Getting a detailed look at human proteins is often key to designing drugs that target them. For decades, drug developers have typically relied on X-ray crystallography for that purpose—a technique that captures the three-dimensional structure of protein crystals by blasting them with X-rays and analyzing the resulting diffraction patterns. This has revealed protein structures in unprecedented detail, sometimes even at an “atomic resolution” of at least 1.2 Å, high enough to discern individual atoms.
X-ray crystallography has been vital to structure-based drug design, whereby drugs are designed based on the chemical structure of target proteins. In-depth studies of protease enzymes used by human immunodeficiency virus (HIV), for instance, have facilitated the development of enzyme-inhibiting drugs to treat HIV. But the technique has a major drawback. It requires that proteins be in a crystallized state, which means subjecting them to certain physical processes to coax them into an ordered lattice, a cumbersome process that doesn’t work for all proteins.
That’s where cryo-electron microscopy (cryo-EM) comes in, a newer technique that doesn’t require crystallization. Rather, protein samples in solution are simply flash-frozen and then bombarded with electrons; images are produced from the resulting patterns of electrons projected onto a detector. Yet the technique was long derided as “blobology” for the relatively low-resolution images it captured. These images, which up until the late 1990s had a resolution of around 20 Å, could show only the overall shape of a protein.
Fortunately, persistent scientists managed to improve cryo-EM’s resolution over time. The development of better hardware for detecting electrons and efficient algorithms to construct a three-dimensional structure from multiple images of a protein eventually triggered a “resolution revolution” in cryo-EM around 2015.1 In 2020, researchers achieved an impressive 1.25 Å structure of the protein apoferritin, demonstrating that at least in some cases, cryo-EM could visualize proteins in atomic detail.2 The technique has “moved biochemistry into a new era,” declared the Royal Swedish Academy of Sciences when it awarded the 2017 Nobel Prize in Chemistry to the scientists Jacques Dubochet, Joachim Frank, and Richard Henderson for their work in developing cryo-EM.3
Among drug makers, cryo-EM has quickly established itself as a popular tool to parse the structures of target proteins, especially the structures of large proteins (such as the membrane-bound proteins that make up the majority of modern drug targets), which are hard to tackle with X-ray crystallography. Cryo-EM has not only expanded the range of proteins that can be targeted through structure-based drug design, but it has also become a key imaging tool in drug development more generally. GEN spoke with several entities at the forefront of this cryo-EM movement—microscope manufacturers, companies providing cryo-EM services, drugmakers, and specialists in data storage—to learn more about how cryo-EM is reshaping drug discovery.
Harnessing cryo-EM for structure-based drug design
Since its founding in 2001, Nanoimaging Services, a contract research organization based in San Diego, CA, has specialized in performing cryo-EM for pharmaceutical and biotechnology companies. Initially, Nanoimaging Services used the technique to image drug delivery vehicles such as nanoparticles to ensure that each batch consisted of same-sized and intact particles, says Giovanna Scapin, PhD, the company’s chief scientific officer.4
Now that higher resolution can be achieved, one of the company’s focuses is determining the structures of the proteins that are being targeted by drug makers. The goal is often to scrutinize how a lead drug candidate binds to a protein, either during the initial design of a drug, or later on, to get a better understanding of how drug and target interact.5
Scapin sees a strong demand for cryo-EM structures of membrane proteins and protein complexes, which are usually too big, too flexible, or available only in quantities too small to allow crystallization. She points out that another advantage of cryo-EM is that it doesn’t require that proteins be crystalized in a specific confirmation. “We can look at the real native state of the protein,” she emphasizes.
The average resolution of cryo-EM structures are in the range of 2 to 2.5 Å. Scapin admits that this is below the typical resolution of X-ray crystallography, but she adds that computational modeling can be used to optimize the atomic models of the proteins. The greatest challenge with cryo-EM, however, lies in preparing the protein samples. It’s often necessary to bind targets to other molecules. For example, highly flexible proteins may need to be stabilized, or proteins that are too small to be sharply imaged on their own may need to become part of something larger. In principle, cryo-EM “can be used for everything,” Scapin declares, “[but] we have to make the protein work for cryo-EM.”
Parsing protein complexes
Lei Jin, PhD, the COO of Wuxi Biortus Biosciences, a contract research organization that specializes in structural analysis of proteins, agrees that sample preparation remains a challenge. Still, the main advantage of cryo-EM is its speed, he points out. He explains that for many popular drug targets, especially large proteins and protein complexes, “it can take crystallography years to get a structure.”
Wuxi Biortus scientists say that many such proteins have been imaged with cryo-EM. That includes, for instance, cyclin-dependent kinase 7 (CDK7), an enzyme involved in cell cycle regulation and a popular target in oncology. At 2.5 Å resolution, scientists have mapped the large complex that CDK7 forms with two other proteins, together with a small-molecule drug designed to inhibit CDK7.6 Such studies “can tell us a lot about the nature of interaction between drug and target,” Jin says. “We can even use this information for further optimization of the small molecule.”
Other cryo-EM subjects include proteolysis-targeting chimeras (PROTACs), large complexes that employ the body’s own protein degradation mechanisms to destroy disease-causing proteins that are hard to tackle with small-molecule drugs. Imaging such complexes with cryo-EM can aid in optimizing of PROTAC molecules, for instance, by making them more specifically targeted to proteins of interest. Cryo-EM studies can also offer unique insights into how established drugs work on new targets, such as the drug remdesivir, which was employed against the novel coronavirus SARS-CoV-2.7
The only main limitation of cryo-EM, as Jin sees it, is that the protein under investigation has to be larger than 80 kDa (or incorporate more than some 700 amino acids) for its structure to be reliably determined—although Jin adds that Biortus scientists have succeeded in pushing the limit down to 70 kDa. He remarks that if a protein is any smaller than that, X-ray crystallography “can do a better job.”
Cryo-EM meets fragment-based drug design
The Cambridge, U.K.-based company Astex Pharmaceuticals began experimenting with cryo-EM in 2016. Along with FEI (a microscope manufacturer based in Hillsboro, OR), the U.K.’s Medical Research Council Laboratory of Molecular Biology, and the University of Cambridge’s Nanonscience Center, Astex participated in an effort to help pharmaceutical companies explore the use of cryo-EM for early-stage drug discovery research.8
“Initially, learning the technique “was a really steep learning curve,” says Pamela Williams, DPhil, Astex’s senior director of molecular sciences. That was especially the case as she and her colleagues were accustomed to the high-throughput, automated workflows developed over years for X-ray crystallography. “The [cryo-electron] microscope is an incredibly technical piece of equipment,” she adds.
In 2020, Williams and colleagues conducted a proof-of-concept study that yielded more encouraging results. This study demonstrated the utility of cryo-EM for Astex’s specialty, fragment-based drug design.9 This method screens small fragments and studies how they bind in isolation to a protein; the fragment can be iteratively expanded to build a molecule that effectively binds to the protein. The study showed that cryo-EM can indeed be used to study the molecular interactions between a small fragment and a protein—in this case, the popular oncology target pyruvate kinase 2—at a resolution comparable to that achieved with X-ray crystallography.
Astex is not pursuing that target, but the company is using cryo-EM to apply fragment-based drug design to membrane protein targets, such as ones involved in central nervous system disorders. According to Williams, cryo-EM could be combined with X-ray crystallography, for instance, when designing drugs that disrupt binding events between two proteins. Cryo-EM could be employed to generate images of the protein complex, whereas subsequent X-ray crystallography could be used to get detailed looks at each protein individually. “The two techniques,” she insists, “have the potential to work alongside each other.”
Tackling cryo-EM’s data storage problem
Björn Kolbeck, PhD, the CEO of Quobyte, a data storage specialist in Santa Clara, CA, has seen that in recent years, many life sciences fields have started generating vast quantities of data, especially cryo-EM data. As cryo-EM is used to accumulate the many high-resolution images of proteins that are needed to achieve a good three-dimensional reconstruction of a protein, the amount of data, Kolbeck notes, is “just much more than other imaging technologies in life sciences.”
In such cases, it’s crucial to have a robust and efficient data storage system that scales with the amount of generated data, allowing the continuous operation of microscopes and the rapid processing of images. But too often, storage is an afterthought to the scientists who use cryo-EM, Kolbeck remarks. Many store their data on individual servers, creating a single point of failure if the system breaks. Small pharmaceutical and biotechnology firms often start out storing their data in the cloud. According to Kolbeck, this approach quickly becomes very expensive and doesn’t work well when data needs to be streamed from an on-site appliance like a microscope.
That’s where Quobyte’s software comes in.10 The system essentially aggregates data from multiple servers into one big storage system. That way, more servers can be added at need, and storage is still available if one server fails or needs to be serviced.
Since the company began to advertise its software for cryo-EM purposes two years ago, it has seen interest from academic groups and drug developers alike. Kolbeck predicts that the demand for data storage will grow as cryo-EM technology continues to evolve. “From a storage perspective,” he states, “it’s only [going to get] worse.”
The future of cryo-EM
Cryo-EM technology has indeed come a long way, says Melanie Adams-Cioaba, PhD, senior director and general manager of pharma at Thermo Fisher Scientific. The company, which is based in Waltham, MA, provides a range of laboratory instruments and supplies—including those for cryo-EM. The company has been selling cryo-EM microscopes since 2016, when it acquired FEI Company, an industry leader in developing electron microscopes. These instruments include the Titan Krios microscope, which captured the iconic, atomic-resolution image of apoferritin in 2020.
Advances in cryo-EM technology have established the technique as the method of choice for hard-to-crystallize targets, and now cryo-EM applications are rapidly expanding, Adams-Cioaba points out. One of the more popular cryo-EM applications is the imaging of large and flexible biologics such as therapeutic antibodies. Cryo-EM, Adams-Cioaba adds, may prove useful “anywhere where we really want to understand the structure of proteins, how proteins interact with other proteins, and how modifications to those proteins may affect how the proteins function.”11
For instance, drug makers can use cryo-EM to help sift through large numbers of antibodies they’ve generated against a particular protein. Such structural analyses can help quickly categorize antibodies by their structure and select the ones that have the greatest therapeutic promise. Historically, structural biology has seldom been employed at such early stages of antibody development. Doing so has been difficult in part because the proteins aren’t easy to crystallize. “But thanks to the faster pace of cryo-EM, the technique allows structural studies to guide antibody development much earlier in the process.”
Adams-Cioaba anticipates that cryo-EM will continue to find new applications in drug development. “The combinations of innovation in the field really just continue to break my barriers of perception,” she declares. “[Cryo-EM possibilities are] limited only by our tenacity and our imagination.”
References
1. Method of the Year 2015. Nat. Methods. 2016; 13: 1.
2. Yip KM, Fischer N, Paknia, et al. Atomic-resolution protein structure determination by cryo-EM. Nature 2020; 587: 157–161.
3. Cressey D, Callaway E. Cryo-electron microscopy wins chemistry Nobel. Nature 2017; 550: 167. DOI: 10.1038/nature.2017.22738.
4. Zhao Q, Potter CS, Carragher B, et al. Characterization of virus-like particles in GARDASIL(R) by cryo transmission electron microscopy. Hum. Vaccin. Immunother. 2014; 10(3): 734–739. DOI: 10.4161/hv.27316.
5. Scapin G, Potter CS, Carragher, et al. Cryo-EM for small molecules discovery, design, understanding and application. Cell Chem. Biol. 2018; 25(11): 1318–1325. DOI: 10.1016/j.chembiol.2018.07.006.
6. Greber BJ, Remis J, Ali S, et al. 2.5 Å-resolution structure of human CDK-activating kinase bound to the clinical inhibitor ICEC0942. Biophys. J. 2021; 120(4): 677–686. DOI: 10.1016/j.bpj.2020.12.030.
7. Yin W, Mao C, Luan X, et al. Structural basis for inhibition of the RNA-dependent RNA polymerase from SARS-CoV-2 by remdesivir. Science 2022; 368(6498): 1499–1504. DOI: 10.1126/science.abc1560.
8. Cambridge Pharmaceutical Cryo-EM Consortium formed by FEI, five pharmaceutical companies, the Medical Research Council and the University of Cambridge. News Medical Life Sciences. Published April 4, 2016. Accessed October 25, 2022.
9. Saur M, Hartshorn MJ, Dong J, et al. Fragment-based drug discovery using cryo-EM. Drug Discov. Today 2020; 25(3): 485–490. DOI: 10.1016/j.drudis.2019.12.006.
10. Velasquez F. What is cryo-EM and how to choose the best storage for it? Quobyte. Published September 26, 2022. Accessed October 25, 2022.
11. Drulyte L, Koester S, Lundberg D, et al. High-throughput cryo-EM epitope mapping of SARS-CoV-2 spike protein antibodies using EPU Multigrid. Thermo Fisher Scientific White Paper. 2022.