Research

Where do "consciousness" and "thoughts" come from?

These fascinating higher-order brain functions emerge as a result of a dynamic "chemical dialogue" among billions of cells—most of which remains invisible nowadays. We develop materials-driven tools by combining synthetic nanomaterials and biological matter to uncover this hidden dialogue.

Our Approach

Over billions of years, nature has evolved extraordinarily sophisticated systems. Meanwhile, human has engineered synthetic materials, such as single-walled carbon nanotubes (SWCNTs), that possess rare properties not found in nature.

When these different materials meet, they can trigger fascinating nanoscale phenomena—such as DNA spontaneously wrapping around a SWCNT in a highly ordered structure.

We leverage these unique interactions to engineer functional tools. By doing so, we create novel measurement platforms that expand our perception and illuminate the hidden dynamics of life.

Research Topics

1. Engineering the Corona with Precision

The performance of nanomaterials is often hindered by structural heterogeneity and thermal fluctuations at their interfaces. We overcome this challenge by leveraging unique, spontaneous interactions at the nano-bio interface to precisely control molecular corona structures at the single-nanoparticle level.

Nishitani et al., Angewandte (2026)
×
Monovalent Functionalization of Carbon Nanotubes
In this work, we addressed the limitation of structural heterogeneity at the nano-bio interfaces by developing a DNA-guided, bottom-up route to synthesize uniform ensembles of SWCNTs with a single molecular tag. The key insight is to treat ssDNA adsorption to SWCNTs as a stochastic selection and binding process: mixing a small fraction of modified ssDNA (m-ssDNA) with unmodified ssDNA (um-ssDNA) yields a binomial distribution of m-ssDNA counts per nanotube whose mode and tails are tunable by the input ratio p (m-ssDNA:ssDNA) (a). Accordingly, the model predicts purity-recovery trade-offs and identifies operating points where above 90% of the m-ssDNA-bearing SWCNTs are singly modified at p < 1% (e.g., 91% at p = 1%) with predicted purity reaching > 97% at p = 0.25% (b). Experimentally, we implemented a magnetic bead-assisted isolation that leverages biotin-streptavidin capture on m-ssDNA and a photocleavable linker for recovery, enabling enrichment of the rare, functionalized population without compromising colloidal integrity. At p = 0.25%, this workflow recovered ∼0.05 μg of nanotubes from a 2.5 μg input (2% overall), with > 80% capture efficiency; these values closely match model predictions for the m-ssDNA-bearing fraction (3.4%) and single-site purity (97.6%).
Monovalent tag occupancy was directly validated at the single-nanotube level using single-molecule total internal reflection fluorescence (smTIRF) microscopy, where we directly quantified the number of fluorophore tag per nanotube by tracing photobleaching trajectories. Under the model-selected condition of p = 1%, step-counting analysis revealed that approximately 90% of the fluorescence trajectories exhibited a single discrete bleaching event (n_steps = 1), corresponding to a narrow pre-bleach intensity distribution, demonstrating a close convergence between experimental single-molecule metrology and the stochastic model predictions (c). This batch-scale, facile workflow converts a major source of heterogeneity into a controllable design parameter, enabling deterministic single-site molecular tagging for quantitative probe metrology, precision cell targeting, and potentially converting SWCNTs into a single-photon emitter for the basis of quantum computing.
(a) Schematic of the stochastic occupancy model where ssDNA strands are sampled stochastically, yielding a binomial distribution of modified strands per nanotube. (b) Predicted single-site purity and theoretical maximum recovery as a function of the input fraction p. (c) Monvalency verification via smTIRF microscopy: representative fluorescence fields of view, a schematized surface immobilization platform, and photobleaching step-number distributions demonstrating predominant monovalency (∼90% single-step events) at p = 1%.

2. De Novo Design of Nanoscale Probes

Imagine a nanoparticle that illuminates only when it encounters a specific biomolecule under a specific condition. By understanding i) the biomolecular recognition at the corona phase (signal input) and ii) how these interactions modulate the optical properties of nanomaterials (signal output), we rationally design nanoscale probes with programmable functionality.

Nishitani et al., PNAS (2025)
×
Fluorescence Energy Transfer of Carbon Nanotubes
To overcome the intrinsic selectivity limits of conventional ssDNA–SWCNT designs, we coupled molecular recognition to a distance-controlled charge-transfer pathway mediated by the redox dye methylene blue (MB). MB reversibly adsorbs to SWCNTs and quenches their NIR emission via photoinduced charge-transfer scheme (a); importantly, this quenching is stable in redox-rich biofluids (including human plasma) and is reversed only upon MB displacement from the nanotube surface. Leveraging this unique mechanism, we tethered MB to the target-binding domain of ssDNA so that target binding physically displaces MB away from the SWCNT, yielding a predictable NIR turn-on. Quantitative analyses revealed a usable FRET-like distance window of ~6.8 nm from the SWCNT surface for the first time and identified an optimal MB loading that maximizes transduction, achieving predictable, reproducible NIR intensity increases of up to 80%.
Guided by these principles, we developed a versatile nanosensor platform for detecting nucleic acids and proteins. For oligonucleotides, MB displacement directly transduced hybridization into NIR intensity changes across multiple target sequences, enabling sequence-specific detection of Tobacco Mosaic Virus (TMV) fragments and distinguishing TMV-infected plant extracts from mock controls (b). Extending beyond nucleic acids, a biotin–MB construct selectively reported neutravidin with ~100% intensity increases, whereas control proteins elicited minimal response (c). Importantly, these nanosensors exhibit selectivity over various molecules that fully turned on conventional nanosensors (d). Together, these results demonstrate a generalizable, rational route to design SWCNT nanosensors in which selectivity derives from programmed molecular motion of a quencher, rather than promiscuous redox chemistry, thereby addressing a central barrier to translation of NIR nanosensors in biological applications.
(a) MB quenches SWCNT fluorescence in a concentration-dependent manner. (b) Conceptual representation of oligonucleotide nanosensors in which hybridization induces MB desorption (top); sequence-specific detection of Tobacco mosaic virus (TMV) RNA in infected plants enabled the detection of virus-infected plants. (c) Translation to protein detection exemplified by robust biotin–neutravidin binding. (d) Nanosensor exhibited selectivity to the target while remaining insensitive to the redox molecules.
Nishitani et al., Angewandte (2024)
×
Engineered Enzyme–Carbon Nanotube Hybrids
Proteins provide defined molecular recognition and are therefore well suited as templates for rational nanosensor design. In this study, we established a robust, facile sonication-based protocol to generate glucose oxidase (GOx)–SWCNT conjugates via noncovalent adsorption, contrary to the previous understanding that such processing would compromise protein function. Surprisingly, the resulting hybrids retained glucose sensitivity and selectivity, producing instantaneous NIR turn-on responses of up to 40% ΔF/F₀ within 1 s of exposure to glucose (Figure 3a).
Furthermore, mechanistic studies using an engineered, non-catalytic GOx revealed a unique signal transduction pathway in fluorescence modulation that is driven by glucose binding alone without enzymatic turnover (Figure 3b). This binding-mediated, non-redox transduction pathway enabled fully reversible, tissue-transparent glucose sensing that neither consumes analyte nor generates toxic byproducts (e.g., H₂O₂)—overcoming a central limitation of conventional enzyme-based glucose sensors for continuous glucose monitoring. Furthermore, these nanosensors exhibited dynamic range of 0.1 mM–3 mM and the limit of detection of 42 μM, both of which are sufficient for a direct in-blood and in-tissue quantification of glucose. Leveraging the non-photobleaching NIR emission of SWCNTs and the high spatiotemporal resolution achievable with these probes, we demonstrated glucose imaging in acute mouse brain slices (Figure 3c). Together this study establish a generalizable platform for generating enzyme-SWCNT nanosensors for a variety of metabolite and neurochemical analytes.
(a) GOx–SWCNT operates as a glucose nanosensor. (b) A non-catalytic nanosensor using engineered GOx remains responsive to glucose, indicating a binding-mediated transduction that does not require glucose oxidation. (c) Translational potential: representative NIR images of a nanosensor-labeled mouse brain slice before and after glucose addition.
Ledesma, Nishitani et al., Adv. Func. Mater. (2024)

3. Visualizing Chemical Dialogues

What if we could visualize the entire process of neurochemical signaling—from the precise moment of a molecule's release into the synaptic space to its dynamic diffusion—not just for one, but for multiple molecular species simultaneously? We integrate our functional probes with advanced optical systems to build translational molecular imaging platforms.

Nishitani et al., Adv. Func. Mater. (2024)

Research Poster

Research Overview

Discover our core research themes, methodologies, and future directions in a single comprehensive poster. Click the button to view or download the high-resolution PDF.

📄 View Poster (PDF)