Nanomedicine Development Centers
Cell Propulsion LabCenter for Cell ControlCenter for Protein Folding MachineryNanomedicine Center for Nucleoprotein MachinesNanotechnology Center for Mechanics in Regenerative MedicineNational Center for Design of Biomimetic NanoconductorsOptical Control of Biological FunctionPhi29 DNA-Packaging Motor for Nanomedicine

NDC for Optical Control of Biological Function

Executive Summary

A major challenge for biomedicine is to develop new ways of determining how proteins operate in cells. Our center is developing methods to rapidly and remotely control proteins using light. We are focusing on ion channels, G-protein coupled receptors and enzymes--which together comprise about 7 percent of the gene products of the human genome and are implicated as causes for many diseases. We are synthesizing two classes of nanoscale photoswitches: 1) ligand photoswitches that attach to proteins and are engineered to remotely control their function with light; and 2) light-gated peptides, which bind to target proteins in a light-dependent manner. A critical challenge is to transfer our methods from basic cell biology to therapeutics. We will focus on animal models of human disease, including using light-gated channels to restore vision to retinae that have lost photoreceptor cells and prevention of cardiac damage from ischemia.


With the explosive advances in structural biology emerging in this post-genomic era, biomedical research is beginning to shift from simply identifying proteins and defining their individual properties to developing new ways of determining their structure, function, and coherent operation in cells. Recent developments in high resolution imaging of intracellular structures have led to extraordinary advances in understanding protein structure and function; however, little progress has been made in directly controlling what the proteins do. Proteins often do not work alone; rather, they function in multi-component molecular machines, and their activities are coordinated in intricate pathways that interact dynamically. To understand how these nanoscale systems work, and to use that information for medical treatments, we must develop the means to study nanoscale systems and structures in living cells. We also need to develop the ability to precisely control their function. An attractive approach is to use light to probe molecular machines in cells. Our center is developing methods for light-induced, rapid and sensitive remote control of proteins in living cells. We are beginning with non-invasive, quantitatively precise control of individual proteins, and will then pursue the remote control of signaling between pairs of proteins. This will bring us closer to one of the central challenges in biology: directly relating molecular events in particular cells to the function of an organ and the behavior of the organism. This holds promise for drug discovery and directly treating disease.

Which Proteins?

We are focusing on 3 classes of proteins: ion channels, G-protein coupled receptors, and enzymes (phosphatases and kinases). These represent huge protein families, and together they comprise about 7% of the human genome. Defects in these proteins or their function contribute to a wide variety of the most devastating diseases, including cancer, diabetes, neurological disorders, blindness, inflammation, and heart disease. Advances in understanding the function of these proteins are hampered by an inability to selectively and reversibly control their function in desired locations and for well-defined periods of time. We are developing novel optical switches for targeted remote control of these proteins both to study their basic biology and to manipulate them in experimental treatments of animal models of human disease. By enabling rapid, reversible, and highly specific switching, these tools should make it possible to overcome two major problems in standard drug delivery: side effects and rapid on-off action.

How We Using Light to Control Protein Function.

An important advance for improved pharmacology is the development of tethered ligands, in which the targeting of a small molecule to a protein of interest depends not only on how well the ligand fits into its binding site, but also on covalently attaching the ligand near the binding site. This "two-key" approach makes it possible to obtain high specificity, even when binding is weak, because the ligand is held by a tether very close to the binding site. We exploited and extended this idea using a synthetic photoisomerizable molecule azobenzene, which is a rigid molecule, about 17 Å long in its trans configuration, but shortens to only 10 Å in its bent, cis configuration. Exposure to long wavelength UV light triggers photoisomerization of azobenzene from the lower-energy, trans form, to its higher energy, cis form.

We use azobenzene in two general ways for controlling protein function. As shown in this Figure (see A), the light-sensitive length change is sufficient to reversibly extend or retract a tethered ligand, thereby allowing or preventing its binding to a regulatory site on the protein. We call such ligand photo-switches "optical nano-pointers". This allows photoactivation or inhibition of channels and receptors without compromising their normal activation mechanisms or native gates. Note that the ligand is bound to the end of the azobenzene which which is covalently tethered to specific cysteines on the protein of interest via a maleimide group (MAL). When exposed to UV light of 380 nm, azobenzene lengthens allowing the ligand to bind. Upon exposure to 500 nm, the cis configuration is restored, and the ligand is effectively pulled off the site.

The second approach involves using the photoisomerization of azobenzene to exert force on the protein with what we call "optical nano-tweezers" (See B in this Figure) in order to generate light-gated peptides. These are photo-switching MAL-AZO-MAL molecules that cross-link between cysteines that are placed at two locations within the protein. Photoisomerization of the azobenzene group changes the overall length of the nano-tweezers, pushing apart the connected regions of the protein when azobenzene lengthens (trans configuration), or pulling them together when the azobenzene shortens (cis configuration). When attached to peptides, nano-tweezers generally denature them when not exposed to UV because the trans state of azobenzene is more stable in the dark. Illumination with 380 nm will bend the nano-tweezers and allows the peptide to assume its native conformation and become active.

The applications that we propose are generalizable to many classes of proteins, providing astonishingly rapid, reversible, and spatially constrained remote control of protein function. These tools will revolutionize the in vivo analysis of cell function, including analysis of neural circuits and the relationship between the molecular events in specific cells and the behavior of an entire organism. The methods also open a new avenue for reengineering cells for targeted "smart-cell" therapy and for high throughput drug screening. Most enticing of all is the long-term goal of using light switches to control the function of individual proteins in specific cells therapeutically to circumvent problems of drug targeting by ensuring that the reagent acts only in the desired location pinpointed by light.

Figure 1
Figure 1.

NDC Goals

A critical challenge is to transfer our methods from basic research in cell biology to therapeutics. Besides the proof-of-concept and development of the tools and model systems, our NDC maintains four major areas of interest of medical relevance:

Cell biological studies of diseases processes in live animals

Combined capabilities for photonic imaging and control will allow increasingly sophisticated studies of disease mechanisms in live animal models. The ability to acquire imaging data, manipulate cells with micron-scale precision based on these data, and then to image the subsequent effects of manipulation will allow online, active intervention to become part of the disease researcher's toolkit and to move beyond passive imaging. For example, laser-based precise control of cellular properties in a high-speed, three-dimensional manner will eventually allow us to simulate (and thus repair) improper cellular dynamics.

Creation of novel therapeutic strategies

Microendoscopy is presently the only minimally invasive imaging modality with the ~1 micron resolution needed to resolve all cell types.Many therapeutic strategies of the future will involve aspects of optical control. This includes light-driven prosthetics for the visual system, light-activated gene therapies, and optically guided surgery for tissue ablation with cellular precision. The capability for minimally invasive delivery of light for imaging and control will be important for all these emerging therapeutic strategies.

Development of new drugs

It is common to test promising pharmaceutical compounds in rodent models of disease. The ability to perform minimally invasive imaging allows extended studies of the cellular effects of such compounds. A new generation of promising pharmaceuticals emerging from work at our NDC will be 'smart', optically activated drugs. Insertion of minimally invasive micro-optics into the body will allow both optical activation of these drugs in deep tissues and regular assessment of their performance at the cellular scale.

Defining Specific Medical Targets: Pathway to Medicine

We will focus on animal models of human disease, including using light-gated channels to restore vision to retinae that have lost photoreceptor cells, treatment of retinal disease due to angiogenesis, and prevention of cardiac damage due to ischemia.

Blinding diseases such as retinitis pigmentosa (RP) and age-related macular degeneration (AMD) result from the progressive degeneration of photoreceptors, namely, rods and cones. Retinal neurons that receive and process signals from the photoreceptors are preserved for years after the onset of blindness, giving hope that visual sensitivity might be restored by allowing the artificial input of information to these surviving cells. Prosthetic devices currently under development stimulate the retinal neurons with electrode arrays driven by images from a video camera. Although this produces some light perception in patients, spatial and temporal resolution are poor because of the relatively large size and imprecise positioning of electrodes in the arrays embedded in the intricate and complex neuronal bed. Furthermore, the non-biological materials used in these devices also present serious challenges of long-term biocompatibility. Using light-activated ion channels as an alternative therapeutic approach will circumvent these issues.


1) Light-gated proteins using nano-pointers

We developed a light-gated glutamate receptor (LiGluR), where light regulation is conferred by a nano-scale photo-switch consisting of a glutamate analog that is covalently tethered to the glutamate receptor through a photoisomerizable azobenzene moiety. We recently defined the photophysics of gating and chemistry of attachment (Gorostiza et al., 2007). Attachment of the synthetic photoswitch occurs by affinity labeling which depends on the isomeric state of the photoswitch. In other words, the glutamate analog is bound to azobenzene and, to form the tether, it needs to be attached to the glutamate receptor. Only the activating state produces the correct geometry to allow binding of the ligand to its receptor. In this state, the maleimide end is positioned for binding to a targeted cysteine completing formation of the tether. Thus, the distribution of photoswitch derivatization can be controlled with patterns of light, even in an otherwise uniform tissue where all cells express the iGluR6 channel containing the cysteine handle. We also learned that photoswitch conjugation is selective for the target channel, and that it does not affect native neurons, despite the fact that they contain several types of native GluR (Szobota et al., 2007). Additionally, optical stimulation can be structured in designed spatial and temporal patterns. Action potentials can be reliably triggered by 1-2 millisecond long pulses of light. We introduced these engineered light-gated LiGluR into sensory neurons in zebrafish larvae. Activation of LiGluR reversibly blocks the escape response to touch. Our studies show that LiGluR provides robust control over neuronal activity, enabling the dissection and manipulation of neural circuitry in vivo.

2) Non-tethered, photochromic agonists (nano-tweezers)

The ability to control the active concentration of neurotransmitters in a spatially and temporally precise manner with caged compounds has revolutionized the study of the central nervous system. In particular, caged glutamate has emerged as a tool for the dissection of both neural circuitry and the fast kinetic events of channel activation. The success of caged glutamate is based upon the irreversible photochemical cleavage of a protecting group and the subsequent release of the native neurotransmitter. Photochromic ligands provide another opportunity for the control of neural excitability. We have developed a non-tethered, photochromic agonist that effectively modulates channel activity upon photoisomerization (Volgraf et al., 2007). This photochromic agonist (GluAzo) consists of a glutamate analog and azobenzene moiety. We can therefore use light to remotely, reversibly activate native glutamate receptors and control neuron activity . The active agonist is highly subtype-specific, possesses good efficacy and affinity, and can be controlled rapidly with the wavelength of light. This success indicates that the design of photochromic agonists can be extended to other proteins and ligands.

3) Engineering Light Sensitive Neurons in the Retina

SPARK channels are genetically encoded light-activated potassium channels. Introduction of these channels into neurons of the retina should allow optical regulation of neuronal action potentials conferring the ability to artificially stimulate the retina using light (Banghart et al., 2004). This could represent a direct treatment for the loss of the photoreceptors (rods and cones) in blinding diseases such as retinitis pigmentosa and age-related macular degeneration.

Figure 2
Figure 2. Photocontrol of RGCs expressing light-activated SPARK channels. A) Loose-patch extracellular recording from a SPARK-expressing RGC. Illumination with 500 nm light promotes firing (green bars); 380 nm light inhibits firing (red bars). B) Quantification of photoswitching in RGCs.

SPARK channels have two components: a genetically-modified Shaker K+ channel and a covalently attached azobenzene. Recall that with no light, the azobenzene is in the trans or elongated conformation and the channel is blocked. Upon exposure to short-wavelength light (380 nm), azobenzene shortens unblocking the channel and allowing K+ current to flow. This hyperpolarizes neurons that contain the SPARK channel, thereby silencing the firing of action potentials. Exposure to long-wavelength light (e.g. 500 nm) returns the molecule to its trans state, depolarizing the neurons and promoting firing.

We expressed SPARK channels in Retinal Ganglion Cells (RGCs) using AAV2, which is injected into the vitreous of the eye (i.e., near the retina). Electrical activity from flat mounts of the retina are recorded with patch clamp to monitor individual channel activity. Figure 2A shows an example of photoswitching in an RGC that is expressing SPARK channels. The neuron fires action potentials at a high frequency when exposed to 500 nm light (green bars). Exposure to 380 nm light inhibits firing. Repeated exposure to the two wavelengths reliably promotes and inhibits firing, with no decrement (Fig. 2B). Thus, SPARK reliably imparts light sensitivity to RGCs.

We have also begun studies to target the engineered, light-gated SPARK channel to particular retinal cell types. Our next goal is to develop gene transfer vectors that can specifically deliver the appropriate excitatory and inhibitory photoswitches to specific subclasses of the output cells of the retina (ON and OFF-ganglion cells).

Finally, experiments using rats to determine whether photoswitching of RGCs in vivo can be perceived and used by the animals for performing a visual-behavioral task. Our behavioral assay is a simple forced-choice test, in which a rat that is placed a water-filled Y-maze must find a translucent submerged platform to escape from the water. The hidden platform is randomly placed in the left or right side of the maze, but it is consistently paired with a visual cue, consisting of an illuminated array of LEDs that emit either 380 nm or 500 nm light.

At first, rats make the correct choice about half of the time, consistent with chance. Then, the animals gradually learn to associate the correct light stimulus with the platform. Once animals learn the task, they quickly generalize this knowledge to other visual associations. Remarkably, subsequent discriminations between UV exposure vs. darkness or 500 nm are learned within 10 trials. The next step is to use this task to test whether blind rats expressing the SPARK channel or LiGluR can perceive RGC photoswitching and use this information alone to guide behavior.



Current NDC supported papers

Contact Information

Center Website

Key Investigators