The Beauty Behind the Beasts

Andreas von Bubnoff visits with Sriram Subramaniam and discusses his work studying the structure of HIV and visualizing HIV-infected cells

In 1987, Harden McConnell, a professor at Stanford University, was out of the country when Sriram Subramaniam, an Indian-born graduate student in his lab, wanted to tell him that there was no longer any need to send a recommendation letter to H. Gobind Khorana at the Massachusetts Institute of Technology (MIT), where he had applied for a postdoctoral position.

Subramaniam had just received a letter of rejection from Khorana. But McConnell could not be reached, and sent the letter to Khorana anyway. Once Khorana read it, he changed his mind and decided to invite Subramaniam for an interview. “[McConnell] must have said nice things [in his recommendation letter],” remembers Subramaniam, now chief of the biophysics section in the laboratory of cell biology at the National Cancer Institute (NCI) in Bethesda. “I went to the interview and then immediately he offered me the position.”

The incident was one of many lucky breaks that he has had in his academic life, Subramaniam says. “[It] was a complete accident.” It would also be the first of several times Subramaniam would switch fields.

In McConnell’s lab, Subramaniam was trained as a physical chemist, studying lipid monolayers at an air water interface. But when his application for a position at Bell labs to continue this type of research had been unsuccessful, he decided to switch to biology. In part, that decision may have been influenced by McConnell, who had done the same. “He was a physicist’s physicist for the things that he did,” Subramaniam says, “but he also became interested in immunology.” Subramaniam chose to apply to work in Khorana’s lab because he realized that research in biological membranes was related to his previous work in lipid monolayers.

Once in Khorana’s lab, he studied bacteriorhodopsin, a light-activated transmembrane protein that bacteria use to pump protons across the membrane to store energy. His non-biological background was quite unusual there, he remembers. “I had never done anything biological,” Subramaniam says. “I was the only physical chemist when I joined [Khorana’s] lab. I think he was very suspicious at first.”

But he had experience in building instruments and that turned out to be quite useful. In Khorana’s lab, Subramaniam spent two years building an instrument to measure the effect light of different wavelengths had on the ability of bacteriorhodopsin to pump protons.

Perhaps his broad background in physics and biology contributed to Subramaniam’s next lucky break. In 1989, Richard Henderson, a structural biologist at the Medical Research Council Laboratory of Molecular Biology in Cambridge, UK, sent the manuscript of a paper he was about to send out for publication to Khorana for comments. The paper used electron microscopy (EM) to reveal the first atomic-resolution structure of bacteriorhodopsin (1). Subramaniam was the only person in Khorana’s lab who responded with comments. “Only he was sufficiently broad minded and interested to reply with some constructive comments on our manuscript,” remembers Henderson. “It was clear he appreciated what we were doing.”

About a year later, Henderson and Gebhard Schertler, another postdoc from his group, spent several days talking to Subramaniam at a Gordon conference and realized “with more certainty that Sriram was quite a capable person,” Henderson remembers. So he tried to persuade Subramaniam to join his group in Cambridge for a second postdoc. Even though Subramaniam had already accepted a position as an assistant professor at Johns Hopkins University in Baltimore, Maryland, he decided to go to Cambridge for six months in 1992. “[Henderson] said he would teach me everything I needed to know about EM in two days,” Subramaniam says. “[It] seemed like a good deal.”

In Cambridge, Subramaniam and Henderson shock-froze bacteriorhodopsin in its “open state,” while it was pumping a proton across the bacterial membrane. They then used the electron beam of an electron microscope to analyze two-dimensional crystals of the protein, in a method called electron crystallography. This resulted in the first study that described the structure of the “open state” (2). “[It] was the first structural description of a proton pump caught in action,” Subramaniam says. It was also Subramaniam’s first experience with EM, which he is still using today—in a different way—to study the structure of the Envelope protein spike of HIV.

After six months in Cambridge, Subramaniam became an assistant professor at Johns Hopkins University after all, to study visual pigments in fruit flies. Again, the project involved something he had never done before: Growing fruit flies in the dark. Not being a biologist, he remembers, “I was very squeamish even to touch a fly, [let] alone growing it in the dark, [but] we figured it out.” And again, his ability to tinker with and build instruments came in handy, this time to build an instrument that could measure the absorption of different wavelengths of light inside mixtures of ground up fly eyes. This work led to the elucidation of the mechanism that allows fruit flies to recycle their light pigments after they have been exposed to bright light (3).

Eventually, Subramaniam lost interest in this biochemical approach and decided he wanted to resume his structural work, so he called Henderson to ask if he could rejoin his lab.

Henderson said yes, and in 1997, Subramaniam again went to Cambridge, where he used EM to further refine the open structure of bacteriorhodopsin he and Henderson had published in 1993.

Again, Subramaniam and Henderson used electron crystallography to analyze two-dimensional crystals of the bacteriorhodopsin open state. But this time, they analyzed the crystals at varying angles, each time measuring the diffraction pattern from the electron beam. Since each time the electron beam destroyed the protein, the challenge was to use computers to merge the information of thousands of individual data sets, Subramaniam says. In other words, the approach averages, or merges, a large number of data that come from the way the electron beam in an electron microscope is scattered when it goes through a biological sample. If this is done with a large number of the same proteins, it becomes possible to deduce a common structure at a relatively high resolution. This work resulted in the first three-dimensional (3D) atomic model of bacteriorhodopsin captured in the process of proton pumping done completely by EM (4).

Later that year, Subramaniam moved to the NCI, his current home. Initially, he continued to use electron crystallography to study other crystallized membrane proteins. But later, he realized that it should be possible to use EM to also understand the structure of proteins that are too variable to form two-dimensional crystals, such as the HIV envelope spike.

This approach, called electron tomography, also analyzes the scattering of electron beams through proteins, but in this case, the proteins don’t have to be crystallized. “We cannot crystallize HIV because each virus is different,” Subramaniam says. “That’s why tomography became the natural choice.” The approach requires the merging and analysis of thousands of data sets from the same protein to get a high-resolution image of the protein.

Last year, this work resulted in the analysis of the structure of the HIV Envelope spike with electron tomography, at a higher resolution than previous studies (5; see Figure 1, below). “[The paper] is generally considered to be the best paper on the HIV trimeric spike structure,” Henderson says. 


Figure 1: 3D Structure of the Trimeric Glycoprotein Spike on Native HIV-1 

Electron tomographic analysis of the native, undecorated Env trimer shows the general shape and arrangement of gp120 monomers in the native spike (5). Image courtesy of Sriram Subramaniam, US National Institutes of Health

Next, Subramaniam plans to use electron tomography to analyze the structure of Envelope spikes of different HIV variants, such as ones that are more or less difficult to neutralize, or of transmitted founder viruses that are responsible for establishing infection. He also wants to take a closer look at the way broadly neutralizing antibodies bind to the HIV Envelope spike.

Slice and view

Electron tomography isn’t the only approach Subramaniam has been using to study biological structures. In 2003, another method that uses EM caught his attention. He realized that ion-abrasion scanning electron microscopy (IA-SEM) could be used to understand the structure of relatively thick biological samples, such as cells, at a high resolution. IA-SEM uses a beam of gallium ions to take off layers of the surface of a sample. That surface can then be analyzed by scanning EM. 

Subramaniam says he first heard about such an instrument when visiting a microscopy manufacturer. At the time, the semiconductor industry used these instruments to cut silicon wafers to see if they contained the right circuits. But Subramaniam realized that the approach could be modified into a “slice and view” strategy to get three-dimensional images of cells, by taking SEM images each time the gallium beam had taken off additional layers of the sample. The images could then be recombined to give a 3D image. “When I learned about the existence of the focused ion beam, it was fairly obvious to me that it could be used iteratively to not just look at it once, [but to] keep looking at it multiple times,” Subramaniam says. “It occurred to me that we could use it to look inside cells.”

Subramaniam says the microscope manufacturer FEI had had the same idea, and so he collaborated with FEI to show that IA-SEM could be used to study biological samples such as yeast cells and tumor tissue (6). Subramaniam hopes the approach will be able to provide 3D images with a higher resolution than other approaches, bridging the gap between light and electron microscopy. In addition, the work results in beautiful images. In 2008, one image of a human melanoma cell won him an honorable mention in the International Science & Engineering Visualization Challenge, a competition of efforts to visualize scientific data sponsored byScience magazine and the US National Science Foundation (see image, below).


Visualizing Cells in 3D 

Describing and modeling the hierarchical organization of molecules, molecular machines, and organelles within the cellular interior is a challenge of fundamental interest in cell biology. To bridge this gap in imaging, Subramaniam and colleagues have been developing ion-abrasion scanning electron microscopy (IA-SEM), a strategy for three-dimensional (3D) imaging of biological specimens that utilizes a focused ion beam to remove material from the surface, followed by a scanning electron beam to image the newly exposed surface. These steps can be iterated to “walk into” the cell at step sizes of ~20 nm, resulting in a stack of 2D-surface images that can be combined to visualize the cell in 3D. This is illustrated in the image above with the segmented rendering of a human melanoma cell (mitochondria, endoplasmic reticulum, and nucleus are shown in red, yellow, and dark purple, respectively). Pictured in the photo below is Sriram Subramaniam (center) discussing IA-SEM measurements with Gavin Murphy, a postdoc in his lab. The ion-abrasion scanning electron microscope is on the right. Image courtesy of Donald Bliss and Sriram Subramaniam, NIH. Photo by Andreas von Bubnoff. 


But IA-SEM can yield more than stunning images. Earlier this year, Subramaniam used IA-SEM to show HIV-filled compartments deep inside macrophages that are connected to the surface through channels that HIV particles appear to use to move to the surface (7; see cover image). He is also using IA-SEM to study virological synapses, the contacts HIV uses to move from infected to uninfected cells. This work, he hopes, will provide clues about how to prevent HIV from moving to a CD4+ T cell and infecting it. “The driving force for the work is [to understand] how HIV [is] delivered to the synapse,” Subramaniam says, adding that cell-bound HIV transfer through a virological synapse is over a thousand times more effective than transfer of cell-free virus. “We want to know why.”

“I think that Subramaniam does great work,” says Tom Hope, a professor of cell and molecular biology at Northwestern University, who also visualizes HIV. “The ion-abrasion work will bring new insights into the 3D structure of cells and how HIV takes advantage of these structures during viral replication.”

Subramaniam’s frequent moves between different fields and techniques make him quite unique, says Henderson. “He is very flexible and very willing to move into new areas and try new things,” Henderson says. “He is fearless about using new techniques.” But for now, Subramaniam plans to stay where he is. “There is a lot to be done,” says Subramaniam, who heads a lab of about a dozen researchers, over half of whom are working on HIV. “I am just beginning to get into HIV. We have barely scratched the surface.”

He works in an office where the door is never closed, right next to his wife Jacqueline Milne, whom he met at Johns Hopkins University (see photo, below). Milne, an associate scientist at NCI, is a cell biologist by training, but now also uses electron microscopic methods to study biological structures. The two don’t have a TV at home and “talk about science as much as we can,” she says. Recently, when Richard Henderson came from Cambridge to visit, they talked about science until the wee hours of the morning. Milne says she went to bed at 2:30 a.m., but her husband and Henderson “kept on cutting fruit cake and drinking coffee until 4:30 in the morning.”

The fact that many conversations revolve around science even appears to affect the words they use in everyday situations. “I am diffusing around,” Subramaniam said during a recent visit, referring to the fact that he can rarely be found at his desk. “He is sort of osmosing,” Milne said.

And with two daughters, age two and 10, they are truly busy. “It’s been busy since our younger daughter was born,” Milne says. “We are still sort of trying to get highly efficient.” One thing that allows Subramaniam to be more efficient is the fact that he can remotely control the electron microscope he uses to analyze HIV particles from anywhere in the world. “[When I] was in India to see my mother,” he recalls, “I had this wireless card. I was doing my email and logging onto the microscope—from a taxi in India!”

1. J. Mol. Biol. 213, 899, 1990
2. EMBO J. 12, 1, 1993
3. Science 266, 1369, 1994
4. Nature 406, 653, 2000
5. Nature 455, 109, 2008
6. J. Struct. Biol. 155, 63, 2006
7. PLoS Pathog. 5, e1000591, 2009