ASCB-Gibco Emerging Leader Prize Essays

Clifford Brangwynne

Clifford Brangwynne
Princeton University

How does the cell organize its contents to achieve biological function? My answer to this question is strongly impacted by decisions I made as an impressionable 18-year-old. When I enrolled at Carnegie Mellon University (CMU) in 1996, I was (like most freshmen) all over the place. I almost majored in Biology, but after taking an introductory lab class where we crystallized lava-hot molten copper alloys, I was sold on Materials Science. I was still fascinated with Cell Biology (attending my first ASCB meeting in 1997), and started working in a biology lab, where I began thinking of the cell as a special kind of material. A number of great early mentors supported my anarchical interdisciplinarity, including Fred Lanni and Vivek Abraham at CMU and Kit Parker and Don Ingber at Harvard. I went on to do a PhD with Dave Weitz, looking at cells through the lens of a field called Soft Matter Physics, which deals with squishy materials like gels, polymeric solutions, and oil/water mixtures, i.e., the materials most akin to living cells. But what does this have to do with intracellular organization?

For most people, imagining the interior of the cell brings up the tidy picture we all learned in high school: vesicle-like organelles such as the endoplasmic reticulum, Golgi, and nucleus that float in a homogenous cytoplasm. But it has become increasingly clear that the cytoplasm is much more interesting and complex than that. Indeed, most intracellular compartments are not vesicle-like, but instead have no enclosing membranes. These structures nevertheless still concentrate specific molecules, often both RNA and protein, into distinct subcellular compartments. Examples include processing bodies, neuronal granules, and germ (P) granules in the cytoplasm and Cajal bodies, nucleoli, and PML bodies in the nucleus. Despite the importance of these membrane-less compartments for a wide array of biological processes, a mechanistic understanding of their assembly— i.e., the rules governing this whole other dimension of intracellular organization—has been lacking.

Over the last several years, my colleagues and I have shown that these structures represent condensed liquid phases of intracellular matter. We had a key breakthrough when I was a postdoc in Tony Hyman’s lab in Dresden, when I showed that P granules exhibit liquid-like behaviors—flowing, coalescing, dripping, and wetting. These liquid-like properties suggested that the dynamic assembly of P granules could represent a type of phase transition. From my undergrad and graduate work, I’d learned the formal mathematical framework of phase transitions, but they manifest in ways quite familiar from everyday experiences, e.g., dew drops condensing on blades of grass, or water freezing into ice. Phase transitions are also well-known to structural biologists, who work with solutions of purified proteins, which can crystallize and also sometimes form droplets. I suspected that P granules, and other membrane-less bodies, represented liquid phase RNA/protein droplets within living cells.

My lab at Princeton has been focused on testing this hypothesis, and trying to figure out the underlying molecular driving forces, as well as what governs droplet size and biophysical properties and the impact of these factors on biological function and dysfunction. We are particularly interested in the nucleolus, a nuclear body that forms around actively transcribing ribosomal RNA genes. We’ve shown that the nucleolus assembles by a phase transition, which is directly linked to the size of the cell. Our work has also highlighted the role of intrinsically disordered protein sequences, which are important in promoting the phase transitions underlying P granules, nucleoli, and an increasing number of other cytoplasmic and nucleoplasmic droplets. These proteins also contribute to droplet “metastability,” i.e., their ability to slowly transition from more liquid-like to more solid-like structures, which could underlie neurodegenerative protein aggregation diseases.

The nucleolus directly interacts with the genome, and thus its assembly and biophysical properties can strongly impact transcription and RNA processing rates. Moreover, there are numerous other nuclear bodies, and the nucleolus is probably just one of many phase-separated droplets assembling throughout the nucleus. Over the next several years, I plan to continue pursuing my interest in phase transitions and the physics of living matter, while digging deeper into the link between RNA/protein droplets and gene regulation. Like regulatory dewdrops on a grassy field of DNA and RNA, these droplets are likely key players in the dynamic flow of biological information. Twenty years ago, I had only a vague notion of the relation between materials physics and biology. But through the alchemy of luck, hard work, and a driving curiosity, I’ve had the opportunity to help define this new field of intracellular phase transitions. Living matter has indeed turned out to be much cooler than any man-made alloys!

Dmitri Kudryashov

Dmitri Kudryashov
Ohio State University

Born in the Soviet Union, educated in Moscow, I started my professional career as a postdoctoral fellow in Emil Reisler’s Laboratory at the University of California, Los Angeles, studying the biochemistry of the actin cytoskeleton. I will always remain grateful to Emil for his wise and gentle mentorship. Because I am open to opportunities and fascinated by the beauty and complexity of life in all its aspects, my research interests have drifted and encompassed studying the hate–love relationships (mostly their “hate” part) between humans and microbes.

Destined to share the planet with microbes, we developed highly sophisticated relationships with their world. Co-evolution of human and microbial species is embedded in our flesh and blood as the immune system. Many of our tiny planet-mates are ultimate killers and the only reason we survive their attacks is because our countless distant relatives died learning how to protect themselves while the remaining learned to recognize the most conserved and essential elements of microorganisms in order to target them with our own efficient killers—immune cells and effector molecules. Among other things this implies that we have to take good care of our planet as there will be no other inhabitable one with microbes as gentle to us as our own (although it might not seem so when we are sick).

Although due to natural limitations we divide ourselves into physicists, chemists, microbiologists, and cell biologists, it is useful to remember that these professional “comfort zones” exist only in our heads. On several occasions, this awareness was a driving power behind our discoveries. Thus application of basic protein folding principles has facilitated understanding mechanisms of bacterial toxin neutralization by human immune peptides, while the realization that bacterial toxins are destined to be either highly potent or largely useless led to a deeper comprehension of their abilities to hijack the actin cytoskeleton.

As an essential component of both the innate and adaptive immunities, actin is a common target for bacterial proteinaceous toxins.
Most toxins amplify their efficiency by acting on either signaling cascades or essential low-abundance elements (e.g., ribosomes). In contrast, toxins targeting monomeric actin—the most abundant cytoplasmic protein—seem to be doomed to inefficiency. We found that ACD family toxins produced by pathogenic and non-pathogenic Vibrio and Aeromonas species overcome this barrier by generating a novel toxicity amplification cascade. ACD toxins covalently crosslink actin into oligomers, which bind with very high affinity to formins, adversely affecting both nucleation and elongation abilities of these proteins. This discovery was unexpected as it was masked under beliefs that the pathogenicity mechanism of ACD (i.e., sequestering of bulk amounts of actin in polymerization-incompetent oligomers) had been fully understood.

Given the remarkable killing potency of bacterial toxins, having an immune mechanism for their prompt inactivation is a matter of life or death. For over a decade, mechanisms of inactivation of various unrelated groups of bacterial toxins by human immune peptides called defensins remained enigmatic. We found that defensins take advantage of marginal thermodynamic stability—an essential feature of many toxins. Defensins cause local unfolding of such toxins, uncovering new regions for proteolysis and potentiating their precipitation through the exposure of hydrophobic interfaces.

In five years we will have deciphered molecular mechanisms driving inactivation of deadly bacterial toxins and viral proteins by defensins. We will reproduce the desired effects of defensins using more stable and more manageable small molecules and evaluate their therapeutic potentials. At the other end of the spectrum, we will learn to use bacterial toxins as spatially and temporally controlled tools of high precision to dissect the function of actin in subcellular compartments such as the nucleus and mitochondria and pave the way to utilizing other toxins for similar purposes. The boldest dream of my group is to create a therapeutic platform by converting bacterial toxins into safe and potent tools selectively targeting cancer cells and not their healthy counterparts. But this may take a little longer than five years.

Hari Shroff

Hari Shroff
National Institute of Biomedical Imaging and Engineering

I’m an engineer who uses optical microscopes to study living cells and embryos. My time divides more or less equally between improving the technology (e.g., more speed and resolution, less phototoxicity) and collaborating with cell and developmental biologists to use the new tools we develop in attempting to answer fundamental questions in biology. I’m especially interested in imaging neurodevelopment in the Caenorhabditis elegans embryo—an area that has been somewhat unexplored due to technical difficulties, especially with optical microscopy. We’ve been able to make headway into this issue using the imaging methods developed in my lab.

Four key events led me to my current research program: 1) I attended the Marine Biological Laboratory’s summer physiology course. There I learned how modern cell biology is done, and how valuable imaging is to this effort; 2) Although as a graduate student I was able to somewhat “self-educate” myself about microscopy, I credit much of my current knowledge to the postdoctoral training I received in Eric Betzig’s lab; 3) When I was interviewing for faculty jobs, I was lucky enough to meet a neurobiologist at Yale, Daniel Colón-Ramos, who impressed upon me the problems (and opportunities) in studying the developing worm embryo; and 4) I was lucky to land a tenure-track job in the intramural National Institutes of Health (NIH) program, where I’m given the freedom to pursue optical tool development.

In Betzig’s lab, I worked for two years on the development of photoactivated localization microscopy (PALM). Since starting my own lab at the NIH, I’ve focused on developing and improving methods that are well suited to imaging live, dynamic phenomena (PALM is still almost exclusively a fixed cell technique). I see particular promise in structured illumination microscopy (SIM) and light-sheet microscopy (LSM).

SIM was invented by Mats Gustafsson ~20 years ago, and is renowned for its resolving power and relative gentleness (light intensities similar to widefield microscopy, which are much lower than those of other super-resolution techniques). My former postdoc Andy York and I improved the depth penetrance of SIM ~10-fold over other implementations, enabling super-resolution imaging in moderately thick samples (up to ~100 um). We also developed an effectively instantaneous implementation of SIM that removes the need to acquire extra images or post-process the data; the raw images are already super-resolved and can be acquired at hundreds of frames per second. My lab is attempting to push the depth penetrance of instant SIM even further by combining it with adaptive optics.

A staff scientist in my lab, Yicong Wu, and I have collaborated with Applied Scientific Instrumentation to make LSM easily usable for samples that can be placed on a conventional glass coverslip, thus improving its accessibility to biologists. Equally significantly, we’ve developed an implementation of LSM that uses a second specimen view to improve axial resolution ~2-fold over a confocal microscope, without significantly compromising either speed or phototoxicity. This dual-view inverted selective plane illumination microscope (diSPIM) has been cloned in ~50 labs around the world.

I’m very excited about my current collaboration with Zhirong Bao (Sloan Kettering) and Cólon-Ramos to construct a 4D atlas of neurodevelopment in C. elegans. The diSPIM’s resolution and speed make it possible to capture cellular dynamics throughout embryogenesis without detectable phototoxicity. Our vision is that in 5–10 years anyone with a computer will be able to navigate the events we’ve cataloged, thereby watching, with subcellular resolution, the processes that orchestrate brain formation.