In a speech given in 1959, Caltech legend Richard Feynman—the embodiment of the brash, know-it-all physicist—described his idea for tackling some of the biggest mysteries in biology. To understand how biological molecules like DNA and proteins work, he suggested, “You just look at the thing.”
The idea sounds obvious. After all, to understand how, say, an engine works, it helps to know what the thing looks like. To know how the molecular machinery of the cell works, it makes sense to look at the molecules in question. Feynman had been referring to the possibility of developing better electron microscopes to directly see these molecules. But even now, the best electron microscopes have their limitations.
So to infer the structure of molecules, chemists and biologists have been using a technique called X-ray crystallography, which allows them to visualize what those molecules look like, albeit indirectly. The tool is powerful, and for the last 100 years, researchers have used it to probe everything from small and simple compounds to increasingly large and complex proteins. Crystallography gave birth to the fields of structural chemistry and biology, both of which can claim much of their origins at Caltech, thanks to the pioneering work of two-time Nobel laureate Linus Pauling (PhD ’25).
Today, Caltech researchers are continuing that legacy, using X-ray crystallography to determine the structures of molecules that are more complicated than even Pauling thought would be possible to analyze. In doing so, the researchers are not only uncovering the molecular processes central to life, they are also developing treatments for HIV and laying the foundation for future drugs.
There are many ways to study a molecule, but if you don’t know its exact structure, your tools can only hint at what it looks like and how it works. It’s as if you’re blindfolded, says biochemist André Hoelz. X-ray crystallography, however, unveils the whole picture. “Someone takes off your blindfold, and you say, yeah it all makes sense,” he says.
Unlike X-ray techniques whose images reveal a broken bone, X-ray crystallography doesn’t produce an actual picture of molecules. It’s an indirect method that involves shooting X-rays at a molecule and measuring how those rays scatter, data that can be used to establish the molecule’s structure. It’s the closest you can get to a photograph.
Molecules are tiny and often treated abstractly, relegated to the diagrams and ball-and-stick models you remember from chemistry class. But with X-ray crystallography, molecules become physical, tangible structures. Biochemist Bil Clemons compares the experience to that of a mountaineer seeing Mount Everest for the first time. “Until you’ve really engaged in it, you haven’t really discovered it,” Clemons says. To see is, in the most visceral sense, to discover.
The Birth of a Field
X-ray crystallography was discovered in 1912, when German physicist Max von Laue realized that the atoms of a crystal—or, more precisely, the electrons in those atoms—can deflect the trajectory of an incoming X-ray. When you blast a crystal with X-rays, the crystal diffracts the incoming beam. Von Laue didn’t fully understand the resulting pattern of the diffracted X-rays, but researchers soon realized that the pattern was a fingerprint that reveals the unique structure of the crystal. He was awarded the Nobel Prize just two years later.
Arthur Amos Noyes, who in 1919 resigned from MIT to join full time the faculty of the school soon to be renamed Caltech, recognized the potential of this discovery. He established an X-ray crystallography lab where research was conducted that led to the first PhD awarded at Caltech, to Roscoe Gilkey Dickinson in 1920. Three years later, when Pauling arrived to work on his PhD with Dickinson, Caltech was already the best place outside of Europe to do crystallography. Following a brief stint abroad after earning his degree, Pauling returned to Pasadena, and his efforts to understand chemical structures and bonds would make Caltech the center of what’s now known as structural chemistry.
The following decades were the golden age of the field, says chemist Doug Rees, whose own PhD adviser at Harvard, Nobel laureate William Lipscomb (PhD ’46), was a student of Pauling’s. “Almost anyone who became anyone in small-molecule crystallography and structural chemistry was here in Pasadena or studied with someone who had been here,” Rees says.
Pauling moved from simple molecules toward the more complex ones involved in biology, using the same technique to help establish the field of structural biology. By the 1930s, thanks to the Nobel Prize–winning work of chemist James Sumner, biologists learned that proteins could be turned into crystals—which meant that crystallography wasn’t just for simple compounds, but also for complex, biological molecules.
With the help of X-ray crystallography, Pauling uncovered the nature of the peptide bond, which holds together amino acids, the building blocks of proteins. In 1951, he and colleagues figured out the structures of alpha helix and the beta sheet, the two major structural components of proteins. The discovery of the alpha helix blazed the trail for understanding more complex molecules and the structure of DNA, Rees says. In fact, Pauling was racing to solve DNA but was beaten in 1953 by James Watson, Frances Crick, and Rosalind Franklin and her X-ray diffraction measurements. Pauling had proposed a triple helix, instead of the correct double-helix structure.
Using X-ray crystallography, the Pasadena and Cambridge groups pioneered the structural approach, which, Rees says, is based on the thesis that, to understand how a molecule works, you have to know what it looks like. “That world view permeates everything that we do,” he notes.
In 1939, Pauling had written that, given how long it took at the time to solve the structures of just a couple of amino acids, it might never be possible to determine the structure of an entire protein. Pauling was brilliant, Rees notes, but about this he was wrong. “Now, with time and technical advances, we’re able to approach problems that were inconceivable in Pauling’s time,” he says.
In the past, the calculations needed to interpret X-ray diffraction patterns and solve molecular structures were the most time-consuming part of crystallography. Now, X-ray technology is constantly improving, with better detectors and with beams that can be controlled with greater and greater precision, allowing researchers to analyze bigger and more complex structures. Computers now take less than a second to do the math that once took weeks by hand. These advances have made crystallography nearly limitless in its potential, Hoelz says. Researchers can study as complex a system as they wish. “We pick a project and we make it happen,” he says.
To aid in their experiments, Caltech researchers have access to the Molecular Observatory, which includes an automated X-ray beam line at the Stanford Synchrotron Radiation Laboratory. The campus also has its own in-house macromolecular X-ray crystallography facility. The Macromolecular Crystallization Laboratory in the Beckman Institute provides automated and robotic facilities (yes, there are actual robots) to help researchers prepare their samples efficiently.
Such automated methods allow researchers to do experiments with much less sample material than before, and thus to solve more challenging problems. According to Clemons, in the roughly 15 years since he started doing crystallography the amount of sample protein needed has dropped by about 20-fold. Still, preparing and crystallizing samples is laborious, and is the hardest part of crystallography today.
In general, there are no rules that tell you how to crystallize proteins. Researchers must resort to trial and error, and even when they do find the right methods, the crystallization process often takes weeks to months.
The basics of crystallization, however, are simple enough. You can make your own crystals at home by dissolving sugar (a sample) in hot water (a solution). As the water cools, the sugar molecules come out of solution and arrange themselves in a regular pattern, forming a crystal. The same general process applies to crystallizing more complicated samples like proteins. But because such molecules are so big and complicated, it’s hard to fit them together in an organized crystal structure. Stacking boxes is easy. But what if you had to stack chandeliers?
Hoelz, for example, is tackling one of the most elaborate structures of them all, a top-shaped behemoth called the nuclear pore complex. This structure acts as the gatekeeper of a cell’s nucleus, which contains the organism’s DNA—the genetic blueprint for all the proteins that enable the cell to function.
Whenever the cell needs a protein to complete a task, a signaling molecule is sent to the nucleus, where the genetic instructions for building that protein are read and sent back out in the form of a molecule called messenger RNA. The protein is then synthesized outside the nucleus in the cytoplasm. The nuclear pore complex ensures that the signals get into the nucleus and the messenger RNA gets out. Many other kinds of molecules go in and out of the nucleus, including up to 40 percent of all proteins, Hoelz says. The nuclear pore complex is the portal through which they all must pass.
Finally understanding the structure of the nuclear pore complex could lead to a new class of cancer or antiviral medicine, Hoelz says. Some viruses, like HIV, hijack this structure to invade the nucleus, integrating themselves with the host’s genome. If researchers can figure out how the nuclear pore complex works, then they might be able to design drugs to block such intruders.
The nuclear pore complex is enormous, containing roughly 10 million atoms. For comparison, the alpha helix that Pauling modeled contains fewer than two dozen atoms. “This is a totally different scale from what people have done in the past,” Hoelz says.
In fact, it’s so big and complicated that you can’t use X-ray crystallography on the whole thing. The center of the complex can move, so if you try to do crystallography on the entire structure, the results are blurry, like when you try to take a picture of a squirming toddler.
So the researchers have to break it down into more than 1,000 components and solve the structure of each, one by one. Electron microscopy then provides a rough picture of the complex’s overall shape, which helps them piece together the puzzle.
Some declared the nuclear pore complex too big to solve, Hoelz says. “People said, ‘You’re crazy. It’ll take you forever to get any structures and any insight.’” But over the last 10 years, he and his colleagues have made tremendous progress, and he says they may be able to figure it all out in another 10 years.
Meanwhile, Rees has already spent over three decades studying the structure of enzymes that help bacteria convert nitrogen—an important nutrient, but inert and useless in its gaseous form—into chemicals such as ammonia, which can be used by living organisms.
In addition, Rees focuses on molecules called membrane proteins. Sitting inside the cell membrane, which separates the cell from the outside world, these proteins act as border control, determining what can enter or exit. We have several thousand types in our body; about 30 percent of the genome encodes for these proteins.
And they’re particularly challenging to study, says Clemons, whose research also focuses on membrane proteins, in particular those that control the comings and goings of other proteins. The layers inside the cell membrane are hydrophobic, meaning that, like oil, they don’t adhere to water. Because membrane proteins are embedded inside these layers, they assume especially complicated structures that enable them to interact with the watery world outside the membrane as well as the oily interior.
Rees looks at membrane proteins called transporters, which control the passage of smaller molecules such as the amino acids, sugars, and other nutrients that come from food. He and his colleagues are trying to understand how transporters efficiently use energy to pump those molecules into the cell. They’re also studying how the transporters themselves are regulated. As border-control agents, they have to know when, for instance, to turn away a particular nutrient if the cell already has enough. Because membrane proteins play such a major role in the cell, many drugs are designed to alter their behavior, Clemons says. Studying them can thus lead to new and better drugs.
Transporters are particularly involved with drug resistance, Rees explains. One feature of cancer cells is that they often produce transporters specialized to pump out anticancer drugs that have entered the cell. By better understanding how these pumps work, researchers may be able to devise ways to inhibit these transporters and allow the anticancer drugs to attack the tumor.
While much of structural biology is basic research that may eventually lead to new treatments for diseases, biologist Pamela Bjorkman is using X-ray crystallography to tackle HIV head-on by examining the structures of antibodies.
Antibodies are proteins that roam the body, hunting for foreign intruders. When an antibody finds such a pathogen, it latches on, preventing the invader from harming cells. A typical mammal can produce more than 10 thousand trillion antibodies to attack all kinds of enemies. But when it comes to HIV, the body has trouble mounting a defense. Although antibodies have two arms with which to grab onto the molecular spikes protruding from the outside shell of a virus, HIV has very few such spikes, making the distance between them hard to bridge. Plus, it mutates, moving the spikes around and changing their shapes, thwarting antibodies that try to attack it.
Still, some HIV patients can develop effective antibodies. Using a variety of tools including X-ray crystallography, Bjorkman is identifying what makes those rare antibodies successful. But because those antibodies are specific to unique strains of the HIV virus, they aren’t able to defend against HIV in other patients. The idea, then, is that by learning how those antibodies work, researchers can engineer new ones able to defeat HIV. For example, Bjorkman and her team recently engineered antibody-based molecules that could grab one spike with both arms, an approach that shows strong promise.
Some antibodies are being tested in human clinical trials, but it’s still too early to know how well they work, Bjorkman says. Lately, however, researchers have developed better ways to isolate unique antibodies. Before 2009, there were only four such antibodies known. Now there’s about a hundred, giving researchers like Bjorkman much more to work with. “There’s just been an explosion,” she says.
A Frontline Tool
For an entire century, X-ray crystallography has continuously propelled science forward, particularly in chemistry and biology. As many as 29 Nobel Prizes arguably involve discoveries related to X-ray crystallography, according to the International Union of Crystallography. Crystallography has even made its way to Mars. In 2012, the Curiosity rover used X-rays to analyze minerals in the martian soil, finding that it’s similar to the volcanic grains found in Hawaii. Right now, crystallography provides great snapshots of molecular structure. But such pictures are only equivalent to still photographs. “My dream in the end is that we have a movie,” Hoelz says. Researchers like him hope to eventually be able to take sequential measurements of biological molecules in action. Then they can play the sequence and watch the molecular motions of life unfold before their very eyes.
X-ray crystallography is still but one technique. Researchers rely on many others, as well as ingenuity and good old-fashioned hard work. But as a single bread-and-butter tool, nothing beats crystallography, Rees says. “For 100 years, crystallography has been the frontline tool for solving structures,” he says. “I think that will always be the case.”
Pamela Bjorkman is the Max Delbruck Professor of Biology and an investigator with the Howard Hughes Medical Institute. Her research is funded by the National Institutes of Health (NIH) and the Bill and Melinda Gates Foundation.
Bil Clemons is a professor of biochemistry. His work is supported by the NIH National Institute of General Medical Sciences.
André Hoelz is an assistant professor of biochemistry. His work is funded by the NIH, the Sidney Kimmel Foundation for Cancer Research, the Leukemia and Lymphoma Society, The V Foundation for Cancer Research, and the Edward Mallinckrodt, Jr. Foundation.
Doug Rees is the Roscoe Gilkey Dickinson Professor of Chemistry and an investigator with the Howard Hughes Medical Institute. His work is funded by the NIH and the Howard Hughes Medical Institute.
Header photo by Louise Scharf