in Fig. 2, upon adding linker DNA, gold colloids form gel-like porous
and amorphous aggregates. Phase separation occurs and eventually the dense
aggregates precipitate out of the solution, and the solution eventually became
clear.
3 Summary
We have presented direct evidence of a liquid-gel phase transition of the DNA-
linked gold colloidal assemblies. Compared to free DNA, a sharp melting tran-
sition is observed for this system. The formation of DNA-linked gold colloids
and the sharpness of the melting transition resemble the phase behavior of
gelation in dilute solution phenomena. We have shown that the scaling prop-
erty of the binary system AB, where two types of particles exists, is different
7

from that of AA, where all the particles are of the same type. The results
shown here indicate that the DNA-linked gold nanoparticles represent a new
class of complex fluids.
∗To whom correspondence should be addressed, email: chkiang@rice.edu.
References
[1] D. Frenkel, Soft condensed matter, Physica A 313 (2002) 1–31.
[2] K. t Leung, B. Schmittmann, R. K. P. Zia, Phase transitions in a driven lattice
gas with repulsive interactions, Phys. Rev. Lett. 62 (1989) 1772–1775.
[3] D. B. Lukatsky, D. Frenkel, Phase behavior and selectivity of DNA-linked
nanoparticle assemblies, Phys. Rev. Lett. 92 (2004) 068302.
[4] A. V. Tkachenko, Morphological diversity of DNA-colloidal self-assembly, Phys.
Rev. Lett. 89 (2002) 148303–1–148303–4.
[5] D. K. Lubensky, D. R. Nelson, Pulling pinned polymers and unzipping DNA,
Phys. Rev. Lett. 85 (2000) 1572–1575.
[6] Y. Zeng, A. Montrichok, G. Zocchi, Length and statistical weight of bubbles in
DNA melting, Phys. Rev. Lett. 91 (2003) 148101–1–4.
[7] G. Zocchi, A. Omerzu, T. Kurabova, J. Rudnick, G. Gr¨uner, Duplex-single
strand denaturing transition in DNA oligomers, arXiv:cond-mat/0304567 v1.
[8] R. M. Wartell, A. S. Benight, Thermal denaturation of DNA molecules: A
comparison of theory with experiment, Phys. Rep. 126 (1985) 67–107.
[9] D. Cule, T. Hwa, Denaturation of heterogeneous DNA, Phys. Rev. Lett. 79
(1997) 2375–2378.
[10] M. E. Fisher, Effect of excluded volume on phase transitions in biopolymers, J.
Chem. Phys. 45 (1966) 1469–1473.
[11] M. S. Causo, B. Coluzzi, P. Grassberger, Simple model for the DNA
denaturation transition, Phys. Rev. E 62 (2000) 3958–3973.
[12] C. A. Gelfand, G. E. Plum, S. Mielewczyk, D. P. Remeta, K. J. Breslauer, A
quantitative method for evaluating ther stabilities of nucleic acids, Proc. Natl.
Acad. Sci. USA 96 (1999) 6113–6118.
[13] F. Naef, D. A. Lim, N. Patil, M. Magnasco, DNA hybridization to mismatched
templates: A chip study, Phys. Rev. E 65 (2002) 040902–1–4.
[14] C. A. Mirkin, R. L. Letsinger, R. C. Mucic, J. J. Storhoff, A DNA-based method
for rationally assembling nanoparticles into macroscopic materials, Nature 382
(1996) 607–609.
8

[15] A. P. Alivisatos, K. P. Johnsson, X. Peng, T. E. Wilson, C. J. Loweth, M. P.
Bruchez Jr., P. G. Schultz, Organization of nanocrystal molecules using DNA,
Nature 382 (1996) 609–611.
[16] S. Y. Park, D. Stroud, Theory of melting and the optical properties of gold/DNA
nanocomposites, Phys. Rev. B 67 (2003) 212202.
[17] V. J. Anderson, H. N. W. Lekkerkerker, Insights into phase transition kinetics
from colloid science, Nature 416 (2002) 811–815.
[18] C.-H. Kiang, Phase transition of DNA-linked gold nanoparticles, Physica A 321
(2003) 164–169.
[19] E. R. Weeks, J. C. Croker, A. C. Levitt, A. Schofield, D. A. Weitz, Three-
dimensional direct imaging of structural relaxation near the colloidal glass
transition, Science 287 (2000) 627–631.
[20] C.-H. Kiang, Single particle study of protein assembly, Phys. Rev. E 64 (2001)
041911–1–3.
[21] C. R. Cantor, P. R. Schimmel (Eds.), Biophysical Chemistry, Part II: Techniques
for the Study of Biological Structure and Function, W. H. Freeman and
Company, New York, 1980.
[22] V. A. Bloomfield, D. M. Crothers, I. Tinoco, Jr., Nucleic Acids, University
Science Books, California, 2000.
[23] L. M. Demers, C. A. Mirkin, R. C. Mucic, R. A. Reynolds, III, R. L.
Letsinger, R. Elghanian, G. Viswanadham, A fluorescence-based method for
determining the surface coverage and hybridization efficiency of thiol-capped
oligonucleotides bound to gold thin films and nanoparticles, Anal. Chem. 72
(2000) 5535–5541.
[24] Y. Sun, N. C. Harris, C.-H. Kiang, Melting transistion of directly linked gold
nanoparticle DNA assembly, Physica A 350 (2005) 89–94.
[25] N. C. Harris, C.-H. Kiang, Disorder in DNA-linked gold nanoparticle assemblies,
Phys. Rev. Lett. submitted.
[26] S. Y. Park, D. Stroud, Structure formation, melting, and optical properties of
gold/DNA nanocomposites: Effects of relaxation time, Phys. Rev. B 68 (2003)
224201.
[27] P.-G. de Gennes, Scaling Concepts in Polymer Physics, Cornell University Press,
Ithaca, 1979, pp. 148–152.
9