Self-Assembly of Rolling Circle Amplification DNA into Nano-particles using Mg²⁺ and ppi

1. INTRODUCTION

DNA exhibits excellent biocompatibility and biodegradability [1]. Therefore, DNA has been widely used as a next-generation biomaterial in various scientific and industrial fields. As highlighted by Dey et al., DNA origami technologies have progressed over the past few decades from one- and two-dimensional structures to complex three-dimensional architectures with precise spatial control and specific functionalities [2]. However, the DNA components of DNA origami have been artificially manufactured. Therefore, the DNA origami structure yield and cost remain challenges for commercialization.

DNA synthesis is a complex and chemically demanding [3]. One of the most well-established methods for DNA synthesis is phosphoramidite chemistry, which is based on solid-phase synthesis. This process involves removing the 5'-DMT protecting group, coupling an activated nucleotide, capping the unreacted sites, and oxidizing the linkage to form a stable phosphate bond. This four-step cycle is repeated for each nucleotide addition step. All protecting groups are removed to obtain the final DNA strand. This requires expensive reagents, precise control, and rigorous purification steps, resulting in high production costs and limitations for large-scale manufacturing [4].

Instead of synthesizing DNA, the DNA amplification principle is utilized to building three-dimensional DNA based structure by inserting specific sequences. Similar to the aptamer sequence, amplified DNA can have specific functions based on the designed sequences [5]. For example, Yang et al. detected platelet-derived growth factor (PDGF) at concentrations below 1 nM using a conformation-switching aptamer that formed a circular conformation in the presence of the target protein, enabling ligation with T4 DNA ligase [6]. Another strategy involves loop-mediated isothermal amplification (LAMP), particularly using Bst DNA polymerase, as a point-of-care diagnostic tool. Taylor et al. highlighted its applications, such as RT-LAMP for SARS-CoV-2 detection, colorimetric readouts, and smartphone-based platforms, demonstrating its potential as a rapid and low-cost alternative to PCR. This synthesis and self-assembly strategy may solve the cost problems associated with large-scale DNA mass production. Therefore, demand is increasing for new technologies that can produce DNA in bulk quantities using designed sequences to meet the demands of advanced bioapplications.

The rolling circle amplification (RCA) method can be used to synthesize long single-stranded DNA for mass production [7]. When the circular DNA is supplied, DNA polymerase continuously synthesizes the DNA strand. A repeat DNA sequence with a specific function is constructed by rolling a circle. RCA has two advantages: its sequence is less difficult to design [8] and it has specific functional sequences, such as self-assembled [9] and G-quadraplex forms [10]. For example, an RCA-based sensing platform technique has been developed for the simultaneous multiplex detection of infectious pathogens such as COVID-19 using a self-assembled DNA hydrogel formed in large quantities on a nylon mesh surface by blocking the microfluidic flow [11]. In another system, a microfluidic platform detects viral pathogens using self-assembled RCA product DNA on the surface of microbeads, thereby reducing the detection time to 15 min [12,13]. Chang et al. employed in-situ DNA-oriented polymerization (isDOP), a system capable of encapsulating the RCA product long DNA while simultaneously loading drugs to treat intestinal inflammation [14].

Although previous studies utilizing the advantages and characteristics of RCA have been continuously conducted, commercialization remains a challenge. One of the techniques closest to commercialization is the synthesis of DNA particles via RCA. In this approach, following the RCA reaction, magnesium (Mg²⁺) and pyrophosphate (ppi) are added to neutralize the negative charges of the DNA backbone, thereby inducing massive intermolecular interactions that lead to the condensation of DNA into particle-like structures [15]. This self-assembly DNA technology has recently attracted interest for its potential applications in fields, such as drug delivery and other biomedical areas [16]. However, its commercialization has been hindered by several challenges, including the low efficiency of DNA synthesis using circular templates during RCA [17], lack of precise control over the concentrations of Mg²⁺ and ppi ions [18,19], and insufficient theoretical understanding of DNA particle formation. Therefore, this study focused on elucidating the theoretical background of DNA self-assembly induced by the addition of Mg²⁺ and ppi following RCA to provide a clearer understanding of the underlying mechanisms and improve the reproducibility of particle formation.

In this study, we used the RCA method to synthesize a long single-stranded DNA (ssDNA) and assembled the it using Mg²⁺ and ppi ions. Optical microscopy, atomic force microscopy (AFM), scanning electron microscopy (SEM), energy dispersive X-ray spectroscopy (EDS), Kelvin probe force microscopy (KPFM), and electrostatic force microscopy (EFM) were used to observe the process of DNA assembly into DNA particles and to analyze the particles. Furthermore, we analyzed the assembly of the DNA using molecular dynamics (MD) simulations and examined the DNA particle assembly circumstances using binding energy and H-bonding analysis. This study theoretically and experimentally analyzed and validated the synthesis of DNA particles via RCA, thereby providing a foundation for optimizing and scaling-up the production of DNA particles.

2. MATERIALS AND METHODS

3. RESULTS AND DISCUSSIONS

3.1 Manufacturing the self-assembled DNA particles

In RCA techniques, DNA-mediated RCA has predominantly been utilized [25]. Although DNA mediators can be used, we selected RNA for this study. DNA-based ligation typically requires the use of T4 DNA ligase, which is known to have a low ligation efficiency and may result in inaccurate ligation. Consequently, in conventional approaches, an additional exonuclease treatment step is required to remove the residual dumbbell-shaped DNA, as these residual DNA fragments could otherwise participate in unintended amplification reactions in subsequent steps. In contrast, the SplintR ligase specifically ligates DNA strands hybridized with RNA and exhibits significantly higher ligation efficiency [26]. Therefore, by using SplintR ligase, the need for exonuclease treatment can be eliminated, effectively minimizing undesired DNA amplification and improving the overall specificity of the RCA process.

As illustrated in Fig. 1, a dumbbell-shaped DNA template was ligated using the mediate RNA and Splint R. The primer DNA was then annealed to a portion of the dumbbell structure and RCA was performed using phi29 DNA polymerase. In place of other high-performance polymerases, such as Bst polymerase, we used phi29 polymerase. For example, in hyperbranched RCA, Bst polymerase can synthesize approximately 1000-fold faster than phi29 polymerase; however, significant tradeoffs exist [27]. The rapid synthesis offered by Bst polymerase often compromises the fidelity of the reaction, and its requirement for temperatures above 60°C complicates reactor design and scalability for bulk synthesis. In contrast, phi29 polymerase operates reliably at approximately 30°C, making it a more suitable choice for long-duration, high-fidelity RCA.

Fig. 1.

Schematic of self-assembled DNA particle. Self-assembled DNA particles were formed from DNA amplified using RCA through the addition of magnesium and pyrophosphate.

When the polymerase is initiated, the primer continuously rotates around the dumbbell-shaped DNA, yielding long and repeating stretches of complementary sequences. Because of their dumbbell shape, the resulting RCA products contained internal complementary sequences, enabling long DNA strands to self-assemble spontaneously.

However, the complementary sequences generated within the RCA products are insufficient to form condensed DNA particles on their own, resulting in a hydrogel-like structure [28]. Although ppi is generated as a byproduct of phi29 mediated DNA synthesis, its concentration is too low to drive the condensation of RCA products into discrete nanoparticles [29]. Therefore, an additional chemical condensation agent is required to facilitate the formation of compact particle-like structures. Several studies have demonstrated the additional use of Mg²⁺ and ppi to induce DNA condensation and particle formation [30,31]. In this study, we utilized Mg²⁺ and ppi to condense the RCA products, allowing the generation of distinct and well-defined DNA nanoparticles.

As shown in Fig. 2, the assembly and ligation steps required for the RCA were confirmed using agarose gel electrophoresis before the initiation of particle formation. Lane 1 shows the annealed dumbbell-shaped DNA, which was designed as a 90-mer single-stranded DNA. Upon annealing, this strand adopts a hairpin-like structure, with some molecules remaining as non-ligated circles, yielding a band slightly above the 50 bp marker. Lane 2 shows the results after incubation with SplintR ligation, where the mediate RNA hybridizes to the dumbbell-shaped DNA and promotes ligation. Ligation increased the molecular weight by approximately half the length of the RNA, yielding a band that migrated slightly higher than the original dumbbell-shaped DNA. Lanes 3 and 4 depict the results of RCA performed in the absence and presence of the mediate RNA, respectively. In Lane 3, no RNA was added; consequently, ligation did not occur, yielding no RCA product despite the addition of the primer. In Lane 4, the primer was annealed to the ligated dumbbell-shaped DNA, enabling phi29 polymerase to perform long, continuous strand displacement synthesis. The resulting high-molecular-weight DNA accumulated at the top of the gel and was unable to pass through the agarose matrix owing to its large size. This confirmed the successful and robust synthesis of the RCA products. Subsequently, Mg²⁺ and ppi were introduced into the self-assembled particles.

Fig. 2.

Gel electrophoresis image of RCA process (ladder: 50 bp ladder, lane 1: dumbbell-shaped DNA, lane 2: Splint R ligation, lane 3: DNA amplified without mediate RNA, lane 4: DNA amplified with mediate RNA.

3.2 Morphological verification of DNA particles

After the RCA reaction, Mg²⁺ and ppi were introduced to assess particle formation, and various microscopic techniques were used for characterization. As shown in Fig. 3, upon the addition of Mg²⁺ and ppi to the RCA product solution, optical microscopy revealed the presence of small, round particles.

Fig. 3.

50X Optical microscope image of DNA particles.

The AFM was performed for a more detailed analysis. As shown in Fig. 4 (A), AFM images confirmed the presence of long, extended DNA strands resulting from the RCA and small nuclei of the nanoparticles [32]. As shown in Fig. 4 (B), the DNA strands combined with Mg²⁺ and ppi, forming self-assembled DNA nanoparticles. This process was observed across multiple scales, from the initial particle morphology arising from the interaction between long ssDNA and Mg²⁺/ppi ions to larger aggregates reaching several micrometers in diameter. The AFM observations in this study directly visualized and confirmed the multiscale self-assembly process, highlighting its significance.

Fig. 4.

AFM images of (A) nucleation of RCA product (in 0–30 nm) and (B) self-assembled DNA particles (in 0–300 nm). (C) SEM image of DNA particles. (D) EDS analysis of DNA particles (Red: carbon, Green: nitrogen, Skyblue: oxygen).

SEM and EDS analyses were conducted for further characterization. The SEM image in Fig. 4 (C) confirms the particle morphology, whereas EDS mapping (Fig. 4 (D)) detected the elemental composition, including C, N, and O. Notably, Mg²⁺ and ppi formed inorganic particles spontaneously, regardless of RCA activity. As shown in Fig. 5, a control sample was prepared without mediate RNA, and Mg²⁺ and ppi were added. The SEM images of both samples revealed no significant differences in particle morphology. Element map analysis describe the component ratios of the DNA nanoparticles with and without mediate RNA, respectively. However, as shown in Fig. 6, EDS data indicated 6% higher C Kα1,2 value in the sample containing mediate RNA, suggesting a higher organic content and confirming the presence of RCA-derived DNA within the particles.

Fig. 5.

SEM/EDS analysis of particles formed with Mg2+ and ppi without mediate RNA. (A) SEM and EDS analysis image.

Fig. 6.

Bar graph of weight percent of C kα1,2 for the Si-wafer, without (w/o) mediate RNA, and with (w) mediate RNA.

KPFM was used to measure the surface charges of the particles. Figs. 7 and 9 show the surface potential of RCA products and their electrostatic interaction with Mg²⁺ and ppi ions using KPFM and EFM. To experimentally analyze the contribution of ionic and nucleic acid, we prepared three sample conditions: (1) Mg²⁺ and ppi ions only, (2) RCA products only, and (3) RCA products mixed with Mg²⁺ and ppi.

The surface potentials of the particles formed with and without the mediate RNA were compared. As shown in Fig. 7, despite no significant difference in particle size (AFM height), the surface charge was most highly charged in the sample containing mediate RNA, supporting the conclusion that the DNA generated via RCA, in conjunction with Mg²⁺ and ppi, contributes to particle self-assembly [33-35].

Fig. 7.

KPFM height and potential images of Mg2+ and ppi ((A) and (B)), RCA ((C) and (D)), RCA with Mg2+ and ppi ((E) and (F)).

As shown in Figs. 7 (A) and (B), particles from the Mg²⁺ and ppi sample was observed. Slightly lower surface potentials were observed at some points; however, other particles did not exhibit this (Fig. 7 (B)). Mg²⁺ and ppi ions were not deionized by ion crosslinking; therefore, we assumed that some points had dominant Mg²⁺ and others dominant ppi ions. Therefore, we applied the average global surface potential values. As shown in Figs. 7 (C)–(F), when the RCA products combined with Mg²⁺ and ppi were analyzed, no significant morphological differences were observed; however, distinct variations in surface potential were detected. As the RCA product was mediated, the surface potential distribution followed a particle distribution.

As shown in Fig. 7 (D), although the RCA products had a high density of negatively charged DNA, the reaction buffer included cations such as Na⁺ and Mg²⁺, which neutralized the negative charge. As larger aggregates showed a higher potential, the overall concentration of mobile ions may have resulted in an elevated surface potential value. As shown in Fig. 7 (F), when Mg²⁺ and ppi were combined with RCA products, ionic crosslinking between the phosphate backbones and ppi led to densely packed, crosslinked structures with highly charge density. As shown in Fig. 8 (A), This high degree of ionic interaction contributed to the higher surface potential observed for RCA + Mg²⁺ and ppi samples.

Fig. 8.

(A) Bar graph of the KPFM potential of the Mg2+ and ppi , RCA with Mg2+ and ppi, and RCA samples. **: p<0.01; ***: p<0.00001 (B) scheme of the KPFM measurement.

Fig. 8 (B) illustrates the working principle of the KPFM. The voltage difference between the cantilever probe and the sample is defined as

$\[ V_{total}=\left(V_{DC}-V_S\right)+V_{AC} cos{ w}_t \]$

(1)

where V_total is the total voltage acting on the cantilever; V_DC and V_AC and are the applied DC and AC voltages, respectively; and V_S is the surface potential of the sample. w_t is the frequency of the AC voltage. The resulting electrostatic force is given by

$\[ F_e=\frac{1}{2}\frac{dC}{dz}{V}_t^2 \]$

(2)

Expanding the equation yields three terms, as shown in Eq. (3). When V_DC = V_S, only the AC term remains. KPFM measures the surface potential by adjusting V_DC such that it compensates for the contact potential difference, thus enabling absolute surface potential measurements [36].

KPFM provides surface potential maps but does not distinguish the charge polarity. As shown in Fig. 9, an EFM analysis was performed to complement this.

Fig. 9.

EFM height and phase images of Mg2+ and PPi ((A) and (B)), RCA ((C) and (D)), RCA with Mg2+ and PPi ((E) and (F)). G. Scheme of EFM measurement.

The electrostatic force is expressed as follows:

$\[ F_e=\frac{1}{2}\frac{dC}{dz}\left\{{\left(V_{DC}-V_S\right)}^2+2\left(V_{DC}-V_S\right)V_{AC}cos{w}_t+{\left(V_{AC}cos{w}_t\right)}^2\right\} \]$

(3)

Although the squared terms are always positive, the cosine term determines the direction (phase) of the force. Thus, the EFM phase shift can be used to determine whether a sample is positively or negatively charged based on the direction of the electrostatic interaction.

As shown in Figs. 9 (A) and (B), particle-like structures with phase shifts exceeding 100 degrees were interpreted as positively charged Mg²⁺ and ppi aggregates. In contrast, Figs. 9 (C) and (D) show that RCA products alone exhibited reduced phase shifts, likely owing to cation-mediated charge neutralization by reaction buffer components, previous reports where zeta potential measurements showed near-zero potential in ion-rich environments [37,38]. Figs. 9 (E) and (F) show negatively charged phase responses of the RCA combined with Mg²⁺ and ppi. These observations support that the RCA + Mg²⁺ and ppi aggregates exhibited strong electrostatic potential and were overall negatively charged owing to ionic crosslinking and high molecular size. This characteristic may be beneficial for applications involving drug delivery, where the surface charge balance affects interactions with biological targets.

Through microscopy, SEM/EDS, AFM, KPFM, and EFM analyses, we successfully demonstrated the DNA particle growth procedures, morphological shape, size, and surface potential differences with or without mediating RNA. For further research, we used MD to explain the reason of self-assembly of RCA product by Mg²⁺ and ppi.

3.3 Molecular dynamics simulation of self-assembled DNA

3.3.1 Binding properties and atomic contact analysis

To investigate the stepwise assembly pathway, we conducted atomic contact analysis with a 0.6 nm cutoff over the 10 ns trajectory. The number of atomic contacts (<0.6 nm) averaged over the final 1 ns of the simulation. The contact numbers were 1528.824 ± 208.671 for ppi–Mg²⁺, 370.251 ± 133.932 for DNA–Mg²⁺, and 129.018 ± 100.766 for DNA–ppi, respectively. These findings suggest that DNA self-assembly occurs via a sequential interaction.

As shown in Fig. 10 (A), in the initial stage, Mg²⁺ and ppi rapidly nucleated into compact ionic clusters, as evidenced by a sharp radial distribution function (RDF) peak with g(r) = 78 at approximately 0.50 nm. This was followed by a bridging phase, in which the phosphate backbone of DNA formed strong coordination with Mg²⁺, shown by an RDF peak with g(r) = 6 at approximately 1.50 nm. In the final phase, DNA associated with the pre-assembled Mg²⁺ and ppi clusters, forming a ternary condensed structure supported by the increasing DNA–ppi contacts over time, with an RDF peak of g(r) = 4.7 at approximately 1.70 nm.

Fig. 10.

(A) RDF profiles for DNA–ppi, DNA–Mg2+, and ppi–Mg2+ interactions. (B) Time evolution of atomic contacts between DNA–ppi, DNA–Mg2+, and ppi–Mg2+

This three-step sequence was corroborated by the time-resolved contact profiles shown in Fig. 10 (B). All three interaction types ppi–Mg²⁺, DNA–Mg²⁺, and DNA–ppi underwent a notable transition at approximately 6 ns. Specifically, the number of ppi–Mg²⁺ contacts began to decline after peaking early, DNA–Mg²⁺ contacts reached their maximum and then start to decrease, and DNA–ppi contacts exhibited a significantly increased starting near 6 ns. These synchronous but distinct transitions underscored a coordinated assembly mechanism, where DNA is progressively incorporated into the preformed Mg²⁺–ppi clusters.

The dynamic trends in atomic contacts decrease in ppi–Mg²⁺ interactions, transient rise and fall in DNA–Mg²⁺ coordination, and sustained increase in DNA–ppi contacts. This is strongly correlated with the progressive folding of DNA. This contact pattern, along with the observed decrease in Rg, confirms that multivalent ion-mediated electrostatic interactions are stepwise structural compactions of DNA in the post-RCA environment.

In this paper, g_AB(r) is the RDF, which is defined as the probability of finding a B atom at distance r from a reference A atom relative to the bulk. We computed g(r) using a histogram of all substrate A to B pair distances over the MD trajectory under periodic boundary conditions, normalized by the spherical shell volume 4πr² dr and bulk number density ρ_B, and averaging over frames (Eq. (4)).

To quantify the spatial distribution of ions around DNA, we calculated the RDF as follows:

$\[ g\left(r\right)=\frac{1}{4\pi r^2\rho }⟨\frac{dn\left(r\right)}{dr}⟩ \]$

(4)

where r is the radial distance from a reference atom, ρ is the bulk number density, and dn(r) is the average number of atoms found in a spherical shell between r and r+dr. A sharp peak in g(r) indicates strong local ordering or preferential binding between interacting species.

With this normalization, g(r) approached 1 for a large r. The first-peak values in this study were ppi–Mg²⁺ at 0.50 nm with g = 78, DNA–Mg²⁺ at 1.50 nm with g = 6, and DNA–ppi at 1.70 nm with g = 47, respectively.

In addition, molecular mechanics Poisson–Boltzmann surface area MM-PBSA analysis [39] revealed a moderate thermodynamic interaction between DNA and ppi, with an average binding free energy of 86.191 ± 45.463 kJ/mol. This positive value indicates that the interaction is not energetically favorable, which corresponds with the observation that both DNA and ppi are negatively charged under physiological conditions. The resulting electrostatic repulsion limits the direct binding between them, which is consistent with the relatively low number of atomic contacts. As divalent cations, may act as electrostatic bridges facilitating indirect stabilization of the DNA–ppi interaction and promoting ternary complex formation.

Altogether, these simulations offer mechanistic insight into the sequential self-assembly of DNA particles under high-salt post-RCA conditions. The findings support a model in which Mg²⁺ and ppi ions initially nucleate into ionic clusters, followed by phosphate-mediated binding of DNA to Mg²⁺, and subsequent growth of condensed particles through complex formation with the Mg²⁺ and ppi matrix [40].

3.3.2 Conformational analysis of DNA and ion distribution

Furthermore, the RDF analysis confirmed these preferential interactions. DNA exhibited sharp coordination with Mg²⁺, especially at phosphate groups, with a peak at 0.21 nm (g(r) = 4.2). In contrast, DNA–ppi interactions were more diffuse, with a broader peak between 0.25 and 0.30 nm (g(r) = 2.2), indicating weaker electrostatic associations. As shown in Fig. 10 (A), the interaction between Mg²⁺ and ppi was particularly strong and localized, with a narrow, high-intensity RDF peak at 0.15 nm (g(r) = 92.0).

Representative MD snapshots at the beginning and end of the simulation (0 and 30 ns) illustrated the morphological changes in ssDNA during compaction (Fig. 11 (A)), whereas a slight decrease in Rg (Fig. 11 (B)) quantified the onset of molecular compaction. These findings underscore the critical role of the post-RCA ionic composition in triggering spontaneous DNA particle formation and offer molecular-level design principles for developing programmable DNA-based nanomaterials.

Fig. 11.

(A) Representative MD snapshots of RCA-derived ssDNA at0 and 30 ns. (B) Time-dependent change in Rg of ssDNA over 30 ns. (C) Time-resolved MM-PBSA binding free energy profile for the DNA–ppi interaction

As shown in Fig. 11 (B), the structural compactness of DNA was evaluated using the radius of gyration (Rg), which slightly decreased from approximately 7.9 nm to 7.7 nm during the simulation. This suggests the onset of structural collapse due to multivalent-ion-mediated condensation.

Rg was calculated using

$\[ Rg=\sqrt{\frac{1}{N}{\sum }_{i=1}^N{\left|r_i-r_{com}\right|}^2} \]$

(5)

where N is the number of atoms, r_i is the position of each atom, and r_com is the center of mass of the DNA strand. A decrease in Rg reflects the progressive structural collapse or compaction of DNA molecules.

Moreover, time-resolved MM-PBSA analysis (Fig. 11 (C)) indicated that the DNA–ppi interaction maintained a moderate binding free energy of approximately 86.191 ± 45.463 kJ/mol throughout the 30 ns simulation. Although the interaction was not strongly favorable in absolute thermodynamic terms, it was consistent with the fact that both DNA and ppi carry negative charges, leading to electrostatic repulsion. Nevertheless, the observed increase in atomic contacts over time suggested that ppi may contribute to the final DNA–Mg²⁺–ppi architecture through a transient and Mg²⁺ mediated electrostatic bridging mechanism, rather than direct stabilization.

Compared with an earlier study on the ppi release mechanism in T7 RNA polymerase [41], which highlighted the rapid dissociation of ppi through transient interactions with Mg²⁺ ions in the active site, our simulation revealed a more structured and hierarchical assembly process. Specifically, the formation of stable Mg²⁺–ppi ionic clusters, followed by progressive coordination with DNA, suggests that ppi in RCA products plays not only a role in release but also in subsequent self-assembly. This distinction underscores the context-dependent behavior of ppi. While ppi is produced rapidly in enzymatic reactions, it contributes to complex formation and structural compaction in condensed-phase ionic systems. Thus, our findings provide complementary insights into the dualistic role of ppi depending on the molecular context and ionic conditions.

4. CONCLUSIONS

In this study, long ssDNA produced by RCA and self-assembled DNA particles by Mg²⁺ and ppi ions were prepared. Padlock DNA was efficiently transformed into a dumbbell-shaped structure with mediate RNA and SplintR ligase, while repetitive DNA sequences were produced via RCA (30°C, Phi29 DNA polymerase). In the presence of Mg²⁺ and ppi, DNA condensed into particulate aggregates, as evidenced by optical microscopy, AFM, and SEM.

The resulting particles were further characterized via surface potential analysis using KPFM and EFM. Samples containing RCA products exhibited a more negatively charged surface than the ionic controls, confirming the incorporation of DNA into the particles. EDS also verified the presence of a carbon-rich organic content in the RCA-derived particles.

MD simulations revealed a stepwise assembly mechanism, in which Mg²⁺ and ppi initially formed ionic clusters, followed by DNA–Mg²⁺ coordination and progressive association of DNA with the preformed Mg²⁺–ppi matrix. Despite the electrostatic repulsion between negatively charged DNA and ppi, then Mg²⁺ acted as an electrostatic bridge, promoting ternary complex formation and structural compaction of DNA. The decrease in Rg and the stepwise progression of the atomic contacts corroborated this hierarchical assembly pathway.

Overall, the combined experimental and simulation results provided mechanistic insights into RCA-derived DNA particle formation and demonstrated the critical role of ionic interactions in directing DNA self-assembly. These results provide a basis for the rational DNA sequence and structure design and large-scale fabrication of RCA-derived DNA based nano-materials for drug delivery, biosensing, and nanoscale devices.

Acknowledgments

This study was supported by the Kumoh National Institute of Technology.

References

• S. Pravin, I. Kataria, Potential applications of folded and unfolded DNA nanocarriers in medicine, J. Drug Deliv. Sci. Technol. 57 (2020) 101729. [https://doi.org/10.1016/j.jddst.2020.101729]
• S. Dey, C. Fan, K.V. Gothelf, J. Li, C. Lin, L. Liu, et al., DNA origami, Nat. Rev. Methods Primers 1 (2021) 13. [https://doi.org/10.1038/s43586-020-00009-8]
• B.K. Davis, Complexity transmission during replication, Proc. Natl. Acad. Sci. U.S.A. 76 (1979) 2288–2292. [https://doi.org/10.1073/pnas.76.5.2288]
• A. Al Ouahabi, L. Charles, J.F. Lutz, Synthesis of non-natural sequence-encoded polymers using phosphoramidite chemistry, J. Am. Chem. Soc. 137 (2015) 5629–5635. [https://doi.org/10.1021/jacs.5b02639]
• Z. Huang, L. Qiu, T. Zhang, W. Tan, Integrating DNA nanotechnology with aptamers for biological and biomedical applications, Matter 4 (2021) 461–489. [https://doi.org/10.1016/j.matt.2020.11.002]
• L. Yang, C.W. Fung, E.J. Cho, A.D. Ellington, Real-time rolling circle amplification for protein detection, Anal. Chem. 79 (2007) 3320–3329. [https://doi.org/10.1021/ac062186b]
• M.G. Mohsen, E.T. Kool, The discovery of rolling circle amplification and rolling circle transcription, Acc. Chem. Res. 49 (2016) 2540–2550. [https://doi.org/10.1021/acs.accounts.6b00417]
• R.R. Garafutdinov, A.R. Sakhabutdinova, A.R. Gilvanov, A.V. Chemeris, Rolling circle amplification as a universal method for the analysis of a wide range of biological targets, Russ. J. Bioorg. Chem. 47 (2021) 1172–1189. [https://doi.org/10.1134/S1068162021060078]
• W. Pan, S. Zhu, L. Chen, C. Chen, C. Xue, R. Wu, et al., Self-folding RCA product into a parallel monolayer DNA nanoribbon and woven into a nano-fence structure by a short bridge strand, J. Colloid Interface Sci. 677 (2025) 30–39. [https://doi.org/10.1016/j.jcis.2024.08.013]
• X. Tang, Y. Wang, L. Zhou, W. Zhang, S. Yang, L. Yu, et al., Strand displacement-triggered G-quadruplex/rolling circle amplification strategy for the ultra-sensitive electrochemical sensing of exosomal microRNAs, Microchim. Acta 187 (2020) 172. [https://doi.org/10.1007/s00604-020-4143-9]
• H. Kim, N. Abbas, S. Shin, A rapid diagnosis of SARS-CoV-2 using DNA hydrogel formation on microfluidic pores, Biosens. Bioelectron. 177 (2021) 113005. [https://doi.org/10.1016/j.bios.2021.113005]
• W. Na, D. Nam, H. Lee, S. Shin, Rapid molecular diagnosis of infectious viruses in microfluidics using DNA hydrogel formation, Biosens. Bioelectron. 108 (2018) 9–13. [https://doi.org/10.1016/j.bios.2018.02.040]
• Y. Ren, K. Ge, W. Lu, X. Xie, Y. Lu, M. Wang, et al., Multivalent DNA flowers for high-performance isolation, detection, and release of tumor-derived extracellular vesicles, ACS Appl. Mater. Interfaces 15 (2023) 55358–55368. [https://doi.org/10.1021/acsami.3c12211]
• Y. Chang, K. Wang, M. Zhou, Y. Miao, X. Wu, Q. Chen, et al., Bioencapsulation of probiotics with self-assembled DNA capsules for the treatment of inflammatory bowel diseases, Chem. Eng. J. 499 (2024) 155527. [https://doi.org/10.1016/j.cej.2024.155527]
• V.N. Danilevich, A.V. Machulin, A.V. Lipkin, T.V. Kulakovskaya, S.S. Smith, A.L. Mulyukin, New insight into formation of DNA-containing microparticles during PCR: the scaffolding role of magnesium pyrophosphate crystals, J. Biomol. Struct. Dyn. 34 (2016) 625–639. [https://doi.org/10.1080/07391102.2015.1040842]
• R. Ye, Y. Wang, Y. Liu, P. Cai, J. Song, Self-assembled methodologies for the construction of DNA nanostructures and biological applications, Biomater. Sci. 12 (2024) 3712–3724. [https://doi.org/10.1039/D4BM00584H]
• B. Joffroy, Y.O. Uca, D. Prešern, J.P.K. Doye, T.L. Schmidt, Rolling circle amplification shows a sinusoidal template length-dependent amplification bias, Nucleic Acids Res. 46 (2018) 538–545. [https://doi.org/10.1093/nar/gkx1238]
• E. Kim, S. Agarwal, N. Kim, F.S. Hage, V. Leonardo, A. Gelmi, et al., Bioinspired fabrication of DNA–inorganic hybrid composites using synthetic DNA, ACS Nano 13 (2019) 2888–2900. [https://doi.org/10.1021/acsnano.8b06492]
• X. Yuan, F. Xiao, H. Zhao, Y. Huang, C. Shao, Y. Weizmann, et al., High-yield method to fabricate and functionalize DNA nanoparticles from the products of rolling circle amplification, ACS Appl. Bio Mater. 1 (2018) 511–519. [https://doi.org/10.1021/acsabm.8b00238]
• M.J. Abraham, T. Murtola, R. Schulz, S. Páll, J.C. Smith, B. Hess, et al., GROMACS: High performance molecular simulations through multi-level parallelism from laptops to supercomputers, SoftwareX 1 (2015) 19–25. [https://doi.org/10.1016/j.softx.2015.06.001]
• W.L. Jorgensen, J. Chandrasekhar, J.D. Madura, R.W. Impey, M.L. Klein, Comparison of simple potential functions for simulating liquid water, J. Chem. Phys. 79 (1983) 926–935. [https://doi.org/10.1063/1.445869]
• J. Huang, S. Rauscher, G. Nawrocki, T. Ran, M. Feig, B. L de Groot, et al., CHARMM36m: an improved force field for folded and intrinsically disordered proteins, Nat. Methods 14 (2017) 71–73. [https://doi.org/10.1038/nmeth.4067]
• S. Nosé, A molecular dynamics method for simulations in the canonical ensemble, Mol. Phys. 100 (2002) 191–198. [https://doi.org/10.1080/00268970110089108]
• W. Humphrey, A. Dalke, K. Schulten, VMD: visual molecular dynamics, J. Mol. Graph. 14 (1996) 33–38. [https://doi.org/10.1016/0263-7855(96)00018-5]
• M.G. Mohsen, E.T. Kool, The discovery of rolling circle amplification and rolling circle transcription, Acc. Chem. Res. 49 (2016) 2540–2550. [https://doi.org/10.1021/acs.accounts.6b00417]
• J. Jin, S. Vaud, A.M. Zhelkovsky, J. Posfai, L.A. McReynolds, Sensitive and specific miRNA detection method using SplintR Ligase, Nucleic Acids Res. 44 (2016) e116. [https://doi.org/10.1093/nar/gkw399]
• N.A. Tanner, T.C. Evans Jr, Loop-mediated isothermal amplification for detection of nucleic acids, Curr. Protoc. Mol. Biol. 105 (2014) 15.14.1–15.14.14. [https://doi.org/10.1002/0471142727.mb1514s105]
• J. Bush, C.H. Hu, R. Veneziano, Mechanical properties of DNA hydrogels: Towards highly programmable biomaterials, Appl. Sci. 11 (2021) 1885. [https://doi.org/10.3390/app11041885]
• E. Kim, S. Agarwal, N. Kim, F.S. Hage, V. Leonardo, A. Gelmi, et al., Bioinspired fabrication of DNA–inorganic hybrid composites using synthetic DNA, ACS Nano 13 (2019) 2888–2900. [https://doi.org/10.1021/acsnano.8b06492]
• H. Zhao, X. Yuan, J. Yu, Y. Huang, C. Shao, F. Xiao, et al., Magnesium-stabilized multifunctional DNA nanoparticles for tumor-targeted and pH-responsive drug delivery, ACS Appl. Mater. Interfaces 10 (2018) 15418–15427. [https://doi.org/10.1021/acsami.8b01932]
• J. Li, X. Liu, J. Wang, Q. Jiang, M. Chen, W. Zhang, et al., Self-Assembled DNA Nanospheres: Design and Applications, Chemistry 5 (2023) 1882–1910. [https://doi.org/10.3390/chemistry5030129]
• S.B. Seo, J.S. Hwang, E. Kim, K. Kim, S. Roh, G. Lee, et al., Isothermal amplification-mediated lateral flow biosensors for in vitro diagnosis of gastric cancer-related microRNAs, Talanta 246 (2022) 123502. [https://doi.org/10.1016/j.talanta.2022.123502]
• S.H. Choi, Y.J. Kim, D.H. Lee, J.H. Lee, S.J. Choi, Dual-hydrogen bond donors for chemical sensors: Transducing chemical interactions into electrical signals, J. Sens. Sci. Technol. 34 (2025) 352–361. [https://doi.org/10.46670/JSST.2025.34.4.352]
• G.S. Shin, D.S. Kim, T. Van Tran, G.J. Oh, S.Y. Hong, H.G. Song, et al., Synthesis and oxidation behavior of Pd-Ir@CeO₂ core-shell nanoparticles for hydrogen gas sensor, J. Sens. Sci. Technol. 33 (2024) 288–297. [https://doi.org/10.46670/JSST.2024.33.5.288]
• D. Lee, H.G. Jung, D. Park, J. Bang, J.H. Hong, S.W. Lee, et al., Biomimetically engineered amyloid-shelled gold nanocomplexes for discovering α-synuclein oligomer-degrading drugs, ACS Appl. Mater. Interfaces 15 (2023) 2538–2551. [https://doi.org/10.1021/acsami.2c14650]
• S. Roh, T. Lee, Y. Kim, S. Oh, G. Lee, Direct observation of surface charge and stiffness of human metaphase chromosomes, Nanoscale Adv. 5 (2023) 368–377. [https://doi.org/10.1039/D2NA00620K]
• J. Austin, D. Fernandes, M.J.A. Ruszala, N. Hill, J. Corbett, Routine, ensemble characterisation of electrophoretic mobility in high and saturated ionic dispersions, Sci. Rep. 10 (2020) 4628. [https://doi.org/10.1038/s41598-020-61624-9]
• Y. Cho, S. Jeong, J. Kang, The role of two-dimensional materials in optical chemical sensing, J. Sens. Sci. Technol. 34 (2025) 324–341. [https://doi.org/10.46670/JSST.2025.34.4.324]
• R. Kumari, R. Kumar, Open Source Drug Discovery Consortium, A. Lynn, g_mmpbsa—A GROMACS tool for high-throughput MM-PBSA calculations, J. Chem. Inf. Model. 54 (2014) 1951–1962. [https://doi.org/10.1021/ci500020m]
• J. Yoo, A. Aksimentiev, The structure and intermolecular forces of DNA condensates, Nucleic Acids Res. 44 (2016) 2036–2046. [https://doi.org/10.1093/nar/gkw081]
• S. Wu, M. Jiang, L. Liu, Mechanism of pyrophosphate ion release in T7 RNA polymerase revealed by free energy simulations, Comput. Theor. Chem. 1094 (2016) 127–132. [https://doi.org/10.1016/j.comptc.2016.09.005]