Ultrafast Biophysical Studies of Proteins: Time-resolved WAXS Studies

时间分辨劳厄结晶学，顾名思义，只能在晶体样品上进行。维持晶体有序的分子间作用力限制了大范围的构象运动，这种灵活性的丧失可能会扰乱甚至抑制蛋白质的功能。尽管如此，劳厄结晶学在获取超快时间尺度上的近原子结构信息方面是独一无二的。另一方面，X射线还可以从允许全构象运动的溶液中的分子中提取结构信息。由于蛋白质溶液中没有长程有序，蛋白质的小角和广角X射线散射(SAXS/WAXS)是弥散的。然而，SAXS/WAXS图案中的衍射环报告了蛋白质的大小和形状。尽管这种结构信息不是原子分辨率的，但它确实提供了一个指纹，可以通过模型与蛋白质结构相关联。因此，SAXS/WAXS指纹随时间的变化可以用来评估哪些模型最好地描述了溶液中的反应途径。 当我们第一次用这项技术研究血红蛋白的四级结构转变时，我们在我们能测量到的最早的时间观察到了WAXS图谱的实质性变化，这表明WAXS图谱对三级结构的变化和四级结构的变化都很敏感。为了证明这一点，我们记录了一氧化碳氧合肌球蛋白(MbCO)光解后的时间分辨WAXS模式，并观察到了泵浦引起的相当大的变化。根据MbCO和脱氧Mb的X射线结构，它们的原子位置的均方根差小于0.5。这种意想不到的能力，能够感觉到如此微小但系统的结构变化是相当令人鼓舞的，并激励我们继续努力开发这种新的实验方法。 我们在ESRF推进这些研究的努力因缺乏波束时间而受阻。因此，在过去的几年里，我们投入了大量的精力来开发在APS进行时间分辨X射线散射研究所需的基础设施。值得注意的是，我们已经成功地将我们的时间分辨能力扩展到SAXS地区，同时进一步扩大了可访问的WAXS范围。为了进一步提高我们的数据质量，我们开发了一种能够5G加速的样品平移台，它可以以高达41赫兹的重复频率在X射线脉冲之间将样品移动到新的位置。使用这一基础设施获得的数据的质量比早先在ESRF获得的数据质量有了很大提高。这一改进是非常有益的，因为X射线散射信号由蛋白质以外的源主导，并且归因于时间分辨结构变化的信号比静态散射信号弱3到4个数量级。要提取这样的小信号，需要非常稳定和可重复的实验方法，以及复杂的计算方法来分析弱散射模式。从这一分析中得到的反馈导致了样品保持器和数据采集协议的重大重新设计。例如，我们的数据分析表明，低于1度的温度波动产生的WAXS信号比由于蛋白质结构变化而产生的信号更大，从而影响了我们分离WAXS散射信号的时间分辨贡献的能力。我们已经通过开发一种新的恒温样品电池支架来缓解这个问题，该支架可以在很大范围内提供非常高稳定性的温度控制。 一旦热问题得到解决，我们发现我们仍然受到X射线吸收引起的辐射分解效应的影响。辐射分解产生的自由基可以结合在一起，产生溶解的气体，最终达到过饱和。溶解的气体可以随机成核，产生微小的气泡，并在散射模式中产生伴随的变化，淹没了感兴趣的时间分辨信号。我们通过优化样品通过毛细管的采样速度缓解了这一问题，并更新了我们的数据采集协议，以便于回收和重复使用我们宝贵的样品，从而使我们能够获得仅含约150 ul 50 mg/ml蛋白质溶液的时间分辨数据集。一旦解决了这些问题，我们发现由于X射线探测器的热平衡(读出过程产生的热量会使CCD芯片升温并改变其本底水平)，散射图案中的随时间变化的变化，并开发了解决此问题的方法，方法是以用于获取时间分辨数据的相同速率定期读取探测器，从而将CCD芯片保持在恒定温度。有了这些最新的创新，我们现在可以提取蛋白质对总散射信号的贡献，并具有接近散粒噪声限制的检测灵敏度。 除了对溶液中蛋白质结构变化的时间分辨研究外，我们还开发了在从0到100摄氏度以上的广泛温度范围内记录蛋白质散射的能力。我们与W.A.Eaton博士合作，研究了一种广泛用于研究蛋白质折叠问题的模型系统--Villin(4kD)的随温度变化的回转半径。SAXS散射强度与体积的平方成正比。由于这种蛋白质相当小，它的散射能力比肌红蛋白弱20多倍，这使得这个项目相当具有挑战性。然而，我们已经开发了一种实验方案，能够提取蛋白质信号并确定其旋转半径Rg作为温度的函数。此外，正在进行的对稀释系列散射模式的分析有助于揭示蛋白质-蛋白质相互作用的本质。

Time-resolved Laue crystallography, as implied by its name, 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. Nonetheless, Laue crystallography stands alone in its ability to acquire near-atomic structural information on ultrafast time scales. On the other hand, X-rays can also extract structural information from molecules in solution where the full range of conformational motion is permitted. Because there is no long-range order in protein solutions, Small- and Wide-Angle-X-ray-Scattering (SAXS/WAXS) from the protein is diffuse. However, diffraction rings in the SAXS/WAXS pattern report on the size and shape of the protein. Though this structural information is not at atomic resolution, it does provide a fingerprint that can be correlated via models with the protein structure. Time-dependent changes of the SAXS/WAXS fingerprint can therefore be used to assess which models best describe the reaction pathway in solution. When we first set out to study the quaternary structure transition of hemoglobin with this technique, we observed a substantial change of the WAXS pattern at the earliest time we could measure, which suggested that the WAXS pattern is sensitive to tertiary structure changes as well as quaternary structure changes. To prove this point, we recorded time-resolved WAXS patterns following photolysis of carbon monoxymyoglobin (MbCO) and observed sizable pump-induced changes. According to X-ray structures of MbCO and deoxy Mb, the rms difference in their atomic positions is less than 0.5 . This unexpected ability to sense such small but systematic structure changes is quite encouraging, and has spurred us to continue our efforts to develop this new experimental methodology. Our efforts to further these studies at the ESRF were thwarted by a lack of beamtime. Thus, we have invested much effort over the past few years to develop the infrastructure required to pursue time-resolved X-ray scattering studies at the APS. Significantly, we have succeeded in extending our time-resolved capabilities into the SAXS region while expanding even further the accessible WAXS range. To further improve the quality of our data, we have developed a sample translation stage capable of 5G acceleration which can move the sample to a new position between X-ray pulses at a repetition frequency as high as 41 Hz. The quality of the data acquired with this infrastructure is significantly improved beyond that attained earlier at the ESRF. This improvement is quite beneficial, as the X-ray scattering signal is dominated by sources other than the protein, and the signal ascribed to time-resolved structural changes is 3 to 4 orders of magnitude weaker than the static scattering signature. To extract such small signals requires a very stable and repeatable experimental methodology, as well as sophisticated computational methods to analyze the weak scattering patterns. Feedback from this analysis has led to a major redesign of the sample holder and the data acquisition protocol. For example, our data analysis showed that temperature fluctuations of less than one degree produce a WAXS signal that is large compared to the signal due to protein structural change, and thereby compromised our ability to isolate the time-resolved contribution to the WAXS scattering signal. We have mitigated this problem by developing a new thermostated sample cell holder that provides temperature control over a broad range with very high stability. Once the thermal issues were resolved, we found that we still suffered from the radiolysis effects that arise from X-ray absorption. Radiolysis produces radicals that can combine to produce dissolved gases that eventually become supersaturated. The dissolved gas can nucleate stochastically to produce tiny gas bubbles and produce a concomitant change in the scattering pattern that overwhelms the time-resolved signal of interest. We have mitigated this problem by optimizing the rate at which sample is drawn through the sample capillary, and have updated our data acquisition protocol in a fashion that facilitates recovery and reuse of our precious samples, thereby allowing us to acquire a time-resolved data set with only about 150 uL of 50 mg/ml protein solution. Once these problems were addressed, we discovered time-dependent changes in the scattering pattern due to thermal equilibration of the X-ray detector (the readout process generates heat that warms the CCD chip and alters its background level), and developed a workaround to this problem by periodically reading the detector at the same rate as that used to acquire our time-resolved data, thereby maintaining the CCD chip at a constant temperature. With these recent innovations, we can now extract the protein contribution to the total scattering signal with near shot-noise-limited detection sensitivity. In addition to time-resolved studies of protein structure changes in solution, we have also developed the ability to record protein scattering over a broad range of temperatures, spanning from 0 to over 100C. In collaboration with the group of Dr. W. A. Eaton, we have studied the temperature-dependent radius of gyration for villin (4kD), a model system extensively used to investigate the protein folding problem. The SAXS scattering intensity scales as the square of the volume. Since this protein is quite small, its scattering power is more than 20 times weaker than that from myoglobin, making this project quite challenging. Nevertheless, we have developed an experimental protocol capable of extracting the protein signal and determining its radius of gyration, Rg as a function of temperature. Moreover, ongoing analysis of scattering patterns from a dilution series is helping to unveil the nature of protein-protein interactions.