Single-walled carbon nanotubes as near infrared fluorescent sensors : characterization, biological and analytical applications

Abstract

Reactive oxygen species (ROS) have emerged as biological signaling molecules, participating in newly discovered cascades that govern cell proliferation, migration, and pathogenesis. A major challenge in understanding these pathways is the lack of detection technologies that allow for spatial and temporal resolution of specific ROS at the cellular level. The goal of this thesis is to design a nanotube sensor platform able to detect and study H2 0 2 signaling fluxes at the cellular level in order to elucidate their role in biological processes. Understanding this role may lead to new therapeutic targets, and improve understanding of biological signaling. Single-walled carbon nanotubes (SWNT) are rolled sheets of graphene and can be either semiconducting or metallic depending on the angle of rolling and the diameter of the tube. Semi-conducting SWNT are one of only a few types of molecules that exhibit band gap photoluminescence (PL) in the near infrared (nIR), making them ideal for detection in biologically relevant media since it avoids biological auto-fluorescence. SWNT are also completely photostable even at high fluence, unlike conventional fluorophores and quantum dot systems, allowing them to serve as nIR single molecule optical sensors capable of long term and stable operations in vitro and in vivo. In this thesis, we show that the 1D quantum confinement of photogenerated excitons in SWNT can amplify the detection of molecular adsorption to where single-molecule discrimination is realizable, even from within living cells and tissues.

(cont.) We have developed a type I collagen film, similar to those used as 3D cell scaffolds for tissue engineering, containing embedded SWNT capable of reporting single-molecule adsorption of quenching molecules such as H₂0₂ . A Hidden Markov Modeling algorithm is utilized to link single-molecule adsorption events detected on the nanotube to forward and reverse kinetic rate constants for many different analytes. The collagen matrix is shown to impart selectivity to H₂0₂ over other ROS and common interferents. We utilized these new single-molecule sensors to study the fluxes of H₂0₂ from A431 human skin carcinoma cells and particularly the local generation rate from Epidermal Growth Factor Receptor (EGFR), a membrane protein and tyrosine kinase that controls cell proliferation among other functions. We show that an array of nIR fluorescent SWNT is capable of recording the discrete, stochastic quenching events that occur as H₂0₂ molecules are emitted from individual A431 and murine 3T3 fibroblasts cells in response to epidermal growth factor (EGF). We also show mathematically that such single molecule detection arrays have the unique property of distinguishing between "near field" and "far field" molecular generation, allowing one to isolate the flux originating from only the membrane protein. Corresponding inhibition experiments suggest a mechanism whereby water oxidizes singlet oxygen at a catalytic site on the receptor itself, generating H₂0₂ in response to receptor binding. An EGFR-mediated H₂0₂ generation pathway that is consistent with all current and previous literature findings has been proposed for the first time and numerically tested for consistency.

(cont.) In an effort to extend this detection to in vivo systems, we investigated how SWNT are uptaken and localized within living cells and as well as their potential cytotoxicity. To this end, we have developed a novel method of studying this problem by tracking the non-photobleaching SWNT in real time by using a single particle tracking method. Over 10,000 individual trajectories of SWNT were tracked as they are incorporated into and expelled from NIH-3T3 cells in real time on a perfusion microscope stage. An analysis of mean square displacement allows the complete construction of the mechanistic steps involved from single duration experiments. We observe the first conclusive evidence of SWNT exocytosis and show that the rate closely matches the endocytosis rate with negligible temporal offset, thus explains why SWNT are non-cytotoxic for various cell types at a concentration up to 5 mg/L, as observed from our live-dead assay experimental results. Further, we studied the cellular uptake and expulsion rates of length-fractionated SWNT from 130 to 660 nm in NIH-3T3 cells using this method. We developed a quantitative model to correlate endocytosis rate with nanoparticle geometry that accurately describes our data set and also literature results for Au nanoparticles. The model asserts that nanoparticles cluster on the cell membrane to form a size sufficient to generate a large enough enthalpic contribution via receptor ligand interations to overcome the elastic energy and entropic barriers associated with vesicle formation.

(cont.) The total uptake of both SWNT and Au nanoparticles is maximal at a common radius of 25 nm when scaled using an effective capture dimension for membrane diffusion. The ability to understand and predict the cellular uptake of nanoparticles quantitatively should find utility in designing nanosystems with controlled toxicity, efficacy and functionality. The development of such single molecule detection technologies for ROS motivates their application to many other unexplored signaling pathways both in vitro and in vivo.