WorksPhysics and Astronomy
Interfacial Packing of DNA Nanostars Regulates Dynamics on Synthetic Cell Membranes
Abstract
DNA nanostructures are emerging as programmable components for engineering synthetic cell membranes, yet how their collective packing and deformability regulate molecular dynamics at membrane interfaces remains poorly understood. Here, we investigate the packing and mobility of DNA nanostars with tunable stiffness on lipid-coated droplets. The negatively charged nanostars spontaneously adsorb onto cationic membranes, and their interfacial packing is changed by the bulk DNA concentration and droplet size, which together determine the number of encapsulated nanostars. Combining fluorescence recovery after photobleaching experiments with coarse-grained simulations, we reveal distinct packing-dynamics relationships for rigid and soft nanostars. For rigid nanostars, diffusion first decreases gradually and then drops sharply with increasing interfacial packing, approaching a dynamically arrested state consistent with jamming-like behavior. In contrast, at equivalent experimental conditions, soft nanostars systematically reach lower interfacial packing fractions and show a weaker decrease in apparent mobility. This behavior is consistent with their weaker membrane affinity, which facilitates adsorption-desorption with the bulk. When this exchange is suppressed in simulations, crowding reduces the lateral diffusion of both nanostar types, but produces distinct dense configurations: soft nanostars become strongly deformed, whereas rigid nanostars largely retain their shape and form interlocked, gear-like arrangements. Nanostar adsorption also impedes lipid diffusion, while differences in membrane affinity result in distinct lipid mobile fractions. These findings reveal how the interplay between nanostar packing and deformability regulates molecular transport at membrane interfaces, providing a physical design principle for tuning lateral fluidity and crowding in artificial cells.
main_v4.pdf
Aqueous droplets surrounded by lipid membranes, including liposomes and lipid-coated droplets, serve as versatile platforms for synthetic biology and artificial cell engineering. [1, 2, 3].
Their functionality critically depends on membrane properties such as stability and fluidity, which are governed by interfacial organization. To tune these properties, polymers and nanoparticles are often adsorbed at membrane interfaces. [4, 5] While such colloidal additives can stabilize the interface, high interfacial coverage can suppress membrane fluidity and permeability, both of which are essential for emulating cell-like behavior. [6, 7, 8] Understanding how interfacial packing regulates membrane dynamics while preserving interfacial stability therefore remains a major challenge.
In two-dimensional colloidal systems, increasing particle density can lead to caging and ultimately jamming, strongly suppressing particle mobility. [9] Similar crowding-induced slowing of diffusion has been reported for rigid particles and macromolecules associated with supported lipid membranes. [10, 11] Yet it remains unclear whether such packing–dynamics relationships apply to deformable soft particles coupled to closed, fluid membranes. Unlike planar substrates, membrane-coated droplets provide a finite, curved, and laterally fluid interface on which adsorbed particles and lipids can rearrange together. Moreover, in a closed compartment, the number of encapsulated particles relative to the available membrane area depends on the droplet size, even at a fixed bulk concentration, which presents a challenge when comparing droplets of different sizes, but also provides a direct means to tune interfacial packing.
DNA nanostructures provide an ideal materials platform for studying and utilizing such packing–dynamics relations: their geometry, anisotropy, and conformational rigidity can be programmed with nanometer precision, providing a versatile route to engineer particles with controlled shape and deformability. [12, 13] Recent studies have shown that, like double-stranded DNA of several kbp or more, [14, 15] nanometer-sized DNA nanostructures can spontaneously adsorb onto cationic lipid membranes, including planar membranes, liposomes, and membrane-coated droplets. [16, 17, 18, 19, 20] Membrane association of DNA nanostructures is also strongly modulated by the ionic environment, particularly by divalent cations. [21] The number of phosphate groups and the conformational rigidity of DNA can strongly influence membrane adsorption, with larger and stiffer structures generally adsorbing more strongly. [22] Previous experiments with DNA nanostructures on supported lipid membranes found density-dependent packing transitions, such as long-range ordered structures for Y-shaped DNA nanostructures.[23, 24] These observations suggest that the interplay between membrane adsorption, rigidity, and interfacial packing of DNA nanostructures may jointly regulate their mobility and the dynamics of the surrounding lipids.
Here, we test this hypothesis by systematically investigating how interfacial packing of DNA nanostructures governs dynamics on synthetic cell membranes, using star-shaped DNA nanostructures (hereafter termed “nanostars”) as model soft colloids. By encapsulating nanostars within cationic lipid-coated droplets at varying droplet radii and bulk nanostar concentrations, we control the number of enclosed nanostars and thereby tune their interfacial packing. Combining fluorescence measurements with coarse-grained molecular dynamics (MD) simulations, we identify a packing-controlled crossover in nanostar mobility that strongly depends on nanostar stiffness. We further show that lipid diffusion is suppressed by nanostar interfacial crowding, where differences in membrane affinity lead to distinct lipid mobile fractions for rigid and soft nanostars. These findings establish DNA nanostars as programmable soft colloids whose collective packing governs both their own mobility and the dynamics of the surrounding synthetic-cell membrane.
Results and discussion
Interactions of individual DNA nanostars with lipid membrane
To investigate how particle deformability influences interfacial packing and dynamics, we designed two types of three-armed DNA nanostars that differ in their conformational rigidity (Figure 1); rigid nanostars contain nucleotides in a double-duplex configuration, whereas soft stars contain nucleotides in a single-duplex structure (see Figure S1 for sequence information). Unlike conventional Y-shaped DNA motifs that may form networks through intermolecular hybridization, [20, 24, 25, 26, 27] the DNA nanostars used here lack sticky ends and therefore behave as isolated nanoparticles. We estimated the molecular footprint, i.e., the instantaneous area occupied by a DNA star, from MD simulations in the dilute limit, yielding and for rigid and soft nanostars, respectively (see Sec. S7 in the Supporting information). The approximately fourfold difference in , despite rigid nanostars containing only 2.6 times as many nucleotides as soft nanostars, indicates a more loosely packed spatial arrangement of nucleotides in the rigid nanostars.

Figure 1. Representative conformations of rigid and soft stars obtained from coarse-grained MD simulations, and schematic illustration of a DOTAP-coated droplet encapsulating DNA nanostars.

Figure S1. DNA sequences used to construct the DNA nanostars (S1-3 for a soft nanostar and R1-3 for a rigid nanostar). Each color corresponds to the same-colored strand shown in the simulation snapshot.
In principle, positively charged lipid membranes attract DNA nanostructures through electrostatic interactions with their negatively charged sugar-phosphate backbone, with stronger adsorption at higher fractions of cationic lipids. [14] However, previous experiments found that -DNA chains desorbed from lipid membranes containing only 10% cationic 1,2-dioleoyl-3-trimethylammonium-propane (DOTAP) by mass when the NaCl concentration exceeded 50 mM. [14] Therefore, we used a synthetic cell membrane composed of a 100 mol% DOTAP monolayer to maximize electrostatic attraction between the membrane and the DNA nanostars.
To estimate the adsorption strength, we modeled the interaction between individual DNA nanostars and the lipid membrane as a combination of screened electrostatic attraction and short-range hydration repulsion. [28] Other contributions to the adsorption free energy, including those associated with counterion release and the entropic cost of water reorganization, were neglected for simplicity. The hydration repulsion, which is commonly employed for hydration interactions between lipid and hydrophilic interfaces, [29, 30] arises from the energetic cost associated with perturbing the interfacial hydration layers between two surfaces. It is described by an exponentially decaying pressure, , where is the surface-to-surface separation between the DNA nanostar and the membrane, is the hydration pressure extrapolated to contact, [31, 32] and is the characteristic decay length of the hydration interaction. [33] The corresponding hydration free energy was calculated as
To estimate the electrostatic contributions to the adsorption free energy, the nanostars were approximated as uniformly charged disks with area and effective charge , where regulates the effective fraction of the nominal DNA charge. The lipid membrane was represented as an uniformly charged planar surface, with surface charge density [34, 35] scaled by an effective charge fraction . Exploiting the rotational symmetry of this disk–plane configuration, we solved the non-linear Poisson–Boltzmann (PB) equation in a two-dimensional axisymmetric geometry. Because both the DNA nanostars and the cationic DOTAP membrane possess high nominal charge densities, their effective charges in solution are expected to be lower than their bare charges. To bracket the physically plausible range of electrostatic interactions, we considered two limiting charge conditions in the PB calculations. To estimate the upper bound, we assumed that both the DNA phosphate groups and DOTAP molecules carry their full nominal charges, so that . For the lower bound, the DNA effective charge fraction was reduced to to account for Manning counterion condensation, [36] whereas the membrane effective charge fraction was set to , chosen to be consistent with the experimentally measured -potential of approximately for DOTAP membranes. [37] Note that these lower values should be regarded as estimates, since Manning theory assumes an idealized line charge, while the exact conversion from -potential to the membrane surface potential requires knowledge of the slip-plane position. [38]
For the upper-bound estimate, the adsorption free energy, , is minimized at a distance of about from the membrane, with and for the soft and rigid nanostars, respectively, indicating effectively irreversible adsorption (Fig. 2a). For the lower-bound estimate, the minimum of shifts to roughly , with substantially weaker adsorption free energies of approximately and for the soft and rigid nanostars, respectively (Fig. 2b). Thus, the soft nanostars are only weakly bound in the lower-bound model, suggesting reversible and highly dynamic membrane association. In contrast, the more negative free energy of the rigid nanostars indicates substantially stronger binding, and hence less frequent adsorption–desorption events.
Adsorption and packing of encapsulated DNA nanostars
In the experiments, we encapsulated the nanostars within cell-sized droplets coated with DOTAP and observed fluorescently labeled DNA nanostars by confocal laser scanning microscopy over a broad range of droplet radii, ranging from to (Fig. 3). Because the molecular footprint of rigid nanostars is approximately four times larger than that of soft nanostars, different DNA concentration ranges were used for the two systems.
Figures 3a,b show selected representative cases (small droplets with and large droplets with ) at two DNA concentrations ( and for rigid nanostars, and and for soft nanostars). The total number of nanostars in a droplet scaled with its volume; hence, the number of stars available for membrane adsorption was determined by and .
To quantify the effective interfacial packing fraction of DNA nanostars from the confocal microscopy data, we determined the relative fluorescence intensity of the membrane-adjacent region (see Figure S2; Sections S1 and S2). Because the first adsorbed layer cannot be distinguished from subsequent wetting layers within the optical resolution of confocal laser scanning microscopy, reflects the relative nanostar concentration near the membrane rather than direct adsorption (Figure S3). Assuming that the DNA concentration inside the droplet is equal to the initial bulk concentration and neglecting the droplet’s curvature,

Figure S2. (a) Experimental determination of from the fluorescence intensity profile of DNA nanostars. The mean background intensity, , was first calculated, and the position where the intensity started to increase above was defined as the outer boundary of the droplet. The inner intensity, , was obtained by subtracting from the mean intensity inside the droplet, and the position where the intensity started to deviate from was defined as the inner boundary of the DNA nanostar adsorption shell. The volumes of each region ( and , respectively) were estimated from the corresponding radii. The volume-weighted mean intensity corresponds to the integrated fluorescence intensity divided by the total volume of the corresponding region. (b) and were determined using ImageJ from pixel-resolved fluorescence intensity data in two-dimensional cross-sectional images. The mean pixel intensity within each region defined in (a) was used as or .

Figure S3. Localization factor determined for rigid (a) and soft nanostars (b) as functions of the inverse droplet radius . The solid lines are guides to the eye only.
with Avogadro constant . To determine whether the estimated interfacial packing is limited by the number of available nanostars within the droplet, we compared with the idealized value obtained by setting in (1), which implies that all nanostars are adsorbed onto the membrane.
For both rigid and soft nanostars, generally increased with increasing droplet radius, , and DNA concentration, , as expected from the increasing number of nanostars per unit membrane area (fig:3c,fig:3d). For rigid nanostars, closely followed the predicted linear dependence, but increasingly fell below as increased. At sufficiently large , saturated at approximately for and for (fig:3c). The soft nanostars likewise exhibited an initial increase in , but the linear regime extended over a much smaller range of . Further, the estimated values were significantly more scattered and reached a smaller plateau value of , far below the ideal limit $\phi_{\mathrm{id}} (Fig. 3d).
To understand the molecular origin responsible for this relatively sparse interfacial localization, we performed MD simulations of DNA nanostars adsorbed to a smooth, flat substrate. In the simulations, nanostars were confined to the surface by a short-range attraction along the surface normal, which ensured complete adsorption throughout the simulations. The number of nanostars, , was kept constant, and the concentration was varied by adjusting the lateral size of the simulation box. The packing fraction from the simulations is then defined as , with and denoting the dimensions of the simulation box. fig:4a,fig:4b show representative simulation snapshots of rigid and soft nanostars, respectively, at different packing fractions . Notably, rigid-star configurations appear more sparse at the same , owing to their less compact double-duplex architecture which has a 35 % lower nucleotide density than the soft stars.
These simulation snapshots already reveal increasing deformation of the nanostars with increasing relative to the dilute limit, with the effect being particularly pronounced for the soft nanostars. We quantified this deformation via the distribution of angles between neighboring arms, , computed from the scalar product of the unit vectors connecting the center of mass of a star to the tip of an arm. In the dilute limit (), the resulting probability distribution exhibits a sharp peak at for rigid stars, whereas soft stars display a much broader distribution spanning to (Figure 4 c,d), confirming their substantially greater conformational flexibility. As increased, broadened for both nanostar types, indicating increased conformational distortion due to intermolecular crowding. Importantly, these changes emerged at packing fractions close to the experimentally observed saturation values for each motif, namely for rigid nanostars, and for soft nanostars (Figure 3 c,d). These results suggest that the experimentally observed saturation of interfacial localization is strongly tied to the significant distortion of the nanostars, consistent with an increasing free-energy penalty for accommodating additional nanostars at the membrane.
Figure 3. (a,b) Confocal cross-section images showing nanostars (green) inside droplets covered with a lipid membrane (red), for (a) rigid and (b) soft DNA nanostars at various DNA concentrations, , and droplet radii, , as indicated. Fluorescence intensity profiles along the dashed lines are shown beside each image. Scale bars indicate 20 m. (c,d) The effective interfacial packing fraction of (c) rigid and (d) soft nanostars, estimated from confocal fluorescence images, as a function of droplet radius . Different colors indicate different , while dashed lines indicate the ideal case, , in which all encapsulated nanostars are localized at the membrane interface ( in (1)).

Figure 4. Representative MD simulation snapshots of (a) rigid and (b) soft DNA nanostars at different packing fractions , as indicated. The colors represent the individual stars. (c,d) Probability distribution of inter-arm angle, , of (c) rigid and (d) soft nanostars at different .
Diffusivity of DNA Nanostars
We now turn to the dynamic consequences of interfacial crowding, which may restrict the lateral mobility of membrane-bound nanostars. To quantify these effects, we performed fluorescence recovery after photobleaching (FRAP) measurements on individual droplets to determine the diffusion coefficients () and mobile fractions () of both nanostar types (Figure S5). Circular regions (radius ) on the membrane were photobleached, and the fluorescence recovery was monitored over time (Figure 5 a,b). Recovery curves were obtained under all conditions and were well described by a single-exponential function with recovery time (see Materials and Methods section):
Figure 5. (a) Representative FRAP curves for nanostars (green circles) and lipids (orange squares) for small and large droplets. The darker-colored curves are the fitted results using (2). experimentally denotes the average fluorescence intensity of the ten pre-bleaching frames. Photobleached regions are indicated by dashed white circles in the images. Scale bars are 10 m. The DNA concentrations are or for rigid stars and or for soft stars.

Figure S5. Diffusion coefficients and mobile fractions of (a) lipids and (b) DNA nanostars on DOTAP membranes covering droplets with varying radii, , measured for two DNA concentrations ( and ).
from which the diffusion coefficients were calculated as
Control experiments showed that unbound nanostars in the droplet interior exhibited complete fluorescence recovery (Figure S4), confirming that the limited recovery observed at the membrane is associated with membrane binding rather than photobleaching artifacts.

Figure S4. A representative FRAP curve for inner rigid nanostars freely diffusing in the interior of large droplets with at .
The extracted values of and reveal that interfacial packing strongly regulates the nanostar mobility, but the trends differed markedly between rigid and soft stars (fig:6a,fig:6b). To visualize the overall trends, we grouped individual measurements (open symbols) into linearly spaced bins in (filled symbols). The diffusion coefficient, , and mobile fraction, , of rigid stars decreased markedly with increasing , indicating a gradual transition into a dynamically arrested state (fig:6a). The concurrent reduction in and further suggests that crowding restricts not only the translational diffusion of individual stars but also the molecular exchange within the membrane-associated nanostar layer. Dynamic arrest is approached at , well below the jamming threshold of hard disks at ; this onset is closer to the jamming range of reported for quasi-2D experiments on cross-shaped granular materials, [39, 40] suggesting that shape anisotropy causes arrest at lower packing fractions.
In contrast, and of soft stars varied only weakly across the experimentally accessible -range (fig:6b). Notably, exhibited substantial variability near the highest packing fraction, . This weaker -response may reflect their deformability, which allows local rearrangement under crowding. Further, reversible membrane binding of soft stars could facilitate transient desorption and re-adsorption events, which would contribute to fluorescence recovery in addition to lateral diffusion along the membrane. To test whether the exchange of bleached membrane-bound nanostars with unbleached nanostars from the bulk increases the apparent mobility of the nanostars, we compared FRAP measurements
performed at two different positions on the droplet membrane (Figure S6). Bleaching at the top surface provides greater access to the surrounding bulk phase, whereas bleaching in the cross-sectional membrane region restricts bulk access while preserving lateral diffusion along the membrane. For soft stars, the mobile fraction was significantly higher at the top surface than in the cross-sectional geometry (, paired -test), consistent with an additional recovery pathway through nanostar exchange with the bulk region. In contrast, FRAP measurements for the rigid stars showed no detectable difference between the two bleached regions (), indicating that their fluorescence recovery is dominated by lateral motion within the membrane-associated layer. These measurements therefore support bulk-mediated membrane exchange as an additional contribution to the apparent mobility of soft, but not rigid, nanostars.

Figure S6. (a) Schematics of FRAP experiments conducted at the top and cross-sectional regions of the droplet. Fluorescence recovery is dominated by reversible adsorption–desorption at the top and by lateral diffusion in the cross-sectional region. (b) Mobile fraction of rigid and soft nanostars at the top and cross-sectional regions. Statistical significance between the cross-sectional and top measurements from the same droplets was assessed using a two-sided paired -test ( values).
To isolate the effect of nanostar–surface attraction in our MD simulations to suppress desorption events. Thus, both rigid and soft stars remained adsorbed and underwent two-dimensional diffusion confined to the surface. Their diffusion coefficient was calculated directly from the mean-square displacement (MSD) of individual nanostars (Figure S8). At low packing fractions (), the soft stars had an approximately 2.6 times higher diffusion coefficient , consistent with their correspondingly lower number of nucleotides, , compared to the rigid nanostars. A similar inverse scaling was observed in previous experiments on -phage DNA adsorbed to planar lipid membranes. [14, 41] As the packing fraction increased, decreased almost linearly with for both rigid and soft nanostars, showing that crowding suppresses their intrinsic lateral diffusion irrespective of stiffness; the weak -dependence of measured experimentally for soft nanostars is therefore consistent with an additional contribution from bulk-mediated exchange.
Figure S8. In-plane mean-square displacement of the center of mass of (a) rigid and (b) soft DNA nanostars at different .
Notably, the normalized diffusion coefficient, , decayed significantly faster for soft stars and reached at a slightly lower packing fraction () compared to rigid stars (). To elucidate the molecular origin of this difference, we examined the internal relaxation of the nanostars via the autocorrelation function of the angle between two arms
For both rigid and soft stars, decays exponentially (fig:7a,fig:7b) and we extracted the characteristic internal relaxation time, , by fitting . For dilute systems, increases only gradually with , but then rises sharply as crowding restricts arm relaxation (fig:7c). This slowing of internal relaxation is more pronounced for soft stars, whose surpasses that of rigid stars at . This crossover occurs near the earlier onset of translational slowdown (fig:7c), linking their translational arrest to the suppression of internal relaxation. The difference in internal relaxation is also evident in the simulation snapshots (Figure 4f,g): rigid stars largely retain their shape and form an interlocked gear-like arrangement, whereas soft stars become strongly deformed to maximize interfacial packing.
Figure 7. (a,b) The arm vector autocorrelation function of (a) rigid and (b) soft stars at different . (c) The characteristic relaxation time , obtained from exponential fits of , as a function of .
Suppressed diffusion of lipids by adsorbed DNA nanostars
Previous studies of protein-rich membranes have revealed a decrease in lateral mobility of both proteins and surrounding lipids. [7, 42] Therefore, we examined whether the diffusion coefficient, , and mobile fraction, , of the lipids decrease with increasing nanostar packing fraction, , and whether this response differs between rigid and soft nanostars. To test this hypothesis, 1 in 1000 membrane lipids was replaced with a dye-labeled analog (Rho-PE), while a small fraction of nanostars was likewise fluorescently labeled (see SXX for the details). We determined and from FRAP experiments (orange lines in Figure 5), and plotted the resulting values against in Figure 8. Note that the experimentally accessible range of for soft stars () is considerably smaller than that of rigid stars (), as discussed above.
Figure 8. (a, b) Experimentally obtained and of lipids as a function of the star area fraction, , for membranes decorated with (a) rigid and (b) soft stars. Symbols and error bars represent mean values and standard deviations, respectively. Faint open symbols indicate the original experimental data.
To establish a baseline, we first characterized the lipid dynamics of bare membranes (), finding and . For membranes decorated with rigid nanostars, gradually decreased with increasing and reached a plateau of at (Figure 8a, left). Concurrently, the mobile fraction decreased almost linearly from to as the nanostar packing fraction increased to (Figure 8a, right). For membranes with soft nanostars, lipid diffusivity decreased sharply by and remained approximately constant at /s (Figure 8b, left). In contrast to the rigid nanostar case, most lipids remained highly mobile () over the experimentally accessible range (Figure 8b, right).
The stronger suppression of lipid diffusivity for the soft nanostars may originate from their approximately 1.5-fold higher projected nucleotide density at the same nominal packing fraction , potentially resulting in more DNA–lipid contacts per unit area. Despite this different sensitivity, the similar plateau value of suggests a common limiting lipid mobility for sufficiently dense DNA layers. The distinct trends in qualitatively mirror those observed for the nanostars themselves (Figure 6), and thus likely reflect differences in the persistence of the DNA–membrane association (see Figure 2 and corresponding discussion).
Figure 2. Adsorption free energy of DNA nanostars at the membrane, , shown for the (a) upper-bound condition () and (b) lower-bound condition ( and ). Blue and pink curves denote rigid and soft nanostars, respectively.
Figure 6. (a, b) Experimentally obtained diffusion coefficient, , and mobility fraction, , of (a) rigid and (b) soft DNA nanostars as functions of packing fraction, . Open symbols show original data, while filled symbols show averaged values from linearly spaced bins; error bars denote the standard deviation. (c) from simulations of rigid and soft nanostars as functions of . (d) , normalized by its value at infinite dilute, , as functions of , from simulations and experiments, as indicated.
Reversible adsorption–desorption of soft nanostars likely makes their interactions with nearby lipids more transient, allowing most lipids to redistribute during fluorescence recovery even though their diffusion is slowed. In contrast, rigid nanostars bind much more tightly to the membrane, thereby preventing the redistribution of nearby lipids over the FRAP timescale.
Conclusions
By combining experiments on membrane-coated droplets with coarse-grained molecular dynamics simulations, we show that interfacial packing and bending rigidity jointly regulate the mobility of adsorbed DNA nanostars and surrounding lipids. The closed droplet geometry couples bulk DNA concentration and droplet size to the number of encapsulated nanostars per unit membrane area, thereby providing direct control over interfacial crowding. Despite using membranes composed entirely of cationic lipids, the experimentally measured nanostar packing fractions remain systematically below the ideal limit of full adsorption. At a given droplet size and DNA concentration, soft nanostars reach significantly lower packing fractions than rigid nanostars, suggesting more frequent bulk-mediated desorption–adsorption exchange. These differences indicate incomplete interfacial localization and are consistent with a weaker effective electrostatic attraction of soft nanostars to the membrane.
FRAP experiments show that the diffusivity and mobile fraction of adsorbed rigid nanostars decrease simultaneously with increasing packing fraction, approaching a dynamically arrested state. In contrast, soft nanostars exhibit a weaker dependence of their apparent mobility on packing fraction, likely because bulk-mediated nanostar exchange provides an additional pathway for fluorescence recovery that partly masks the suppression of lateral diffusion. By excluding this exchange pathway in our simulations, we find that crowding suppresses the intrinsic lateral diffusion of both rigid and soft nanostars, albeit through distinct crowded configurations: rigid nanostars largely retain their shape and form interlocked, gear-like arrangements, whereas soft nanostars undergo substantial deformation and exhibit earlier suppression of internal relaxation.
Notably, nanostar adsorption also suppresses lipid mobility; soft nanostars cause a sharper decrease in lipid diffusivity, which we attribute to their 1.5-fold higher potential nucleotide–lipid contacts per unit membrane area at the same nominal packing fraction. In contrast, the mobile fraction of lipids qualitatively mirror those of the nanostars and reflect their different adsorption–desorption dynamics: the mobile fraction decreases with increasing packing fraction of rigid nanostars, but remains nearly constant for the more transiently adsorbed soft nanostars. Together, these results establish DNA nanostars as programmable components for tuning interfacial crowding and lateral membrane mobility through molecular architecture, offering a physical design principle for controlling membrane dynamics in synthetic cells.
Materials and Methods
Materials
1,2-dioleoyl-3-trimethylammonium-propane (DOTAP; chloride salt) was purchased from Avanti Polar Lipids. Lipids were dissolved in chloroform (Nacalai Tesque), and lipid films obtained after solvent evaporation were redissolved in hexadecane (Nacalai Tesque) as the oil phase. Sodium chloride (NaCl; FUJIFILM Wako Pure Chemical Co.) and Tris–HCl buffer (pH 8.0; Nacalai Tesque) were dissolved in ultrapure water (Invitrogen) to prepare the buffer solution for soft DNA nanostars. Tris-Acetate-EDTA buffer (pH 8.3; Nacalai Tesque) with Magnesium Acetate Tetrahydrate (Nacalai Tesque) added was used to prepare the buffer solution for rigid DNA nanostars. For DNA nanostars, lyophilized DNA oligonucleotides (Eurofins Genomics Japan) were dissolved in ultrapure water and added to the aqueous phase as appropriate. DNA stock solutions were prepared at 2 mM and stored at -20 °C until use. All materials were used as received without further purification.
DNA nanostars preparation
We designed DNA oligomers based on previously reported architectures to assemble DNA nanostars. [27, 43, 44] At high temperatures, the DNA oligomers remain dissociated due to entropic stabilization, whereas upon cooling they hybridize via Watson–Crick base pairing to form DNA nanostars. Two DNA nanostar architectures were designed: a rigid nanostar with arms consisting of two double-stranded DNA (dsDNA) helices and a soft nanostar with arms consisting of a single dsDNA helix. DNA nanostars display a single melting transition at and remain as isolated motifs. The melting temperatures were for rigid and for soft nanostars, respectively.
For self-assembly of rigid and soft nanostars, the constituent DNA oligomers were separately mixed at equimolar ratios and diluted in their respective buffers (20 mM Tris–HCl, pH 8.0, 350 mM NaCl for rigid nanostars; 20 mM Tris–HCl, pH 8.0, 350 mM NaCl for soft nanostars). The solution was heated to for 10 minutes and then slowly cooled to at a rate of using a T-Gradient thermocycler (Biometra, Germany). For fluorescence analysis, one of the constituent DNA oligomers was labeled with 6-carboxyfluorescein (FAM) for the soft nanostar and fluorescein isothiocyanate (FITC) for the rigid nanostar.
Droplet preparation
Water-in-oil droplets encapsulated by lipid monolayers were prepared as follows. DOTAP was first dissolved in chloroform at a final concentration of 10 mM. An aliquot (50 µL) of the lipid solution was transferred to a Durham tube, and the solvent was evaporated under a gentle stream of nitrogen gas to form a dry lipid film at the bottom of the tube. Hexadecane (500 µL) was then added, and the mixture was sonicated at approximately for 90 minutes to obtain a lipid-in-oil solution with a final lipid concentration of 1 mM. Approximately 0.1 mol% of the total lipids was replaced with Rho-PE for fluorescence analysis. The solution was subsequently cooled slowly from to room temperature (approximately ) with intermittent vortex mixing to ensure homogeneous lipid dissolution.
To prepare artificial cells with DNA nanostars localized beneath the membrane, of the aqueous solution containing nanostars was added to of the DOTAP-in-hexadecane solution after sonication. Here, to minimize the ion difference in the buffers, the mixture of soft nanostars are diluted with the buffer for rigid nanostars. Since both systems contained divalent cations in the buffer, differences in membrane–DNA electrostatic interactions were minimized. Throughout this study, the DNA nanostar concentration, , denotes the final concentration after dilution, which was assumed to be identical in the droplets. The mixture was then gently tapped to induce droplet formation without bulk emulsification. The resulting droplets were transferred onto a silicone-coated glass-bottom dish (Matsunami) to prevent adhesion to the glass surface and were used for subsequent observations and measurements.
Fluorescence intensity analysis
Droplets with DNA motifs were observed by confocal laser scanning fluorescence microscopy (Olympus IX83 with FV1200; Olympus). Fluorescent images of FAM and FITC, which were excited by a mercury lamp or a laser ( for DNA nanostars and for lipid membrane), were obtained using fluorescence filter sets (U-FBNA for mercury lamp; to for laser; Olympus). The pinhole size was fixed to be about . The obtained images were analyzed by the National Institutes of Health ImageJ software.
FRAP
Fluorescence recovery after photobleaching (FRAP) experiments [45] were performed using the same confocal microscopy setup as described above. Circular regions of interest with a radius of were photobleached by applying a high-intensity laser pulse. Fluorescence recovery was then monitored over time under the same imaging conditions as before photobleaching. The fluorescence intensity in the bleached region was corrected for background fluorescence and normalized by the pre-bleach intensity. The pre-bleach intensity, , was defined as the average fluorescence intensity immediately before photobleaching. The intensity immediately after photobleaching was denoted as , and the plateau intensity after recovery was denoted as . The mobile fraction was calculated as the fraction of fluorescence recovered relative to the fluorescence lost by photobleaching, as .
Simulation Model
We employ the coarse-grained oxDNA2 model, [46, 47], where single-stranded DNA is represented as sequences of rigid-body nucleotides, interacting via effective bonded and non-bonded potentials. [48, 49] All molecular dynamics (MD) simulations were conducted using the oxDNA package with GPU acceleration. [50] All simulations were conducted in the ensemble with an implicit solvent, where is the number of nucleotides and is the volume of the simulation box. The fundamental simulation units of length, mass and energy were set as , , and , respectively. The temperature was kept constant at using a Langevin thermostat with translational friction coefficient with being the intrinsic simulation timescale. Electrostatic interactions between DNA chains are modeled using Debye-Hückel theory, with the concentration of implicit monovalent salt chosen as . The equations of motion are integrated using a time step of .
DNA nanostars were prepared manually using the oxView software [51, 52] from the single-stranded DNA sequences (see Figure S1 in the SI). We placed DNA nanostars on an attractive flat substrate, modeled using a standard Lennard-Jones potential acting on all nucleotides with interaction strength . This potential acts only on the direction normal to the substrate, thus neglecting lateral friction from the substrate. In all simulations, the dimension normal to the substrate was fixed at . The number of stars was kept constant as , and for rigid and soft nanostars, respectively. The simulation box dimensions were chosen as for rigid nanostar system and for soft nanostar system. To vary the area fraction of DNA nanostars, we incrementally applied two-dimensional deformation to the simulation box, compressing both and by 1% at each step and then relaxing the system for 30.3 ns before the next deformation. Before measuring nanostar diffusivity, the system was further equilibrated for 303.0 ns at the target .
To determine the diffusion coefficient of DNA nanostars at different , MD simulations were performed for time steps, corresponding to s. The trajectory was sampled every time steps, yielding 1,000 frames for analyzing the in-plane mean-square displacement of the nanostars’ center-of-mass, . The long-time self-diffusion coefficient of the nanostars was then determined by fitting (Figure S8). We have repeated the simulations three times, starting from different initial configurations, to improve the statistics.
Data availability
The data that support the findings of this study are available from the corresponding authors upon reasonable request.
Author contributions
K. M.: Conceptualization (equal); data curation: (equal); formal analysis (lead); investigation (equal); methodology (equal); validation (equal); visualization (lead); writing – original draft (lead); writing – review and editing (equal).
T.Y.: Conceptualization (equal); data curation: (equal); formal analysis (supporting); investigation (equal); methodology (equal); software (lead); validation (equal); visualization (supporting); writing – review and editing (supporting).
M.Y.: Conceptualization (equal); funding acquisition (equal); project administration (equal); resources (equal); supervision (equal); writing – original draft (supporting); writing – review and editing (equal).
A.N.: Conceptualization (equal); funding acquisition (equal); project administration (equal); resources (equal); supervision (equal); visualization (supporting); writing – original draft (supporting); writing – review and editing (equal).
Conflicts of interest
There are no conflicts to declare.
Supporting information
The Supporting Information is available free of charge at [URL]. Section S1: Experimental determination of nanostar localization using the relative fluorescence intensity, ; Section S2: Estimation of the experimental packing fraction; Figure S1: DNA sequences of the rigid and soft nanostars; Figure S2: Representative fluorescence intensity profiles across the droplet center and schematic of the droplet cross section used to determine ; Figure S3: Dependence of on droplet radius; Figure S4: Representative FRAP recovery curves for freely diffusing nanostars in the droplet interior; Figure S5: Droplet-size dependence of the diffusion coefficient, , and mobile fraction, , of DNA nanostars and lipids; Figure S6: Experimental FRAP results obtained at different positions on the droplets; Figure S7: Estimation of instantaneous area of DNA nanostars; Figure S8: Mean-square displacement (MSD) of rigid and soft nanostars at different .

Figure S7. Representative DNA nanostar configurations and their projections onto the -plane (black points) for (a) rigid and (b) soft nanostars. The blue and pink lines indicate the outer boundaries of the nanostars, and is defined as the area enclosed by these boundaries.
Acknowledgements
We thank Takamasa Sakai for helpful discussions. This work was partially funded by the Japan Society for the Promotion of Science (JSPS) KAKENHI (grant numbers 22H01188, 24H02287 (M.Y.), the Japan Science and Technology Agency (JST) (grant numbers FOR-EST, JPMJFR213Y; CREST (JPMJCR22E1) (M. Y.)), the World-Leading Innovative Graduate Study Program for Advanced Basic Science Course (WINGS-ABC) at the University of Tokyo (K. M.), the Deutsche Forschungsgemeinschaft (DFG, German Research Foundation) through Project No. 470113688, and by the Klaus Tschira Foundation through Project No. 00.050.2024/ID 25347 (A. N.).
Supporting Information for Interfacial Packing of DNA Nanostars Regulates Interfacial Dynamics on Synthetic Cell Membranes
Kazutoshi Masuda; Takahiro Yokoyama; Miho Yanagisawa; Arash Nikoubashman
supplementary.pdf
Number of pages: 6
Number of figures: 8
DNA sequences
Estimation of the experimental and its dependence on
Because not all encapsulated nanostars necessarily localize at the membrane interface, we introduced an experimentally determined localization factor, , which represents the fraction of the total fluorescence signal associated with the membrane-adjacent region. As illustrated in Figure S2, was estimated from the fluorescence intensity profile together with the droplet geometry, and then used as a correction factor to experimentally estimate the interfacial packing fraction. The total fluorescence signal from a droplet can be expressed as the sum of the local fluorescence signals
where , , and are the mean fluorescence intensities of the whole droplet, the membrane-adjacent shell region, and the inner region, respectively. The corresponding volumes of those regions are
with and being the radii of the whole droplet and the inner region, respectively.
The membrane-associated fraction of fluorescence was then defined as
Since the confocal cross-section images are much thinner than the typical droplet radii, we neglect spherical curvature effects. Experimentally, was obtained as the mean pixel intensity of the inner region in two-dimensional cross-sectional fluorescence images after background subtraction. Because the shell region is thin, was calculated from the mean intensity of the entire cross section, , and the mean intensity of the inner region, . Assuming that the fluorescence intensity is uniform within the shell and inner regions, respectively, the total fluorescence intensity in the cross-sectional image is given by
Therefore, was calculated as
For rigid nanostars, increased approximately in proportion to only for large droplets and then reached a plateau at (), indicating saturation of membrane localization (fig:S3a). In contrast, for soft nanostars, increased almost linearly with over the entire droplet-size range examined, consistent with the expected surface-to-volume scaling of a spherical droplet (fig:S3b). This difference is consistent with the larger molecular footprint, , of the rigid nanostars, leading to membrane saturation in smaller numbers of nanostars than in soft nanostars.
Estimation of the experimental
The experimental packing fraction of DNA nanostars on the droplet membrane, was estimated from , , and . The number of membrane-associated nanostars was approximated as
where is Avogadro’s constant. The corresponding surface packing fraction was then calculated as
where is the projected area occupied by a single DNA nanostar on the membrane. When is expressed in , in , and in , this expression becomes
Additional experimental results
Estimation of instantaneous area of DNA nanostars
We characterized the size of the DNA nanostars by their instantaneous area, , defined determined from single-nanostar simulations. The value of was obtained by projecting the nucleotide positions onto the -plane, identifying the exterior nucleotides by Delaunay tessellation on the projected coordinates (Fig.S7), and then calculating the area enclosed by the resulting nanostar boundary. We took the average over different configurations for both rigid and soft DNA nanostars.
Mean-square displacement of DNA nanostars from MD simulations
References
- [1]Podolsky, K. A.; Devaraj, N. K. Synthesis of lipid membranes for artificial cells. Nat. Rev. Chem. 2021, 5, 676–694.
- [2]Yanagisawa, M.; Watanabe, C.; Yoshinaga, N.; Fujiwara, K. Cell-size space regulates the behavior of confined polymers: from nano-and micromaterials science to biology. Langmuir 2022, 38, 11811–11827.DOI
- [3]Yanagisawa, M.; Fujiwara, K. Beyond Simplifications: Overlooked Physics of Macromolecular Behaviors in Living Cells. Macromolecules 2025, 58, 9557–9566.DOI
- [4]Simovic, S.; Barnes, T. J.; Tan, A.; Prestridge, C. A. Assembling nanoparticle coatings to improve the drug delivery performance of lipid based colloids. Nanoscale 2012, 4, 1220–1230.
- [5]Umar, A. K.; Wathoni, N.; Zothantluanga, J. H.; Das, S.; Luckanagul, J. A.; others Liposome-polymer complex for drug delivery system and vaccine stabilization. Heliyon 2022, 8.DOI
- [6]Ramadurai, S.; Holt, A.; Krasnikov, V.; Van Den Bogaart, G.; Killian, J. A.; Poolman, B. Lateral diffusion of membrane proteins. J. Am. Chem. Soc. 2009, 131, 12650–12656.
- [7]Javanainen, M.; Hammaren, H.; Monticelli, L.; Jeon, J.-H.; Miettinen, M. S.; Martinez-Seara, H.; Metzler, R.; Vattulainen, I. Anomalous and normal diffusion of proteins and lipids in crowded lipid membranes. Faraday Discuss. 2012, 161, 397–417.
- [8]Houser, J. R.; Busch, D. J.; Bell, D. R.; Li, B.; Ren, P.; Stachowiak, J. C. The impact of physiological crowding on the diffusivity of membrane bound proteins. Soft Matter 2016, 12, 2127–2134.DOI
- [9]O’hern, C. S.; Silbert, L. E.; Liu, A. J.; Nagel, S. R. Jamming at zero temperature and zero applied stress: The epitome of disorder. Phys. Rev. E 2003, 68, 011306.
- [10]Horton, M. R.; Höfling, F.; Rädler, J. O.; Franosch, T. Development of anomalous diffusion among crowding proteins. Soft Matter 2010, 6, 2648–2656.DOI
- [11]Liu, P.; Beltramo, P. J. Effects of crowding on the diffusivity of membrane adhered particles. Soft Matter 2023, 19, 7708–7716.DOI
- [12]Hong, F.; Zhang, F.; Liu, Y.; Yan, H. DNA origami: scaffolds for creating higher order structures. Chem. Rev. 2017, 117, 12584–12640.DOI
- [13]Lee, J.; Kim, J.; Posnjak, G.; Ershova, A.; Hayakawa, D.; Shih, W. M.; Rogers, W. B.; Ke, Y.; Liedl, T.; Lee, S. DNA origami colloidal crystals: opportunities and challenges. Nano Lett. 2024, 25, 16–27.DOI
- [14]Maier, B.; Rädler, J. O. DNA on fluid membranes: a model polymer in two dimensions. Macromolecules 2000, 33, 7185–7194.DOI
- [15]Herold, C.; Schwille, P.; Petrov, E. P. DNA condensation at freestanding cationic lipid bilayers. Phys. Rev. Lett. 2010, 104, 148102.
- [16]Langecker, M.; Arnaut, V.; List, J.; Simmel, F. C. DNA nanostructures interacting with lipid bilayer membranes. Acc. Chem. Res. 2014, 47, 1807–1815.DOI
- [17]Zhan, P.; Peil, A.; Jiang, Q.; Wang, D.; Mousavi, S.; Xiong, Q.; Shen, Q.; Shang, Y.; Ding, B.; Lin, C. et al. Recent advances in DNA origami-engineered nanomaterials and applications. Chem. Rev. 2023, 123, 3976–4050.DOI
- [18]Samanta, A.; Baranda Pellejero, L.; Masukawa, M.; Walther, A. DNA-empowered synthetic cells as minimalistic life forms. Nat. Rev. Chem. 2024, 8, 454–470.DOI
- [19]Fang, Y.; Chen, X.; Qi, Q.; Lin, M.; Liang, Q. DNA Origami and Its Applications in Synthetic Biology. Adv. Sci. 2026, 13, e13357.
- [20]Masuda, K.; Yanagisawa, M. Programming Nonlinear Interfacial Mechanics of Synthetic Cells: Lipid Geometry and DNA Nanostructures. Small Science 2026, 6, e70321.
- [21]Morzy, D.; Rubio-Sánchez, R.; Joshi, H.; Aksimntiev, A.; Di Michele, L.; Keyser, U. F. Cations regulate membrane attachment and functionality of DNA nanostructures. J. Am. Chem. Soc. 2021, 143, 7358.DOI
- [22]Morzy, D.; Tekin, C.; Caroprese, V.; Rubio-Sánchez, R.; Di Michele, L.; Bastings, M. M. Interplay of the mechanical and structural properties of DNA nanostructures determines their electrostatic interactions with lipid membranes. Nanoscale 2023, 15, 2849–2859.
- [23]Avakyan, N.; Conway, J. W.; Sleiman, H. F. Long-range ordering of blunt-ended DNA tiles on supported lipid bilayers. J. Am. Chem. Soc. 2017, 139, 12027–12034.
- [24]Dong, Y.; Yao, C.; Zhu, Y.; Yang, L.; Luo, D.; Yang, D. DNA functional materials assembled from branched DNA: design, synthesis, and applications. Chem. Rev. 2020, 120, 9420–9481.
- [25]Sato, Y.; Sakamoto, T.; Takinoue, M. Sequence-based engineering of dynamic functions of micrometer-sized DNA droplets. Sci. Adv. 2020, 6, eaba3471.
- [26]Sato, Y.; Takinoue, M. Pioneering artificial cell-like structures with DNA nanotechnology-based liquid-liquid phase separation. BPPB 2024, 21, e210010.
- [27]Kurokawa, C.; Fujiwara, K.; Morita, M.; Kawamata, I.; Kawagishi, Y.; Sakai, A.; Murayama, Y.; Nomura, S.-i. M.; Murata, S.; Takinoue, M. et al. DNA cytoskeleton for stabilizing artificial cells. Proc. Natl. Acad. Sci. U. S. A. 2017, 114, 7228–7233.
- [28]Cherstvy, A. G.; Petrov, E. P. Modeling DNA condensation on freestanding cationic lipid membranes. Phys. Chem. Chem. Phys. 2014, 16, 2020–2037.
- [29]Rand, R.; Parsegaian, V. Hydration forces between phospholipid bilayers. Biochim. Biophys. Acta - Biomembr. 1989, 988, 351–376.
- [30]Pera, I.; Stark, R.; Kappl, M.; Butt, H.-J.; Benfenati, F. Using the atomic force microscope to study the interaction between two solid supported lipid bilayers and the influence of synapsin I. Biophys. J. 2004, 87, 2446–2455.
- [31]Simon, S.; Fink, C.; Kenworthy, A.; McIntosh, T. The hydration pressure between lipid bilayers. Comparison of measurements using x-ray diffraction and calorimetry. Biophys. J. 1991, 59, 538–546.
- [32]Mukhina, T.; Hemmerle, A.; Rondelli, V.; Gerelli, Y.; Fragneto, G.; Daillant, J.; Charitat, T. Attractive interaction between fully charged lipid bilayers in a strongly confined geometry. J. Phys. Chem. Lett. 2019, 10, 7195–7199.
- [33]Marrink, S. J.; Berkowitz, M.; Berendsen, H. J. Molecular dynamics simulation of a membrane/water interface: the ordering of water and its relation to the hydration force. Langmuir 1993, 9, 3122–3131.
- [34]Zhao, W.; Gurtovenko, A. A.; Vattulainen, I.; Karttunen, M. Cationic dimyristoylphosphatidylcholine and dioleoyloxytrimethylammonium propane lipid bilayers: Atomistic insight for structure and dynamics. J. Phys. Chem. B 2012, 116, 269–276.
- [35]Gao, Y.; Li, M.; Chen, B.; Shen, Z.; Guo, P.; Wientjes, M. G.; Au, J. L.-S. Predictive Models of Diffusive Nanoparticle Transport in 3-Dimensional Tumor Cell Spheroids. AAPS J. 2013, 15, 816–831.
- [36]Manning, G. S. Limiting laws and counterion condensation in polyelectrolyte solutions I. Colligative properties. J. Chem. Phys. 1969, 51, 924–933.
- [37]Miatmoko, A.; Asmoro, F. H.; Azhari, A. A.; Rosita, N.; Huang, C.-S. The effect of 1, 2-dioleoyl-3-trimethylammonium propane (DOTAP) Addition on the physical characteristics of β-ionone liposomes. Sci. Rep. 2023, 13, 4324.
- [38]Measurement and interpretation of electrokinetic phenomena. J. Colloid Interface Sci. 2007, 309, 194–224.
- [39]Zheng, H.; Wang, D.; Barés, J.; Behringer, R. Jamming by compressing a system of granular crosses. EPJ Web Conf. 2017; p 06014.
- [40]Stannarius, R.; Schulze, J. On regular and random two-dimensional packing of crosses. Granul. Matter 2022, 24, 25.
- [41]Maier, B.; Rädler, J. O. Conformation and self-diffusion of single DNA molecules confined to two dimensions. Phys. Rev. Lett. 1999, 82, 1911.
- [42]Jeon, J.-H.; Javanainen, M.; Martinez-Seara, H.; Metzler, R.; Vattulainen, I. Protein crowding in lipid bilayers gives rise to non-Gaussian anomalous lateral diffusion of phospholipids and proteins. Phys. Rev. X 2016, 6, 021006.
- [43]He, Y.; Chen, Y.; Liu, H.; Ribbe, A. E.; Mao, C. Self-assembly of hexagonal DNA two-dimensional (2D) arrays. J. Am. Chem. Soc. 2005, 127, 12202–12203.
- [44]Sun, X.; Hyeon Ko, S.; Zhang, C.; Ribbe, A. E.; Mao, C. Surface-mediated DNA self-assembly. J. Am. Chem. Soc. 2009, 131, 13248–13249.
- [45]Axelrod, D.; Koppel, D.; Schlessinger, J.; Elson, E.; Webb, W. W. Mobility measurement by analysis of fluorescence photobleaching recovery kinetics. Biophys. J. 1976, 16, 1055–1069.
- [46]Poppleton, E.; Matthies, M.; Mandal, D.; Romano, F.; Šulc, P.; Rovigatti, L. oxDNA: coarse-grained simulations of nucleic acids made simple. J. Open Source Softw. 2023, 8, 4693.
- [47]Snodin, B. E. K.; Randisi, F.; Mosayebi, M.; Šulc, P.; Schreck, J. S.; Romano, F.; Ouldridge, T. E.; Tsukanov, R.; Nir, E.; Louis, A. A. et al. Introducing improved structural properties and salt dependence into a coarse-grained model of DNA. J. Chem. Phys. 2015, 142, 234901.
- [48]Ouldridge, T. E.; Louis, A. A.; Doye, J. P. K. Structural, mechanical, and thermodynamic properties of a coarse-grained DNA model. J. Chem. Phys. 2011, 134, 085101.
- [49]Šulc, P.; Romano, F.; Ouldridge, T. E.; Rovigatti, L.; Doye, J. P. K.; Louis, A. A. Sequence-dependent thermodynamics of a coarse-grained DNA model. J. Chem. Phys. 2012, 137, 135101.
- [50]Rovigatti, L.; Šulc, P.; Reguly, I. Z.; Romano, F. A comparison between parallelization approaches in molecular dynamics simulations on GPUs. J. Comput. Chem. 2015, 36, 1–8.
- [51]Poppleton, E.; Bohlin, J.; Matthies, M.; Sharma, S.; Zhang, F.; Šulc, P. Design, optimization and analysis of large DNA and RNA nanostructures through interactive visualization, editing and molecular simulation. Nucleic Acids Res. 2020, 48, e72–e72.
- [52]Bohlin, J.; Matthies, M.; Poppleton, E.; Procyk, J.; Mallya, A.; Yan, H.; Šulc, P. Design and simulation of DNA, RNA and hybrid protein–nucleic acid nanostructures with oxView. Nat. Protoc. 2022, 17, 1762–1788.






