Periodic Reporting for period 1 - CryoSTEM (Phase Contrast STEM for Cryo-EM)
Periodo di rendicontazione: 2022-06-01 al 2024-11-30
How do we see what goes on inside the living cell? An entire factory of biochemical components works together to energize, build, repair, and ultimately reproduce a minimal living unit. Many of these components can be isolated or reconstructed in the lab, outside the cell, but their coordination in space and time requires a study in place. A wide variety of microscopes have been invented for this purpose, beginning with the light microscope that we encounter in school years. In terms of pure resolving power, the champion is the electron microscope. The high vacuum of the microscope is not a natural environment for cells and tissues, however. The first challenge is to preserve the specimen in a state as close as possible to the native condition. This is the realm of cryo-microscopy. The biological sample, still embedded in water, is frozen instantly into a glass without forming crystals of ice. The second challenge is that the concentrated electron beam used for imaging is an ionizing radiation that can cause severe damage to the delicate organic specimen. Optimizing the information gained before the damage accumulates, the useful contrast per exposure, is the name of the game. The CryoSTEM project explores and develops a new approach to cryo-electron microscopy: phase contrast by scanning transmission EM. Traditional STEM creates an image based on scattering of electrons, as charged particles. Phase contrast is based on interference of the electrons, as waves. Using sophisticated electron detectors, on one hand, and advancing image processing, on the other, we learn to exploit the wave properties of the electrons, as in a hologram from light. Moreover, since the components of a cell are much smaller than the volume they fill, we image in three dimensions by tomography. The approach is similar to medical CT at the hospital, but at the nanometer scale of biological molecules. Phase contrast methods by CryoSTEM offer a new horizon in the range of studies accessible by electron microscopy.
The first part of the work program has been to establish the hardware and control software needed to record phase contrast STEM tomograms. This has been achieved in two configurations. One uses segmented electron detectors to produce differential phase contrast, DPC, by measuring the deflection of a narrow beam that passes through the specimen while it scans. The other uses a very fast camera to record the entire pattern of electron scattering at every pixel. Such cameras have become available only very recently. The approach is known as 4D STEM, for a 2D scattering (i.e. diffraction) pattern recorded at every 2D coordinate (i.e. pixel) in the x,y plane. To both of these, we have added a third axis in tilting the specimen for tomography. Commercially-available solutions could achieve only a fraction of the required tasks.
The second direction of the work so far has been to explore the most efficient ways to extract useful contrast from the recorded signals. The optimal methods for processing two dimensional data from ultra-thin specimens need to be modified when reconstructing a volume in 3D. Many combinations can be made, and as yet there is no single accepted protocol. In fact there may never be just one, because different ways of processing the same original dataset provide complementary information. There is an enormous flexibility and a single image may not convey all the available information. The full 4D STEM data are very heavy, however, so we also explore the optimization of configurations that can be implemented with simpler detector schemes such as DPC.
The platform we have established is very general and can be applied to a wide variety of specimens. Method development efforts employed cryogenically-preserved bacteriophage, and the lessons are applied immediately to study of intact cells. Indeed, a primary advantage of the scanning approach with respect to conventional electron microscopy is for application to much thicker specimens, which relates to energy losses from the illuminating electrons in their interaction with the material. Using a new generation electron spectrometer, we explored the quantitative nature of electron scattering over the angular range relevant to phase contrast, and identified measurable parameters by which to distinguish different materials. This is an extension of the classic Z (elemental atomic number) contrast in STEM, which can be implemented with the new 4D STEM tools and incorporated into tomography.
The second direction of the work so far has been to explore the most efficient ways to extract useful contrast from the recorded signals. The optimal methods for processing two dimensional data from ultra-thin specimens need to be modified when reconstructing a volume in 3D. Many combinations can be made, and as yet there is no single accepted protocol. In fact there may never be just one, because different ways of processing the same original dataset provide complementary information. There is an enormous flexibility and a single image may not convey all the available information. The full 4D STEM data are very heavy, however, so we also explore the optimization of configurations that can be implemented with simpler detector schemes such as DPC.
The platform we have established is very general and can be applied to a wide variety of specimens. Method development efforts employed cryogenically-preserved bacteriophage, and the lessons are applied immediately to study of intact cells. Indeed, a primary advantage of the scanning approach with respect to conventional electron microscopy is for application to much thicker specimens, which relates to energy losses from the illuminating electrons in their interaction with the material. Using a new generation electron spectrometer, we explored the quantitative nature of electron scattering over the angular range relevant to phase contrast, and identified measurable parameters by which to distinguish different materials. This is an extension of the classic Z (elemental atomic number) contrast in STEM, which can be implemented with the new 4D STEM tools and incorporated into tomography.
The current state of the art in electron cryo-tomography uses conventional wide-field transmission EM imaging, with phase contrast introduced by imaging slightly out of focus. This approach is superbly suited for study of isolated macromolecules at near-atomic resolution, but the mathematical models of contrast transfer depend on assumptions that are poorly justified for thick samples such as cells. In addition, energy losses in electron scattering produce a haze in the images whose removal by energy filtering imposes another limitation on specimen thickness. These limits lead to the practice of cutting ultra-thin cryo-sections or lamella prior to imaging. While resolution is very impressive, thinning the sample obviously removes much of the 3D context. Moreover, the high resolution obtained in the conventional state of the art depends on averaging the images obtained from thousands (or even tens or hundreds of thousands) of identical molecular particles. The more flexible STEM contrast methods have the potential to circumvent, or at least attenuate these limitations. Our efforts, therefore, represent an extension of the contribution of cryo-EM from structural biology toward cell biology. Coinciding with the recent success of protein structure prediction, putting the molecules into their cellular context takes on a new significance. We can also anticipate another area of impact in the material sciences, where the focus on efficient contrast generation with minimal electron exposure extends the realm of analysis to radiation-sensitive specimens such as polymers and hybrid organic-inorganic materials.