Samuel Andrew Hires - US grants

Project Summary: This proposal outlines the development of a fundamentally new optogenetic technology capable of flexibly manipulating the activity of thousands of neurons contributing to the dynamic activity of distributed neural circuits with single neuron resolution. No method that currently exists even remotely meets the need of flexible, selective control of thousands of neurons distributed across large volumes of the brain. Filling this methodological gap is a central research objective of the BRAIN Initiative, because doing so will transform our ability to investigate how the nervous system encodes, processes, utilizes, stores, and retrieves information. The overall objective for this application is to acquire critical structural knowledge of photoactive states of a red-shifted channelrhodopsin and use these to engineer a photoselectable channel prototype that demonstrates the potential of our approach for future development in behaving animals. This would allow opsin-expressing neurons to be flexibly selected, activated, and deselected with light. By leveraging new structural knowledge, we anticipate that we can develop a fundamentally new approach to optogenetics that takes us beyond genetically targeted control and into an era of functionally targeted, flexible control of any neural ensemble. The aims of our research are to obtain the first atomic structures of red-shifted channelrhodopsin mutants in three channel states, engineer a three-state ReaChR mutant with high open conductance and optimized action spectra, and demonstrate reversible photoselective control of neurons in vivo with PReaChR prototypes. We anticipate that completion of these aims will yield the following expected outcomes. First, it will produce new knowledge of the underlying structural transformations between channelrhodopsin photostates that will enable efficient computational design of photoselectable optogenetic tools. Second, it will produce the first examples of photoselective channelrhodopsins useful for neural excitation. Third, it will assess the utility of these new opsins for flexible control of distributed sets of neurons. Collectively, these will provide a roadmap to extending the transformative new trait of photoselectabilty to a wide range of existing optogenetic tools for excitation, inhibition and modulation of neural activity. Further research in this direction should ultimately enable flexible control of spatially complex distributions of neurons in head-fixed and freely moving animals during behavior, a key to furthering our understanding of the intricate neural dynamics that underlie our thoughts, feeling, and actions and how circuit dynamics are disrupted by neurological disorders.

Project Abstract Fragile X Syndrome (FXS) is a leading inheritable cause of mental impairment. There is no known cure for FXS or treatment that reverses the collective pathology. There is a fundamental gap in our knowledge of how FXS causes mental impairments through alteration of neural circuitry. The long-term goal of this research is to develop an understanding of FXS that links learning impairments to specific changes in neural circuits. Characteristic symptoms of FXS include reduced intellectual abilities, learning deficits, and hypersensitivity to sensory stimuli. FXS arises from a loss-of-function in the FMR1 gene; mice lacking a functional fmr1 gene exhibit several phenotypes similar to FXS. Fmr1 mutant mice display an intriguing deficit on both the gap cross task, a freely-behaving whisker-dependent tactile learning task, as well as a head-fixed whisker-dependent tactile learning task. The central hypothesis is that impairments in tactile learning are driven by reduced dendritic spine stability and hypersensitive touch responses in primary somatosensory cortex resulting from attenuated activity of somatostatin-expressing (SOM) interneurons. Experiments in this proposal will determine the extent to which loss of the fmr1 gene disrupts spine stability, tactile learning, and circuit dynamics during task performance. Guided by our strong preliminary data, we will pursue this hypothesis in two related specific aims. In Aim 1, longitudinal two-photon in vivo imaging is combined with an automated head-fixed whisker- dependent tactile learning task to evaluate if reduced activity of SOM interneurons in fmr1 mutant mice decreases dendritic spine stability and impairs learning. In Aim 2, sophisticated electrophysiology is combined with high-speed tracking of whisker position during this same head-fixed object localization to quantify the extent to which tactile discrimination and cortical representations of afferent sensory activity in somatosensory cortex are abnormal fmr1 mutant mice and if attenuated function of SOM interneurons contributes to this deficit. This approach is particularly innovative because the synaptic changes that underlie learning are measured longitudinally throughout task acquisition. Furthermore, breaking from the anesthetized status quo, the cortical circuit dynamics that represent touch are quantified during active perceptual behavior. The proposal is significant because it vertically advances our knowledge of FXS mechanisms across levels of analysis, from synapse to circuit to behavior. Additionally, it opens new horizons for these advanced techniques to be applied to other cortical layers and brain regions to build a comprehensive understanding of neural circuit defects in a premier FXS model system. This proposal squarely meets the key mission objectives of the NINDS and NIMH to provide detailed and integrated knowledge of how the function of synapses and circuits is disrupted in neurological disorders. Ultimately, the resulting improved understanding of circuit dysfunction has the potential to lead to therapies that improve the quality of life for the roughly 1 in 5,000 people born with Fragile X Syndrome.

PROJECT SUMMARY Tactile perception is an active, exploratory process that emerges through a cycle of sensory input driven by purposeful motion. Consider a baseball pitcher, stroking the seams of the ball, reorienting and refining the grip to deliver a wicked curveball. In nearly every sensorimotor action, our body and mind has been trained to efficiently execute precision movements to extract relevant tactile cues for the task at hand. The ability to execute these movements, receive tactile sensation, and construct sensorimotor perception is impaired in millions of Americans following stroke or spinal injury. Understanding neural circuit mechanisms that underlie sensorimotor learning and integration could guide creation of more effective treatments for these and other impairments, and thus have a major positive impact on public health. The immediate goals of this proposal are to develop new technologies to link the activity in sensorimotor circuits to purposeful motion, tactile sensation, and perception, and to apply these tools to identify how movement and these circuits are refined to identify relevant tactile features during skilled actions. A successful outcome of this research would generate fundamental knowledge of the neural code for touch, for how sensory and motor signals are integrated at the level of neural circuits, and of general mechanisms of cortical circuit refinement across learning of skillful behavior. Furthermore, this work would produce a suite of new technologies for observing and manipulating neural circuit dynamics in real-time that can be adapted to other sensory modalities (e.g. vision, hearing) and animal models of nearly any neuropsychiatric disorder. The overall scientific question we will address: how does the sensorimotor system learn to encode and identify task-relevant tactile features and filter out irrelevant features? We will train mice to identify the distance or angle of a pole where both features vary. We will quantify touch forces and the refinement of the motor program throughout learning using a novel motion tracking system. We will map activity patterns of identified cell-types in sensorimotor circuits of cortex across learning using calcium imaging and electrophysiology. We will build a system that translates observed sensorimotor signals into predicted patterns of neural activity that represent object features. We will deploy new optogenetic tools to perturb activity patterns in closed-loop with tactile exploration to identify activity patterns that drive sensorimotor perception, distinguish between models of circuit refinement, and test two theories of the functional consequences of learning on cortical dynamics.

Neuromodulators are essential signaling molecules that regulate many neural processes, including cognition, mood, memory, and sleep, through their influence on brain circuits. Monitoring the release and distribution of neuromodulators in behaving animals is critical for understanding the diverse functions of these molecules. A major impediment to developing this understanding is the lack of tools that can monitor these compounds at the temporal, spatial and concentration scales relevant to these brain processes. Filling this technological gap is one of the most pressing needs in neuroscience research. Our proposal directly bridges this gap by developing a platform of new tools for chronic, non- invasive monitoring of neuromodulators at millisecond, subcellular, and nanomolar resolution. Genetically-encoded fluorescent indicators for calcium and glutamate have transformed investigation of dynamic brain processes in the major model systems, including worms, flies, rodents, and increasingly primates. Building on our prior experience in developing these tools, we now propose to build a new suite of GPCR-activation-based (GRAB) genetically-encoded fluorescent indicators for neuromodulators. Our preliminary data shows we can generate GRABs with >500% fluorescence change and nanomolar affinity in mammalian cells. We propose to further develop and validate these prototypes in cultured neurons, flies, rodent brain slices, anesthetized and behaving mice to maximize their utility. In Aim 1, we will develop GRAB indicators for acetylcholine, serotonin, and norepinephrine by iteratively screening libraries that systematically vary in insertion site, linkers, cpGFP sequence, and FP-GPCR protein surface interface. The dimensions of optimization will be dF/F, membrane surface expression, affinity, and non-disruption of endogenous signaling. Our targeted performance levels are >10x dF/F, nanomolar range affinity and <10 millisecond on-rates in vitro. In Aim 2, performance of top candidate GRAB indicators from the in vitro screen will be validated following long-term expression in drosophila olfactory system, in brain slice, in anesthetized and behaving mouse cortex. Feedback from these experiments will guide iterative optimization in Aim 1. Successful completion of our Aims will yield a suite of powerful molecular constructs, cell-type specific viral tools and technical approaches that will be broadly disseminated to the neuroscience community. The GRAB indicators can be easily integrated with existing mouse models of human mental disorders. Since these probes for neuromodulators are well-suited for a wide range of preparations, and a large number of investigators, they will have a multiplicative impact on our understanding of neural circuit function and dysfunction when combined with other advances supported by the BRAIN Initiative.

Neuromodulators are essential signaling molecules that regulate many neural processes, including cognition, mood, memory, and sleep, through their influence on brain circuits. Monitoring the release and distribution of neuromodulators in behaving animals is critical for understanding the diverse functions of these molecules. A major impediment to developing this understanding is the lack of tools that can monitor these compounds at the temporal, spatial and concentration scales relevant to these brain processes. Filling this technological gap is one of the most pressing needs in neuroscience research. Our proposal directly bridges this gap by developing a platform of new tools for chronic, non- invasive monitoring of neuromodulators at millisecond, subcellular, and nanomolar resolution. Genetically-encoded fluorescent indicators for calcium and glutamate have transformed investigation of dynamic brain processes in the major model systems, including worms, flies, rodents, and increasingly primates. Building on our prior experience in developing these tools, we now propose to build a new suite of GPCR-activation-based (GRAB) genetically-encoded fluorescent indicators for neuromodulators. Our preliminary data shows we can generate GRABs with >500% fluorescence change and nanomolar affinity in mammalian cells. We propose to further develop and validate these prototypes in cultured neurons, flies, rodent brain slices, anesthetized and behaving mice to maximize their utility. In Aim 1, we will develop GRAB indicators for acetylcholine, serotonin, and norepinephrine by iteratively screening libraries that systematically vary in insertion site, linkers, cpGFP sequence, and FP-GPCR protein surface interface. The dimensions of optimization will be dF/F, membrane surface expression, affinity, and non-disruption of endogenous signaling. Our targeted performance levels are >10x dF/F, nanomolar range affinity and <10 millisecond on-rates in vitro. In Aim 2, performance of top candidate GRAB indicators from the in vitro screen will be validated following long-term expression in drosophila olfactory system, in brain slice, in anesthetized and behaving mouse cortex. Feedback from these experiments will guide iterative optimization in Aim 1. Successful completion of our Aims will yield a suite of powerful molecular constructs, cell-type specific viral tools and technical approaches that will be broadly disseminated to the neuroscience community. The GRAB indicators can be easily integrated with existing mouse models of human mental disorders. Since these probes for neuromodulators are well-suited for a wide range of preparations, and a large number of investigators, they will have a multiplicative impact on our understanding of neural circuit function and dysfunction when combined with other advances supported by the BRAIN Initiative.

PROJECT SUMMARY How sensory and motor signals are integrated in the brain to produce perception of object location remains poorly understood. Primary somatosensory cortex (S1) is a candidate site for sensorimotor integration that underlies object localization. Mouse S1 is a powerful system in which to uncover general principles and specific circuit implementations of sensorimotor integration that shape perception of object location. Revealing these will provide fundamental knowledge of healthy cortex function from which processing disruptions from stroke, spinal injury, and other neurological disorders may be more fully understood. The long-term goal of this work is to understand cellular and circuit mechanisms underlying tactile perception. This proposal focuses on how S1 integrates sensory and motor signals during active touch behaviors. Head-fixed mice can determine the angular position of objects by active exploration with a single whisker. Sophisticated neural processing underlies this simple behavior, which makes it an excellent model system for dissecting circuit mechanisms of somatosensory integration. Several competing models exist for how the brain solves this task. They differ in the type, origin, and integration location of sensorimotor signals used. Distinguishing between these models is critical for understanding the role internal motor signals in cortical circuits play in construction of tactile perception. Prior studies failed to do so because of limitations in task design and quantification of behavioral variation. This proposal overcomes these limitations with innovative approaches that include an improved localization task, high-speed sensorimotor tracking, cell type- specific electrophysiology, calcium imaging, sophisticated decoding models, and closed-loop optogenetics. The overall objective of this proposal is to distinguish between sensorimotor integration models by quantifying behavior, identifying candidate codes for object location in S1, how these are constructed, and their influence on perception. Our central hypothesis is that object location is encoded by the set of excitatory neurons activated by touch in L5B of S1, and that object location tuning in L5B cells requires both thalamic input and motion-subtracted touch signals from L4 of S1. We further hypothesize that M1 input amplifies L5B activity without affecting object location tuning. This hypothesis is supported by our preliminary data including cell-type and layer-specific recordings in S1 and optogenetic circuit manipulation during object localization. The hypothesis will be tested by pursuing three Specific Aims. 1) Identify candidate codes for object location in S1 neurons. 2) Identify the origin of signals contributing to object location tuning in S1 neurons. 3) Test object localization models with closed-loop optogenetic manipulation of S1 circuits. The contribution of the proposed research will be significant because it will generate detailed knowledge about the neural dynamics in S1 that underlie touch perception, uncover general principles of sensorimotor integration and specific cortical circuit implementations of that integration during behavior, and package it all into a publically accessible resource.  

Neuromodulators regulate addiction, attention, cognition, mood, memory, motivation, sleep, and more through their influence on brain circuits. Classic tools for measuring neuromodulation in the brain have poor spatial and temporal resolution. This has hampered the discovery of the diverse and complex functions neuromodulation plays during behavior. Over the past few years, new indicators for imaging neuromodulator dynamics have begun to dismantle these barriers. However, all existing neuromodulator indicators have significant limitations. The goal of this proposal is to optimize our GPCR-activation-based (GRAB) genetically-encoded fluorescent indicators of four major neuromodulators: dopamine (DA), acetylcholine (ACh), norepinephrine (NE), and serotonin (5-HT). We will make their responses bigger and more specific, create red versions for multiplexed imaging, and make them easier for end-users to successfully deploy in vivo. In Aim 1, we will optimize GRAB indicators for DA, ACh, NE, and 5-HT by iteratively screening libraries via high-content confocal imaging and FACS. We will vary insertion site, linkers, cpGFP, FP-GPCR protein surface interface, and thermostabilizing GPCR residues on a range of chimeric GCPR sensor backbones. Library generation will be prioritized by computational prediction of function from GPCR structures. The dimensions of optimization will be brightness, dF/F0, ligand selectivity, affinity, and non-disruption of endogenous signals. Top hits will be validated following long-term expression in mammalian brain slice and behaving mice. Our targeted performance levels are: 1000x ligand selectivity across all neuromodulators (3rd gen), >5x SNR improvement over 2nd generation indicators in vitro and in vivo (3rd gen), and reliable single-trial subcellular resolution of graded responses with in vivo 2-photon imaging of cortex during behavior for all neuromodulators (4th gen). In Aim 2, we will use the same approach as Aim 1 to develop and validate in vivo 1st and 2nd generation red GRABs for the same neuromodulators to enable simultaneous imaging of multiple signals. Our targeted performance levels for second generation, spectrally orthogonal red GRABs are 10x dF/F in vitro, >50% dF/F in vivo responses. We will also engineer out any photoactivation of red GRAB fluorescence, demonstrate multiplexed imaging and optogenetic stimulation with zero opsin excitation crosstalk from imaging light. In Aim 3, we will optimize GRAB packaging and distribution for maximum end-user ease of use. We will quantify the best FPs for in vivo coexpression with GRABs, engineer viral-genetic strategies for robust, brain- wide GRAB expression from systemic AAV injection, and make cre-reporter mouse lines for the best green GRAB of each neuromodulator. Optimized plasmids, AAVs, and mice will be broadly disseminated. Successful completion of our Aims will yield an optimized suite of powerful molecular tools packaged for maximum utility and ease of use. Since these probes are well-suited for a large number of investigators, they will have a multiplicative impact on our understanding of neural circuit function and dysfunction.