Planar oligomerization of reconfigurable gold nanorod dimers

Chiral self-assembly of nanoparticles is an area of research that is rapidly advancing geared toward the fabrication of unique nanostructures with intrinsic properties.^1–3 Anisotropic plasmonic nanoparticles, such as gold nanorods (AuNRs), have been of particular interest recently because of their tunable chemical and optical properties based on their unique anisotropic geometry.^4,5 Despite tremendous advances, effective manipulation of AuNRs for deliberately designing chiral complex remains a challenge. This challenge arises from the need to control not only the number and positions of the nanoparticles but also their orientations.

Specifically, DNA origami provides a groundbreaking nanofabrication method utilizing precisely defined 3D conformations with nanoscale resolution.^6–9 This technique offers unique platforms to build complex, hierarchical, and hybrid nanophotonic devices based on AuNRs, enabling a comprehensive understanding of light–matter interaction processes and the attainment of tailored optical properties and functionality.^10,11 On the other hand, DNA origami presents a unique advantage in the controlled reconfiguration of chiral plasmonic nanostructures via strand-displacement reactions.^12–14 Through the finely tuned optical response, these well-designed plasmonic components could significantly amplify and reveal changes in environmental information.^15–17 AuNR dimers with “fingers crossed” structure are well known for being the simplest structures exhibiting the phenomenon of plasmon coupling and generating intense plasmonic circular dichroism (CD).¹⁸ In addition, this chiral plasmonic system can be easily adjusted by controlling the tilt angle between the two AuNRs, on which diverse dynamic chiral plasmonic structures have been realized.^19–21 Though there have been both experimental and theoretical studies on the transferring and coupling of multiple static chiral centers,^22–26 relatively little attention has been given to dynamic chiral interactions of chiral centers.^27,28

Herein, we report planar oligomerization of reconfigurable plasmonic chiral centers based on DNA nanotechnology. Each chiral center is formed by assembling two AuNRs on opposite planes of a dynamic DNA origami platform. These chiral centers can further assemble with each other with precise number and arrangement by capping orthogonal connecting strands. The chiral center in the assembled superstructure can be separately reconfigured to generate desired diastereomers, giving rise to characteristic chiroptical signals. We find that the CD spectra of these chiral AuNR superstructures are not simply given as a sum of the CD spectra of each AuNR dimer, as coupling of chiral centers can obviously affect the overall CD signal.

All the chemicals were commercially obtained and used without further purification. Gold (III) chloride tetrahydrate (HAuCl₄ · 4H₂O, ≥99.9%), silver nitrate (AgNO₃, ≥99%), hexadecyltrimethylammonium bromide (CTAB, ≥98%), potassium bromide (KBr, ≥99%), and tris(2-carboxyethyl) phosphine (TCEP, ≥98%) were purchased from Sigma Aldrich. Sulfuric acid (H₂SO₄, 95.0%–98.0%), sodium borohydride (NaBH₄, 98%), ascorbic acid (≥99%), and sodium dodecyl sulfate (SDS, 99%) were purchased from Alfa Aesar. The p8064 scaffold DNA was purchased from tilibit nanosystems^® GmbH (Garching, Germany). Non-thiolated DNA staple strands were purchased from GENEWIZ (Suzhou) Co., Ltd. and thiolated DNA strands of high-performance liquid chromatography (HPLC) grade were purchased from Sangon Biotech (Shanghai) Co., Ltd. Agarose G-10 was purchased from Biowest.

The DNA origami template was designed with caDNAno (Fig. S1). For self-assembly of the plate-arm DNA origami (PADO) core structure, 10 nM of the p8064 DNA scaffold strand and a staple mixture containing 100 nM of each staple strand were mixed in the folding buffer (pH = 8.0), which contained 10 mM MgCl₂, 5 mM NaCl, 1 mM Tris-HCl, and 1 mM EDTA. The mixture was then annealed under the following conditions: 85 °C for 5 min; from 84 to 65 °C, −1 °C/1 min; from 64 to 60 °C, −1 °C/5 min; from 59 to 50 °C, −1 °C/60 min; from 49 to 37 °C, −1 °C/20 min; from 36 to 25 °C, −1 °C/5 min; and held at 25 °C. For self-assembly of PADO multimers, connecting staples were mixed with the pre-assembled PADO core structure at a ratio of 10:1. The mixture was then annealed as follows: 47 °C for 1 min; from 47 to 25 °C, −1 °C/3 min; and held at 25 °C. The PADO units with paired connecting staples were then mixed in stoichiometric proportion and held at 40 °C for assembly. The DNA origami templates were purified by agarose gel electrophoresis (1%) in the ice bath at 70 V for 2 h. The gel was characterized by gel imager and the target bands were cut and chopped with a sharp knife on a sealing film to release the DNA origami templates. The sample solution was collected by extrusion with a thick cover-glass. The purified DNA origami templates were then adsorbed onto the TEM grids and stained with 1% uranyl formate. Then, the samples were characterized with TEM imaging using Hitachi HT7700 TEM (100 kV).

The AuNRs (43 × 11 nm²) were synthesized through seed-mediated growth following the reported protocol.²⁹ Functionalization of the AuNRs with thiolated DNA was carried out following the butanol dehydration method.³⁰ The AuNRs functionalized with DNA were purified to remove excess free DNA strands by centrifugation and the precipitate was redissolved in 0.02% SDS-0.5×TBE buffer. Twice centrifugation was carried out at a rate of 10 000 g for 20 min. The purified AuNRs were mixed with unpurified DNA origami templates at a ratio of 5:1 and annealed as follows: 35–30 °C, −1 °C/120 min; 29–25 °C, −1 °C/60 min; held at 25 °C. The sample was purified by agarose gel electrophoresis (1%) in the ice bath at 70 V for 2 h. Then, the target bands were cut and squeezed into solution for characterization.

The CD spectra were obtained using an Applied Photo-physics Chirascan-Plus CD spectrometer. Measurements were carried out at a wavelength range of 500–1000 nm at 25 °C in a 10 mm long cell. All the products were diluted to 0.5 nM with 0.5×TBE-10 mM Mg²⁺ buffer. The scanning speed was 100 nm/min. The baseline was corrected with 0.5×TBE-10 mM Mg²⁺ buffer. The concentrations of driving strands were set to 20 µM. For each addition, 0.3 µl of driving strands was added. The volume of fuel strands doubled upon addition of complementary driving strands to the sample. For details on the specific process of adding driving strands, please refer to the supplementary material, Table 1. The CD spectra were measured 5 min after the addition of driving strands at room temperature.

Figure 1(a) presents a schematic of the plate-arm DNA origami (PADO) template. PADO is composed of a rotatory arm and a rectangle plate that are hinged at the center via scaffold DNA linkage. The arm is 63 nm long and 15 nm wide, and the plate is 50 nm long and 49 nm wide, assuming the helical pitch and diameter of B-form DNA are 0.34 nm/bp and 2.5 nm, respectively. According to functional distinction, PADOs can be divided into a core motif and four edge motifs (Fig. S2). The core motif can be assembled with a constant set of staple strands, which undertakes reconfiguration of the template. The edge motifs, at the ends of the arm and plate, can hybridize with connecting strands for hierarchical assembly [Fig. 1(b)]. In order to precisely control the number and arrangement of the PADO units, we designed “Mini-scaffold” strands to generate multiple sets of orthogonal and directional connections (Fig. S3). The “Mini-scaffold” strands are hybridized to the extended staples strands on one end of these PADO units, forming 5 bp sticky ends that are complementary to the 5 bp sticky ends of staples extended from the other PADO. Multiple 5 bp sticky ends form a group to connect two PADO units. Since the sequence of the 5 bp sticky end is customizable, multiple sets of orthogonal interactions can be generated readily. In addition, there are no flexible segments at the junctions, so the relevant position of two PADO units can be fixed. These two factors together ensure precise control over the number and arrangement of PADO units in oligomers. For experiments, sequence-specific connecting strands were hybridized to pre-assembled PADO core motifs via a quick annealing process (from 47 to 25 °C, −1 °C/2 min). Subsequently, different PADO units with paired connecting strands were mixed in stoichiometric proportion and incubated at 40 °C overnight for assembly. Following this process, we constructed a plate dimer, a plate trimer, an arm dimer, and a 2 × 2 tetramer as templates for the chiral assembly of AuNRs. All of these structures exhibit consistent shapes as expected and demonstrate excellent spatial orientation control capabilities, laying the foundation for guiding the precise assembly of AuNRs [Figs. 1(c)–1(g)]. It should be noted that the arm is intentionally designed to be longer than the plate, which enables their assembly along the direction of the arm. The gel image shows the high yield of these structures (Fig. S4).

To assemble AuNRs into the chiral center, staple strands with extended overhangs were placed along the arm and at the middle of the plate to capture complementary DNA-functionalized AuNRs via hybridization (Fig. S1). A fivefold excess of AuNRs was mixed with the unpurified PADO template and incubated through an annealing process for assembly. The sample was purified by agarose gel electrophoresis and then characterized by TEM (Figs. S5–S8). As shown in the Fig. 1(h), two AuNRs are assembled onto the arm and plate of PADO and demonstrate a well-tuned configuration, forming the PADO chiral center.

With three pairs of DNA hybridizations (N/L/R-arm and N/L/R-plate strands) between the arm and plate, we can fix the relative orientation of the arm and plate, and switch it by adding specific driving strands via toehold-mediated strand-displacement reactions (Fig. S9). Thereby, the configuration of the two AuNRs is set to N or L or R, generating the corresponding CD signals [Figs. 2, S10(a), and S10(b)]. When the PADO template is set at the N state, two AuNRs are arranged in parallel, showing overlapping configurations in top view. When the two AuNRs are fixed to either the L or R state, they form a “fingers crossed” structure with an angle of ∼38° and correspond to the left-handed or right-handed configuration, respectively [Figs. 2 and S11(a)–S11(c)]. In all three configurations, the surface-to-surface distance of the up and bottom AuNRs is around 22 nm and does not change during reconfiguration [Fig. S11(d)]. During the oligomerization of PADO chiral centers in a 2D plane, the center-to-center distance along the direction of the plate is 56 nm, and the distance along the direction of the arm is 70 nm [Figs. S11(e) and S11(f)]. The distances ensure that the reconfiguration process is not blocked by steric hindrance. Through sequence screening, we designed three sets of fixing and driving strands and avoided undesired crosstalk between each set of strands. For each set of PADO chiral center, we obtained its CD spectra at the L state and R state and found that they have similar CD spectra [Figs. S10(b)–S10(d)].

With these results in hand, we constructed the reconfigurable plasmonic structure containing two chiral centers. As shown in the Fig. 3(a), two sets of PADO templates were assembled with AuNRs separately and purified by agarose gel electrophoresis. The two chiral centers were then assembled side by side with plate edges to form a dimer structure. With a second gel purification, the PADO chiral dimer was obtained and confirmed by TEM characterization [Fig. 3(b)]. It should be noted that during the assembly process, the initial state of the two PADO chiral centers was fixed to N state to avoid the impact of arm rotation on assembly. The state of each chiral center could be independently switched to L or R states by adding the corresponding driving strands, forming dimers with different configurations. These dimers were named by the states of chiral centers from left to right, as NN, LL, LR, RL, and RR. In the experiment, we set NN as the initial configuration and obtained the CD spectra of all four configurations from one sample by switching the two chiral centers among three states.

During dimerization of the PADO structure, right-handed CD signals are observed [Fig. 3(c)], resulting from coupling between the AuNR dimers, which are about 50 nm apart. Similar CD signals are also observed in PADO dimers with AuNRs only at the arm or plate (Fig. S12), ascribed to the innate preference that was commonly observed in other DNA origami based chiral plasmonic systems.³¹ Such coupling weakens when the AuNR dimers are further apart. For example, we constructed a spaced structure in which two AuNR dimers are separated by an empty PADO unit, resulting in a distance of about 100 nm. The generation of a right-handed signal was observed during the assembly process and the signal is weaker than that generated by two adjacent chiral centers (Fig. S13). Furthermore, we explored the coupling between chiral centers of different states. For the pair of enantiomers of LR and RL, each of them could also be regarded as a plasmonic mesomer that contains two chiral centers with a plane of symmetry, meaning they are achiral structures. They reflect similar CD intensity to that of NN [Fig. 3(d)], while the two detached chiral centers L and R show no CD intensity in solution (Fig. S14). Meanwhile, the CD spectrum of NN is slightly blue-shifted compared to that of LR and RL, because of more apparent anti-bonding hybridized plasmon modes in NN [Fig. 3(d) inset figure]. In-phase, anti-bonding plasmon resonances of AuNR longitudinal modes are dominant in N owing to the similar orientations of AuNRs, thus leading to higher plasmonic energy compared to the L (Fig. S15) (i.e., blue shift of absorption spectra). When the two chiral centers are switched to the same chiral state, the CD spectrum of RR exhibits a much stronger right-handed CD signal than a single right-handed state R, while LL showed weaker signal than L, as a result of chiral coupling between the two chiral centers. When two PADO units are assembled along the arm, we also observed changes of the CD spectra and a very weak left-handed CD signal generated during assembly (Fig. S16).

We then further fabricated a more complex plasmonic system, which contains three independently reconfigurable chiral centers, to explore the impact of the arrangement of chiral centers on the overall chirality. Following the approach mentioned above, the PADO chiral trimer could be reconfigured by selectively switching each chiral center, forming eight configurations, including LLL, RLL, LRL, LLR, RRL, RLR, LRR, and RRR [Fig. 4(a)]. The assembly process of the trimer structure was similar to that of the dimer structure, and the TEM image confirmed the well-assembled trimer structure [Fig. 4(b)]. Three AuNR dimers were arranged parallelly on the template. The CD spectra of eight configurations are summarized in Fig. 4(c) along with the theoretical calculations [Fig. 4(d)]. Not surprisingly, LLL and RRR show the strongest left-handed and right-handed CD signals. Due to the coupling of three chiral centers, the initial CD spectrum shows a stronger right-handed signal than NN (Fig. S17). Except LLL, the other configurations all show right-handed signal. The coupling of chiral centers could obviously affect the overall CD signals, which is far away from the sum of the CD intensity of individual chiral centers.

For this plasmonic system with three chiral centers, the overall CD is also influenced by the arrangement of these centers. RRL, RLR, and LRR share the same species of chiral centers with two R chiral centers and one L chiral center but in different arrangements. For RRL and LRR, two identical chiral centers are adjacent, while the other one is on their side. In this case, they present nearly the same CD spectrum, relatively strong right-handed signals, while when the L chiral center is located in the middle of two R chiral centers, the CD signal drops slightly. For the other group, LLR, LRL, and RLL, the CD spectra follow the same rule. In contrast, our previous systems with an arrangement of four AuNRs along the helical axial direction showed a slight difference.²⁰ When the L chiral center is located in the middle of two R chiral centers, the CD signal drops significantly.

In addition to assembling PADO chiral centers along the direction of the plate to explore the interactions between parallelly arranged chiral centers, 2D PADO array structures provided more powerful templates for the construction of complex chiral structures. Herein, we selected the 2 × 2 tetramer as the template to assemble chiral centers in various arrangements. As shown in the Fig. 5(a), the four PADO units are sequentially labeled as I, II, III, and IV, and AuNRs could be arbitrarily assigned to a certain PADO in the tetramer, generating four two-chiral center structures and a four-chiral center structure. TEM characterization confirmed that the structures were well assembled and the arrangement of AuNRs was consistent with the design [Figs. 5(b) and S18–S21]. It is worth noting that in contrast to the aforementioned dimer and trimer structures, configuration switching of the 2 × 2 tetramer was constrained by the overall structure. As a result, the four constituting PADO units could only switch between N, L, and R states simultaneously, as illustrated in Fig. 5(c). In addition, the length of the arm was long enough for reconfiguration of the tetramer structures. We tested the CD spectra of all of the chiral tetramers in the three states [Figs. 5(d), 5(e), and S22].

As shown in Fig. 5(d), both two-chiral center structures and four-chiral center structure will generate coupling between chiral centers, resulting in different CD spectra compared to single chiral center in the N state. The chiral center of I + II is placed along the plate and that of I + III is placed along the arm, while those of I + IV and II + III are arranged along the diagonal. As mentioned above, the CD spectrum of I + II is right-handed while the others are left-handed, resulting from coupling between chiral centers with corresponding arrangement. Specially, the four-chiral center structure exhibits a similar right-handed chiral spectrum. The strong coupling observed along the direction of the plate with the shortest center-to-center distance dominates the overall CD. During reconfiguration, the chiral interactions of these AuNR dimers are still obvious [Figs. 5(e) and S22].

In conclusion, we successfully realized planar oligomerization of AuNR dimers. The concept of 2D pattern of AuNR dimers may be of general interest in the field of reconfigurable self-assembled plasmonic devices. We carefully studied the chiral interactions of chiral plasmonic superstructures with different arrangements and amounts of chiral centers. Our experiment demonstrated that strong coupling existed in two chiral centers even with a distance of 50 nm, and the overall CD could not be simplified to the summation of separate chiral centers. Overall, we demonstrated that the PADO can be used as a versatile and deterministic template for fabricating reconfigurable plasmonic devices to manipulate chiral light–matter interactions in future.

See the supplementary material associated with this article for more information.

This work was financially supported by the National Natural Science Foundation of China (Grant Nos. 21934007 and 22302227), the Strategic Priority Research Program of the Chinese Academy of Sciences (Grant No. XDB36000000), the China Postdoctoral Science Foundation (Grant No. 2022M712325), the Natural Science Foundation of Jiangsu Province (Grant No. BK20230232), and the Jiangsu Funding Program for Excellent Postdoctoral Talent.

The authors have no conflicts to disclose.

Yihao Zhou: Data curation (lead); Formal analysis (equal); Methodology (lead); Visualization (lead); Writing – original draft (lead). Jinyi Dong: Formal analysis (equal); Funding acquisition (supporting); Investigation (lead); Supervision (equal); Writing – review & editing (equal). Qiangbin Wang: Conceptualization (lead); Funding acquisition (lead); Project administration (lead); Supervision (lead); Writing – review & editing (lead).

The data that support the findings of this study are available from the corresponding authors upon reasonable request.