Ultrafast Biophysical Studies of Biomolecules at the APS

Anfinrud, Philip

Abstract

Enzymatic reactions exhibit remarkable selectivity and efficiency, the likes of which are rarely achieved in bench-top chemical reactions. While it is clear that the biochemical prowess of an enzyme arises from its highly-ordered structure, the detailed mechanism by which it functions has proven elusive. This is because enzymes are not simply static macromolecules that host an active site, as depicted by their crystal structure; rather, they are dynamic molecules whose choreographed motions can gate the transport of substrate to and from the active site and can modulate over time the activity of that site. To develop a mechanistic understanding into how enzymes function, it is essential to study this choreography of life with structurally-sensitive methods capable of ultrafast time resolution. In this report, we focus on time-resolved X-ray diffraction studies that employ the pump-probe method. Briefly, a laser pulse (pump) photoactivates or thermally excites a biomolecule, after which a suitably delayed X-ray pulse (probe) passes through the sample and records a diffraction pattern on a 2D detector. Thanks to capital equipment investments made by the NIH, the BioCARS beamline is capable of delivering approximately ten billion X-ray photons into a small focused spot (25 microns) in a very short duration pulse (100 picoseconds). The methods developed to take advantage of this high-flux X-ray beamline include time-resolved Laue crystallography and time-resolved SAXS/WAXS. Time-resolved Laue crystallography takes advantage of a polychromatic X-ray pulse, which can produce thousands of reflections in a single shot, and boosts substantially the rate at which time-resolved diffraction data can be acquired. The information needed to determine the protein's structure is encoded in the relative intensities of the diffraction spots observed. Since the structural information contained in a single Laue diffraction image is incomplete, repeated measurements at multiple crystal orientations are required to produce a complete set of data. Prior studies have generally required a small number of very large, homogeneous crystals, which has limited the application of this methodology to a handful of proteins. The impact of time-resolved Laue crystallography would be boosted significantly if we were to develop methods capable of acquiring high S/N diffraction images from a large number of relatively small crystals (30-35 microns), rather than small number of large crystals. To that end, we continue our efforts to develop novel microfluidic methods for growing crystals and delivering them in a fashion that can be automated. Briefly, we have developed an alternating drop microfluidic mixer that has been used to grow more than 1000 uniformly sized lysozyme crystals ( 30-35 microns) in a 1-m long glass capillary. A microfluidic crystal delivery system based on a home-built multi-axis syringe-pump tower is being developed with an aim to automate crystal delivery to the BioCARS 14IDB beamline. This effort also includes the development of a high-speed diffractometer capable of rapidly and precisely positioning crystals at the intersection of the laser and X-ray beams. While much progress has been made, more work remains to be done.
Our aim to automate the acquisition of X-ray diffraction images from thousands of crystals without user intervention will hopefully be realized within the coming year. Time-resolved Laue crystallography, as its name implies, can only be performed on crystalline samples. The intermolecular forces that maintain crystalline order constrain large amplitude conformational motion, and this loss of flexibility may perturb or even inhibit the function of a protein. Though Laue crystallography stands alone in its ability to track structural changes in proteins on ultrafast time scales with near-atomic spatial resolution, it is crucial to also study structural dynamics of biomolecules in solution where the full range of conformational motion is permitted. Without external alignment forces, biomolecules in solution are randomly oriented, and the structural information contained in their orientationally-averaged diffuse scattering pattern is one dimensional. Nevertheless, it is well known that the SAXS region of the diffraction pattern reports on the size and shape of the biomolecule, while the WAXS region is sensitive to secondary and tertiary structure. Time-resolved SAXS/WAXS scattering patterns therefore provide 'fingerprints' that can be correlated with protein structure via molecular models, and can assess which models best describe reaction pathways in solution. Progress in this area requires close connections between experiment and theory. Our time-resolved SAXS/WAXS diffractometer currently employs a secondary K-B mirror pair to focus the X-ray beam onto the sample capillary with independent control of the the vertical and horizontal dimensions, a very small beamstop ( 0.5 mm), and a large area (340x340 mm), high-speed (up to 10 Hz) X-ray detector. With the sample-detector distance set at 186 mm, scattering data can be acquired over a broad range of q (momentum transfer) spanning 0.02 to 5.2 inverse Angstroms, which corresponds to spatial resolution down to 1.2 Angstroms. When the X-ray source operates in hybrid mode, nearly ten billion, 12-keV photons are delivered to the sample in each 125-ps X-ray pulse. With this infrastructure, integrating the signal from 200 X-ray pulses is sufficient to produce high dynamic range scattering images. To mitigate the adverse effects of radiation damage during X-ray exposure, the capillary containing the protein solution is rapidly translated over a 20-mm span using a home-built, high-speed diffractometer that is based on 1-micron resolution linear motor translation stages capable of more than 1-g acceleration. Thanks to a closed-loop circulation system, about 150 microliter of protein solution is sufficient to acquire a high signal-to-noise ratio data set. A stepper-motor-driven peristaltic pump pushes a fresh volume of solution into the capillary during each return stroke of the linear translation stage. The flow generated by peristaltic pumps vary as a function of rotation angle, and since the pump head consists of four rollers, is interrupted four times per revolution. This variable and interrupted flow can lead to systematic errors in the scattering data acquired. To mitigate this problem, we developed a novel means that linearizes the flow and minimizes the pulsation effects that plague peristaltic pumps. We also recently developed a new capillary holder and high-precision temperature controller that allow us to characterize structural changes over a broad range of temperatures spanning from approximately -20 to 120 degrees Celsius. Thanks to the relatively small spot size that can be generated with the secondary K-B mirror pair, it is possible to focus a 1 mJ infrared pulse down to a dimension small enough to heat samples in a glass capillary by more than 10 degrees Celsius. Using a high-precision temperature controller to set the temperature just below a protein's unfolding temperature, this magnitude T-jump is sufficient to trigger unfolding of a protein, and allows us to investigate the dynamics of protein unfolding with unprecedented spatial resolution. The time-resolution achieved is currently limited by the duration of the infrared laser pulse, which is about 5 ns. As our time-resolved SAXS/WAXS methodology becomes more precise and easier to use, we expect it to become an ever more important complement to time-resolved Laue studies and time-resolved optical spectroscopy studies of biomolecules, and will help provide a structural basis for understanding how biomolecules function.