Michael Roukes, Caltech's Frank J. Roshek Professor of Physics, Applied Physics, and Bioengineering, has been awarded the National Institute of Health's (NIH) Director's Transformative Research Award.
This award, granted to only six scientists nationwide this year, "supports individuals or teams proposing transformative projects that are inherently risky and untested but have the potential to create or overturn fundamental paradigms," according to the NIH.
The Transformative Research Award, established in 2009, is one of four awards in the NIH High-Risk, High-Reward Research program. Roukes's five-year award will enable him to continue to develop novel nanotechnologies that can provide high-throughput, single-molecule analysis of the proteome—the population of all proteins within an organism.
We sat down with Roukes recently to discuss his research program, an especially intriguing one given that Roukes is a physicist.
What led you to apply for the Transformative Research Award?
Most grants are reviewed by teams, and there's a tendency to fund more tried-and-true approaches. The NIH created the high-risk, high-reward grants to fund "moonshot" projects, such as those that build the tools necessary to get the data that make it possible for us to see things that have eluded us so far.
What is the nature of your project?
It's another phase of the quest I've been on for the last 20 years: to enable deep profiling of the proteome. Today, we have perfected the necessary equipment to sequence the entire human genome, our DNA, but efforts to survey the proteome, the collection of all proteins in a biological system, lags far behind.
This is partly because of the massive number of proteins we're talking about. A mammalian cell has about 3 billion proteins. Of those 3 billion proteins, there are about 15,000 basic types. One type may be expressed in a hundred million copies, while another protein may be expressed in only a few copies. You have this incredible range of concentrations that must be spanned to understand everything that's going on in the cell's machinery—that is, in the proteome.
You might think that if there are only a few copy numbers of a protein, it may not be doing anything of significance in the cell. But this is completely incorrect. These rare proteins can signal the presence of foreign matter threatening the cell's viability, or control cellular processes to ensure homeostasis to keep the cell functioning. So, it's very important to see the full complement of proteins present, irrespective of their copy numbers, not just the most prevalent ones.
Essentially, it's as if we have a huge haystack that's mostly hay but spiked with a few needles, and these rare needles—proteins—are what we need to find.
How do you propose to find the "needles," the less common proteins?
What we need to do is to count every piece of hay in that stack, all 3 billion, to ensure that we don't lose the rare and important proteins. You can use separation methods, for example, involving beads that stick to some of the most prevalent proteins and then remove them, thus filtering the sample so that the rare proteins are a larger fraction of what's left. But all proteins are sticky, and the rare species end up being removed along with the more prevalent proteins. Really, in proteomics there are no reliable amplification techniques such as those utilized in genomics, so the surest way to enable deep profiling of the proteome it to just measure everything.
Have the more common proteins been well analyzed already?
Yes, absolutely. But there's still a lot of what we might call "biological dark matter" out there, if you will, which has evaded our discovery, so today we really don't know all of the components of the proteome and how they provide biological function. The task at hand for us is to use nanotechnology, which is well matched to the size scale of individual molecules, and massively parallelize detection using arrays of these nano-sensors so that many, many proteins can be analyzed rapidly and simultaneously.
We've calculated that if we can identify several million proteins per second, then we can analyze the billions of proteins in the proteome of a biological sample in a reasonable amount of laboratory time, in tens of minutes. The problem is that several million identifications per second far outstrips anything that's possible today. The predominant technique for identifying proteins is mass spectrometry. A mass spectrometer is basically a single-channel instrument. Each instrument monitors one processing path at a time, and each instrument is hugely expensive—typically more than a million dollars. The challenge is to put together thousands of miniaturized mass spectrometers working in parallel at the nanoscale. That's what we're working on!
Is there anything in particular you are hoping to learn about the proteome?
Yes, the award is to enable another step forward with this new ultrasensitive, single-molecule-level technology. And that is to not just survey all the proteins that are in a cell but to actually show where they reside within the cell. We call this spatial proteomics. Our aim in this new project is to scan with subcellular resolution to find where the proteins are located in the cell.
This has many applications in the life sciences as well as in medicine. For example, a tumor is usually a very heterogeneous thing. There are many different cell types in a tumor, and just a few are the ones that drive the disease forward, affecting the tumor's ability to mutate and metastasize and create body-wide disease. If we can resolve biological processes at the single-cell level, including the cell's manufacture and distribution of internal proteins, we can focus our attention on the most problematic cell types.
You trained as a physicist. How did you get interested in biology?
I consider myself an experimental physicist. I have always been involved in precision and quantum measurement. When I came to Caltech in 1992, I wanted to find a way to use physics to help people, so I thought I'd try to apply some of the emerging techniques of nanoscience and precision measurement to open new frontiers of measurements in biology. I started with just a couple of SURF [Summer Undergraduate Research Fellowship] students doing a small side project. This was great fun, and by the early 2000s, I was already carrying out funded collaborations with folks in Caltech's Division of Biology [now the Division of Biology and Biological Engineering].
I remember one day in particular, sitting in one of my main collaborator's conference rooms in the basement of Beckman Institute with a cross-disciplinary group that included a theoretical physicist, a surface chemist, a cell biologist, a developmental biologist, and a mathematician. I thought then, this is exactly where I want to be for the rest of my career. We all bring different things to the table, and this allows us to address questions that none of us could successfully address on our own. It's really a wonderful place to be!
Roukes' co-investigators in this effort are Alexander Makarov of Thermo Fisher Scientific, Julia Laskin of Purdue University, Kenneth Shepard of Columbia University, and Amir Safavi-Naeini (MS/PhD '13) of Stanford University; and, at Caltech, Tsui-Fen Chou (research professor of biology and biological engineering), John Sader (research professor of aerospace and applied physics), and Jeff Jones (senior scientist, proteomics).