RESEARCH @ THE UNIVERSITY OF ILLINOIS, URBANA-CHAMPAIGN
Aptamer Chiral Arrangement on DNA Origami Control Cellular Outcomes
We combined DNA origami and test the hypothesis that different spatial ligand patterning can dictate cellular uptake behavior. We designed rigid DNA origami nanotubes and decorated them with multiple identical aptamers positioned at defined locations on the tube surface, allowing us to construct left-handed (L-CAP) and right-handed (R-CAP) chiral architectures without altering chemical composition. Structural fidelity and handedness were validated through computational design and high-resolution structural characterization, while receptor engagement and uptake were evaluated using fluorescence microscopy and quantitative cellular internalization assays. By correlating structural chirality with receptor organization at the plasma membrane, we observed that only L-CAP configuration effectively induces receptor dimerization, leading to robust endocytic internalization. The mirror-image R-CAP structure failed to induce dimer formation and largely remained surface-associated. This systematic, geometry-controlled approach allowed us to directly link three-dimensional ligand arrangement to biological function, thereby experimentally validating our hypothesis.
_edited_edited.jpg)
Figure: Left-handed Chiral Aptamer Pattern (L-CAP) on DNA-Origami exhibits enantioselective receptor dimerization and cellular uptake.
Our findings have important implications for precision nanomedicine. Chirality-encoded DNA nanostructures offer a powerful strategy for selective drug delivery, diagnostics, and receptor-specific signaling control, while minimizing nonspecific uptake. More broadly, this work establishes geometry as a programmable biological signal, opening new directions for designing therapeutics that regulate cell behavior through nanoscale spatial organization rather than chemistry alone (Read the full text here).
Engineering DNA Nanoarchitectures for Precision Cancer Therapeutics
While innovative cancer therapies such as vaccines and CAR-T cells are gaining momentum, their variable efficacy and uncertain long-term effects have sustained chemotherapy’s central role in cancer treatment. However, conventional chemotherapy is often limited by off-target toxicity and poor tumor specificity, particularly in diffuse malignancies like acute myeloid leukemia (AML), where relapse is frequently driven by treatment-resistant leukemic stem cells (LSCs).
Figure: Schematic overview of the design and function of the Designer DNA Architecture–Drug Conjugate (DDA-DC).
In our recent work, we identified a distinct biomarker combination—CD117 and CD123—selectively expressed on AML LSCs, and developed single-stranded DNA (ssDNA) aptamers to target them. Notably, some of these aptamers independently induced apoptosis in AML cells (Kasumi-1) by activating intrinsic cell death pathways. Leveraging their DNA-binding capability, these aptamers also facilitated efficient delivery of the chemotherapeutic drug daunorubicin without requiring complex chemical modifications. To enhance selective targeting, we engineered a Designer DNA Architecture (DDA) loaded with daunorubicin-conjugated aptamers specific to CD117 and CD123, forming DDA-Drug Conjugates (DDA-DCs). These constructs demonstrated precise targeting and effective elimination of AML cells in both ex vivo and in vivo models, while reducing the required daunorubicin dosage by over 500-fold ex vivo and 10-fold in vivo. Building on these results, we are now developing advanced DDA-DCs for solid tumors, using AND Gate logic, we are engineering pH-responsive systems (Read the full text here).
Honeycomb Shaped DNA Nanostructure for Influenza Neutralization
To address the urgent need for broad-spectrum antivirals against rapidly mutating influenza A virus (IAV) strains, we developed a modular, programmable platform based on a honeycomb-shaped designer DNA nanostructure (HC-DDN). By leveraging the principles of geometry-matched multivalency, we precisely arranged hemagglutinin (HA)-targeting ligands in trimeric clusters to mirror the native spatial distribution of the virus's own HA trimers. Our research demonstrates that this spatial organization dramatically enhances the binding avidity and neutralization potency of three distinct classes of binders: host defense peptides (Urumin), single-domain nanobodies, and DNA aptamers.Through molecular dynamics simulations and surface plasmon resonance (SPR) assays, we discovered that our multivalent constructs achieve a 100-to-1000-fold increase in binding affinity compared to their free, monomeric forms, reaching sub-picomolar to picomolar levels. Specifically, we found that while free Urumin only inhibits H1 subtypes at micromolar doses, our HC-Urumin construct achieves over 98% neutralization of both H1N1 and H3N2 subtypes at nanomolar concentrations. Similarly, our HC-Nanobody construct achieved greater than 99% inhibition of viral entry, significantly outperforming free nanobodies in both murine and porcine respiratory epithelial models.
Figure: Schematic overview of the design and function of the design and working principle of the Honeycomb DNA Nanostructure feature multiple binder types.
In our biological assays, we observed that these multivalent platforms provide superior cytoprotection, improving cell viability by up to 55% relative to free ligands across various IAV subtypes. Furthermore, our in vivo studies using a murine H1N1 infection model showed that treatment with HC-Urumin effectively reduced disease severity and decreased mortality rates from 50% in the free-peptide group to just 20%. Importantly, we confirmed that our DNA scaffolds are safe and biocompatible; they do not induce a detectable anti-DNA antibody response or impair the host's ability to generate virus-specific adaptive immune responses.We believe this "plug-and-play" strategy offers a versatile foundation for the rational design of next-generation antivirals. Because our platform is modular, we can easily customize it with different payloads to target a broad spectrum of rapidly evolving respiratory pathogens beyond influenza, including RSV and coronaviruses (Read the full texts here and here).
APIPred: Accelerating Aptamer Discovery Using Machine Learning Tools
In the summer of 2022, I was provided the opportunity to mentor four undergraduates visiting our laboratory. One small caveat- they were computer science majors! Amalgamating their expertise in machine learning and our knowhow of aptamer selection, we asked a question- can machine learning tools aid in aptamer discovery? To overcome the limitations of traditional SELEX, we first developed APIPred, a novel XGBoost-based machine learning algorithm designed to predict aptamer-protein interactions with high precision. By integrating multiple feature extraction methods, such as k-mer frequencies for aptamers and pseudo amino acid composition for proteins, we achieved a benchmark accuracy of 96.5%. We successfully validated this model through molecular docking and SPR assays, confirming its ability to identify robust binders for targets like interleukin-23 and streptavidin. (Read the full text here).
.png)
Figure: APIPred's machine learning modules, predictive capabilities and experimental validation.
Building on this foundation, we evolved our technology into APIPred Web 1.0, a unified web platform that provides a user-friendly interface for the broader research community. This platform expands our original pipeline by integrating constraint-aware library generation—enforcing biological limits on GC content and base repeats—with parallelized secondary-structure analysis using ViennaRNA. To handle high-throughput demands, we optimized the backend with vectorized batch processing to efficiently rank up to a million candidates per request. We demonstrated the platform's practical utility by rapidly generating specific binders for the CD64 protein, whose efficacy we confirmed via flow cytometry. Collectively, our scalable ecosystem streamlines the transition from target sequence to validated binder for diverse biomedical applications. (Read the full text here)