Self-Assembly of Geometry-Based DNA Origami-Histone Protein Hybrid Nanostructures for Constructing Rationally-Designed Higher-Order Structures

The emergence of hybrid DNA−protein hybrid nanostructures in recent years has expanded the application of DNA nanotechnology. Previous studies reported the integration of proteins into DNA nanostructures by sequence-dependent interactions or chemical modifications of DNA, which limit the design flexibility of hybrid nanostructures. Here we report the construction of sequence-independent geometry-based DNA−protein hybrid nanostructures using the self-assembly of single-stranded (ss) DNA and histone proteins. We demonstrate that nucleosome-like ssDNA-histone complexes are integrated into various shapes of DNA origami at specific predefined locations. We further show the possibility of using these hybrid nanostructures as the building blocks for designing more complex nanostructures. Our finding would facilitate the development of geometry-based sequence-independent predesigned DNA−protein hybrid nanostructures for more flexible and straightforward construction of higher-order structures.


■ INTRODUCTION
Biological nanomaterials are defined as materials with size within the nanoscale range of 1−100 nm in at least one dimension, including antibodies, nucleic acids, lipids, hemoglobin, and viruses. These materials are biocompatible and biodegradable and are often capable of self-assembly due to their inherent characteristics. 1 DNA nanostructures, an example of a biological nanomaterial, are artificial nanoscale structures composed of DNA. In addition to their advantage of small size, the programmability of the four building blocks of DNA provides a tool to synthetically design various DNA shapes and vehicles using long single-stranded DNA (ssDNA) and linker staple strands. 2−4 However, this programmability limits the complexity of the constructs that can be constructed using DNA. This prompted the idea of incorporating proteins into nanostructures (i.e., hybrid DNA−protein nanostructures) to add structural flexibility and functional variations.
DNA nanotechnology was enriched by integrating proteins using specific sequences 5−7 and nonsequence-based 8−10 interactions between DNA nanostructures and proteins. Hybrid DNA−protein nanostructures contribute to the advancement of applications of DNA nanostructures by serving as functional, 11,12 spatial, 13,14 and structural 8,10 platforms in bioimaging 15,16 and in biological molecule and drug delivery. 17,18 Several proteins were successfully integrated with DNA nanostructures by DNA sequence-dependent interac-tions 5,7 or chemical modifications of DNA. 8,10 However, the self-assembly of DNA−protein hybrid nanostructures that do not rely on these mechanisms has been challenging. Sequenceindependent geometry-based DNA−protein hybrid nanostructures would expand their application since this allows for more straightforward assembly and a more flexible design of hybrid nanostructures.
We recently found a spontaneous sequence-independent assembly of single-stranded (ss) DNA and histone proteins into nucleosome-like hybrid nanostructures (HDs), 19−21 which indicated the possibility of designing a geometry-based assembly of DNA−protein hybrid nanostructures using this interaction. 20,21 Here we report the successful integration of the HD with varied shapes of two-dimensional (2D) DNA nanostructures at specific locations by conjugating ssDNA sticky ends to DNA origamis. We also show that the obtained nanostructures serve as building blocks for designing and constructing complex nanostructures. In this study, we characterized the integration of self-assembled HD into three different custom-designed DNA origamis at specific locations ( Figure 1a); a multilayer cuboid DNA origami with an ssDNA overhang end at the edge (CDO), multilayer cuboid with aperture DNA origami with two ssDNA overhangs at the opposite sides of the open aperture (CADO), and single-layer square DNA origami with an ssDNA overhang at the edge (SDO). These DNA origamis allow us to investigate the effect of the spatial positions of the HD, binding modes of the HD to the DNA origamis, and the thickness of DNA origamis on the formation of the hybrid nanostructures. We characterized assembled nanostructures using atomic force microscopy (AFM), cryogenic electron microscopy (Cryo-EM), dynamic light scattering, and agarose gel electrophoresis (see Materials and Methods section). We confirmed the structural integrity of the DNA origamis throughout the formation of DNA−protein hybrid nanostructures by AFM or agarose gel electrophoresis (Figure 1a, Figure S1). Our findings indicate that geometrybased DNA-histone protein hybrid nanostructures are wellsuited for the flexible and straightforward construction of higher-order nanostructures. ■ RESULTS AND DISCUSSION Experimental Design. In order to construct hybrid nanostructures of DNA origamis and HDs, we developed a protocol for a sequential assembly of the HDs and their integration into the DNA origamis ( Figure 1b). Our previous study demonstrated that each self-assembled HD consists of a histone octamer and a total of 300-nucleotide (nt) ssDNA. 19 An ssDNA sticky end (25 nt) also needed to be added to integrate the HD into the DNA origami. Thus, in this study, we used two 175-nt ssDNA that consist of 150-nt polythymine (polyT) for the assembly of the HDs and a 25-nt sticky end (see Materials and Methods section for the exact ssDNA sequence). The interaction of DNA with histones is typically described as nonsequence-specific, 22 allowing for the assembly of HD complexes using any DNA sequence. However, to minimize the formation of secondary structures arising from intrachain self-stacking during self-assembly at 4°C, polyT ssDNA sequences were used. PolyT ssDNA has the advantage of weaker self-stacking compared to other repetitive DNA sequences such as polyguanine, polycytosine, or polyadenine, which can fold into non-B-DNA structures such as quadruplexes, i-tetraplexes, or A-motifs, respectively. 23 Furthermore, we used H1-depleted native calf thymus core histones (H2A, H2B, H3, and H4, see Materials and Methods section for the purification procedures) rather than recombinant histones for the assembly of the HDs, as our previous study showed that post-translational modifications of the histone proteins are essential for the efficient self-assembly of the HDs. 19 The DNA origamis (CDO, CADO, and SDO) were modified with one or two 25-nt protruding handles complementary to the sticky ends of the 175-nt ssDNA (see Materials and Methods for the exact sequence of the 25-nt handles).
Optimization of the Assembly Conditions of Histone-DNA Complexes. To assemble the histone proteins and 175nt ssDNAs into the HDs, we mixed the histones and ssDNA at a 24:1 molar ratio (Figure 2a, see Materials and Methods for the detailed procedures of the assembly of the HDs). The ratio (i.e., 24:1) was determined using the fluorescence quenching of Cy5 dyes conjugated to ssDNA (i.e., the Cy5 fluorescence was most quenched at the optimum molar ratio to form an ssDNA-histone nucleosome-like structure). 19 This ratio is denoted by the quenching cutoff ratio (QCR, see Figure S2). While we were able to experimentally determine the QCR for a varied length of Cy5-ssDNA up to 100-mer, it was impossible to experimentally determine the QCR for a Cy5-ssDNA longer than 100-mer because fluorescence quenching of the Cy5 dyes did not occur upon the assembly of the nucleosome-like structure (i.e., number of Cy5 dyes in the complex is too small to cause fluorescence self-quenching). We found that the QCR increases exponentially with the length of ssDNA. Thus, we extrapolated the QCR for 175-mer ssDNA that we used to assemble the histone-ssDNA complex in this study. The QCR for the 175-nt ssDNA was extrapolated from the selfquenching result obtained for a Cy5-conjugated 50-nt ssDNA. 19 A dry phase AFM image of the histone proteins ( Figure 2b) showed that the majority of the histone proteins have a height of around 3 nm (Figure 2c). On the other hand, an AFM image of the mixture of the ssDNA and histones clearly showed larger particles (Figure 2d). The mean height of the particles (3.8 nm, Figure 2e) agrees well with the previously reported height of nucleosome 21 and nucleosome-like histone-ssDNA complexes 19 determined by dry phase AFM.
We found the highest efficiency of the formation of the HDs at a low salt concentration (1 mM Tris base, Figure S3). 19 The formation of the HDs at higher salt concentrations (200−400 mM Mg 2+ ) was inefficient ( Figure S3), which resulted in the inefficient formation of the HD-DNA origami complexes (see below, Figure S4). A similar salt effect was previously reported for nucleosome particles (i.e., a complex of double-stranded DNA and histone proteins). 24,25 The Assembly of DNA Origami−HDs Nanostructures. We next integrated the HDs into DNA origami. To assemble the complexes, we mixed the DNA origamis and the HDs at a 1:1 or 1:2 molar ratio (see Materials and Methods for the detailed procedures of the assembly of the complexes). The AFM imaging of the study was conducted in the liquid phase to preserve the native structures and eliminate the possibility of the coexistence of DNA origami with the histone due to the drying effect. Considering the small lateral size of the histone complex (10−15 nm), typically a scan size of 500 nm × 500 nm is necessary to properly visualize the structure. In this study, AFM images were captured in a 500 nm × 500 nm frame size to allow us to clearly distinguish between the HDs complex bound to the DNA origami and the coexistence of free HDs near the DNA origami ( Figure S5). Figure 1a (bottom right) shows a liquid-phase AFM image of the 2.6 nm thick rectangular shape SDO. Upon the assembly of the HDs and SDOs, we observed the appearance of a 2.5 nm-height protruding structure at the center of the short side edge of SDOs (arrowheads in Figure 3a,b). The locations of the protruding structures exactly match the position of the 25-nt protruding handles conjugated to the SDO (Figure 1b). The obtained AFM images clearly demonstrate that the protruding structures observed for the HD-SDO complex (Figure 3a,b) are absent in SDO ( Figure S6). The height of the protruding structure agrees well with that of HDs and nucleosome particles measured by liquid-phase AFM ( Figure S7). These results exclude the possibility that the 25-nt ssDNA attached to SDO appears in the AFM images that we captured in this study. Note that measurements with AFM regularly show a reduction of the heights due to the interaction of the probe with the sample. 26 Together, these results strongly suggest that the HD is integrated into the SDO at the preprogramed position through the interaction between the 25-nt complementary strands. The result also indicates that only one 25-nt sticky end on the HD binds to the SDO, leaving another sticky end unbound (see below). The protruding structures were observed in 66% of the SDOs (92/140, Figures 3c, S8) upon the addition of the HDs, demonstrating the efficient formation of the HD-SDO complex. We note that the small fraction of  the bound SDO observed without HDs is due to structural defects of the DNA origami, which are difficult to distinguish from HD-SDO complexes. Additionally, we ran the SDO assembled with HDs on gel electrophoresis ( Figure S9). The band appears to be shifted on the gel after the addition of HDs, which indicates a higher molecular weight of the structure in comparison to that of SDO alone. This corroborates the results obtained from the AFM imaging, where protruding structures were observed on the SDO upon the addition of HDs, and is consistent with the successful integration of HDs into the DNA origami structures at the preprogrammed positions through the interaction between the 25-nt complementary strands, as revealed in the AFM images. The presence of the shifted band in the gel electrophoresis provides additional evidence supporting the efficient formation of the HD-SDO complex and reinforces the validity of the findings obtained in the study.
We also estimated the percentage of CDO complexed with the HD using a gel shift assay ( Figures S9 and S10). We observed a shift and broadening of the band upon mixing CDO and the HDs. While the intensity profile of the CDO (i.e., DNA origami only) band can be fitted nicely to a Gaussian function (Figure S10a), the fitting of the CDO + HDs band needed two Gaussian components ( Figure S10b), which suggests the presence of two components in this sample (i.e., CDO-HD and CDO). The relative contribution of the high molecular weight band (i.e., CDO-HD complex, 67%) agrees very well with the percentage of SDO that is complexed with the HDs estimated by the AFM experiments (Figure 3c).
These results strongly suggest that the HDs are efficiently integrated into DNA origami. Note that a split of the band in the gel assay is not expected for the CDO + HD sample because the molecular weight difference between CDO and CDO-HD complex is very small (approximately 2%, see Supporting Information).
DNA Origami-HD Assembly in a Confined Space. The feasibility of integrating the HDs in a more spatially confined location of DNA origami was investigated using 8 nm thick cuboid DNA origami with a 9 × 15 nm size open aperture (CADO) (Figures 1a and 4a). The size of the aperture is just enough to accommodate an HD. AFM image of the CADOs (Figure 1a bottom middle) confirmed that the height profile of CADO is accurately measured. Upon the assembly of the HDs and CADOs, we clearly observed a spot at the aperture in the captured AFM images (Figure 4b,c), which is absent in the AFM images of the CADOs. We also observed the change in the height profile and the intensity at the aperture upon assembly of the HDs and CADOs (Figure 4d). The analysis of the height profiles of a large number of CADOs and HD-CADOs (approximately 30 samples) demonstrated a statistically significant change upon the assembly (Figure 4e). The dynamic light scattering experiment on CADO and HD-CADO showed changes in the size and zeta potential upon the assembly ( Figure S11). All of these results strongly indicate the integration of the HDs into the aperture of CADOs through the interaction between the two 25-nt protruding handles conjugated to the aperture and the 25-nt sticky ends on the HDs. An HD in the aperture was directly visualized using immunoelectron microscopy by attaching 5−10 nm size gold nanoparticles to H2B or H3 histones (see Materials and Methods for the details of sample preparation). A cryo-EM micrograph of the HD-CADO complex immunolabeled against H2B histone shows that a high contrast spot of the gold nanoparticle colocalizes with CADO at the preprogramed position (i.e., center aperture, Figure 4f). We also observed a similar colocalization for the H3 histone-labeled sample ( Figure S12). These results demonstrate that at least one heterotopic tetramer composed of H2A-H2B bound to H3− H4 was localized in the center aperture. We note that a statistical analysis of the HD-CADO complex formation and the spatial colocalization of H2B and H3 based on the cryo-EM data was impossible because the high-throughput cryo-EM imaging experiment was technically challenging. Nevertheless, the cryo-EM images and the statistical analysis of the AFM images strongly suggest the successful integration of the HDs into CADOs at the spatially confined preprogramed position.

Geometry-Based Direct Assembly of DNA Origami-Histone Proteins into Hybrid Nanostructures.
We showed in the previous sections that the HDs can be integrated into the various shapes of DNA origami (i.e., SDO, CDO, and CADO) at preprogramed positions. Although the histone proteins and ssDNA self-assembled into the HDs in a sequence-independent geometry-based manner, the HDs were integrated into the DNA origami by annealing the 25-nt handles on the DNA origami and the ssDNA on the HDs (i.e., sequence-dependent assembly). Thus, we attempted to construct fully geometry-based, self-assembled hybrid nanostructures of DNA origami and histone proteins.
We preassembled DNA-origami and 175-mer ssDNA, which was followed by geometry-based assembly of ssDNA (that is a part of DNA origami) and histone proteins (i.e., DNA−protein complexes were self-assembled through a pure geometry-based interaction). In this protocol, the annealing of the handles on SDO and the 175-mer ssDNA complementary to the handles were initiated at a high annealing temperature (50°C) to avoid nonspecific secondary structures prior to assembling the HDs complexes. We found that 35% of SDO formed SDO-HD complexes through this pure geometry-based interaction ( Figures 5, S13). While the efficiency of the SDO-HD complex formation is not as high as that of the original assembling method, probably due to the nonspecific electrostatic attraction between the histone and DNA, we unambiguously demonstrated the fully geometry-based self-assembly of DNA origami and HDs.
To pave the way for future applications, we investigated the stability of the assembled DNA−protein hybrid nanostructures. We found that more than 80% of the assembled nanostructures maintained their integrity for 1 month under suitable storage conditions (4°C in the dark, Figures 6 and  S14).
Higher-Complexity DNA−Protein Hybrid Nanostructures. As stated above, a 25-nt sticky end is left unassociated when HD forms the complex with SDO that has only one protruding handle at the edge, indicating the possibility of constructing more complex HD-DNA origami self-assembled nanostructures. We tested this hypothesis by assembling HDs and 14 nm thick 28 × 61 nm size cuboid DNA origami that has a 25-nt handle at the edge (CDO, Figure 7a). We used CDO for this experiment because of its higher structural rigidity compared with SDO. We assembled HDs and CDOs in a way similar to that of SDO and CADO. While the presence of the HDs in the complexes is difficult to judge from the AFM image due to the large height difference between the HD and CDO (Figure 7b), the cryo-EM images of the immunolabeled sample clearly showed the colocalization of HD and CDO at the programmed position similar to the HD-SDO complex ( Figure  7b). Their colocalization was also suggested by the shift of the molecular weight in the agarose gel electrophoresis upon the assembly (Figures S9, S10). Interestingly, we also observed

ACS Applied Nano Materials www.acsanm.org
Article head-to-head dimers of CDO mediated by an HD (Figure 7c). Given the position of the protruding handle on the CDO and the two sticky ends on the HD, the observation strongly suggests that the two CDOs are bridged by the HD. This finding suggests that the self-assembled histone-ssDNA complexes can be used as building blocks of more complex DNA−protein hybrid nanostructures.

■ CONCLUSIONS
We presented in this study the self-assembled DNA−protein hybrid nanostructures guided by the geometry-based selfassembly of histone proteins and ssDNAs. Our approach is independent of DNA sequences and does not require any chemical modifications of DNA. Therefore, it could provide greater flexibility to the design and construction of higherorder DNA−protein hybrid nanostructures. Our results demonstrated that the structures of the complexes are strictly controlled by the locations of the ssDNA handles conjugated to the DNA origamis and the ssDNA sticky ends on the nucleosome-like histone-ssDNA complex (i.e., HD) and that the proper arrangement of these components enables the construction of more complex structures (e.g., head-to-head dimer). Even more complex nanostructures could be obtained using the building block that we developed in this study by alternating the locations and numbers of protruding handles conjugated to DNA origamis and by changing the total number of sticky ends on HDs via the self-assembly of HDs with different lengths of ssDNAs. For example, longer handles (i.e., the dsDNA arms formed between the DNA−protein hybrid nanostructures) are expected to form more loose structures, whereas shorter ssDNA strands (i.e., a larger number of sticky ends on HDs) would result in the assembly of a larger number of DNA origami on HDs. Our findings would facilitate the development of new rational, geometry-based predesigned complex DNA−protein hybrid nanostructures. In addition to the application in three-dimensional (3D) nanostructures, the hybrid nanostructures we developed in this study could be applicable to a biosensor for histone modification, because histone acetylation weakens the electrostatic interaction between the histone proteins and DNA and, thus, could be detected by measuring the efficiency of the assembly of ssDNA-histones complexes. 27 The cell membrane permeability of nucleosomes implies the potential application of the DNA− protein hybrid nanostructures reported in this study for drug delivery by designing hybrid enclosures as vectors. 28 The hybrid nanostructures reported in this study may also be useful for studying the structural organization and dynamics of nucleosomes. 29,30 ■ MATERIALS AND METHODS Purification of Core Histone. The acid extraction method was used to purify the calf thymus core histones (Sigma-Aldrich). 4% perchloric acid (Sigma-Aldrich) was used to precipitate core histones and then stored overnight at 4°C. The stored core histones were then centrifuged at a maximum speed of 15 000 rpm for 1 h at 4°C. After decanting the supernatant, the pellet was washed twice with 1 mL of 4% perchloric acid and centrifuged at a maximum speed of 15 000 rpm for 5 min at 4°C. After decanting the supernatant after each washing step, the core histone was washed twice with 0.2% HCl in acetone. Followed by washing the core histone twice with 100% acetone under the same conditions. A total of six washes were performed, and the pellet was air-dried at room temperature for 20 min before being resuspended in ultrapure water. The resuspension was transferred to a spin filter (Sartorius -VIVASPIN2 5000 MWCO) and washed with cold distilled water seven times (centrifugation at 15 000 rpm, 1 h, 4°C). Excess salt was removed from the suspension, and the core histones were concentrated during the centrifugation. UV−vis absorption spectra were measured by using a spectrophotometer (U-3900, Hitachi).
The sequences used in the core-histones/ssDNA are as follows: 50mer ssDNA: 5′/Cy5/TTTTTTTTTTTTTTTTTTTTTTTTTTTT-TTTTTTTTTTTTTTTTTTTTTT 3′. 175-mer ssDNA: 5′  T T T T T T T T T T T T T T T T T T T T T T T T T T T T T T T T T T T -TTTTTTTTTTTTTTTTTTTTTTTTTTTTTTTTTTTTTT  TTTTTTTTTTTTTTTTTTTTTTTTTTTTTTTTTTTTTTT-T T T T T T T T T T T T T T T T T T T T T T T T T T T T T T T T T T  TTTTCTTTCTTTCTTCTCTTCTTCTTCTT 3′. DNA Nanostructure. DNA origamis (Tilibit Nanosystems) were modified with 25-nucleotide handles, the sequence as follows: 5′ AAGAAGAAGAAGAGAAGAAAGAAAG 3′. Dilution of DNA origamis to a final concentration of 5−20 nM was achieved using sample buffer (1 × 5 mM Tris-base, 1 mM ethylenediaminetetraacetic ACS Applied Nano Materials www.acsanm.org Article acid (EDTA), 5 mM NaCl, and 5 mM MgCl 2 dissolved in ultrapure water). Self-Assembly of DNA Origami−HD Hybrid Nanostructures. Following the self-assembly of HD nanostructures, the reaction mixture (DNA origami and HD nanoparticles) was incubated at 4°C for 1 h. The ratio of 1:1−1:2 DNA origami/core histones was used in all of the experiments.
Geometry-Based Self-Assembly of DNA Origami−HD Hybrid Nanostructures. Annealing of the handle of DNA origami and ssDNA was performed at 50°C prior to adding the histone protein. The ratio of 1:2 DNA origami/core histones was used in the experiments.
Agarose Gel Electrophoresis. Agarose gels (2%) were prepared by dissolving 2 g of agarose (GoldBio agarose LE) in 100 mL of 1× TAE (Omega Bio-Tek) buffer. TAE buffer (1X) supplemented with 6 mM MgAc 2 (Sigma-Aldrich) was used as the running buffer. The gel was stained using an ethidium bromide solution (0.025 μg mL −1 ). All the samples were supplemented with 10 mM MgAc 2 before adding the gel loading dye purple (New England Biolabs). A 5 μL aliquot of 1 kb DNA ladder (GoldBio) was used as a reference, and the wells were loaded with 10 μL of the samples (10 nM). The gels were run with a constant voltage of 90 V for 2 h and adjusted to 20 V for 24 h upon loading DNA origami-HD nanoparticles hybrid nanostructures. Image of the gel was acquired using the Vilber E-Box gel imaging system.
AFM Imaging. For imaging in a dry state, 10 μL of the DNA origami sample (5 nM) and AFM buffer (5 mM Tris-base, 1 mM EDTA, 5 mM NaCl, 5 mM MgCl 2 , and 1 mM of NiCl 2 ) was deposited onto a freshly cleaved mica surface. After 10 min of incubation, the mica surface was rinsed three times with 200 μL of water and dried overnight in a desiccator. AFM measurements were performed using a Bruker Dimension Icon AFM in intermittent contact mode with an FESPA-V2 probe (Bruker) and cantilevers with a frequency of 50−100 kHz, and a spring constant of 2.8 N m −1 was used. Images were collected with a scan size of 0.5 × (0.5−2) × 2 μm 2 with a resolution of 256 × 256 pixels. To avoid highly aggregated HD nanoparticles, the samples were centrifuged for 15 min, and supernatants were imaged.
For AFM imaging in liquid, 10 μL of the samples was deposited onto freshly cleaved mica surfaces within a liquid cell. After 10 min of incubation, the liquid cell was filled with 0.5 mL of 0.5X TBE buffer containing 200 MgAc 2 and 40 mM NiCl 2 . For liquid AFM imaging of histone protein, 0.5X TBE buffer was used without any supplementation. AFM measurements were performed using Bruker Dimension Icon AFM in ScanAsyst mode with a SCANASYST-FLUID+ probe (Bruker), and cantilevers with a frequency of 100− 200 kHz and a spring constant of 0.7 N m −1 were used. Images were collected with a scan size of 0.5 × (0.5−2) × 2 μm 2 with a resolution of 256 × 256 pixels. All of the images were refined using Gwyddion software. Particle heights were determined manually or by using a script written in Matlab. For imaging histone monomers in the liquid, the mica surface was treated with air plasma.
Cryo-EM. Samples at a concentration of 20 nM were used. The reaction mixtures of DNA origami and HD nanoparticles were incubated at 4°C for 1 h at a ratio of 1:1−1:2 DNA origami/core histones. A small volume of the sample was deposited onto holey carbon grids and blotted with filter papers (for 3 s). The grid was plunged into liquid ethane using a Vitrobot (FEI) instrument to preserve the hydrated sample in a thin vitreous ice layer. A side-entrytype cryo-transfer specimen holder (Gatan 626) was used to load the grid into the TEM instrument. All sample preparation steps were carried out at the temperature of liquid nitrogen.
Dynamic Light Scattering. The size distribution and zeta potential of DNA origami and DNA origami-HD nanoparticles were measured with a DLS analyzer (Malvern Zetasizer Nano ZS). A 50 μL sample was used to measure the size distribution at 25°C. The same sample was diluted in water to 1 mL for the zeta-potential measurement.
The effect of small molecular weight of histone on the overall molecular weight of DNA origami (Text S1), stability of the CADO and CDO DNA origamis evaluated by agarose gel electrophoresis ( Figure S1), optimization of the mixing ratio of histones and ssDNA ( Figure S2