Designer DNA nanostructures for viral inhibition

The challenges underlying current coronavirus disease 2019 treatment and rapid diagnostic development are already well known from previous encounters with newly emerging pathogens (e.g., the 2009 H1N1 pandemic^1,2). Inhibition and treatment of virus infections typically relies on neutralizing antibodies (NAbs) that target virus surface-specific epitopes mainly in a one-to-one fashion³. Production of NAbs can be triggered by vaccination or active virus infection in the host. However, safe and effective vaccines normally take years to develop for an emerging virus. Therapeutic antibodies can be administered in response to viral infections. However, producing antibodies for treatment is very costly and time consuming. Importantly, NAbs may induce unwanted antibody-dependent enhancement of infection^4,5 (for example, with dengue virus (DENV) vaccine), where antibodies induce increased viral infectivity in vivo. Viruses present unique spatial patterns of antigens on their surfaces⁶. Such patterns facilitate multivalent binding of the virus to host cells for enhanced pathogenic infectivity. Based on this naturally occurring multivalent virus-cell binding mechanism, creating polyvalent virus entry blockers is a promising and practical approach to producing potent inhibitors of virus infections.

Development of the DNA star antiviral platform

Our group recently designed and synthesized a ~43 nm diameter star-shaped DNA architecture, called a ‘DNA star’ to multivalently bind to viral epitopes to efficiently inhibit virus infection. We designed the antiviral DNA star to specifically target complex epitopes on the DENV envelope protein domain 3 (ED3). The five-point DNA star provides structural rigidity to display ten dengue ED3 targeting aptamers in a 2D pattern precisely mirroring the complex spatial arrangement of DENV ED3s⁷. Each DENV (~50 nm diameter) can bind up to two DNA stars, one on each hemisphere. The DNA star demonstrated high virus-binding avidity and specificity to DENV and was a highly potent DENV inhibitor in human blood with a half-maximal effective concentration (EC₅₀) of 2 nM (~7,500-fold more effective than the monovalent aptamer). We anticipate that our strategy can be tailored to target any viral epitope pattern to combat emerging and re-emerging viruses by generating the requisite ligand patterns on customized designer DNA nanostructures (DDNs).

The DNA star was designed to afford local structural flexibility through a single-stranded DNA (ssDNA) region on each of the five internal edges and unpaired thymine bases (Ts) at the five internal junctions. The ssDNA regions offer limited but sufficient local structural flexibility to ensure effective binding to DENV under a range of physiological conditions and temperatures⁸, in which the pattern of DENV envelope proteins may change slightly. Additionally, the unpaired Ts introduced at the junctions allow bendability of the DNA star, enhancing its ability to wrap around and inhibit a virus particle. We found that more closely matching the geometry of the antigen pattern, by adding more triangles into the structure to create five- or six-point DNA stars whose vertices precisely match the interspatial pattern of ED3 clusters, results in stronger virus binding avidity. In theory, more points can be added to the current five- or six-point DNA star scaffold to increase avidity. However, larger structures assemble in a lower yield and require more DNA oligonucleotides, thus increasing cost. For a new virus target, an optimal DDN size can be determined by testing the virus binding avidity and antiviral activity of various DDN sizes. An optimal DDN construct can then be selected by collectively considering cost, assembly yield and antiviral performance of the candidate DDNs. The cost of making a DDN platform can be largely reduced if the DNA oligos are synthesized at a larger scale (>1 μmol from existing commercial sources such as IDTDNA or Sangon) or produced in a microbiome using molecular cloning^9,10,11,12.

The DNA star platform is a complex tile-based DDN that consists of multiple branched DNA junctions assembled from short DNA oligonucleotides^13,14,15. Despite growing interest in DNA origami-based biological applications^16,17,18,19, to date there are no comprehensive protocols to guide researchers through the process of creating tile-based DDNs for antiviral applications. Here we describe how to design tile-based DDNs, specifically antiviral DNA stars. The process begins by identifying antigen spatial patterns on virus particles using available cryogenic electron microscopy (cryo-EM) data. A DNA star with antigen-targeting ligands is then designed to mirror this spatial pattern. The binding avidity between the DNA star and intact viral particles is investigated using surface plasmon resonance (SPR) spectroscopy, and antiviral efficacy of DNA stars is evaluated using plaque-forming assays. While sample preparation strategies for atomic force microscopy (AFM) imaging of DNA origami nanostructures^16,17,20 are well established, we discuss the adjustments required to image small tile-based DDNs like the DNA stars. Furthermore, generic guidelines for employing SPR in studying individual ligand–protein interactions have been discussed in recent review articles^21,22,23,24. However, these scenarios do not recapitulate the polyvalent interactions between DNA stars and intact virus particles that we outline here. In addition, plaque-forming assays are commonly used to screen small molecule-, peptide- and protein-based antiviral drugs^25,26, but we have adapted the standard experimental conditions here to evaluate the potency of DNA star-based virus inhibitors. In our procedure, we provide a detailed step-by-step protocol for virus surface antigen pattern analysis, DNA star inhibitor design, synthesis, characterization and evaluation of DNA star cytotoxicity and antiviral efficacy.

Comparison with other methods

One strategy in the development of antivirals is to create materials designed to bind to a virus to prevent it from interacting with and entering host cells. Multivalent virus entry blockers have been previously constructed by linking epitope-targeting ligands to a synthetic scaffold to improve binding avidity^27,28,29. We recently designed dendrimer conjugates and demonstrated that matching average ligand spacing with the distance of viral epitopes is a key determinant for effectively inhibiting influenza viral infection in mice²⁷. However, synthetic scaffolds (polymers, dendrimers, nanofibers, inorganic nanoparticles, lipid nano-emulsions, etc.) exhibit substantial drawbacks as part of an antiviral ‘drug’, such as toxicity and limited control over the scaffold shape, ligand spacing and ligand valency^30,31,32,33.

Nucleic acid-based scaffolds are able to overcome some limitations of synthetic scaffolds because they can be designed into biologically stable and biocompatible 2D and 3D platforms while controlling ligand spacing, valency and spatial arrangements with nanometer precision^{7,34,35,36,37,38,39,40,41,42,43,44,45}. The DDN platform acts as an ideal template to display multiple binding motifs. For instance, a DDN platform has been used to arrange ligands with well-defined patterns to direct and regulate receptor activation, and thus invasiveness, of breast cancer cells⁴⁶. A DDN platform was also used to precisely display antigens to study optimal distances for bivalent antibody binding⁴⁷. Recently, virus-like DDN nanoparticles were created to display clinical vaccine immunogens to determine the impact of DDN particle size and rigidity, and immunogen spacing on B-cell receptor activation, and thus elucidate DDN-based vaccine design principles⁴⁸.

In terms of stability and biocompatibility, DDNs have exhibited blood circulation lifetime of up to 24 h (ref. ³⁸) and are eventually removed by the liver³⁸ and kidney^40,49. Certain DNA sequences can drive functional immune responses and thus can be used as vaccine adjuvants^50,51. Three-dimensional DNA origami structures can be designed to mimic virus-like particles and thus can exhibit relatively high immunogenicity^37,52. Coating DNA origami with biocompatible ligands (i.e., PEGylated lipids³⁷, BSA⁵², PEGylated oligolysines^53,54) increases the DDNs’ in vivo stability by reducing nuclease degradation and/or low salt denaturation, and also can be used to attenuate the immune response activated by these structures. This molecular coating strategy can be used to tune the in vivo lifetime and immune activity of a DDN for specific biomedical applications. Please refer to recent review articles for extended discussion on DDNs’ biostability and biocompatibility^55,56,57.

The building blocks of DDNs, DNA oligonucleotides, can be affordably produced at scale^{9,10,11,58,59}. They are thermally stable for hours or days in solution^7,60 and can be stored for months as a dried powder¹⁰. In response to emerging viruses and mutants, new high-affinity virus-targeting aptamer sequences can be readily and economically produced and evolved through the in vitro selection process, called systematic evolution of ligands by exponential enrichment (SELEX)^61,62,63. DDNs and aptamers have already found applications in biomedicine and biosensing^{7,38,64,65,66,67,68,69,70,71,72,73,74,75,76,77,78,79}. Additionally, other antigen-binding ligands such as antibodies or nanobodies can be decorated on customized DDN scaffolds to target viruses or to serve as coating molecules to enhance DDN in vivo stability and compatibility.

Application of the DNA star platform

The threat of emerging and re-emerging infectious diseases has emphasized the need to develop robust surveillance technologies and novel therapeutics. The DNA star platform presented here has been effectively used to both detect and inhibit DENV. We anticipate that our platform can be tailored and tuned to target antigens displayed on other viruses. The DNA star can be used as a multivalent sensor for diagnostic testing, or as a therapeutic candidate potentially capable of inhibiting in vivo viral entry into host cells (Fig. 1a). The DNA star is designed to interact with the intact/infectious virions with partially or fully preserved surface antigens that interact with host cell receptors. Thus, it may be possible to use a DNA star sensor to distinguish between infectious and noninfectious forms of a virus and minimize false positive test results that can be generated by nucleic acid-based tests such as PCR. Additionally, DNA structures customized to target different viruses can be designed to display distinct fluorescent readouts for multiplexed diagnostics.

Beyond the potential clinical applications in diagnostics and antiviral therapeutics, it may also be possible to use the DNA star strategy for the following five purposes. (1) Virus titer determination: a fluorophore-labeled DNA star may be used to quantify virus titers of stocks, clinical samples or environmental samples, for instance, by sorting the DNA star–virus mixture using flow virometry technique^80,81, a derivative of flow cytometry. (2) Purification: DNA star may be incorporated into resins to improve purification of viral particles manufactured in cell culture (e.g., to purify virus particles that are used in gene therapy) (Fig. 1b). (3) Imaging and tracking: fluorescently labeled DNA stars may be used as a molecular tag to track how the intact/infectious virions behave in biological fluids and environments (Fig. 1c). Furthermore, the specificity provided by newly customized DDNs can be tailored to other scenarios such as single-cell imaging. The architecture of cells and distribution of cell surface receptors is typically altered as cells transition between different developmental stages and physiological states. We can use the availability of these cell surface receptors as targeting moieties to perform single-cell analysis on multiple populations of cells. (4) A molecular ruler: for an emerging virus that has one or more previously known ‘cousins’ from the same virus family/subfamily, a candidate DDN can first be designed to match the antigen spatial pattern of a cousin virus to get an approximate structure for the emerging virus before the cryo-EM data are obtained. The candidate DDN can be turned into a ‘molecular ruler’ by presenting affinity ligands at varied distances compared with the initial design to elucidate the optimal antigen pattern on the outer surface of the emerging virus (Fig. 1d). Identifying the distribution and arrangements of viral antigens may assist in vaccine and antiviral drug development. (5) Binder selection: a newly customized DDN can serve as a synergetic platform to template multiple DNA libraries for the simultaneous evolution and selection of specific viral antigen-binding aptamers in a multivalent and pattern-matching interaction with the virus particles (Fig. 1e).

Overall, we envision that this protocol can be used to develop customized DDNs that can be tailored to address a broad range of biological and virological questions. Implementation of this protocol requires a multidisciplinary team of structural biologists, chemists and virologists, and may be useful to researchers in the fields of drug design, biology, materials science, bioengineering, nanotechnology and chemistry.

Challenges of applying the ‘DNA star strategy’ to an emerging virus

Designing an antiviral DNA star is dependent on our ability to identify unique patterns of antigens displayed on the surface of a viral particle. When using our DNA star strategy to tackle an emerging virus, two key pieces of information are needed: knowledge of antigen arrangement on the outer surface of the virus (e.g., from cryo-EM data) and availability of antigen-binding ligands (e.g., antibodies, nanobodies or aptamers). Most emerging viruses have one or more previously known ‘cousins’ from the same virus family/subfamily (i.e., severe acute respiratory syndrome coronavirus and severe acute respiratory syndrome coronavirus 2). They differ in antigen sequences but have the same or similar antigen arrangement. Thus, the DDN design can start with a known antigen pattern for a similar virus. The DDN design can then be further optimized based on the confirmed patterns from cryo-EM data. It normally takes months to obtain antibodies from recovered patients or to engineer nanobodies that can specifically bind to the antigen of an emerging virus. In contrast, it is much easier and faster to derive and produce aptamers that target an emerging virus antigen and its mutants with at least submicromolar avidity. Our DNA star-based virus capture probe allows for polyvalent and spatial pattern-matching interactions, dramatically improving virus-binding avidity compared with individual monomeric aptamers⁷. As a result, a modest affinity aptamer can be turned into an excellent binder when presented on a multivalent DDN structure.

As described in this protocol, live DENV is used under biosafety level (BSL)-2 containment to measure DNA star antiviral efficacy. However, highly virulent and pathogenic viruses such as severe acute respiratory syndrome-associated viruses must be handled under BSL-3+ containment. Therefore, when targeting these highly virulent viruses, we suggest initially screening candidate antiviral DNA star constructs using pseudoviral assays (i.e., pseudomodels of severe acute respiratory syndrome coronavirus 2 (ref. ^82,83,84)) before working with the live virus.

Limitations of the DNA star platform

The in vitro DENV inhibition performance of the DNA star platform shows promising therapeutic potential. However, biostability of the DNA star, which is composed of unmodified DNA oligos, must be improved before moving into in vivo assays and potential future clinical use. Chemically modified component DNA oligos can be used to improve the DNA star’s in vivo nuclease resistance and biostability⁵⁷. Unpaired thymine nucleotides introduced at each vertex can be UV crosslinked⁴² to improve the DNA star’s lifetime in vivo. Additionally, the aptamer targeting ligand can be substituted with antibodies or peptides to improve virus-binding affinity. Protein or peptide coatings on the DNA star can also improve its biostability and repress potential immunogenicity, as observed for other DNA nanostructures^37,52.

When applying the DNA star platform to biosensing (which is not outlined in the Procedure here), the sensor relies on the strong interaction between the patterned aptamers on the DNA star and antigens on the surface of the virus⁷. For DENV, the envelope proteins on the virus surface act as steady anchors that, when bound to the DNA star, can effectively stretch the hairpin loops in the molecular beacon-like motifs on the DNA star, which can trigger the release of a quencher dye and restore a fluorescent detection signal. However, unlike DENV antigens that are immobile, spike glycoproteins on viruses such as coronaviruses, influenza and human immunodeficiency virus have flexible stems and mobile roots, which cannot serve as steady anchors to stretch the hairpin loops of DNA star sensors to report viral binding. Thus, the mechanism by which the detection signal is generated with the current DNA star strategy cannot be applied to viruses or cells whose surface proteins are flexible and mobile. For these targets, it may be possible to integrate different signal generation and detection technologies (e.g., lateral flow assay⁸⁵, light scattering imaging⁸⁶, aptamer switching⁸⁷) with the DNA star platform to enable rapid, ultrasensitive and inexpensive disease detection.

Overview of the procedure

Here we describe how to develop DDNs for inhibition of DENV. This protocol comprises five main stages as illustrated in Fig. 2. The procedure starts with identifying unique spatial patterns of antigens on a viral particle surface using available cryo-EM data of dengue virions (Fig. 2a). After the virus antigen spatial pattern is determined, the next step is to design a DNA nanostructure mirroring this pattern using a combination of in silico and our in-house DNA design principles (Fig. 2b). Component DNA oligos are purified by denaturing gel electrophoresis, and then the DDN is self-assembled via thermal annealing. The structural formation of a DDN is characterized first by nondenaturing gel electrophoresis and then by AFM (Fig. 2c). Interactions between DDNs and intact virus particles are analyzed using SPR spectroscopy to determine the binding avidity (Fig. 2d). For DENV, 3-(4,5-dimethylthiazol-2-yl)-2,5-diphenyl tetrazolium bromide (MTT) and plaque-forming assays are used to evaluate DDN’s cytotoxicity and antiviral efficacy, respectively (Fig. 2e). While not specifically outlined in this Procedure, Fig. 3 illustrates the ‘mix-and-read’ strategy for DDN-based virus sensing.

Experimental design

This protocol should be accessible to researchers with training in structural biology, DNA nanotechnology or biochemistry. The experiments must be conducted under suitable biosafety containment based on the virus of interest. The DNA star protocol is broken down into subsections, which should be adaptable to other DNA nanostructures: (a) purification of DNA oligos, (b) DDN self-assembly and gel/AFM characterization, (c) purification of the DDN complex, (d) SPR assays and (e) plaque-forming assays. Preparation of control samples for experiments throughout the Procedure are outlined in ‘Reagent setup’. DDN stability can be monitored by gel electrophoresis of the DDN samples after incubating in physiological environments up to 24 h (ref. ⁵⁷).

Identification of virus antigen spatial patterns (Steps 1–5)

Advances in the field of structural biology have made it possible to develop methods such as cryo-EM that provide us with the ability to visualize proteins and protein complexes at up to atomic resolution^88,89. Cryo-EM has been used to determine the arrangement of proteins in macromolecular assemblies such as the dengue viral particle, which has enabled an understanding of the mechanism of viral infection^48,49. The structure of the dengue viral particle (PDB: 1p58) determined by Zhang et al.⁹⁰, was used to investigate the distribution of the ‘QHGTI’ amino acid sequence motif in the ED3 across the entire viral particle (Fig. 4a). Considering that the 1p58 biological assembly contains 12 five-point stars, with each point on a star representative of three monomers of the envelope protein (Supplementary Fig. 1), we first needed to identify individual monomers containing neighboring epitopes that were surface accessible before determining the distance separating them. The ‘QHGTI’ motif was previously identified as part of a conserved region in B- and T-cell epitopes of dengue envelope glycoprotein that could be used to inhibit viral infection⁹¹. Once these individual epitopes were identified, we next determined unique patterns of the distribution of epitopes accessible on the surface of the viral particle (i.e., not buried) (Fig. 4b,c). To accomplish this, we measured the distance between the Cα of the middle residue ‘G’ in neighboring ‘QHGTI’ motifs. The arc length was then measured to determine the surface distance between individual epitopes in each ED3 cluster and in between ED3 clusters (Fig. 4d and Fig. 5a).

Design of the DNA star (Steps 6–12)

Our analysis of the DENV cryo-EM data⁹⁰ has focused on the ED3 viral epitope because it is a major interaction point for DENV NAbs that strongly inhibit DENV infection^{92,93,94,95,96,97}. This motivated us to develop a DNA-based viral inhibition strategy targeting the ED3 epitopes. QHGTI sequence motifs in ED3 were found to organize into a repeating pattern on the DENV surface with alternating trivalent or pentavalent clusters (Fig. 5a). By connecting the ED3 trivalent and pentavalent cluster sites linearly (centered on a pentavalent cluster), we identified a star shape, consisting of an interior pentagon connected to five exterior triangles, which we hypothesized would provide an optimal multivalent scaffold for high DENV binding avidity (Fig. 5a). The orthodromic center to center distances between trivalent–trivalent and trivalent–pentavalent ED3 clusters are respectively 15.3 nm and 14.3 nm.

Based on structural information, we designed a DNA star scaffold, with vertices matching the spatial pattern of the ED3 clusters on DENV surface. Since the size of the DNA star was smaller (~43 nm diameter) and simpler than a typical DNA origami structure⁴³, we used DNA tile nanostructure design principles (e.g., minimizing sequence symmetry) and assembly strategies^{7,13,98,99,100,101,102}. Specifically, the five-point DNA star consists of five geometrically identical isosceles triangles. The orthodromic center-to-center distances between trivalent–trivalent and trivalent–pentavalent ED3 (or other antigen) clusters determine the lengths of the isosceles triangle internal and external edges as 52 bp and 42 bp long, respectively (Fig. 5b). As the design steps are further illustrated in Fig. 6 and elaborated in Steps 6–12 of the Procedure, the DNA star is constructed using 25 ssDNA oligonucleotides: five ‘scaffold’ strands, ten ‘edge’ strands, five ‘close’ strands and five ‘fix’ strands (Fig. 5c). The specific sequences we used were obtained in SEQUIN program^13,103, in which we were able to input the structural information and generate the DNA sequences that assemble into a structure to match the DENV surface antigen geometric arrangement. The ‘scaffold’ and ‘edge’ DNA constitute internal edges of the isosceles triangles. The ‘close’ and ‘edge’ DNA form external edges of the isosceles triangles. The five ‘fix’ strands connect the five isosceles triangles together to assemble a full star structure. The DNA star is assembled of five rigid four-way crossover structures at the five pentagon corners. Linking them together without providing flexible ssDNA linkers may cause torsion in the whole structure, thus deforming canonical B-DNA duplex to make the DNA star structure unstable. In this regard, measures must be taken to provide the crossovers in the DNA star with more flexibility to relieve such torsional stress. For this purpose, several noncomplementary thymine bases (free Ts) can be added at the corner of internal pentagon crossovers and at the five tips of the star (indicated by the green lines in Fig. 5c,d). Different numbers of free Ts at the crossover define different angles between its two-flanking duplex: adding more free Ts results in a smaller crossover angle^104,105. Thus, the optimal number of free Ts can be determined based on the crossover angles required to form a DNA nanostructure. For our (five-point) DNA star, the three respective angles of scaffold/edge–edge/fix, edge/fix–edge/fix, and edge/close–edge/close duplexes, in theory, are 51.6°, 148.8° and 76.8°. These angles dictate the number of free Ts used at the respective corners, which are 7, 3 and 7 nt. After adding the free Ts at the crossovers, the DNA star structure becomes more flexible, allowing it to better wrap around a virus particle. If applying the DNA star strategy to target a different virus, we can repeat the entire DNA nanostructure design procedure, starting with identifying unique spatial patterns of antigens on a viral particle surface using available cryo-EM data of the virion of interest, then design a DNA nanostructure mirroring this pattern, and automatically program the DNA sequences for such a DNA nanostructure using the SEQUIN program.

The DNA star is functionalized for inhibition and detection of DENV by conjugating an ED3-targeting aptamer at each of the ten vertices of the DNA star scaffold to form a DNA star–aptamer complex that superimposes the spatial pattern of ED3 clusters on DENV surface (Fig. 5d). To turn the star–aptamer complex into an effective DENV inhibitor and sensor, we introduced a ssDNA region into each of the five internal star edges to form a hairpin structure with 7 bp stem and 6 nt loop. The hairpins unzipping into ssDNAs stretches each internal edge of the DNA star to a distance that fits the spacing between adjacent trivalent ED3 clusters. These hairpins play a dual function. First, for DENV inhibition, they provide the overall rigid DNA star scaffold with local structural flexibility to ensure binding and inhibition of DENV under various physiological environments and temperatures⁸. DENV typically enters cells first through electrostatic interactions with the host cell plasma membrane and glycosaminoglycans. After binding to the DENV surface with high avidity, the DNA star physically and electrostatically (through interactions between DENV surface and negatively charged DNA scaffold) traps and prevents virions from interacting with and entering host cells. Second, for DENV detection (which is not a focus of this protocol), hairpins provide the DNA star with structurally switchable motifs to report the presence of DENV in a sample. Specifically, fluorophore and quencher carrying strands (e.g., FAM-BHQ1 pair or Cy3-BHQ2 pair) are hybridized to each inner edge of the star so they are brought together to flank the hairpin by canonical Watson–Crick base pairing, much like a molecular beacon motif. However, unlike a molecular beacon that generates a fluorescence signal for target ssRNA or ssDNA hybridization, on the DNA star hairpins are pulled apart as a result of strong aptamer–ED3 binding. Such potent interactions can separate fluorophores from quenchers to restore fluorescence as a signal readout.

Denaturing PAGE for oligonucleotide purification (Steps 13–41)

DNA oligo synthesis using phosphoramidite chemistry^106,107,108 typically generates truncated products that can interfere with formation of DNA nanostructures. A critical prerequisite for ensuring the success of DNA star assembly is to obtain purified full-length products of each of the component DNA oligos. Denaturing polyacrylamide gel electrophoresis (PAGE) can separate full-length products from shorter species (including n-1 species) with high efficiency. PAGE purification is effective for separating unmodified oligos from unwanted truncated products. The component DNA oligos for the self-assembly of the DNA star range from 40 to 80 nt. Different lengths of DNA strands can be resolved by altering the concentration of polyacrylamide gel, where a higher gel percentage is used to purify shorter DNA oligos. After electrophoresis, the desired DNA band is identified using UV visualization, excised, and eluted from the gel using elution buffer (Buffer 3) and concentrated using ethanol precipitation. Precipitates of the DNA oligos are suspended in nuclease-free water, and DNA concentration is quantified using an absorbance measurement at 260 nm. In comparison with other purification methods, like HPLC, denaturing PAGE offers high-resolution separation of full-length products from shorter side products with length differences as small as a single nucleotide. Several samples can be run simultaneously without requiring any special equipment.

Assembly of the DNA star (Steps 42–46)

Forming a DNA star with high yield relies on three factors: a close to 1:1 stoichiometric ratio among all component DNA strands, a suitable annealing buffer and an appropriate thermal annealing protocol. To construct the DNA star (100 μL in each PCR tube at a final concentration ranging from 0.1 to 0.4 μM final concentration of DNA star complexes for different applications), stoichiometrically equivalent quantities (scaffold strands: edge strands: fix strands: close strands = 1:1:1:1) of all component strands are thoroughly mixed in 1× TAE-Mg²⁺ (Buffer 5) annealing buffer. The mixture is then heated in a standard thermal cycler to 95 °C and incubated for 5 min before it is slowly cooled down from 95 °C to 20 °C at a rate of 0.15 °C per min over 8 h (an overnight protocol). Samples are incubated for 5 min at a high temperature (80–95 °C) to break all previously formed Watson–Crick base pairs to ensure that all component DNAs can hybridize with their correctly designed partners. A shorter annealing protocol may work as well, but we have not tested it for the formation of DENV-targeting DNA stars⁷.

Nondenaturing gel electrophoretic characterization and purification of the DNA star (Steps 47–67)

Nondenaturing (native) gel electrophoresis is a reliable method for quickly characterizing the structural formation and yield of a DNA nanostructure assembly. There are two gel matrices to choose from based on the size and complexity of DNA nanostructures. Nondenaturing agarose gel electrophoresis (AGE) is normally used to characterize DNA assemblies with very high molecular weight or complex nanostructures such as 2D and 3D DNA origami. Nondenaturing PAGE is normally used for DNA nanostructures such as DNA tile motifs with a molecular weight <1,000 base pairs (bp). Since certain 3D nanostructures are too big to travel through the polyacrylamide gel matrix, they may not enter the PAGE. Thus, higher-percentage AGE (e.g., 3%) can also be used to characterize the formation of tile-based DNA motifs, but with lower-resolution molecular weight separation (20–30 bp) among different DNA species. The size of the full DENV-targeting five-point DNA star is 615 bp (or ~470.80 kDa). Its length and width in 2D space are ~39.26 nm and ~37.34 nm, respectively. Because the DNA star has a relatively low molecular weight and small size, and the molecular weight difference between the partial and full DNA star is small, we use a 4% nondenaturing PAGE gel to characterize the yield of the structure formation presented here. The electrophoresis is run in 1× TAE-Mg²⁺ buffer (Buffer 5) at 60 V for 90 min. The gel running buffer contains magnesium ions, which can generate heat during electrophoresis. Therefore, we put the gel box either in a 4 °C cold room or in an ice bath to make sure the gel box is not overheated, so it does not denature the DNA star complex. Post electrophoresis, gels can be stained with DNA intercalating reagents such as GelRed, ethidium bromide or Sybr dyes. The resultant PAGE of the partial or full DNA star typically shows a distinct and dominant band corresponding to each of the DNA complexes, confirming that DNA star structures are assembled in high yield. It is worth noting that byproducts with lower molecular weight may form as a result of imperfect stoichiometric ratios among the component DNA oligos. After confirming the yield by PAGE gel, the desired DNA stars may be purified as intact complexes using a high percentage (i.e., 3%) nondenaturing AGE gel and dialysis for DNA star elution.

AFM characterization of the DNA star (Steps 68–73)

AFM imaging can show the entire morphology of a DNA assembly at <5 nm resolution. AFM is particularly useful for visualizing large DNA 2D arrays and DNA origami nanostructures. For more fragile tile-based DNA star nanostructures, special care is needed during AFM sample preparation. This includes modifying the mica substrate surface with positively charged (3-aminopropyl)triethoxysilane (APTES) to facilitate sample deposition of negatively charged DNA¹⁰⁹, and using ‘soft tapping in fluid’ mode on the AFM. The AFM images of the DNA stars in this protocol are obtained on a Bruker MultiMode VIII microscope in 1× TAE-Mg²⁺ buffer (Buffer 5) with SNL-10 probe.

Purification of the DENV (Steps 74–132)

Purified virus is required for SPR analysis of the interaction between viral particles and the DNA star complexes. We provide a DENV purification protocol that is streamlined for laboratories that do not have access to a gradient maker. This protocol is optimized for small-scale virus preparation for up to 60 mL supernatant of DENV-infected cells^110,111, typically from four T-75 cell culture flasks (called T-75 flask herein). DENV is a human pathogen and can cause serious illness. Thus, proper precautions should be taken to ensure the safety of everyone in a BSL-2 laboratory. Researchers need to consult with their institution’s safety office with regard to appropriate personal protective equipment (PPE) and handling precautions. Also, we highly recommend consulting the latest edition of Biosafety in Microbiological and Biomedical Laboratories (BMBL)¹¹² by the Centers for Disease Control and Prevention (CDC) and the National Institutes of Health (NIH) about regulations on handling biohazardous materials.

SPR analysis of the interaction between viral particles and the DNA star complexes (Steps 133–141)

SPR spectroscopy is based on an optical biosensor technique that measures molecular binding events at a metal surface by detecting changes in the local refractive index¹¹³. It is a powerful tool to measure biomolecular interactions in real time and in a label-free environment. SPR assays can typically answer the following questions in regard to ligand–ligand interaction: (i) How specific is an interaction? (ii) How strong is an interaction, and what is the binding affinity (K_D)? (iii) How fast is an interaction, and what are the association and dissociation rate constants (k_a and k_d)? (iv) What are the thermodynamic parameters (enthalpy (∆H) and entropy (∆S)) for an interaction? (v) What is the biologically active concentration of a specific molecule in a sample? In this protocol, intact virions are immobilized on the SPR chip surface to characterize their interaction with DNA star nanostructures for their relative and absolute binding affinities. This setup recapitulates the scenario for DNA stars interacting with intact virions for virus inhibition. The interactions between viral particles and DNA stars are analyzed using the SPR instrument BIAcore 3000.

MTT assay to measure DNA star cytotoxicity (Steps 142–155)

To assess the cytotoxicity of DNA star nanostructures, MTT assay is used to measure cell viability and proliferation. In the MTT assay, cellular oxidoreductase/dehydrogenase enzymes in living cells reduce the yellow tetrazolium dye MTT to an insoluble formazan, which has a purple color, which can be easily quantified by measuring absorbance at 570 nm on a spectrophotometer or plate reader. The absorbance is directly proportional to the number of viable cells. In this assay, cytotoxicity is evaluated by measuring the inhibition of cell proliferation^114,115. This measurement of cell proliferation and viability in terms of reductive activity is widely accepted as a reliable way to examine the cytotoxicity effect of drugs or compounds^116,117. Serially diluted DNA star samples as well as controls with only the dilution buffer are added to the Vero cell monolayer in triplicates. Average of the triplicate reading for each sample can be used to calculate the percentage of cytotoxicity with the following equation: % cytotoxicity = (100 × (control − sample) / control).

Plaque reduction test (antiviral assay) (Steps 156–166)

The plaque reduction neutralization test is considered the ‘gold standard’ in quantifying the titer of neutralization antibody for a virus. Here we use a modified plaque reduction neutralization test assay, called the plaque reduction test, to measure the inhibition of plaque formation efficiency by an antiviral drug or compound^{118,119,120,121}. The solution of antiviral compound is diluted serially and mixed with a known concentration of virus. After a period of incubation allowing the drug to act on the virus, the mixture is added to cell monolayers for assessment. The assay conditions can be further modified to assess the possible mode or timing of action. The assay is typically performed in triplicate, and the arithmetic mean is used to calculate the percentage reduction in plaque numbers by statistical analysis. A dilution buffer without any test compound should always be included to measure the background of the assay conditions. The concentration of test compound to reduce the number of plaques by 50% (EC₅₀) or 90% (EC₉₀) on a dose–response curve, when compared with the medium or buffer controls gives a measurement of the compound’s antiviral efficiency.