A Giant Tool to See the Tiniest Things

Cryo-EM microscope unleashes 'a tsunami' of new protein structures
Share Share This
May 9, 2019

In a back room in the farthest corner of the Shared Materials Instrumentation Facility in Fitzpatrick-CIEMAS, past the rooms of scientists working in white, head-to-toe cleanroom suits, and adjacent to a work space smattered with computer monitors and engrossed, perplexed-looking graduate students, sits Duke University’s shiny new baby. It’s a tool scholars hope can help solve some of science and society’s most head-scratching biological problems.

(Scholars: To learn more visit http://smif.pratt.duke.edu/node/55048 or send an email to krios@duke.edu

The multimillion dollar cryo electron microscope (cryo-EM for short) is a dark gray, 15-foot-tall monolith, standing regally in the center of its own comparatively small, spare room. Its sturdy square body, covered in smooth, featureless panels, reaches upward like a giant Jenga tower. However, the panels are actually seven doors that swing out to allow scientists access to different parts of the remarkably busy and complex machine inside. 

Inside, cords wrap around a coffee-can sized cylindrical chamber and extend in every direction to all sorts of other metal boxes and blinking lights. Structural biologist Mario Borgnia opens the door panel closest to the room’s entrance and motions to show where he would place a nanocab -- a small box that Borgnia thinks of as a “taxicab” for samples of protein -- into the microscope. The insertion of the nanocab activates a mechanized arm that places the sample in the autoloader.  Another arm places the sample in the column, where the real magic happens. 

A 300,000-volt beam of electrons will shoot down through the sample so that highly-sensitive electron detectors at the bottom of the column can take a picture of the interaction that occurs.

The system actually takes hundreds of thousands of molecular images, that are then classified and averaged by powerful software to create a 3D image and ultimately, a model of the protein.

Zoom In on Cryo-EM image
Proteins, such as this human apo-ferritin molecule, are the stuff of life. We need to see them in great detail, in this case, down to 1.6-Angstrom-resolution. 
(Image: Alberto Bartesaghi)

Being able to “see” proteins --  life’s crucial building materials, found everywhere in the body -- can help determine how they are used. And recognizing protein structure and function is essential for scientists trying to design better drugs to tackle some the world’s most devastating diseases, including HIV, cancer, and Alzheimer’s.

“We talk about proteins every day in relation to our diet. What most people do not realize is that proteins come in different shapes and sizes,” explains Borgnia. “Proteins are like the plastic of living organisms. We produce proteins in all shapes, just like with plastic objects. Long strings of a particular protein, collagen, give tensile strength to our skin; smaller arrow-like proteins, called antibodies, attach to viruses and bacteria and help us fight disease.”

Borgnia motions to the quieter work space with the bank of monitors.

“The computer is in this room, but we can view the images on the monitors in the next room,” he shouts over the noise of two loud computer servers in his Argentinian accent. The whole room sounds like a subway station in which a train is perpetually whistling by.

In the monitor room, Borgnia explains that the imaging of a sample is just the endgame, the result of lots of hard work to prepare the sample. 

First, the protein sample must be cooled to preserve it against the powerful electrons passing through, explains Borgnia. Though after enough blasts, it will indeed disintegrate. Borgnia likes to compare the process to a fully-leafed tree being hit by an autumnal gust of wind.  With the first gust of electrons, a few leaves fall.  With another gust, more leaves fall, and after multiple gusts, the tree is nearly bare, and then it’s hardly even there, no longer recognizable.

The tricky part is that each sample must be frozen, embedded within a substance that will not make it more difficult to see under the microscope.

The protein sample is deposited on a lentil sized thin copper disc that is perforated by hundreds of square windows. The disc is coated with a thin film of carbon that is patterned to create micrometer-sized circular holes. When the protein containing solution is spread across the holes it creates a liquid film similar to soap water stretched across a bubble wand.  The sample is then dunked into propane that has been liquefied by cooling it with liquid nitrogen.  When rapidly cooled to this temperature, water forms glass-like ice -- a phase at which ice does not have a defined structure.  Unlike regular crystalline ice, glassy ice is transparent to electrons helping create contrast with the embedded proteins. To prevent the transition to crystalline ice, the sample must be kept in a very specific temperature range, at or below -220 degrees Fahrenheit (the temperature of liquid nitrogen is -321 F). This is the “cryo” part of cryo-EM.

Once frozen, the sample is strengthened with a copper ring. Up to 12 grids with rings can be inserted into a cassette that is then placed in the nanocab for the robot arms to handle. To obtain a specimen suitable for imaging, the chemical environment of the protein needs to be carefully selected. This trial and error process is guided by screening multiple specimens in an electron microscope. The Molecular Microscopy Consortium, directed by Borgnia, provides access to screening microscopes that share the same specimen loading system. Once the proper specimen preparation conditions are found, one single ring imaged in the Duke Krios can provide millions of molecular images.

Although cryo-EM has been around since the 1970s, recent advances in detector technology, software algorithms, and computing have revolutionized the field. The 2017 Nobel Prize in chemistry was awarded to three scientists for the advancement of cryo-EM to allow for high-resolution structure determination of biomolecules in solution.

About two years ago, Duke University, in partnership with the National Institute of Environmental Health Sciences (NIEHS) in Research Triangle Park and the University of North Carolina at Chapel Hill, formed the Molecular Microscopy Consortium which allows scientists from these institutions to use three cryo-EMs, one at each facility. Borgnia was recruited to organize and lead the consortium in 2016. 

Borgnia trained in Argentina and Israel before he moved to the United States in 1996 to begin a post-doc at Johns Hopkins University with Peter Agre, a molecular biologist who won the 2003 Nobel Prize in Chemistry for his discovery of aquaporin water channels. Fascinated by the structural basis of cellular transport systems, Borgnia started working for the National Cancer Institute in 2001 with Sriram Subramaniam, where he learned cryo-EM and developed automated pipelines for image processing.

Cryo-EM model
This image of a human protein was built from 1,000 scans of more than 315,000 individual molecules, and then assembled in a computer. (Credit:  Alberto  Bartesaghi)

Duke’s Titan Krios cryo-EM microscope has already been used to produce more than a dozen high resolution images since it was deployed in April 2018.  A team of scientists at Duke University -- including Richard G. Brennan, PhD, department chair and James B. Duke Professor of Biochemistry, Barton F. Haynes, MD, the Frederic M. Hanes Professor of Medicine and director of the Duke Human Vaccine Institute, and Robert J. Lefkowitz, MD, James B. Duke Professor of Medicine -- advocated for bringing the technology to Duke, and Nancy Andrews, MD, PhD, former dean of the School of Medicine, and her team recognized the opportunity and championed it, winning the approval and support of Chancellor A. Eugene Washington.

“The School of Medicine and the Chancellor are excited to bring this technology to Duke with an initial investment of almost $10 million,” said Raphael Valdivia, Professor of Molecular Genetics and Microbiology and the former Vice Dean for Basic Science who oversaw the launch of this new initiative.

The School of Medicine was fortunate to find a home for the machine in the Shared Materials Instrumentation Facility in Fitzpatrick-CIEMAS, a facility under the direction of Duke engineering professor Nan Jokerst and facility director Mark D. Walters.

The new machine has helped to attract some of the school’s brightest new recruits, who rely on the technology for their research programs.

School of Medicine researcher Alberto Bartesaghi, who came to Duke in 2018, focuses on developing computational approaches to advance the capabilities of cryo-EM. Trained as an electrical engineer, he entered the biomedical field through a position with the NIH, incidentally in the same lab as Borgnia, where he spent 13 years optimizing the imaging of proteins and data processing with cryo-electron microscopy.

“The NIH was one of the first places to get a high-end electron microscope,” said Bartesaghi, an associate professor of biochemistry in the School of Medicine. “Back in the day, the technology was very different from what it is today.”

Although several factors have contributed to the cryo-EM revolution, including improvements in automation and better cameras, Bartesaghi is most excited about the remarkable advances in image analysis that have occurred in the past decade. He thinks the cryo-EM facility will yield a great return at Duke.

Structural biologists using X-ray crystallography -- a time-honored technique that involves crystallizing a molecule and then measuring how X-rays diffract when the crystal is hit -- may find that cryo-EM is an effective alternative to study hard-to-crystallize proteins, said Bartesaghi.

“Until recently, there were many important biomolecules where the biochemistry was well understood but researchers couldn’t determine their structures because they had a hard time crystallizing them,” said Bartesaghi. “Thanks to the recent technological advances in the field, all of a sudden these targets became tractable by cryo-EM and that has created a tsunami of new structures.”

Priyamvada Acharya, an associate professor of surgery in the School of Medicine and director of the division of structural biology at the Duke Human Vaccine Institute, came to Duke in 2018 after working at the Vaccine Research Center at NIH for twelve years. The last three of those years she was an NIH embedded researcher at the Simons Electron Microscopy Center, New York Structural Biology Center. Acharya was also one of the first twelve recruits brought to Duke through the Translating Duke Health Initiative.

Acharya uses cryo-EM in her search for a vaccine against HIV. Using the new Titan Krios, she can obtain high-resolution images revealing nanoscopic spikes on the surface of the virus. These spikes protruding from the virus attach to the membrane of a human cell and fuse it with the viral membrane, allowing a package of genetic material to penetrate into the cell. Soon, the genome of the virus takes over the cell machinery and thousands new viruses bud at the surface of the cell, each inside a tiny bubble of membrane. Because the virus is now hidden inside a membrane from the human cell, the body doesn’t recognize it as an invader. The only hope is for the immune system to recognize the spikes, which are the only protein that the virus has to expose because they are needed to recognize and bind to the next target cell.

Acharya wants to know everything she can about  HIV-1 entry so she can  develop an HIV-1 vaccine. “Cryo-EM helps you rapidly figure out the fine details of intermolecular interactions, thereby giving you the tools to manipulate it,” she said.

In addition to attracting new recruits, the cryo-EM facility has been a valuable resource to faculty members who have called Duke home for decades, allowing them to take their research to the next level without leaving campus.

Robert J. Lefkowitz, MD, the James B. Duke Professor of Medicine who won a Nobel Prize for identifying g-protein coupled receptors, or GPCRs, is using cryo-EM to obtain molecular snapshots of these receptors as they bind to various proteins. In particular, his lab focuses on receptors that regulate cardiovascular physiology, such as blood pressure and heart rate. More than 600,000 Americans die every year from cardiovascular disease, making it the leading cause of death for both men and women, and accounting for almost one in four deaths.

“We believe that studying the structure of GPCRs that regulate cardiovascular function has the potential to aid in the discovery of better medicines for the treatment of heart diseases,” said Anthony Nguyen, an MD/PhD student who works in Lefkowitz’ lab. “Cryo-EM has been critical to our research.”

Cryo-EM has allowed for the direct visualization of the GPCR of interest, said Nguyen, speeding up the rate at which a structure can be solved, as evident by the large influx of GPCR cryo-EM structures within the last year alone. 

“These G-protein coupled receptors are membrane receptors and are challenging to crystallize -- you’d have to do some biochemical tricks to get them to behave,” Nguyen said. “Cryo-EM allows you to bypass some of that.”

Contact: