UC Berkeley

G protein-coupled receptors (GPCRs) are an extremely important class of membrane receptors that convert extracellular stimuli into intracellular signals through interaction with G proteins. These receptors are intimately involved in most physiological processes and for this reason are the largest drug targets in biology. In the nervous system a wide range of GPCRs function in nearly all subcellular locations, including the synapse, where they modulate cellular excitability, neurotransmission, synaptic plasticity, and behavior. An ultimate goal for the understanding of GPCR neurophysiology is to reconstruct a functional map of when and where GPCRs are activated and how this activation affects larger scale network outputs like behavior. However, due to the limitations of classical techniques, such as pharmacology and transgenic approaches, it has been difficult to decipher the role of individual GPCRs in specific cell types with temporal precision.

In order to gain a foothold toward understanding the contribution of different GPCRs in physiological functions, I developed a means of optically controlling individual GPCRs. This work focused on the metabotropic glutamate receptors (mGluRs) which are crucial neuronal GPCRs that respond to the major excitatory neurotransmitter, glutamate. There are eight different subtypes of mGluRs that have unique, but overlapping expression profiles and distinct G protein coupling and regulatory properties. Since pharmacological agents often can't distinguish between subtypes, can only be poorly targeted spatially, and are slow to apply and remove, we used a chemical optogenetic approach to individually agonize or antagonize mGluRs with light. This approach was based on a previous body of work from the Isacoff lab and is the basis of chapters 2 and 3. I show the molecular engineering and characterization of these tools, their initial characterization as tools for optical control of neuronal activatity, and validate their in vivo function in zebrafish and mice. Light-activated mGluRs, or "LimGluRs", are both a useful tool that are applicable in many contexts, but also an important test case that should serve as a model for the development of optical control over other GPCRs.

Structurally, GPCRs share a common 7 transmembrane domain structure but show divergence in extracellular N-terminal domains and intracellular C-terminal domains. Recent breakthroughs in X-ray crystallography have led to a greater structural understanding of how GPCRs bind ligands, activate, interact with G proteins, and oligomerize. However, functional experiments, which are required to gain a more complete understanding, have been hampered by the lack of high resolution techniques to probe these same processes. mGluRs are particulary interesting in this biophysical context because of their dimeric arrangement and large extracellular ligand binding domains (LBDs) that indirectly couple glutamate binding to G protein activation. In chapters 4 and 5, I use a combination of optical techniques including LimGluRs to measure oligomerization, structural dynamics, and function of mGluRs. While Fӧrster resonance energy transfer (FRET) is an established technique for probing of protein structure, it has recently been greatly enhanced in power by its application at the single molecule level. Single molecule FRET (smFRET) allows for individual receptors to be measured which allows for the observation of distinct states and their transitions, without the obscuring effects of averaging. Using intersubunit smFRET experiments on the extracellular ligand binding domains of mGluRs I found that these receptors visit three distinct conformations that have dynamics which determine receptor activation properties. Using another single molecule fluorescence technique based on counting fluorophore bleaching steps, I was able to show that mGluRs form homo- and heterodimers in living cell membranes. Furthermore, using a variety of perturbations I found that mGluRs have a covalent and non-covalent dimer interface within their LBDs that is complemented by a weak interface at the trans-membrane domains. Using LimGluRs, which allow for individual subunits within a dimer to be liganded, I demonstrate cooperativity that is dependent on receptor subtype and dynamics. All of these studies have given insight into the molecular biophysics of mGluRs and should serve as the basis for future studies on other GPCRs.

A final goal of the study of GPCRs is to understand the responses of their downstream signaling targets, including ion channels. One ion channel subfamily of particular interest for this objective are the K2P potassium channels which classically function as leak channels to maintain cellular resting potential. K2P channels have been shown to be highly regulated by different extracellular and intracellular signals, including GPCR activation. Despite the extreme regulation of these channels by a variety of signals, it has been difficult to determine which forms of regulation happen endogenously on individual channel subtypes. This is especially complicated by the fact that there is very limited pharmacology to individually block K2P channels. To overcome this, I developed, using a similar design to LimGluRs, a means of optically blocking the K2P channel, TREK1. Using a conditional expression system, native TREK1 channels were manipulated in hippocampal neurons where it was shown that they are non-canonical targets of GABAB receptor activation. This work is reported in Chapter 6 and provides the basis for future understanding of the physiological contribution of individual K2P channels. Finally, in Chapter 7 using a variety of approaches I show that TREK channels are specifically regulated by the enzyme phospholipase D through direct interaction. Using the photoswitchable conditional subunit system, this regulation was shown to occur in hippocampal neurons where it may explain some of the long term effects of alcohol exposure.