Cell Propulsion Lab
Executive Summary
Living cells are extremely sophisticated sensor-actuator devices that detect specific environmental cues, process this information, and generate specific mechanical responses such as growth, shape change, or directed movement. These processes are controlled by networks of signal transduction and cytoskeletal proteins that form a dynamic, self-organizing system. Our long-term goal is to understand the fundamental engineering principles underlying these nanoscale molecular control systems. If we understand these principles, we may be able to alter these systems to "reprogram" cells such that they carry out new therapeutic functions. Alternatively, we may be able to build artificial, biologically-inspired cell-like, nanoscale assemblies that can perform targeted therapeutic functions. In the long-term, the ability to precisely program cells or synthesize molecular assemblies with cell-like behaviors would have broad, revolutionary, therapeutic potential. As our primary testbed, our center is focusing on understanding how cells achieve signal-guided movement (chemotaxis), with the goal of learning how to reprogram or build cells that can function as smart search and delivery vehicles.
Introduction
Fantastic Voyage
One of the most popular and captivating visions of nanomedicine is that epitomized by the 1966 film "Fantastic Voyage," in which nanorobots could be injected into our bodies to hunt down pathogens and tumors, or restore damaged organs and nerves. Of course, this was science fiction, but is this really the future? Maybe, but probably not as envisioned in this film. For example, it is unrealistic to build a minature robot operating at the nanoscale that exactly resembles the macroscopic mechanical machines that we encounter everyday (Figure 1a). This is because the behavior of materials at these dimensions differs significantly from behavior at larger scales -- molecular structures operate in unique and interesting ways. Nonetheless, something akin to therapeutic microscale robots does already exist — our own cells are complex mechanical and sensing devices that can carry out highly complex tasks (Figure 1b). For example, our immune cells can detect infection, move to these sites, and respond by destroying the infectious agent either by phagocytosis or by secreting countermeasures such as antibodies. Cells can form new repair structures, such as clots and bone. These are only a few of the many tasks that our cells, which are about 10-40 microns in diameter, are genetically programmed to carry out.
-
- Figure. 1. Visions of nanomedicine. (a) artist rendition of an imaginary nanorobot injecting a payload into a red blood cell. (b) Natural cells, such as this neutrophil sending out protrusions to attack invading bacteria, are capable of functioning as therapeutic agents, except that we do not know how to reprogram them to carry out new or modified functions.
Employing cells for targeted therapeutic purposes, however, faces several fundamental barriers. Real robots are useful because we can program them to carry out specific tasks that we desire. We can also program them with highly specific sets of instructions, so that this task is carried out only when, where, and how we specify. What we lack in the case of cellular "robots" is this ability to reprogram and precisely control the behavior of cells. Thus, we cannot modify cells to carry out novel tasks that they normally do not perform. Natural cells cannot always function as we wish — our immune cells can be tricked by evasive tumors or pathogens; certain nerve cells fail to regenerate — and it is these missing functions that are therapeutically most critical. Our inability to program cells also prevents us from specifying that a cell only carryout its function under a very precise time and set of conditions. The execution of a task at the wrong time and place might do no good and could be harmful. An essential element of a robot's utility lies in our ability to precisely control the robot. Thus, the overall goal of our center is to learn how cellular sensor-actuator systems are designed, and how they could be reprogrammed.
Directed cell movement as a testbed
As a testbed for understanding how to reprogram cells we are focusing on directed cell motility. Many cells can detect specific extracellular signals, process this information, and use the information to activate complex mechanical programs, such as directional cell movement or morphing into a new shape. Such programs are critical in immune cell movement and phagocytosis; they are also used in developing cells like neurons as they send out axons over great distances towards specific targets.
Pathway to Medicine: Programmed search and delivery cells
If we could reprogram cells to detect and move to novel signals, we could generate powerful search and delivery "robots". It would be extremely useful to have cells that could search the body for an elusive target, such as a microscopic tumor. Engineered smart cells could in principle be designed to detect and integrate a host of different information to pinpoint the tumor, including detection of generic environmental conditions (such as low oxygen, which is typical in tumors) and detection of tumor specific molecules. These could be integrated and linked to the cytoskeletal movement machinery to allow for efficient search of the tumor. Once at site, the cells could be used to deliver a number of payloads, ranging from dyes for imaging of the tumor, to chemotherapeutics that would kill the tumor.
How can we program cellular robots?
Macroscopic robots are controlled by electronic and mechanical subdevices that are linked by wires (Fig. 2a). What controls cellular behavior is, at the surface, very different. Cells contain a complex, self-organizing network of regulatory and mechanical molecules. Thus, to reprogram cells, we need to understand how this intracellular nanoscale machinery works, how it is wired, and how it can be manipulated. The overriding goal of our center is to learn how to program cellular systems, or more appropriately, cell-like systems, so that we can build novel and precisely controlled therapeutic agents. Ultimately, to fulfill this vision, we need to learn how to build and program cells or cell-like assemblies in which we can flexibly and precisely tune what they sense and how they respond.
While understanding how to program cells is a daunting task, genomic research over the last decade has revealed that biological control circuits are surprisingly modular — the same molecular toolkits are used repeatedly to generate diverse biological functions. This suggests that there is a fundamental logic to how complex biological tasks are programmed, and that we have the potential to understand and exploit this programming logic. For a behavior like directed cell motility, environmental cues are sensed by receptor molecules, this information is processed by signaling networks — in this case primarily by kinase/phosphatase and GTPase networks — which, in turn, engage and activate the actin cytoskeleton, which functions as a mechanical actuation system that ultimately moves the cell. Thus these cellular systems have functional modules that are analogous to those found in macroscopic robots (Fig. 2b), although these cellular control modules are composed of nanoscale molecules. Our goal is to understand how these molecular modules work and how they are linked together. Then, perhaps we can link them in novel ways to build cells that show novel, targeted behavior. Interestingly, this kind of relinking of components resembles a major mechanism in the natural evolution of new cellular responses.
-
- Figure 2. Both robots (a) and cells (b) are built from components and subsystems with distinct modular functions (e.g sensing, processing, mechanical output), although the composition nature of these components is very different. Different robots can be built by linking these components in different ways. Similarly it may be possible to build cells with novel behaviors by linking cellular subsystems in novel ways (as appears to happen in evolution).
NDC Goals
We are pursuing the following specific aims:
Understand the design principles underlying the natural molecular sensor-actuator systems used to generate directed cell motility
We are studying fundamental mechanisms of the signaling and actin cytoskeletal systems that are used to build circuits for cell movement and shape changes. We are interested in understanding how subtle gradients of signals are interpreted by the cell to yield directed, asymmetric activation of the actin cytoskeleton.
Determine ways to relink and reprogram cellular signaling subsystem components
Cell signaling systems are built from receptors, kinases, phosphatases, GTPases, and other regulatory components. How are these components wired in specific ways, and how can we rewire them? A growing body of evidence suggests that many of these components are functionally linked via specific protein interaction modules. These modules can organize multiple proteins into physical complexes and can also gate the activity of signaling enzymes. Thus, we are trying to use a toolkit of protein interaction domains to link signaling components in new ways to generate new cellular behaviors. We are also developing engineered protein-protein interaction modules that are optimized for cellular engineering (i.e. have minimal cross-talk with endogenous molecules).
Determine ways to regulate force generation systems
Ultimately to bring about movement, the signaling systems must activate the actin cytoskeleton in a spatially precise manner, and in a way that yields force and protrusion of the membrane. We are investigating how regulated polymerization systems can be used to perform spatially directed work in a cell-like environment. We are investigating the use of non-actin polymers to perform work in a cell. We are also investigating whether hybrid materials (synthetic nanomaterials linked to biological regulatory components) could be used to generate nanostructures that assemble in a regulated manner to perform mechanical work at the nanometer to micron scale.
Apply this understanding to build cells or cell-like particles with novel motility control that can be used for targeting within an animal
We are taking both a top-down and bottom up approach towards building minimal motile particles that can be used as search and delivery vehicles. In the top-down approach, we are trying to reprogram motile cells (neutrophils) so that they move towards novel signals. We are also trying to generate cytoplasts (cell fragments lacking nuclei) that can move towards specified targets. In the bottom-up approach, we are trying to reconstitute minimal cell-like molecular assemblies that can show regulated movement or shape change, from either biological molecules or from nanomaterial components that are regulated in a biologically-inspired manner. These cells, cell-fragments, or assemblies, could potentially be used as search-and-delivery vehicles in the body. We will test the particles engineered using both approaches for targeting functions within a mouse.
Accomplishments
To date we have made significant progress in characterizing the design principles of natural motility control systems, and have made significant first steps in learning how to manipulate the relevant signaling and force generating subsystems.Characterization of natural cell motility systems
- Wave-like protrusion at the leading edge. We have recently shown that the signaling and actin nucleating molecules operating at the leading edge of a motile neutrophil are organized into a wave generating circuit module. Thus, protrusion of the membrane at the leading edge appears to occur from a series of point-like wave sources. This wave-based mechanism appears to allow motile neutrophils to rapidly detect physical barriers and to move around them, even if it means moving slightly against the chemotactic gradient allowing netrophils to effectively negotiate the complex physical barriers in the body. [Weiner, et al, 2007]
- Mechanical properties of the actin network. Movement and shape change is driven by the actin cytoskeleton, but many of the basic mechanical properties of the actin network are poorly characterized. Our center has made great progress in using nanodevices to probe and characterizes the physical properties of actin networks both in vitro and in cells, revealing unique stress behaviors that resist compression. These properties may be important for the ability of the actin network to generate effective protrusion and shape change. [ Shaevitz & Fletcher, 2007; Chaudhuri et al, 2007]
Reprogramming cell signaling networks controlling actin motility
- Engineering new GTPase switches and using them to alter control of cell morphology responses. Rho family GTPases (Rho, Rac, and Cdc42) are master regulators of the actin cytoskeleton. Activation of Rac generates lamellipodial protrusions, Cdc42 generates filopodial (actin microspike) protrusions, while Rho generates actomyosin contraction (observed at the trailing edge of a motile cell). These GTPases are in turn controlled by upstream activators, known as guanine nucleotide exchange factors (GEFs), which catalyze exchange of bound GDP for GTP (the GTP-bound state is the active form). Most morphological responses are mediated by a uniquely regulated set of GEF's. Thus, we are working on understanding how diverse GEF regulation is achieved, and how novel GEF regulation could be engineered. By learning to control GEF's in novel ways, we could control the output of the actin cytoskeleton. In many cases, GEF's are allosterically regulated, and this mechanism involves autoinhibition of a core catalytic GEF domain by regulatory domains. We have shown that by engineering new regulatory and catalytic domain chimeras, we can generate GEF proteins that are controlled by novel inputs. Using these types of engineered signaling molecules and putting them into cells, we have shown that we can build new morphological control circuits. For example, we can engineer a cell, which normally shows no response to a small molecule signal, to now detect that molecule and respond by inducing actin protrusions. This represents one of the first demonstrations that it is possible to reprogram simple morphological behaviors in cells. [Yeh, et al, 2007]
- Redesigning the interface of signaling proteins to build ideal components for cellular engineering. For cellular engineering, we ideally want to use protein components that interact with the endogeneous components of the cell exactly as we wish. Thus, in many cases, if we are using endogenous proteins for engineering, we want to remove unwanted crossreactions. We are developing computationally guided methods for redesigning protein interfaces so that a protein that binds multiple partners can bind only to specific, desired partners but will not bind to undesired partners. This type of approach is useful to redesign GEF-GTPase interfaces, thus generating new GEF-GTPase pairs that do not crossreact with competing endogenous partners. These types of modified components should be ideal for cellular engineering. [Humphries & Kortemme, 2007]
Reconstitution/Engineering of regulated force generation systems in vitro
- Reconstitution of controlled actin polymerization in vesicles. We have had success reconstituting actin control systems within giant unimellar vesicles and demonstrated that specific phosphoinositide signals can be used to spatially control polymerization in these reconstituted systems. [Liu & Fletcher, 2006]
- Characterization of novel actin-like polymer systems. Although bacteria lack actin, they do have actin-homologs. This is a useful characterizing the ParM bacterial actin homolog which is used to drive plasmid segregation to separate sides of a dividing bacteria. ParM can polymerize to generate force and plasmid segregation, but only does so when nucleated by the ParR protein bound to the ParO operator on the plasmid. Mullins has been able to reconstitute this system in vitro from purified components and can use it to drive movement of latex beads. These studies and comparison of this system to the actin cytoskeleton, help to reveal the basic elements required for a polymerization-based force generating system. We would like to use ParM as an alternative force generating system in engineered cells. [Garner et al, 2007]
References
Background
- Bio FAB Group , Baker D, Church G, Collins J, Endy D, Jacobson J, Keasling J, Modrich P, Smolke C, Weiss R. "Engineering life: building a fab for biology. Sci Am. 2006 294:44-51.
- Brian J Yeh and Wendell A Lim, "Synthetic biology: lessons from the history of synthetic organic chemistry, " Nature Chemical Biology 3, 521-525 (2007).
- Dueber JE, Yeh BJ, Bhattacharyya RP, Lim WA. "Rewiring cell signaling: the logic and plasticity of eukaryotic protein circuitry." Curr Opin Struct Biol. 2004,14(6):690-9.
- Fletcher DA, Theriot JA. "An introduction to cell motility for the physical scientist," Phys Biol. 2004 Jun;1(1-2):T1-10.
- Welch MD, Mullins RD. "Cellular control of actin nucleation." Annu Rev Cell Dev Biol. 2002;18:247-88.
Current NDC papers
- Yeh BJ, Rutigliano RJ, Deb A, Bar-Sagi D, Lim WA. "Rewiring cellular morphology pathways with synthetic guanine nucleotide exchange factors." Nature. 2007;447:596-600.
- Garner EC, Campbell CS, Weibel DB, Mullins RD. "Reconstitution of DNA segregation driven by assembly of a prokaryotic actin homolog." Science. 2007, 315:1270-4.
- Shaevitz JW, Fletcher DA. "Load fluctuations drive actin network growth." Proc Natl Acad Sci U S A. 2007 Oct 2;104(40):15688-92.
- Chaudhuri O, Parekh SH, Fletcher DA. "Reversible stress softening of actin networks." Nature. 2007 Jan 18;445(7125):295-8.
- Liu AP, Fletcher DA. "Actin polymerization serves as a membrane domain switch in model lipid bilayers." Biophys J. 2006 Dec 1;91(11):4064-70.
- Weiner OD, Marganski WA, Wu LF, Altschuler SJ, Kirschner MW. "An Actin-Based Wave Generator Organizes Cell Motility." PLoS Biol. 2007 Aug 14;5(9):e221
- Humphris EL, Kortemme T. "Design of multi-specificity in protein interfaces." PLoS Comput Biol. 2007 Aug 24;3(8):e164.
Contact Information
Key Investigators
- Wendell Lim (PI), University of California San Francisco (lim@cmp.ucsf.edu)
- R. Dyche Mullins (Co-PI), University of California San Francisco (dyche@mullinslab.ucsf.edu)
- Dan Fletcher(Co-PI), University of California Berkeley (fletch@berkeley.edu)
