Live Imaging with Genetically Encoded Physiologic Sensors and Optogenetic Tools

Abstract

Barrier tissues like the epidermis employ complex signal transduction systems to execute morphogenetic programs and to rapidly respond to environmental cues to promote homeostasis. Recent advances in live imaging techniques and tools allow precise spatial and temporal monitoring and manipulation of intracellular signaling cascades. Leveraging the chemistry of naturally occurring light-sensitive proteins, genetically encoded fluorescent biosensors have emerged as robust tools for visualizing dynamic signaling events. In contrast, optogenetic protein constructs permit laser-mediated control of signal receptors and effectors within live cells, organoids, and even model organisms. Here, we review the basic principles underlying novel biosensors and optogenetic tools and highlight how recent studies in cutaneous biology have leveraged these imaging strategies to illuminate the spatiotemporal signals regulating epidermal development, barrier formation, and tissue homeostasis.

INTRODUCTION

As the first line of defense at the body’s surface, the skin is exposed to a variety of environmental insults. Cells within the skin must continually integrate intra- and extracellular cues to respond to these perturbations, whether through proliferation, migration, differentiation, death, or inflammation. A more complete understanding of cellular adaptations to physical, chemical, and biological signals promises to illuminate the mechanisms underlying epidermal homeostasis and how their dysfunction leads to cutaneous disease. Genetically encoded fluorescence-based biosensors (proteins that emit light in response to certain physiologic signals) have emerged as unparalleled tools for visualizing and quantifying complex signaling events across time and space to understand physiology within live cells (Zhou et al. 2020). Coupled with advances in the acquisition speed, light sensitivity, and resolution of microscopes, these tools now permit sub-cellular imaging in live 3D organoids and even model organisms (Kasatkina and Verkhusha 2022), thus redefining what is possible in dermatologic research. Moreover, the expanding repertoire of fluorophores with non-overlapping excitation and emission wavelengths can allow biosensor multiplexing to simultaneously monitor multiple signals. Moreover, optogenetic actuators, which are genetically encoded and light-activated protein effectors, grant investigators precise control of physiologic effectors within intact cells (Endo and Ozawa 2017). Here, we review the latest advances in biosensor design and the principles underlying optogenetic manipulation of cell signaling. We also highlight recent studies applying these tools to make key discoveries in dermatologic science, including progress in intravital imaging and computational modeling to map signaling networks in complex biological systems like the skin.

PRINCIPLES OF GENETICALLY ENCODED BIOSENSORS

Biosensors are typically modular, chimeric proteins composed of a sensor domain fused to a reporter domain (Greenwald et al. 2018; Zhou et al. 2020). Sensor domains are often derived from endogenous proteins that participate in native signaling; the sensor binds specifically to its ligand, whether a signaling molecule (e.g., calcium), a post-translational modification (e.g., phosphorylated tyrosine), or a protein partner. The sensor module can be tethered via flexible peptide linkers to reporter domains, often fluorescent proteins, which provide a real-time visible readout for sensor activity. Biosensors can be engineered to localize to the cytoplasm or to organelle membranes. Changes in the conformation and/or localization of the sensor domain in response to a specific signal is coupled to fluorescence intensity of the reporter domain; this allows spatial and temporal tracking of signaling with sub-cellular resolution.

Over the last two decades, sensor domains have been engineered to reflect the activity of diverse signaling pathways. The use of endogenously derived sensor domains allows for detection of signaling under physiologically relevant conditions and time scales (Greenwald et al. 2018). On the other hand, de novo protein design and high-throughput peptide screening platforms have rapidly expanded the repertoire of synthetic sensor domains available to augment the sensitivity and specificity of biosensors (Klima et al. 2021; Rappleye et al. 2022). Similarly, novel fluorescent protein reporter domains have been developed through circular permutations (e.g., adding a peptide linker to separate and re-orient N- and C-terminal domains) to increase their intensity, to enhance on-off cycling speed, or to make their activation dependent upon pH, intramolecular interactions, or proximity to a partner protein (i.e., fluorescence resonance energy transfer, FRET) (Greenwald et al. 2018; Lin et al. 2019; Zhou et al. 2020). FRET-based methods were previously reviewed in this journal (Broussard and Green 2017) and will not be discussed here. Common biosensor design strategies are depicted in Figure 1 and examples are discussed below. For a comprehensive summary of biosensors, we direct the reader to an excellent review (Greenwald et al. 2018).

Figure 1. Common biosensor design strategies.

a) Translocation-based biosensors contain dynamic sensor domains that change subcellular localization (i.e., from the cytoplasm to the nucleus) in response to various cellular signals, allowing fluorescent visualization of signal-induced changes in protein localization. b) Split fluorescence protein (FP)-based biosensors employ self-assembling FP fragments that are linked to protein(s) of interest i.e., dimeric binding partners. Signal-induced changes in the sensor domain unit bring the two FP fragments together resulting in fluorescence reconstitution. c) Circularly permuted FPs are sandwiched between sensing units composed of two or more interacting proteins. Switch-like conformational changes in the sensing unit trigger closure of the cpFP resulting in enhanced fluorescence intensity. d) pH-sensitive FPs, which, for example, become quenched in acidic environments, can be fused to organelle membrane-localization signals or proteins of interest such as endosomal or autophagosomal makers enabling the detection of pH dynamics within specific subcellular compartments. e) Fluorescence resonance energy transfer (FRET)-based biosensors rely on the distance-dependent transfer of energy between two compatible fluorophores (i.e., CFP and YFP) such that the emission spectrum of the donor FP overlaps with the excitation spectrum of the acceptor FP. Conformational changes in the sensing unit or protein-protein interactions bring the two FPs in close proximity, allowing energy transfer between the donor and acceptor fluorophores and an increase in FRET intensity.

IMAGING CALCIUM WITH GENETICALLY ENCODED SENSORS

Calcium (Ca²⁺) is a ubiquitous second messenger and a central driver of keratinocyte differentiation (Elsholz et al. 2014). Its signaling influences broad-ranging biological processes from proliferation to transcription to intercellular adhesion. Calcium plays an essential role in skin homeostasis and epidermal barrier function, and disturbances in Ca²⁺ signaling have been implicated in common dermatologic disorders like psoriasis as well as rare genetic dermatologic disorders like Darier and Hailey-Hailey disease. Thus, there has been long-standing interest in the temporal and spatial fluxes in calcium within the epidermis under normal conditions and in disease. Prior to the advent of genetically encoded fluorescent sensors, investigation of calcium dynamics relied on synthetic calcium-sensitive fluorescent dyes. Early studies using these dyes provided fundamental insights into keratinocyte biology, including characterization of the Ca²⁺ distribution within the epidermis (Celli et al. 2010); however, these dyes needed to be delivered into cells, which can be inefficient and difficult to control, and offered low spatiotemporal resolution, potential toxicity, and poor compatibility with in vivo models. Genetically encoded Ca²⁺ indicators (GECIs) overcome these challenges.

Several GECIs utilizing chimeras of calcium-binding and fluorescent protein domains have been engineered, but GCaMP sensors have become the most widely utilized (Figure 2) (Zhang et al. 2021; Greenwald et al. 2018). In GCaMP, the calcium-binding domain of calmodulin (CaM) is coupled to the CaM-binding peptide, M13, of myosin light chain kinase yielding a core sensing unit that undergoes large switch-like conformational changes upon calcium binding (Figure 2a). These Ca²⁺-dependent structural changes are translated to a reporter composed of a green fluorescent protein (GFP) having its N- and C-termini separated by a peptide linker (i.e., circularly permuted) to minimize baseline fluorescence. Binding of the sensor domain to calcium triggers deprotonation of GFP, leading to a marked increase in fluorescence intensity. Since its original design, GCaMP has undergone iterative updates to improve functionality (e.g., signal-to-noise ratio, kinetic parameters, sub-cellular localization) with jGCaMP8, engineered for improved fluorescence rise times and in vivo performance, being the latest addition (Zhang et al. 2021).

Figure 2. GCaMP biosensor design for calcium sensing and applications in keratinocyte biology.

a) Schematic representation of the GCaMP structure where a central circularly permuted enhanced GFP (cp-eGFP) is sandwiched between the M13 fragment of myosin light chain kinase (M13) at the N-terminus and Calmodulin (CaM) at the C-terminus. Ca²⁺-dependent binding of CaM to M13 triggers conformational rearrangement of cp-eGFP and fluorescence reconstitution. b) Keratinocytes are chemically coupled through the assembly of intercellular gap junctions (connexons), which enable the propagation of Ca²⁺ waves between neighboring cells within local signaling networks, which can be visualized in real-time using GCaMP sensors. c) Representative time-lapse fluorescence images of primary normal human epidermal keratinocytes (NHEKs) expressing GCaMP6 following treatment with the SERCA inhibitor, thapsigargin (at t=0s), which induces rapid translocation of calcium from the ER into the cytosol. Scale bar = 10 μm.

GCaMP sensors have been utilized to make important advances in cutaneous biology research. Most recently, transgenic mice expressing GCaMP6 in epidermal keratinocytes using a Keratin 14 promoter-driven Cre recombinase were imaged using intravital two-photon microscopy (Moore et al. 2022). Analysis of the mice using a novel deep-learning tool, Geometric Scattering Trajectory Homology (GSTH), mapped epidermal Ca²⁺ dynamics in vivo with unprecedented resolution across the tissue. This approach revealed that epidermal stem cell subpopulations in the G2 phase of the cell cycle exhibit intracellular Ca²⁺ cycling and coordinate long-range calcium signaling across the epithelium via gap junctions, thereby orchestrating epidermal regeneration (Figure 2b). A similar study using live epidermal explants from GCaMP3 reporter mice identified age-related impairments in intracellular Ca²⁺ signaling in aged keratinocytes (Celli et al. 2021). These studies underscore the broad utility of GECIs in unraveling mechanisms of epidermal homeostasis and dysfunction.

GECIs can be targeted to organelles to expand their biological applications. ER-GCaMP6 localizes to the endoplasmic reticulum (ER) lumen using the N-terminal signal peptide of calreticulin and a C-terminal KDEL retention signal (de Juan-Sanz et al. 2017). This tool can provide mechanistic insights into the regulation of ER calcium in keratinocytes, which is disrupted in Darier disease, caused by mutation of the sarco/endoplasmic reticulum Ca²⁺ ATPase, SERCA2 (Elsholz et al. 2014) (Figure 2c). In addition to GECIs, a suite of translocation-based biosensors for other calcium-dependent second messengers (e.g., inositol triphosphate, diacylglycerol) and FRET-based biosensors to monitor protein kinase C activity may be of particular interest to the investigative dermatology community (reviewed in Greenwald et al. 2018). Beyond calcium signals, FRET- and translocation-based biosensors have been used to study the propagation of other crucial regulators of keratinocyte differentiation. For example, biosensors allowed investigators to map the dynamics of Erk signaling in vitro and in vivo in the epidermis and to identify novel chemical modulators of this pathway (Hiratsuka et al. 2015 and 2020; Goglia et al. 2020).

MONITORING pH WITH FLUORESCENT BIOSENSORS

Genetically encoded fluorescent tools have been developed to reflect intracellular pH, which changes during epidermal barrier formation (Greenwald et al. 2018; Zhao et al. 2019). The generation of pH-sensitive biosensors is aided by the inherent acid sensitivity of GFP fluorescence (Chin et al. 2021; Greenwald et al. 2018), which depends upon the protonation state of a tyrosine residue within its chromophore (Figure 3a-b). In the ground state, the weakly fluorescent neutral (protonated) form of GFP predominates; however, upon excitation, the deprotonated, anionic state is stabilized by a hydrogen-bonding network, resulting in a rise in fluorescence intensity. As a result, the chromophore of GFP is quite sensitive to pH and mutations that modify or disrupt the hydrogen-bonding network produced GFP mutants with varying pH sensitivity.

Figure 3. pH- and REDOX-sensitive fluorescent biosensors.

a) Intensiometric pH sensors employ a single pH-sensitive FP which, for example, display reduced fluorescence intensity upon acidification. Ratiometric pH sensors contain two or more tandem FPs with unique spectral properties and pH sensitivities. Upon acidification, the pH-sensitive fluorophore is quenched, while the pH-insensitive fluorophore serves as a stable reference point enabling more robust quantification of pH changes. b) pH-responsive biosensors can be coupled to endogenous proteins to enable compartment-specific detection of pH changes. Fusion of a ratiometric pH sensor to the lysosome-associated membrane protein 1 (LAMP1) allows quantification of lysosomal acidification/maturation. c) Schematic representation of the dual-functional fluorescent biosensor, pHaROS, comprising a REDOX-sensitive iLOV domain and a pH-sensitive mBeRFP. d) Green-light illumination of the photosensitizing fluorescent protein, KillerRed, triggers rapid photobleaching and generation of reactive oxygen species including singlet oxygen and superoxide. KillerRed can be fused to endogenous proteins or organelle-localization signals to allow compartment-specific oxidative damage induced by laser illumination.

pH-sensitive fluorescent proteins have allowed monitoring pH within specific intracellular compartments (e.g., endosomes or lysosomes) as well as the trafficking of proteins between these structures. pH-sensitive fluorophores can be targeted to select organelles using routing sequences, as from lysosomal-associated membrane protein 1 (LAMP1) (Figure 3b) (Chin et al. 2021). This approach could be used to assess cutaneous lysosome dysfunction, which underlies certain inherited dermatologic disorders. Moreover, tandem fluorescent proteins facilitate ratiometric quantification of the pH of cellular compartments. By labeling a protein with both GFP, which is acid-quenched, and red fluorescent protein (RFP), which is acid-resistant, one can follow the fate of the protein as it is routed into autophagosomes, then lysosomes for degradation (Chin et al. 2021; Greenwald et al. 2018). Localization in neutral compartments like the cytosol yields a dual fluorescence signal; however, translocation to an acidic compartment inactivates GFP leaving a single signal from the acid-resistant RFP (Figure 3a). This tandem fluorophore can be appended to any protein of interest, for example, microtubule-associated proteins 1A/1B light chain 3B (LC3) to monitor the fusing of autophagosomes with lysosomes to quantify autophagic flux.

pH modulation is especially important during epidermal cornification. While pH-sensitive fluorescent probes have been previously used to examine the pH gradient within the uppermost layers of the epidermis (Hanson et al. 2002), pH biosensors have more recently illuminated this process with remarkable spatial and temporal resolution. As granular layer keratinocytes mature into corneocytes, they undergo dramatic structural reorganization culminating in a form of cell death, termed corneoptosis, involving formation of cornified cell envelopes and degradation of organelles (Avecilla and Quiroz 2021; Simpson et al. 2021). Leveraging the fluorophores mNectarine (pK_a 6.9) and SEpHluorin (pK_a 7.6), which rapidly quench at pH 6.3, Garcia et al. uncovered a dramatic intracellular pH shift during the granular-to-cornified cell transition that was necessary for the dissolution of keratohyalin granules and initiation of corneoptosis in cultured keratinocytes (Garcia et al. 2020). These findings were consistent with those of Matsui and colleagues, who developed a ratiometric pH sensor composed of pHVenus (pH-sensitive) fused to mCherry (pH-resistant) and similarly observed rapid cytoplasmic acidification during granular cell death in vivo (Matsui et al. 2021). Expanding upon earlier work by Murata and colleagues (Murata et al. 2018), pH-sensitive and -resistant GCaMP-based sensors revealed a gradual rise in intracellular Ca²⁺ that preceded keratinocyte acidification. This was followed by a sustained Ca²⁺ increase that persisted through corneoptosis, which was mediated by the transient receptor potential V3 (TRPV3) ion channel. These studies underscore the power of biosensor imaging to quantify physiologic signals in live tissues and show how they can be combined to dissect the temporal and spatial relationship of multiple signals.

STUDYING OXIDATIVE STRESS WITH BIOSENSORS

Alterations in oxidative stress have been associated with skin aging, oncogenesis, and inflammatory diseases. The development of redox-sensitive protein reporters has allowed real-time visualization of intracellular oxidative stress, which can drive cell signaling and tissue injury. Light, oxygen, or voltage-sensing (LOV) domains have been attractive tools for biosensor design due to their stability across a range of pH and temperature. Since they also contain a blue-light-sensitive flavin-based chromophore, which uses flavin mononucleotide (FMN) as a co-factor, these proteins can be exquisitely sensitive to redox shifts. Multiparametric biosensor engineering was recently employed to generate a dual pH and redox-sensitive reporter (pHaROS) (Figure 3c) (Zhao et al. 2019). Fusion of a LOV domain to a pH-sensitive fluorophore, blue light-excited red fluorescent protein (mBeRFP), generated a probe that can simultaneously detect changes in pH and redox status. Such multiparametric reporters promise to illuminate crosstalk and integration of the multiple signaling pathways supporting skin development and homeostasis. For example, multi-functional biosensors could be used to interrogate redox changes in relation to intracellular pH and calcium shifts in the setting of mitochondrial degradation during early cornification (Simpson et al. 2021).

While fluorescent proteins have been used to monitor redox changes, their unique chemistry can also be leveraged to generate reactive oxygen species (ROS) upon light illumination (Bulina et al. 2006). Most fluorescent proteins generate little phototoxicity; however, mutagenesis efforts identified photosensitizers, which produce damaging singlet oxygen and superoxide upon light excitation. Of these, the engineered dimeric red fluorescent protein, KillerRed, was found to possess the highest phototoxicity, three orders of magnitude above GFP (Figure 3d). KillerRed can be targeted to distinct subcellular compartments to induce localized oxidative stress and organelle damage following green light illumination (Moore et al. 2021). This tool can be multiplexed with redox-sensitive biosensors to simultaneously control and visualize oxidative stress responses with robust spatial and temporal resolution.

OPTOGENETIC CONTROL OF INTRACELLULAR SIGNALING

While genetically encoded biosensors afford real-time visualization of cell signaling, optogenetic tools arose from the desire to directly influence cellular biology. Optogenetics leverages genetically encoded, photosensitive domains called actuators to manipulate protein activity using light illumination (Endo and Ozawa 2017; Rappleye and Berndt 2019). Traditional optogenetic systems relied on light-gated ion channels, such as channelrhodopsins, to control ion flux and membrane potential dynamics; however, newer photoactivatable modules such as cryptochromes, phytochromes, and LOV domains have greatly expanded the breadth of biologic processes that can be controlled using these tools by enabling fusion to endogenous effector proteins of interest. These photoactivatable modules can be used to manipulate intracellular processes and signaling pathways with exquisite spatial and temporal control using lasers. Photosensors exhibit a broad range of responses to light, enabling the design of actuators that can induce protein conformational rearrangements, aggregation, or dissociation (Figure 4).

Figure 4. Optogenetic manipulation of intracellular calcium dynamics.

a) Optogenetic dimerization systems employ photosensitive proteins capable of undergoing light-induced homo- or hetero-dimerization to facilitate clustering of endogenous proteins of interest. b) Light stimulation of cytochrome photoreceptor 2 (Cry2) initiates homo-oligomerization of activated proteins, allowing for precision control of oligomerization-dependent cellular processes. c) Schematic representation of OptoSTIM1-mediated endogenous CRAC/Orai1 Ca2+ channel activation. Blue light illumination triggers Cry2 mediated oligomerization of the STIM1 cytosolic domain and CRAC channel opening. d) Schematic representation of a light-operated Orai1 channel (LOCa) engineered via the introduction of a LOV2 domain within a cytoplasmic loop of constitutively active Orai1. Blue light-induced conformational changes within the LOV2 domain trigger Orai1 channel opening and Ca²⁺ influx.

Many signaling processes are regulated by the assembly and disassembly of protein complexes. Precise control of these pathways can be achieved using actuators to induce clustering of protein monomers upon light exposure. A variety of plant- and microbe-derived LOV and cryptochrome domains with the ability to homodimerize in response to blue light have been utilized to control dimerization-dependent signaling. For example, light-induced homo-dimerization has been achieved with the Arabidopsis thaliana photosensor, cryptochrome 2 (Cry2), which self-aggregates upon blue light exposure (Figure 4a-b) (Endo and Ozawa 2017). Cry2 can also form light-inducible dimers with a second protein, cryptochrome-interacting basic-helix-loop-helix 1 (CIB1). Thus, dual Cry2- and CIB1-based actuators can be used to induce protein hetero-dimerization. These versatile tools have enabled investigators to spatially and temporally control a variety of cellular processes: building signaling complexes, assembling transcription factors to influence gene expression, or coupling organelles to motors to translocate them along microtubules (Ballister et al. 2015).

CONTROLLING ION FLUX WITH OPTOGENETIC TOOLS

An optogenetic strategy for controlling intracellular calcium was recently devised based on light-induced protein clustering. At steady state, cytoplasmic Ca²⁺ is kept low through ER sequestration and export across the plasma membrane (PM) (Elsholz et al. 2014). Release of ER calcium triggers STIM1 (stromal interaction molecule 1)-mediated opening of Ca²⁺ release-activated Ca²⁺ (CRAC) channels, including ORAI1, in the PM to augment calcium influx. To activate CRAC channels, STIM1 forms oligomers upon depletion of ER calcium, allowing it to translocate from the ER to the PM. Fusion of the cytoplasmic domain of STIM1 to Cry2 generated an optogenetic actuator (OptoSTIM1) that drives light-induced oligomerization of STIM1 to control Ca²⁺ influx in both in vitro and in vivo settings (Figure 4c) (Kyung et al. 2015). OptoSTIM1 can be multiplexed with GCaMPs to permit simultaneous control and visualization of calcium dynamics in vivo.

To gain more direct control of cellular ion flux, channelrhodopsins, light-activated ion channels, have garnered significant interest as optogenetic tools (Rappleye and Berndt 2019). While these proteins display varying degrees of ion selectivity and slower conductance rates compared to other channels, structure-guided protein engineering promises to improve channelrhodopsin ion selectivity and conductance. Overcoming these limitations, a single component light-operated Ca²⁺ channel (LOCa) was developed by inserting a blue-light sensitive plant LOV domain within an intracellular loop of ORAI1 (He et al. 2021) (Figure 4d). Similar to OptoSTIM1, LOCa is capable of controlling Ca²⁺ entry in both cultured cells and animal models. Compared to OptoSTIM1, however, LOCa displays slower activation and deactivation kinetics, highlighting the intricacies of optogenetic tool design.

Toward elucidating epidermal biology, OptoSTIM1 or LOCa could be combined with pH-sensitive biosensors to query how calcium may triggers intracellular pH shifts during cornification. These tools could also be helpful for modelling rare disorders like Darier or Hailey-Hailey Disease, which are caused by mishandling of intracellular Ca²⁺. Similarly, a photoactivatable TRPV3 channel could enable control of the final stages of keratinocyte differentiation via inducible calcium influx or could be used to model the cutaneous channelopathy Olmsted syndrome (Blaydon and Kelsell 2014). Optogenetic tools could also be applied to other ion channels like TRPV1, implicated in itch, or SCN9A, mutated in primary erythromelalgia. In addition, light-induced oligomerization of connexins to form gap junctions might illuminate the pathophysiology of keratitis-ichthyosis-deafness syndrome (Blaydon and Kelsell 2014). As a final example, heterodimerization-based actuators designed to link organelles to autophagy receptors to target them for lysosomal degradation could help to elucidate the mechanisms driving organelle elimination during keratinocyte differentiation and dysfunctions in this process in disorders of cornification (D’Acunzo et al. 2019; Simpson et al. 2021).

Although optogenetic tools offer a unique ability for investigators to not only observe, but to control cellular biology, they have only recently entered the mainstream of cutaneous biology research. We hope this review highlights the remarkable versatility of these experimental tools, which have the potential to revolutionize our understanding of skin physiology and the pathophysiology of dermatologic diseases.

PRACTICAL CONSIDERATIONS

Biosensor selection:

Signaling processes can display marked variability in their temporal kinetics and even the same pathway can show variable kinetics depending on the cell type, organism, or signaling context. Thus, the features of a biosensor should be carefully reviewed when designing an experiment including its sensitivity, kinetic parameters, spectral properties, and pH/thermal stability (Greenwald et al. 2018; Lin et al. 2019; Zhou et al. 2020). While some biosensors have been re-engineered to enable faster rise times, this can compromise sensitivity and vice versa. Investigating pathways with slower kinetics may require biosensors with better photostability to minimize photobleaching during long-term imaging.

When designing multiparametric experiments to visualize multiple signals simultaneously, biosensor pairs must carefully be selected to ensure compatible biochemical properties and non-overlapping spectra to allow signal segregation. Fluorescent proteins are undergoing optimization to reduce the spectral band they occupy while new fluorescent proteins and imaging modalities are being developed to expand the spectral range available for multiplexing. The latter effort was advanced by the discovery of far-red and near-infrared (NIR) fluorescent proteins (Greenwald et al. 2018; Zhou et al. 2020). NIR-based biosensors have been essential for deep-tissue and intravital imaging by providing robust tissue penetration while minimizing the autofluorescence and light-scattering associated with shorter wavelength sensors.

While using sensor domains from endogenous proteins has enabled rapid biosensor development, these proteins may remain capable of interacting with their normal binding partners, which could disrupt endogenous signaling or induce cytotoxicity. Indeed, GCaMPs were recently shown to interfere with L-type calcium channels (Ca_v1) in neurons, resulting in distorted Ca²⁺ signaling and gene expression (Yang et al. 2018). Ca_v1 interference was prevented by adding an apoCaM-binding motif to GCaMP yielding GCaMP-X. Various alternative strategies exist for limiting such “off-target” effects, including biosensor sequestration through membrane localization tags and structure-guided engineering of sensor domain binding surfaces to eliminate unwanted protein interactions. As well, the use of inducible promoters enables more precise control of the timing and dosage of biosensor expression.

Gene delivery and expression:

To leverage the many advantages of biosensors and optogenetic actuators, the genetic elements that encode these tools must be effectively delivered into the cells, tissues, and organisms of interest, which can be challenging. In cell culture and organotypic models, modified viral vectors have been used for stable gene delivery. For multiparametric studies, cells can be transduced with several viral vectors, ideally with distinct selection markers to generate stable multiply-transduced cell lines. However, viral vectors often display a bias towards transcriptionally active sites, which can result in off-target effects. This can be overcome using the non-viral Sleeping Beauty transposon system, which relies on a transiently expressed transposase to integrate the transgene at random genomic sites (Izsvák and Ivies 2004). Non-viral gene delivery strategies include transfection with liposome- or cationic polymer-based reagents or electroporation; however, these methods typically only allow transient transgene expression and are often less efficacious, especially in primary cells.

Use of biosensors and actuators is feasible in animal models using viral transduction, microinjection, electroporation or transfection of a DNA construct into zygotes or early embryos to make transgenic mice (Kasatkina and Verkhusha 2022). If targeting to specific loci is desired, CRISPR-based reagents can be leveraged to allow precise insertion of the transgene. To probe signaling dynamics in selected cell types, tissue-specific promoters and Cre-lox systems can be used to drive biosensor expression, for example, using the well-established basal layer Keratin 14 promoter or suprabasal Involucrin promoter (Garcia et al. 2020; Moore et al. 2022). Drug-inducible promoters (e.g., tetracycline-on) can be used if temporal control of biosensor or actuator expression is desired. Interestingly, a light-inducible dimerization module has been applied to the Cre recombinase to afford precise control of transgene expression based on focal application of a particular wavelength of light (Kasatkina and Verkhusha 2022). Transgenic animal lines can be crossed to generate progeny carrying two or more biosensors or actuators for multiparametric experiments. Somatic transgene expression can be achieved using viral vectors injected into live mice or in utero. Finally, intraepidermal plasmid injection into the dorsal skin of hairless mice enabled transient biosensor expression in keratinocytes (Matsui et al. 2021). These techniques can be used in transgenic biosensor mice to allow multiplexed experimental design in vivo.

Imaging and data processing:

Recent advances in live imaging techniques have expanded the breadth of biological questions that can be probed using genetically encoded fluorescence-based sensors and optogenetic actuators. While confocal microscopy (laser scanning or spinning-disk) continues to be a gold standard in fluorescence imaging, multiphoton imaging allows deeper tissue visualization in vivo. Two or more long-wavelength photons are emitted by a femtosecond pulsed laser to simultaneously excite a fluorophore, as opposed to a single, higher-energy photon used for fluorescence. This enables illumination using NIR wavelengths, which reduces light scattering and autofluorescence, thereby allowing imaging through thick tissues with thinner optical sections and minimal phototoxicity. Two-photon excitation microscopy has already enabled several important discoveries in skin science (Ipponjima et al. 2020; Matsui et al. 2021; Moore et al. 2022). For a more comprehensive review of multiphoton imaging of the skin, we direct our readers to a dedicated review (Obeidy et al. 2018).

Advanced complexity of biosensors and imaging modalities have increased the need for software and computational models capable of handling the data generated from long-term live imaging in multiple planes. Data processing has been particularly challenging for the analysis of complex signaling networks in heterogenous cell populations across large tissues over time. An unsupervised, deep learning framework termed Geometric Scattering Trajectory Homology (GSTH) was recently developed to assist with analysis and modeling of signaling dynamics from large image datasets (Moore et al. 2022). Although this method was developed to study Ca²⁺ dynamics in epidermal stem cells, it can be applied to other tissue types and signaling processes. Such computational tools will be crucial to make sense of large-scale datasets derived from biosensors and actuators, especially in multiparametric analyses, to elucidate the spatiotemporal integration and crosstalk between signaling pathways across three-dimensional tissues.

CONCLUDING REMARKS

Genetically encoded fluorescent biosensors and optogenetic tools have rapidly expanded investigators’ ability to interrogate the regulation of complex signaling in real-time in native biological contexts. In this review, we discussed biosensor and actuator design strategies for visualizing diverse intracellular signaling mediators like calcium, monitoring changes in pH and oxidative stress, as well as manipulating intracellular proteins to control their localization, interactions, or activation. By introducing cutaneous biology researchers to the available toolkit of genetically encoded biosensors and optogenetic actuators, we hope to expand the notion of what is possible in cutaneous biology research by adapting these cutting-edge techniques to the skin. To aid our readers, we have summarized the tools and techniques discussed in this review in Supplementary Tables 1 and 2. For a more comprehensive review of biosensor selection and experimental design, we direct readers to a recently compiled online database, the Fluorescent Biosensor Database (FBDB), cataloging over 750 variants (Greenwald et al. 2018). These tools have already provided important insights into the mechanisms of epidermal regeneration and differentiation, but the additional potential applications of biosensors and optogenetic tools to dermatologic science are boundless. Further advances in multiparametric analyses and computational modeling for intravital imaging will permit a more complete understanding of in vivo physiology. Moreover, using these tools to unravel the complex regulation of tissue development, organ homeostasis, and disease pathogenesis will pave the way for novel diagnostic and therapeutic strategies, including for the many skin diseases in need of novel treatments.

SUMMARY POINTS.

Genetically Encoded Fluorescent Biosensors

Advantages:

• Enable quantification of the spatiotemporal kinetics and dynamics of a variety of signaling processes under physiologic conditions.

• Modular design strategies make it easy to develop new biosensors as well as to improve or modify the properties of existing sensors.

• Localization signals, functional domains, and cell-type specific promoters allow targeting to specific subcellular compartments, proteins, cell types, and tissues.

• Multiplexed experimental approaches permit investigation of crosstalk between signaling pathways.

• NIR-based sensors and multiphoton imaging can be used for deep tissue imaging in vivo.

Limitations:

• Genetic elements encoding the biosensor must be delivered into cells, tissues, or organisms.

• Native reporting units may retain biologic activity complicating quantification of signaling kinetics and dynamics.

• Some sensors require access to highly specialized microscopes and laser lines.

Optogenetic Systems

Advantages:

• Light-inducible systems enable precise control of specific signaling processes with high spatial and temporal resolution both in vitro and in vivo.

• Modular design principles allow the generation of actuators that can be activated or deactivated using laser illumination.

• Optogenetic tools can be multiplexed with fluorescent biosensors to simultaneously manipulate and visualize signaling processes with sub-cellular resolution.

Limitations:

• Genetic elements encoding the actuator must be delivered into cells, tissues, or organisms.

• Bypasses native signaling mechanisms, which could compromise physiologic relevance.

• Cells and animals must be protected from light (including ambient light) before the start of an experiment to avoid premature photoactivation.

• May be associated with phototoxicity or thermal activation.

• Some optogenetic tools require access to highly specialized microscopes and laser lines.

• Multiplexed studies require careful consideration of optogenetic actuator and biosensor pairs to ensure non-overlapping spectral properties.

Abbreviations:

(Ca²⁺)

Calcium ion

(CRAC channels)

Ca2+ Release-activated Ca2+ channels

(CaM)

Calmodulin

(DAG)

Diacyl glycerol

(ER)

Endoplasmic Reticulum

(FMN)

Flavin Mononucleotide

(FRET)

Fluorescence Resonance Energy Transfer

(GECI)

Genetically Encoded Calcium Indicator

(GSTH)

Geometric Scattering Trajectory Homology

(GFP)

Green Fluorescent Protein

(IP₃)

Inositol triphosphate

(Ca_v1)

L-type calcium channels

(LOV)

Light, Oxygen, or Voltage Domain

(LOCa)

light-operated Ca2+ channel

(LAMP1)

Lysosomal-Associated Membrane Protein 1

(LC3)

Microtubule-associated Proteins 1A/1B Light Chain 3B

(mBeRFP)

Monomeric Blue Light-excited Red Fluorescent Protein

(NIR)

near-infrared

(PM)

Plasma Membrane

(PKC)

Protein Kinase C

(STIM1)

Stromal Interaction Molecule 1

(TRPV3)

Transient Receptor Potential V3