Resolution Revolution
A note from the guest editor:
The first story of this issue of the Science Journal focuses on a sample of the truly outstanding research that is being enabled by the new Titan/Krios electron microscope that brings the powerful tool of cryo–electron microscopy to our faculty and students.
The story, however, really began several years before the Titan/Krios arrived on campus, when the faculty and the University-wide research administration decided to make a concerted effort to significantly upgrade our rather limited capabilities in super-resolution microscopy. Teams of faculty came together from across the STEM colleges to seek National Science Foundation (NSF) support for some of the most expensive pieces of equipment, such as the NIKON N-SIM/N-STORM microscope system, which permits super-resolution microscopy of living materials. Penn State is permitted to submit only two proposals per year, and because the 80-plus-page Major Research Instrumentation grant proposals are due to the NSF during the second week of January, these teams of faculty worked primarily throughout the Thanksgiving and Christmas breaks to put together competitive proposals. Although the NSF funds less than 10% of the proposals submitted each year, our faculty’s proposals are awarded at significantly higher than the national average. And now our faculty, graduate students and undergraduates are using core microscopy facilities that are world class!
Interestingly, it already appears that the research enabled by the Titan/Krios extends beyond the current capabilities of the instrument. CryoEM currently allows investigators to visualize larger (120kd and up) proteins and protein- protein interactions with ease and precision that were unfathomable even a few years ago. As described in the following story, this has led to an ongoing revolution in our understanding of basic molecular biology. Nevertheless, many of the most interesting proteins are smaller.
Scott Lindner and Susan Hafenstein — both in the Department of Biochemistry and Molecular Biology—have already developed a new sample preparation technology that will allow existing cryo–electron microscopes to visualize proteins and other biological materials that are an order of magnitude smaller and to do so at a resolution that is less than 10 Angstroms (View the Intellectual Property article). This technology holds the potential not only to further the revolution in our basic understanding of molecular processes but also to accelerate the discovery and development of new therapeutics for disease, which often involve the interaction of small molecules. —Andy Stephenson
Resolution Revolution
When you sit down to watch your new 4K television for the first time, suddenly that standard- definition re-run of Star Trek doesn’t look so futuristic. The leaps in scientific and technological knowledge responsible for the impressive advance in video quality since the starship Enterprise took its first voyage on network television in the 1960s are worthy of Starfleet. But recent leaps in image quality have not been limited to photography and film. A massive advance in how images are captured within electron microscopes has heralded what is being called the “resolution revolution,” allowing researchers to see tiny viruses, proteins, and other large molecules with astounding clarity—to the point where individual atoms have become visible.
This advance has revolutionized an imaging technique called cryogenic electron microscopy (cryoEM), which now allows researchers to quickly create high-resolution three-dimensional models of intricate biological, chemical, and synthetic structures.
“It’s like putting on glasses for the first time,” said Susan Hafenstein about the improvement. “With cryoEM, we have gone from a general idea of structure to crystal-clear detail.”
This improved look at biological structures is an important step in understanding their function and how they work, and the scientific community is abuzz with the exciting potential of this updated imaging technique. Indeed, the efforts of three scientists involved in the breakthrough developments to cryoEM were recognized with the 2017 Nobel Prize in Chemistry.
To capitalize on the major advances to cryoEM, Penn State unveiled the Cryo–Electron Microscopy Facility, a joint venture between the Huck Institutes of the Life Sciences and the Materials Research Institute, in 2017. Soon after the construction of the facility’s feature piece of equipment, a unique Titan/Krios microscope, Hafenstein, who has a joint professorship with the Penn State College of Medicine and the Eberly College of Science, moved her lab from the Penn State Milton S. Hershey Medical Center to the University Park campus to serve as director of the facilities.
As director, Hafenstein consults with researchers across the University about using the new microscope and interpreting the resulting data. This allows Penn State researchers to investigate questions as varied as how certain proteins are involved in the transmission of the malaria parasite; how sticky proteins called amyloids cluster into potentially damaging plaques in the brain; and how an enzyme called ATPase interacts with other enzymes to stimulate the copying of genetic material in bacteria. The unique microscope is also poised to attract outstanding new faculty eager to make use of the equipment.
“The advances to the instrument make it much more accessible to scientists,” said Hafenstein. “A few specialists help collect the data on the microscope, but the method is otherwise widely achievable by researchers. It doesn’t take years of specialized training to use.”
Hafenstein also takes advantage of the facilities to gain a new perspective in her own research, which explores how viruses enter host cells.
“We have already used cryoEM to look at the capsid, the outer protein shell, of both human papillomavirus (HPV) and coxsackievirus B3 (CVB3) to see how these viruses bind to receptors on human cells,” said Hafenstein. “This binding is a critical step in the infection process, and understanding capsid structure could eventually lead to therapies that prevent this binding.”
Hafenstein ultimately hopes to identify similarities in how different kinds of viruses enter a host cell, as these similarities may reveal good targets for future drugs.
“There are all kinds of steps that the virus has to go through to enter and infect a human cell, but we don’t know what those are for most viruses,” said Hafenstein. “In the past we were never really able to see this entry process. Now with cryoEM we can, and that’s pretty exciting.”
CryoEM is not actually the first imaging technique to breach atomic-level resolution, but the nuances of how other techniques work makes each suited to imaging certain kinds of materials. For example, a technique called nuclear magnetic resonance (NMR) imaging is particularly suited for resolving atomic structures of small molecules in solution. Another popular technique called X-ray crystallography is well suited for determining atomic models of macromolecules and complexes. CryoEM is poised to help ll in the gaps, especially for larger molecules or when it is challenging to prepare samples for other techniques.
How CryoEM Works
CryoEM has evolved considerably over the past 30 years, but the general process is fairly straightforward: Many images of a specially prepared sample are taken using the microscope, and then the images are sorted and analyzed to reconstruct a 3D image. Recent innovations have spanned the entire process—including advancements to sample preparation, the electron detector in the microscope, and to automation and data processing.
A researcher begins by preparing a sample, which has many copies of the particle to be imaged floating around in a liquid. A small portion of this solution is placed onto a tiny mesh-like grid, which is in turn placed into a machine that blots it mostly dry.
“This is a delicate process,” said Tracy Nixon, professor of biochemistry and molecular biology, who uses the imaging facilities in two separate research projects. “With blotting, you have to get the right thickness. If it’s too thick, you can’t see through the sample. If it’s too thin, the particles you are trying to image become distorted and won’t provide an accurate view of their structure. Once our samples are prepared, we hand them off to Carol Bator, a research technologist at the Penn State imaging facilities, to take over this careful work.”
The grid is then plunged into liquid nitrogen, which vitrifies the sample, or freezes it so rapidly that the molecules retain their natural shapes. This special method of freezing, which was introduced to electron microscopy in the early 1980s by Nobel laureate Jarcques Dubochet, put the “cryo” in cryoEM.
“When water freezes into ice, the water molecules form in well-organized crystals that cause the whole volume to expand,” said Nixon. “That creates enough pressure and force that molecules can be damaged, so freezing and thawing can be bad for proteins or a cell. With vitrification cation in cryoEM, we end up with a thin film of frozen liquid that is essentially glass, and the molecules maintain their shape.”
This special method of freezing also allows a sample to be frozen mid-movement, so vitrifying samples at various stages of a certain process can offer a glimpse into how that process plays out, like how an enzyme changes its shape while it catalyzes a chemical reaction or how proteins bind to and form complexes with other molecules.
Then the grid is placed into the microscope, where the machine takes over the delicate process of moving the sample onto a microscope stage. The microscope then shoots electrons one-by-one through the sample. This wave of electrons from the microscope interacts with the electron clouds of the atoms in the sample, recording their location and density.
“Upon arriving at the end of the microscope, the electron waves are recorded by a detector, revealing what they ‘learned’ as they passed through the sample in the form of ‘contrast,’” said Nixon. This contrast appears as a 2D shadow-like image of the sample. “Because the signal from a single electron wave is very weak, the detector records the information present in many transmitted electrons, finally making a visible image of the macromolecule. This process is repeated hundreds to thousands of times, producing many images of the frozen particles.”
This type of imaging lends itself to a form of 3D image reconstruction called single-particle reconstruction, which relies on the fact than many copies of a single type of particle have been locked into the frozen sample at a variety of different orientations.
“When you collapse that whole structure onto a picture, you can see that some of them are on end, some show one side, some show a different side, and so on,” said Nixon. “The software helps us pick out and sort these different views so they are all in the same orientation. By averaging the many photos of each view, you get better resolution, and when you collect tens of thousands of those and put them together, you can reconstruct a 3D image. Today, much of this process is done with automated algorithms.”
Single-particle reconstruction is helping Nixon and his collaborators see the details of the structure of a plant enzyme called cellulose synthase. Cellulose synthase helps create cellulose, the primary component of plant cell walls that is being targeted as a renewable biofuel.
“There is energy in sunlight that is captured by the plant cells that make cell walls,” said Nixon, “but it’s very hard to get at the reduced carbon, that energy in the cell walls, to use it as a renewable resource. Our goal is to understand the cell wall and how it forms, with the idea that, if we understand the machinery for how it is made, we may eventually be able to engineer plants down the road to make it more accessible. What we have right now are computational guesses as to what this machinery looks like.”
Nixon’s collaborators created a model of the most likely structure of a complex of proteins involved in the creation of cellulose based on the lower- resolution views they already had.
“Our current view doesn’t allow us to see the individual proteins in the complex; it’s just kind of a blob,” he said. “The new instrument will hopefully This allows Penn State researchers to investigate questions as varied as how certain proteins are involved in the transmission of the malaria parasite; how sticky proteins called amyloids cluster into potentially damaging plaques in the brain; and how an enzyme called ATPase interacts with other enzymes to stimulate the copying of genetic material in bacteria. The unique microscope is also poised to attract outstanding new faculty eager to make use of the equipment.
Nixon is also keen to take advantage of another mode of imaging available on the microscope, called cryo-tomography, which provides a bigger-picture view of a sample. With tomography, the grid is placed on the microscope in the same way, but then the microscope stage is tilted. By tilting the sample—up to 60 degrees in both directions—the researcher can obtain views from many different angles.
“It’s like a CAT scan,” said Hafenstein. “If you take lots of images of somebody’s brain from different viewpoints you have enough information to recreate it in 3D.”
Whereas single-particle approaches provide a high-resolution view of many copies of something relatively small, like sections of a protein or a full protein, tomography provides a view in context, like how a protein is connected, or bound, to other proteins in a complex or to a membrane.
The combination of techniques available on Penn State’s Titan/Krios microscope is valuable to Nixon. “Using both single particle and tomography approaches, we hope to get a better view of not just the individual proteins, but how they work together in complexes,” he said. “This will improve our understanding of the molecular machine that builds cell walls.”
Breaching the Atomic Barrier
The first image of a protein at atomic resolution using cryoEM was produced in 1990, after Nobel laureate Richard Henderson made considerable improvements to the speed and sensitivity of the electron detector. Combined with improvements made by Nobel laureate Joachim Frank to analyzing photographs and improvements to the use of vitrification, this advance set the stage for the latest revolution. In 2013, a further advance to the electron detector allowed regular imaging that reveals individual atoms, and since then researchers have flocked to make use of and improve cryoEM.
But other techniques, like NMR and X-ray crystallography, also provide atomic-level resolution, and Penn State has facilities on campus dedicated to both of these methods. Each technique, however, has its own strengths and challenges.
Although the popular method of X-ray crystallography has played a major role in the understanding of biological structures over the past several decades, the process of creating a structural model with this technique can take considerable time. For example, the 2009 Nobel Prize in Chemistry was awarded to three researchers who successfully mapped the structure of ribosomes, subcellular organelles involved in protein synthesis, which consist of hundreds of thousands of atoms, using X-ray crystallography.
“They determined the structure of the ribosome around the year 2000, but the project itself started back in the 1970s,” said Katsuhiko Murakami, professor of biochemistry and molecular biology at Penn State. Murakami has been using X-ray crystallography to better understand the structure and function of RNA polymerase, an enzyme that produces RNA from a DNA template during the transcription process, and how it interacts with other factors that regulate transcription in bacteria and archaea.
To use X-ray crystallography to visualize the 3D structures of RNA polymerase and the complexes it forms, Murakami must first purify and crystallize his samples. But this process can be difficult, because crystallization of macromolecules requires a large amount of a well-organized and uniform sample, where all molecules are in the same configuration.
“RNA polymerase is a large macromolecule that is difficult to crystallize,” said Murakami. “When a transcription factor binds to it, it becomes larger and more complicated, making it even more difficult to crystallize. But in the case of cryoEM single- particle reconstruction, you can deal with such complexities.”
Murakami is using cryoEM as an alternative approach to visualize 3D structures of macromolecules—structures that in some cases he was unable to visualize using X-ray crystallography. CryoEM can tolerate a less uniform sample because crystallization is not required and the analysis process can separate out any molecules that are in different conformations within the sample. And for cryoEM, larger is better for processing images and determining structure.
“We actually have high-resolution X-ray crystal structures of most of the individual macromolecules I’m working on right now,” he said. “But with cryoEM we can see the entire structure of a macromolecule complex, like RNA polymerase and the bacterial transcription factor RapA, and thus investigate how the individual molecules are interacting with each other. Because of this, we have an opportunity to study how RNA polymerase communicates with other factors during the transcription process and can get a better idea of the biological function of these factors.”
Murakami is excited by the potential of cryoEM, but he sees it as a complement to other techniques, not a replacement—at least for now. According to Murakami, preparing samples for cryoEM can also be challenging because, just like with X-ray crystallography, not all samples produce a good image.
And because cryoEM is an evolving eld, analysis can also be time-consuming. With X-ray crystallography there is a standard pipeline for processing images, but this is not the case for cryoEM, where new programs and software are developed almost every week to test the limits of this new technology. These new developments to programs and software, which are all open-source, can be frustrating because there is not always a clear end point to data analysis. But it also means a researcher can write a software patch to do something unique for their lab, which could improve resolution of an image. Thus, the limits of resolution for a particular sample are not always immediately clear.
Optimizing the Facilities
Just as 4K video cameras produce massive files that require more storage space and new, more sophisticated players to play the new files, cryoEM has presented new challenges for data processing, transfer, and storage.
“I am amazed at how much data is produced,” said Murakami. “With X-ray crystallography, I can store data for over 200 samples on a one-terabyte hard drive. But with cryoEM, I need four terabytes of space for data from just one sample. As a result, I started using the Penn State computer cluster to process my data and eventually needed to build a new, faster computer to keep up.”
To help researchers keep pace with their data, Penn State has implemented a high-speed fiber-optic network that allows data transfer from the microscope to the researcher ten times faster than before. They have also allocated additional storage space for researchers using the microscope.
“Penn State has made a huge investment in infrastructure,” said Hafenstein. “We needed to set it up so the data coming off the Krios has someplace to go, and someplace to go as fast as it is being created.”
To make the most of the new facilities, Hafenstein, Nixon, Murakami, and other researchers on campus meet regularly in a working group, where they share the newest ways to analyze data and discuss the challenges of particular projects. They are also exploring how to overcome the apparent limits to cryoEM, to enhance the imaging potential of the new facilities. For example, Hafenstein is working with Scott Lindner, assistant professor of biochemistry and molecular biology, to explore new ways to optimize sample preparation to allow imaging of smaller samples using cryoEM (View the Intellectual Property article).
In addition to promoting new lines of communication among researchers in the working group, the installation of the imaging facilities has fostered opportunities for new collaborations across the University. Materials scientists, for example, can now take advantage of cryoEM—which may allow high-resolution images of more fragile “soft” materials—as well as the expertise of life scientists who frequently interpret this kind of data.
Unique to the Penn State Titan/Krios, the microscope is also capable of performing spectroscopy, which relies on the interaction of electromagnetic energy with compounds and elements inside the material being imaged. This unusual fusion of imaging techniques opens up new lines of inquiry for life scientists and materials scientists alike.
“We can take a biological sample and look at the structure, but we can also identify the inorganic elements in the sample,” said Nicole Hackenbrack, a graduate student in Hafenstein’s lab who also studies virus entry into host cells. “We are interested in merging the two imaging techniques in ways that have not really been done before. For example, the viruses we study have calcium ions all over them. Can we see the calcium ions? Do they play a role in virus entry? Do they interact with the host cell? We’re hoping to explore these questions as we continue our work on the Titan/Krios.”
With the implementation of the new facilities, Penn State is poised to attract new faculty. The imaging facilities have also facilitated collaborations with researchers from other institutions, including at the University of Pennsylvania and industry groups, who are eager to make use of the equipment and knowledge at Penn State.
The unique nature of Penn State’s imaging facilities, combined with the impressive advances to cryoEM in the past decades, has positioned Penn State to make significant contributions to the field of structural biology and beyond.
“There are 90 of these microscopes in the world,” said Hafenstein, “but because of the unique combination of imaging technologies, ours is one and only.”