DNA-based self-assembly: theory, simulations and experiments
DACAM, Vienna, Austria
DNA interactions are a subtle combination of different (e.g. pairing, stacking, electrostatic, steric or chiral) forces that gives rise to a wealth of possibilities in terms of intensity, directionality, reversibility and specificity of the so-obtained total net interaction. This richness has led in the last years to an impressive development of research in various directions [Bellini12]. Among the most notable is the one known as DNA nanotechnology: the rational design of DNA base sequences to produce their controlled and selective assembly into desired structures with the required properties [Seeman10]. Nowadays, structures with virtually any shape can be designed, and the level of control is such that DNA strands can be selectively bent at the molecular level as in the so-called DNA origami [Shih04,Rothemund06]. All-DNA superstructures of different kind have been obtained, including DNA liquid crystals, crystals, linear aggregates and hydrogels. On the other hand nanoparticles have been functionalized with DNA oligonucleotides and successively used for the self-assembly of larger structures [Mirkin96,Alivisatos96]. This is also an extremely active area of research, as mirrored by the more than 5000 citations collected by [Mirkin96,Alivisatos96] since 1996. However, much work has still to be done to achieve a satisfactory understanding of the collective behavior for both all-DNA and DNA-coated particles. For instance, it is quite surprising that despite this fervent activity the experimental crystallization of DNA-coated particles was successfully achieved only in 2008 [Nykypanchuk08,Park08]. In addition, most of the possibilities offered by DNA in the field of soft matter are still unexplored, for example in connection with the realization of the so-called patchy particles [Sciortino11]. All this suggests that a stronger link between theorists, computational physicists and experimentalists would be highly desirable to progress faster toward the realization of the materials of the future. In this framework, numerical simulations represent a precious tool. Simulating all-DNA constructs and DNA-coated particles is a challenging feat, especially when one is interested in the collective behavior of a large number of entities (see below for more details). This workshop aims to collecting a large number of experts in different DNA-research fields to allow for a wide and up-to-date overview of the available simulation techniques and for a presentation of the most recent available experimental results.
Self-assembly is the spontaneous formation - through free energy minimization - of reversible aggregates from basic building blocks. The size of the aggregating units, e.g. simple molecules, macromolecules or colloidal particles, can vary from a few angstroms to microns, thus making self-assembly ubiquitous in nature and of interest in several fields, including material science, soft matter and biophysics. Self-assembly is one of the most promising routes for the realization of novel materials, because tuning shape, valence, flexibility and mutual interactions of the individual building blocks can finely control the physical properties of these new colloid-based materials. In addition, these properties can be tailored for replicating on different time- and length-scales the behavior of atomic systems. In fact, one of the dreams of soft matter scientists would be to use such wide portfolio of physical and chemical interactions to realize materials with novel properties and functions i.e. the materials of the future.
Biological interactions - that are characterized by specificity, directionality, reversibility and tunability - open new ways for realizing particles with peculiar properties. Among the most promising candidates is DNA, the molecule of life, which in the recent years has been increasingly used as soft matter glue for various purposes. In particular, self-assembly of particles involving DNA has attracted the attention of theoreticians and experimentalists for its potential in soft matter science. Such particles can be obtained by functionalizing with DNA strands the surface of normal colloidal particles (typically gold, silver or polystyrene) [Glotzer10] or one can use self-assembly of single DNA strands to obtain all-DNA constructs with various properties of interest such as rigidity or geometry/shape [Seeman03]. As a genetic material in all life forms, DNA biological functions include storing and encoding genetic information. However, from materials science and engineering perspectives, DNA physical and chemical properties render it extremely useful for constructing materials with various functions. DNA constructs can be self-assembled from complementary DNA strands through Watson-Crick base pairing and can also be stabilized via enzyme-catalyzed assembly through covalent bonds. The base sequence of single-stranded DNA can be directly controlled and the resulting double-stranded DNA aggregate is very rigid, moreover the free energy change per base pair for the hybridization process is roughly few times the thermal energy. All these features make DNA an ideal candidate for the “bottom-up” development of novel nanomaterials [Berti08,Condon06] .
In this workshop we will focus mostly but not exclusively on the phase behavior of all-DNA constructs and DNA-functionalized particles that is very relevant for the soft matter community [Bellini12].
A first case of interest is that of DNA duplexes that form as a result of hybridization of short complementary sequences and whose shape can be approximately assimilated to that of cylinders [Nakata07]. Short DNA duplexes due to coaxial stacking interactions between their blunt ends can also self-assemble into weakly bonded chains [Nakata07,Zanchetta10]. Such a reversible physical polymerization is enough to promote the mutual alignment of these chains and the formation of macroscopically orientationally ordered nematic liquid crystal (LC) phases. A better physical understanding of these stacking interactions is very important because they are often invoked as the prebiotic route to explaining the gap between the random synthesis of elementary carbon-based molecules and the first complex molecules, possibly RNA oligomers, which are capable of catalyzing their own synthesis. In this respect, the self-assembly of DNA duplexes acts as an amplifier of the intermonomeric interactions, enabling the study of the effects of minor molecular modification (e.g., oligomer terminations) on base stacking and thus providing a suitable way to access and quantify hydrophobic coaxial stacking interactions. Finally the LC ordering of nucleic acids provides a new model of reversible aggregation leading to macroscopic ordering in which the strength of the intermonomer attraction can be modified by changing the duplex terminals (blunt-end stacking or pairing of overhangs). In two recent publications the reversible physical polymerization and collective ordering of DNA duplexes have been investigated both theoretically and numerically [DeMichele12a,DeMichele12b].
This system is also an excellent candidate to explore the role of chirality in self-assembly. The LC ordering arising from short DNA duplexes self-assembly is usually cholesteric as a consequence of the duplex chirality, which stems from the sugar groups of the DNA duplexes [Zanchetta10]. Chirality can play an active role in the self-assembly process, driving it towards selected structures or favoring the coupling with orientational order to promote the polymer organization under confinement [Marenduzzo10]. Moreover, the propagation of chirality from molecule to thermodynamic phases provides an extraordinary amplification of changes occurring at the microscopic level [Katsonis12]. The chirality of DNA has been the object of recent interest for different reasons. It has been proposed that the chirality of local interactions, which determine the geometry and stability of DNA-DNA crossover, affects the physical properties and then the global topological state of supercoiled DNA [Timsit10]. The local chirality has been claimed to play a key role also in the organization of DNA inside bacteriophages [Marenduzzo10].
Another remarkable experimental fact in the case of short DNA duplexes self-assembly is that, despite the right-handedness of the DNA helix, both left-handed and right-handed cholesteric phases were found depending on the sequence, the length and the nature of oligomer ends, while l-DNA (B-DNA with a number of base pairs more than 130) exhibits a left-handed phase. This result is quite surprising because the same handedness would be expected for perfect, right-handed helices with the same periodicity and the same charge distribution. The molecular origin of the cholesteric organization of solutions of DNA duplexes is controversial. Simple packing effects between hard, perfect helices would lead to a right-handed cholesteric phase for l-DNA, in contrast to the experimental findings [Straley76]. Yet, electrostatics is expected to play a role in the interactions between DNA duplexes, which are highly charged polyelectrolytes. Sophisticated models were developed, which have provided new insights into the subtle effects of interactions between helical charge distributions [Kornyshev07, Cherstvy08, Tombolato05, Frezza11].
A class of very interesting and technologically relevant all-DNA constructs are branched molecules such as T-shaped, Y-shaped, and X-shaped DNA [Li04]. Such branched DNA molecules can in fact be designed and synthesized in such a way that each arm of the DNA molecule possesses a sticky end, thus allowing them to hybridize and ligate with each other. In this way it is possible to construct more complicated dendrimer-like structures, like DNA nanobarcodes [Li05], in a controlled fashion. Furthermore, a single DNA building block can be modified with a variety of functional moieties, either isotropically or anisotropically, providing multi-functional property. In addition, these branched DNA molecules can further form 2D and 3D networked structures such as hydrogels by using various physical and chemical methods [Um06]. These networked DNA structures with improved functions provide new possibility in several nanotechnology applications [Roh11]. Even more complex structures can be created, like the so-called DNA-origami [Rothemund06] where DNA is folded into nanoscale shapes and patterns exploiting the selectivity of Watson-Crick base pairing.
X-shaped and Y-shaped DNA molecules designed with sticky ends in order to form reversible bonds between them can be viewed as neat model systems for patchy particles, i.e. colloids interacting with an anisotropic potential.
Anisotropy of the interaction potential between colloidal particles opens up several new directions for designing novel materials, the so-called anisotropy axis envisioned in the Glotzer and Solomon recent review article [Glotzer04]. In the last years the phase diagrams and percolation thresholds of patchy particles have been extensively investigated both theoretically and numerically. In these recent studies it has been found that the number of bonding sites per particle (its valence) is the key parameter controlling the location of the liquid-vapor critical point. If the average valence approaches the value of two, the phase-separation region shrinks to a very narrow range of densities around zero and low temperatures can be reached without encountering the phase separation at all. These arrested (glassy) states at arbitrarily low density are what can be called an ideal gel [Bianchi06]. More recent studies [Tavares09] have shown that when patches of different types can compete in asymmetric ways, very unconventional phase diagrams can appear, whose experimental evidence is still missing.
Finally, DNA can also be grafted onto the surface of colloidal particles in order exploit also in this case the highly specific properties of DNA to assemble nanomaterials. Attaching a DNA oligonucleotide at one end to colloidal particles provides a “sticky patch” (i.e. a binding site) to such particle that will interact only with a patch made of a complementary oligonucleotide [Chaikin09].
These DNA-functionalized colloidal particles can be designed in such a way that their interactions can be finely tuned and controlled making it possible to create specific self-assembled ordered structures [delaCruz12]. A prominent example is provided by gold nanoparticles functionalized by double stranded DNA grafted on their surface [Glotzer10, Bellini12].
The state of the art concerning simulation of DNA is evolving very rapidly, stimulated by the advances in DNA nanotechnology [dePablo11]. Several models of DNA are available in the literature, ranging from fully atomistic representations, in which all atoms (solvent included) are considered explicitly to highly coarse grained ones, in which many atoms are modeled as few beads linked together in a worm-like chain.
The full-atom approach, implemented with force fields such as AMBER [Pearlman95] or CHARMM [Mackerell95], is the most detailed way of investigating the inner structure of DNA molecules [Orozco03].
At first glance such approach can be considered the best, because it provides more chemical detail. Anyway, because of the huge number of degrees of freedom that have to be taken into account (since one has to consider also water and salt), atomistic simulations, even for small systems such as short duplexes, are computationally demanding and hence the length scales and time scales, which can be investigated, are usually quite restricted.
Indeed thermodynamic investigation of processes such as hybridization transitions in atomistic models is possible only using sophisticated numerical techniques and extensive computational resources, and only for a very limited number of nucleotides [Maffeo12].
In coarse-grained models, the description of DNA is simplified by reducing the number of degrees of freedom and considering effective interactions between coarse-grained units in place of all-atoms interactions and the challenge is to include the minimum level of details which is able to capture all the physical properties of biological DNA.
The parameters of the force fields employed in the coarse-grained models are usually chosen according to either a “bottom-up” approach [Morriss10] or a “top-down” one [Ouldridge09].
In a "bottom-up" approach the parameters of the effective force fields are tuned by direct comparison with either atomistic simulations or data from crystal structures, while in a "top-down" approach the force fields are designed to provide a reasonable description of a range of large-scale properties (such as melting temperatures of helices) when compared to experiments.
“Bottom-up” approaches are usually well-suited for obtaining structural/conformational properties of double stranded DNA [Dans10] but usually they fail to reproduce single strand behavior, DNA hybridization and in general thermodynamic properties of double stranded DNA like melting temperatures. Conversely “top-down” approaches are more suited for studying assembly transitions and duplex thermodynamics, and the range of their applications is very wide.
Indeed they have been used to study duplex renaturation/denaturation [Drukker01], hairpin formation [Kenward09], Holliday junction formation [Ouldridge09], duplex thermodynamics [Sambriski09], overstretching and gelation of DNA-coated nanoparticles [Largo07].
Elisa Frezza ( Université de Paris ) - Organiser & speaker
Roberto Cerbino ( Università degli Studi di Milano ) - Organiser & speaker
Cristiano De Michele ( "Sapienza" University of Rome ) - Organiser & speaker
Francesco Sciortino ( Sapienza, University of Rome. ) - Organiser & speaker