Understanding the broader implications and functions of membrane potential in the brain,heart, and body requires accurate and simultaneous membrane-associate voltage measurements on a large array of single cells. Achieving high spatial resolution and millisecond time resolution with minimal perturbation to live cells is a long-standing challenge, especially within the field of neurobiology. Direct imaging of membrane potential changes with fluorescence-based voltage indicators has become a powerful and most popular approach to mapping electrical changes yet several challenges addressing indicator speed, brightness, sensitivity, and localization remain. The Miller lab focuses on addressing these issues using a relatively new class of small molecule voltage indicators, VoltageFlours, that depend on a photo-induced electron transfer (PeT) mechanism. The intramolecular electron transfer modulates fluorescence intensity based on the surrounding electric field and results in sub-microsecond response kinetics with high voltage sensitivity. As with all organic indicators, the cellular spatial resolution in dense tissue samples is limited since VoltageFluors indiscriminately stain all membrane types. Our group has worked on addressing this limitation by combining VoltageFluors with genetically encoded self-labeling enzymes like the well-known Halo- and SNAP-tag proteins. Portions of this dissertation will focus specifically on the SNAP-tag covalent tethering approach with fluorescein and rhodamine-based VoltageFluors. The remainder of this thesis is primarily focused on a new branch of voltage imaging using bioluminescence which has never been explored within our group. Fluorescent indicators are dependent on externally applied illumination leading to unavoidable light-induced damage, especially at shorter wavelengths, which alters cell physiology and confounds data. There are relatively few examples of bioluminescent indicators relative to fluorescent indicators, and even fewer that are voltage sensitive. This dissertation will explore the development of the first ever quenching bioluminescent voltage indicator, limitations to single cell bioluminescent voltage imaging, and other potential approaches to functionalize bioluminescence at the membrane.

Fluorescent probes are noninvasive tools that can be used to map the electrical activity and synaptic communication of large networks of neurons. Fluorescent probes that bind Ca2+ or measure voltage changes translate activity into fluorescence signals with high spatial and temporal resolution. Current probes used for measuring neuronal activity utilize relatively short wavelengths of light. This dissertation discusses new synthetic methods to access sulfonated voltage sensitive dyes (VSDs) that are red-shifted and exhibit improved performance in cells, including greater photostability, voltage sensitivity, brightness, and signal to noise ratios. As most small molecule VSDs nonspecifically label cells, limiting their use in more complex tissues, chemical-genetic targeting approaches to target VSDs dyes for increased contrast and reduced background fluorescence will be discussed. We synthesize sulfonated carbofluorescein VoltageFluors (carboVFs) which utilize photoinduced electron transfer as a trigger to sense voltage in cells. CarboVFs display red-shifted profiles of >560 nm and high voltage sensitivities (up to 31% ΔF/F per 100 mV). The best performing carboVF, carboVF2.1(OMe).Cl, can be incorporated into an enzymatic, fluorogenic targeting strategy to specifically label cells. By replacing the oxygens in carboVFs with dimethylaminos, we accessed far-red TMCRh voltage indicators which possess excitation and emission profiles >620 nm, display improved photostability and reduced phototoxicity, and still maintain good voltage sensitivities (up to 18% ΔF/F per 100 mV). In particular, TMCRh.OMe rivals or outperforms reported far-red silicon rhodamine VSD, BeRST 1, when comparing signal to noise and ΔF/F in neurons, and has the potential to be used as a bright reporter of absolute membrane potential in non-excitable cells using fluorescence lifetime imaging. In addition to derivatizing the carbofluorescein scaffold to obtain new VSDs, this dissertation discusses a new synthetic route to access sulfonated silicon rhodamines and progress towards applying this strategy to carboxylated silicon rhodamines. Silicon rhodamines which bear a carboxyl instead of the typical sulfonate membrane anchor are important for the genetic-targeting of rhodamine-based VSD dyes to specific cells. We report the development of a genetically-targeted silicon rhodamine voltage indicator, isoBeRST-Halo, through a covalent labeling strategy using the HaloTag system. IsoBeRST-Halo maintains the high voltage sensitivity of the untargeted isoBeRSTs, while achieving high selectivity in cultured cells and cortical brain slice. To address the limitations of fast voltage imaging speeds and allow for the comprehensive mapping of neuronal networks, we sought to develop Ca2+ integrators based on the fluorogenic spirobenzopyran scaffold, calcium ligands, and a photoactivatable group. Various modifications to the spirobenzopyran scaffold will be discussed, as well as the synthesis and initial characterization of new calcium ligands. Together, this work demonstrates the power of synthetic chemistry to achieve highly selective, bright, and functional reporters of neuronal activity.

Fluorescent reporters of membrane potential are important tools in neuroscience, offering a non-invasive, real-time optical readout of neuronal activity. Voltage sensitive dyes of the VoltageFluor family report neuronal activity with high sensitivity and sub-millisecond response times, but lack the cell-type specific targetability of genetically encoded tools, placing certain constraints on their applicability in complex environments (i.e. tissue slice, in vivo, sub-cellular localization).

The work presented in this thesis addresses these problems through two distinct strategies. In Chapters 1 and 2 I describe the development of weakly-fluorescent, photoactivatable “caged” VoltageFluors. Spatial control of illumination allows for release of VoltageFluor at the single cell and subcellular levels in cultured mammalian neurons, and subsequent recording of evoked action potentials. This caging strategy has proven to be generalizable, and Chapter 3 and Appendix 1 detail the synthesis of enzymatically-activatable probes. In Chapter 4 I present a strategy for the selective labeling of cells with a VoltageFluor derivative, which harnesses the known covalent interaction between a short peptide, SpyTag, and its cognate binding protein, SpyCatcher. I synthesized a series of VoltageFluor-SpyTag conjugates, or VoltageSpy dyes, which specifically label the outer membranes of cells expressing SpyCatcher at the cell surface, with minimal off-target labeling, and without requiring a wash step. I demonstrate the application of this system to imaging spontaneous activity in cultured mammalian neurons, as well as imaging evoked activity in axons and dendrites. Appendix 2 describes efforts to apply VoltageSpy dyes in mouse brain slice. Finally, Chapter 5 presents a general strategy for minimizing phototoxic effects sometimes encountered when imaging with VoltageFluor dyes, using either bath applied triplet state quenchers or an intramolecularly-stabilized VoltageFluor derivative. Cyclooctatetraene was found to be an effective protective agent – a result which may prove broadly useful in imaging VoltageFluor dyes, improving the applicability of the fluorogenic and covalently-targeted tools presented in Chapters 1-4.

The ultimate goal of neuroscience is to correlate the activity of neurons and neuronal networks to higher order behaviors and cognition. This colossal task necessitates a variety of optical tools that can both measure and modulate neuronal activity. This dissertation describes the design, synthesis and characterization of a variety of such tools. In order to interrogate the electrical activity of neuronal circuits directly, we synthesized tetramethylrhodamine-based voltage reporters, or RhoVRs, that possess excitation and emission profiles in the green to orange region of the visible spectrum and use photoinduced electron transfer (PeT) as a trigger for voltage sensing. Two RhoVRs with particularly attractive photophysical properties were developed: RhoVR 1, which possessed the highest voltage sensitivity of any RhoVR (47% ΔF/F per 100 mV), and RhoVR(Me), a less sensitive (13% ΔF/F per 100 mV) but far brighter RhoVR (4-fold brighter than RhoVR 1). In order to facilitate the use of RhoVRs in complex tissues such as brain slice or in vivo, we developed genetically targetable RhoVRs using the HaloTag system (RhoVR-Halos). RhoVR-Halos maintained the fast kinetics and high sensitivities typical of RhoVRs (up to 34% ΔF/F per 100 mV) while labeling HaloTag expressing cells with high selectivity in culture, brain slice and in vivo. In addition, RhoVRs possess high two-photon cross sections (up to 190 GM), making them excellent candidates for two-photon voltage imaging. While voltage indicators allow for the study of fast voltage dynamics in real time, they necessitate very fast imaging speeds that are not well suited for mapping activity across large neuronal networks. In order to facilitate neuronal activity mapping, we developed a Ca2+ integrator MethylAzid-1 (MA-1). MA-1 acts as a coincidence detector for both light and Ca2+, allowing a temporally precise “snapshot” of [Ca2+] to be taken. MA-1 undergoes a 25 nm shift in absorbance upon chelation of Ca2+ (Kd = 270 nM) that allows for the selective photolysis of Ca2+-free MA-1 with 400 nm light. The extent of photolysis can be visualized post hoc through either “click chemistry” or the fluorogenic reduction of remaining MA-1. In addition to tools which report neuronal activity, we synthesized a new class of photocages based on 6-aminobenzofurans (BFCs) that can be used to selectively release caged compounds, such as neurotransmitters, upon irradiation with UV light. BFCs operate through a photo-induced elimination reaction that generates an extended azaquinone-methide from the excited state. Initial studies revealed BFCs are strong absorbers of UV/Violet light (λmax = 365 nm, ε = 30,000 M-1 cm-1) that readily photolyze (Φphotolysis up to 0.18) to release their caged cargo. Together, this work demonstrates the power of optical tools to study neuronal systems ranging from sub-cellular domains to large networks of neurons and lays the groundwork for their application in complex tissues such as brain slice or in vivo.

Comprehensively mapping and recording the electrical inputs and outputs of multiple neurons simultaneously with cellular spatial resolution and millisecond time resolution remains an outstanding challenge in the field of neurobiology. Development of fluorescence indicators for direct visualization of membrane potential changes offers a powerful method for probing voltage dynamics in neurons with unprecedented spatial and temporal resolution. Despite the promise of voltage imaging, key challenges of speed, sensitivity, brightness, and localization remain. The Miller lab has developed a new class of small molecules dyes, VoltageFluors (VF), that rely on photoinduced electron transfer (PeT) and display voltage-sensitive fluorescence due to modulation of the rate of PeT within the VF scaffold by the transmembrane electric field. This approach allows for fast, sensitive and non-invasive recording of neuronal activity in cultured mammalian neurons and in ex vivo tissue slices in single trials. One major limitation of small-molecule dye imaging is the inability to target the dye to specific cells of interest, which significantly erodes signal to noise and cellular resolution. To solve this problem, we have worked to combine VF dyes with a genetically encoded component to enable high-contrast imaging in defined neurons. We first attempted a fluorogenic approach, in which the parent VF dye is chemically modified to be minimally fluorescent and non-voltage-sensitive and must be enzymatically activated prior to imaging. We designed a couple of dyes which can only be activated by a cell-surface pig liver esterase (PLE) and observed enhanced contrast in both HEK cells and cultured neurons with good sensitivity. We have also shown that our genetically targeted voltage dyes outperform purely genetically encoded voltage indicators. In addition, we are exploring other enzyme/substrate pair for fluorogenic activation purpose. The second approach is covalent labeling of voltage sensitive dye to cells of interest using a cell-surface HaloTag enzyme. We have successfully demonstrated selective membrane staining in mouse brain slices and live animal using this strategy. Current efforts focus on applying this method for targeted voltage imaging in live animals to probe specific neuroscience questions.

Small-molecule synthetic tools are immensely useful in biological experimentation. However, these small molecules are not inherently targeted to genetically defined cell populations, thus hindering their use in complex living tissues. In my dissertation work, I adapted and developed two novel genetic constructs that traffic small molecule tethering proteins HaloTag1 and SNAPf 2 to the extracellular surface in cell lines and Drosophila brain tissue. I then applied these novel targeting systems toward two distinct applications: functional voltage imaging in live Drosophila brains using synthetic voltage-sensitive dyes and registration of live-fly brains to standard anatomical templates for comparison with existing databases.Voltage imaging in intact brains offers the tantalizing promise of watching, in real-time, the electrical changes that underlie physiology. To enable voltage imaging in Drosophila melanogaster, we combined a chemically synthesized rhodamine voltage reporter (RhoVR) with a genetically encoded, self-labeling enzyme, HaloTag. We generated a Drosophila reporter line that expressed the HaloTag enzyme on the extracellular surface. We validated the voltage sensitivity of this approach in cell culture before driving expression of HaloTag in specific neurons in flies. We showed that selective labeling of synapses, cells, and brain regions can be achieved with RhoVR-Halo in larval neuromuscular junction (NMJ) and whole adult brains. We validated the voltage sensitivity of RhoVR-Halo in fly tissue via dual-electrode/imaging at the NMJ, showing the efficacy of this approach for measuring synaptic excitatory post-synaptic potentials (EPSPs) in muscle cells, and performed voltage imaging of carbachol-evoked depolarizations and osmolarity-evoked hyperpolarizations in projection neurons and in interoceptive suboesophageal zone neurons in fly brain explants following in vivo labeling. The turn-on response to depolarizations, fast response kinetics, and two-photon compatibility of chemical indicators, coupled with the cellular and synaptic specificity of genetically-encoded enzymes will make RhoVR-Halo a powerful complement to neurobiological imaging in Drosophila. The Drosophila neurobiology community has developed a multitude of open-source neuron anatomy databases for comparative analysis of anatomy across samples. Unfortunately, fixation and permeabilization of the sample are often required which precludes their use in live tissue. Our novel extracellularly targeted tethering platforms comprised of SNAPf and HaloTag fusion proteins allowed for multispectral, in vivo anatomical analysis at the single-cell and whole-brain levels. We labeled the neuronal expression patterns of various Gal4 driver lines in live Drosophila brains recapitulating histological staining. Expressing SNAPf pan-neuronally, we registered brains to an existing anatomical template. From directly registered live brain tissue, we performed bridging registrations and a neuronal morphology similarity search (NBLAST)4. We predict that these extracellular platforms will become a valuable complement to existing anatomical methods and prove useful for future genetic targeting of other small molecule probes, drugs, and actuators.

Since the original synthesis of fluorescein in 1871, the use of xanthene fluorophores has remained ubiquitous in biology, medicine, and chemical biology. The wide utility of fluorescein and related xanthene dyes is due, in part, to their high brightness and wide range of colors available with simple modifications to the terminal and bridgehead atoms of the xanthene fluorophore core. By comparison, substitution of the pendant carboxylate has remained relatively underexplored. The identity of the substitution at this 3- position can have profound effects on the properties of xanthene fluorophores. We envisioned installation of a biologically relevant phosphonate functionality may provide access to fluorophores with unique properties, such as enhanced water solubility and opportunities for functionalization. Herein we will discuss the first syntheses of novel phosphonated xanthene fluorophores, their associated spectroscopic properties, and the utility of these dyes for both intracellular imaging and membrane potential sensing in living cells.

Membrane potential, Vmem, is the voltage across a cellular plasma membrane, produced by differences in ion concentration across the semipermeable membrane. Changes in Vmem over millisecond timescales transmit the electrical signals of action potentials in neurons and cardiomyocytes. Slower changes in Vmem over minutes, hours, or days affect processes such as differentiation, and set the resting membrane potential, which can affect neuronal and cardiac excitability. Vmem may also vary across substructures within cells, like organelles or dendritic spines, to enable localized electrical signaling. Forming a complete picture of the role of Vmem across these varied time and length scales requires techniques that can measure the value of Vmem in millivolts and reliably report both the dynamic changes in voltage and the absolute millivolt value of Vmem at rest.In this dissertation, we develop fluorescence lifetime imaging microscopy (FLIM) of the VoltageFluor (VF) small-molecule voltage-sensitive dyes as a method for optically reporting absolute Vmem, and then expand the method to additional dye scaffolds and cellular targets. We review existing tools and methods for recording Vmem, highlighting the need for additional optical techniques alongside electrophysiology. Initially using the VoltageFluor VF2.1.Cl, we demonstrate that its fluorescence lifetime, τfl, can report Vmem in multiple cell lines with voltage resolution at a biologically relevant level (5 mV RMSD voltage changes, and 19 mV RMSD for Vmem in a single trial in HEK293T cells). We optically read out the Vmem of hundreds of cells and report voltage distributions consistent with the values measured electrophysiologically, with 100-fold improvement in throughput over electrophysiology. To expand this FLIM technique to enable optical determination of Vmem in more complex cultures, such as neurons, we investigate two chemical-genetic targeting strategies. The first, VoltageSpy, has a 7-fold improvement in voltage resolution over the existing genetically encoded voltage indicator (GEVI) systems for reporting Vmem using FLIM. The second, VF-HaloTag, has a 10-fold improvement in resolution over the same GEVI FLIM indicator. VF-HaloTag also enables FLIM imaging of individual neurons and even individual dendrites with good spatial and lifetime resolution. We also develop a red-shifted, carborhodamine-based voltage indicator with comparable sensitivity to VF2.1.Cl. This red-shifted indicator is robust to photoxicity and photobleaching, while exhibiting a monoexponential fluorescence decay, requiring fewer photons for a reliable determination of τfl. This allows both rapid and extended FLIM imaging, allowing us to resolve cardiac action potentials in lifetime, and image sensitive cells for minutes at a time. Finally, we utilize FLIM as a tool for examining structural modifications to the VF scaffold. We report a consistent trend across three dye series (six dyes), demonstrating the effect of altering the position of the molecular wire in the VF scaffold on τfl. This overall work showcases the utility of FLIM to optically read out voltage, and shows how the technique can be expanded to measure voltage at various spatial and time scales, broadening the toolbox for optical readout of Vmem.

Light can be used to provide an optical readout of relative changes in membrane potential in a minimally invasive manner with high spatiotemporal resolution via the use of fluorescent voltage reporters. These reporters localize to cell membranes and are composed of three primary elements: a fluorophore, a molecular wire, and an electron rich aniline donor. The donor modulates the fluorescence of the reporter in response to changes in the membrane potential of a cell through a photo induced electron transfer (PeT) process. The existing color palette of these reporters is limited to the visible range of light (< 700 nm). Near-infrared (NIR) fluorescent dyes are favorable since the use of NIR light avoids cellular autofluorescence, is less damaging than visible light, and can be used simultaneously with other tools in the visible spectrum such as calcium indicators and optogenetic actuators. To take advantage of these properties, we synthesized and characterized four NIR voltage reporters using a phosphine oxide rhodamine fluorophore with a maximum absorbance at 704 nm and peak emission at 723 nm. These NIR phosphine oxide rhodamine voltage reporters, poRhoVRs, exhibit fractional changes in fluorescence ranging from 13-43% per 100 mV with good signal to noise ratios. We have successfully used poRhoVR to monitor spontaneous neuronal activity in cultured neurons as well as activity evoked through field stimulation or via optogenetic activation. We have also used poRhoVR in mouse retina in conjunction with GCaMP6f, to monitor pharmacologically induced changes in activity, while offering greater spatial resolution than possible with multi-electrode arrays. These poRhoVR dyes can easily be modified to be genetically targetable by appending chemical handles for proteins such as HaloTag.

Voltage sensitive fluorescent indicators have been a useful method for monitoring cardiac electrophysiology for nearly 50 years, complementing traditional methods of electrophysiology such as patch clamp techniques. The impacts fluorophore phototoxicity on cell health have long been a concern in fluorescent imaging. VoltageFluors (VF dyes) are a class of voltage sensitive indicators that rely on photoinduced-electron transfer to sense changes in membrane potential. While our lab has made significant strides towards expanding the wavelengths of indicators offered and improving indicator voltage sensitivity, efforts to reduce fluorophore phototoxicity in the context of VF dyes was largely unexplored. The work presented in this dissertation explores a generalizable method for reducing the phototoxicity of our VF dyes, with the tethering of our probes to the triplet state quencher cyclooctatetraene (COT). Chapter 1 presents the development of VF-COT, a green, xanthene-based indicator intramolecularly tethered to COT. VF-COT can report cardiac action potentials for twice as long as its parent dye, without significant impacts on electrophysiology parameters. Chapter 2 presents BeRST-COT, a near-infrared, silicon rhodamine-based VF that demonstrates the generalizability of reducing phototoxicity with an intramolecularly tethered COT molecule. BeRST-COT reports cardiac electrophysiology for 2.7- times as long as its parent dye and reduces the relative production of singlet oxygen during imaging sessions. Chapters 1 and 2 present work utilizing cardiac cells towards efforts of better characterizing ultra-stable VF indicators. Appendix 1 aims to utilize VF dyes and fluorescence lifetime imaging to better quantify cardiac electrophysiology parameters and lays out the work completed and the rationale for future work in this project.