Research

Single-Molecule Spectroscopy and Advanced Imaging

1. High-Resolution Single-Molecule Spectroscopy and Imaging

In order to follow structural changes of a functioning biological macromolecule in real time, one needs experimental methods that allow one to make statistical meaningful measurements on the order of milliseconds. On this time scale, however, only a limited number of photons are available because the fluorescence comes from only one molecule. The stochastic photon counting process makes it even more challenging to measure the molecular state quantitatively. Commonly, further averaging and filtering over the detected photons are necessary. The subjective choice of filtering parameters and extent of averaging directly impacts on experimental time resolution and interpretation. In order for fluorescence single-molecule spectroscopy to reach its full potential, quantitative and reliable methodologies are needed.

Focusing on studies of immobilized single molecules, our first contribution is a detailed study of the basic relationship between time resolution and measurement uncertainty using principles from information theory [BiophysJ 2004]. This leads to the successful derivation of a theoretical bound on the number of photons needed to make a meaningful measurement, which in turn, allows us to develop an efficient and practical algorithm to extract the most amount of information from the photon-limited single-molecule data. The first key concept introduced in this contribution is that measurement of physical parameter is made on a photon by photon basis. This approach affords a more objective way of extracting useful information from the noisy single-molecule data compared to the commonly used methods that require subjective choice of smoothing parameters. Moreover, the photon-by-photon approach offers the highest time resolution possible. The second key concept is that time-dependent molecular characteristics are measured without any presumed kinetic model. They are model free methods. These concepts are important because they allow the results to be interpreted in the most informative yet least prejudiced way such that new discoveries can be made with experimental evidence that bears statistical significance. These ideas are extended to treating cases where molecular state change occurs on a time scale faster than experimental resolution. Our new method allows one to accurately locate precipitous emission intensity changes with an uncertainty of only a few photons [JPCB 2005]. Using this and other well-developed statistical methods, we show that the emission states of individual quantum dots are continuously distributed [NanoLett_2006]. Determination of kinetic parameter from these changepoints is relatively straightforward. Applying to biological macromolecules, for example, our method will allow one to pin-point the instance at which a ligand binds to an enzyme, and allow us to investigate if and how the enzyme relaxes following ligand binding. At this point, we have developed the essential tools that permit us to follow fast molecular dynamics precisely; but a robust way of quantitatively measuring conformational distributions is still needed. The difficulty lies in the fact that, in contrast to ensemble-averaged experiments, each single-molecule measurement may carry an uncertainty that far exceeds the extent of the true distribution. This is addressed in our third contribution, where we develop an information-based approach to quantitatively recover the true conformational distribution in a model-free manner [JPCA_2006].

Individual molecules can also be studied while they diffuse through the confocal detection volume. One of the difficulties involved in this type of experiments, however, is the determination of signal amid photon-counting noise. A commonly used approach is to further average the noisy time trace by binning, followed by placing a threshold to discriminate signal from background. The choice of smoothing parameters and the placement of the threshold may impact on the efficiency with which the information-rich region can be harvested, among other potential complications. We introduce a procedure that operates on the data sequence photon by photon, thereby relieving the incertitude in choosing binning-thresholding parameters. We characterize this procedure by detecting the two-photon emission bursts from diffusing single gold nanoparticles. The results support our burst-finding procedure as a reliable and efficient way of detecting and harvesting photon bursts from diffusing experiments [JPCB 2005].

2. Spectroscopy of Single Particles Freely Moving in 3D

Ultimately, one would like to follow the reactivity of a molecular-scale probe in real time and to directly correlate the chemical or biochemical processes with the 3D spatial location of the probe. In order to do this kind of experiments, a new single-molecule spectrometer with 3D single-particle tracking (SPT) capability has been constructed [APL_2006]. Advanced statistical methods with analytical expressions that allow quantitative extraction of diffusion coefficients and accurate localization of dynamic changes in the diffusive behavior from single-particle tracking experiments are also being developed [JPC_2006]. It is shown that 3D-SPT is capable of measuring the hydrodynamic size of non-fluorescenct nanoparticles in water quantitatively and rapidly [JPC_2007]; the size distribution thus obtained is directly comparable to that from TEM. This apparatus also allows one to do single-particle dynamic light scattering. The related theoretical framework is being developed concurrently [JPC_2007]. A new imaging concept that stems from these advances is to guide 3D confocal microscopy by individual fluorescent nanoprobes [OptLett_2007]. The new spectroscopic and imaging capabilities and concepts that we have developed will allow one to attain an unprecedented level of understanding of the microscopic origin of a complex system (see [CPL_2008] for a recent review).

3. Chemical Imaging

Collaborating with Professor Ron Shen in the Physics Department, we are developing a new imaging method based on molecular chirality using sum-frequency generation (optically active sum-frequency generation imaging, or OA-SFGi). In addition to the advantageous flexibility provided by the sum-frequency generation scheme, it has the following distinct characteristics. Since OA-SFGi uses molecular chirality as the contrasting mechanism, it is oblivious to the ubiquitous water that is often the major source of background. In analogous to circular dichroism, it is also sensitive to conformation of molecules or their aggregates. If the pump frequencies are tuned to be in resonance with molecular vibrations, it is possible to achieve chemical resolution, attaining spatial resolution by different chemical species [JACS 2006]. In addition to further developing the technique, we are also applying this new method to imaging individual cells [Ultrafast_2007].

Structure-Function Dynamics of Macromolecules

High-resolution single-molecule spectroscopic techniques have been developed to quantitatively address the functional consequences of structural fluctuations in macromolecules. These new techniques are being applied to examine:

1. The functional consequences of thermal fluctuations: Adenylate Kinase from Escherichia coli

A quantitative description for important features that characterize structure-function dynamics of a working enzyme has been provided. These features include: (a) molecular motions that facilitate or hamper catalytic events in a biomolecule; (b) a free-energy landscape relating the uptake of different substrates and biochemical reactivity of enzymes; and (c) time scale of structural relaxation following ligand binding and release [PNAS_2007]. In addition to providing further understanding how an enzyme works in general, we hope that careful characterizations of these features will help to create predictive theoretical models.

2. The dynamics of energy coupling in AAA+ motor proteins: NtrC (from Escherichia coli) and NtrC1 (from the thermophile Aquifex aeolicus)

In this project, we study two other fundamental aspects of structure-function dynamics, namely, allosteric regulation and energy coupling. In the former, chemical modification to a certain site of a protein is translated to modified functionality at the distant reactive site. In the latter, binding or hydrolysis of high energy-content molecules such as ATP or ADP is converted to mechanical work. Both have been generally thought as through conformational rearrangements. The molecular mechanisms and their time scale, however, remain largely unknown. In collaboration with Professors Tracy Nixon, David Wemmer,and Sydney Kustu, we are investigating these two important processes using as model systems the NtrC protein from E. coli and the NtrC1 protein from the thermophilic A. aeolicus; both proteins interact with the bacterial σ54 factor and regulates the activity of bacterial RNA polymerase (RNAP).

Complex Systems

Two new kinds of imaging and spectroscopic techniques are being developed. Spectroscopy of single particles freely moving in 3D affords a direct assessment of the correlation between the local dynamics and the 3D point locations that is the critical link between microscopic processes and macroscopic phenomena. Chemical imaging uses time-resolved non-linear spectroscopy for the study of location-dependent chemical contents and the dynamics they exhibit. Together, these two new approaches offer new opportunities in our fundamental understanding of the dynamics in complex systems. Specific topics that are related to this area of research include:

1. Multi-component molecular machineries: mechanism of cellulose degradation by cellulosomes

We are collaborating with Professor Jamie Cate to use the new capabilities we developed to study how complex molecular assemblies function, in particular, how cellulose particles are degraded by cellulosome.

2. Optical probes for imaging complex environment

We study the fundamental photophysics of individual CdSe quantum dots and their derivatives [NanoLett_2006], aiming to find new ways to utilize their unique optical characteristics as probe for local properties [JACS 2004]. For example, using high-resolution single-molecule spectroscopy, we have shown that individual quantum dots can be used to sense the local temperature [NanoLett_2007] and chemcial environment [JACS_2007], [JPCB_2008]. These efforts include collaborations with Professor Paul Alivisatos, Professor Liwei Lin, and Dr. Carl Hayden.

3. Micro-electro-mechanical systems (MEMS) control

We are working with Professor Liwei Lin to develop ways to control the microscopic environment [APL_2007], [NanoLett_2007].