Opticool: Cutting-edge transgenic optical tools

Abstract

Only a few short decades have passed since the sequencing of GFP, yet the modern repertoire of transgenically encoded optical tools implies an exponential proliferation of ever improving constructions to interrogate the subcellular environment. A myriad of tags for labeling proteins, RNA, or DNA have arisen in the last few decades, facilitating unprecedented visualization of subcellular components and processes. Development of a broad array of modern genetically encoded sensors allows real-time, in vivo detection of molecule levels, pH, forces, enzyme activity, and other subcellular and extracellular phenomena in ever expanding contexts. Optogenetic, genetically encoded optically controlled manipulation systems have gained traction in the biological research community and facilitate single-cell, real-time modulation of protein function in vivo in ever broadening, novel applications. While this field continues to explosively expand, references are needed to assist scientists seeking to use and improve these transgenic devices in new and exciting ways to interrogate development and disease. In this review, we endeavor to highlight the state and trajectory of the field of in vivo transgenic optical tools.

Introduction

Biological inquiries utilizing live organisms offer superior relevancy to cell culture and fixed tissue studies in unraveling the expansive mysteries which remain in cell, developmental, and disease biology because the research is being performed in the most functionally relevant environment. A panoply of tools has been developed to interrogate these perplexities of nature, but none as exciting as the growing number of optical tools which have recently become available for use in live organism studies.

As such, this review will focus on those optical tools that can be readily used in transgenesis and will not cover the tools for interrogating biological processes via fixed samples, nonoptical means, or by using tools that require external biochemical treatments. In the first instance, there are several available extensive reviews covering advances in chromatin architectural characterization techniques [1,2], fluorescent in situ hybridization (FISH)-related methods [3,4], and transcriptomics [5,6] and proteomics [7,8]. Similarly, improvements in multiomics [9], mass spectrometry [10,11], DamID and BioID [12–14] have been recently chronicled. Exciting new in vivo dye, drug, and other external treatment techniques, while beyond the scope of this inquiry, are exploding in utility in basic and translational molecular biology fields.

Transgenic models recapitulate the natural settings of biological inquiries while reducing the frequency of required optimization steps and autonomously reproducing experimental toolkits. Here, we endeavor to compile a reference for anyone seeking to understand or develop new transgenic models expressing optical tools for visualization, quantification, and/or manipulation of subcellular and tissue-level molecular processes.

1 Advanced tools for visualizing subcellular functions in vivo

Fluorescent transgenic protein detection tags

While GFP has only been used as a transgenic tool for less than 30 years [15], transgenic fluorescent markers for molecular biology have advanced to the point that mere labeling of proteins with even some of the most optimized fluorophores [16,17] is sometimes regarded as banal with growing criticisms of these fluorescent proteins’ (FP) limitations (Fig 1A). Maturation rates of FPs have been a particularly popular target of recent optimization studies to ameliorate unavoidable delays in protein detection of target-FP fusions [18]. However, such limitations are now being avoided altogether through an alternative approach whereby FPs are ubiquitously expressed separately from the target protein but directed to the target protein through tag binding in vivo. In this way, fluorescence can be detected immediately upon translation of the target protein.

Fig 1. Tools for visualizing subcellular components in vivo.

(A) Sequencing of GFP in 1992 allowed for creation of fluorescently labeled, transgenic labels and set in motion the progress of the field. (B) SunTag:FP fucion proteins bind to the SunTag scaffolding to brightly label POIs. (C) LlamaTag binds directly to specific FPs to rapidly label POIs. (D) Stem loop binding proteins are used to label RNA sequences. (E) elF4a is a two-part coat protein dimer that binds its specific stem loop structure and allows for split FP background elimination. (F) Pumilio allows for labeling of RNA sequencing through a genetically encodable 8 bp-binding region which can be customized. It has been improved to work in pairs attached to split FPs to eliminate background fluorescence and increase sensitivity. (G) dCas9 can be used to label DNA sequences in vivo through gRNA targeting. (H) Zinc fingers can be genetically engineered to target DNA sequences in vivo and label them with FPs. Blue arrows represent time in decades since the sequencing of GFP in 1990, left, to the present, right. FP, fluorescent protein; POI, protein of interest.

These secondary attachment fluorescent systems are generally termed tags [18–21], with one of the most popular transgenic tag systems being Sun-tag [19,21]. The Sun-tag system makes use of 2 short peptides with high affinity to one another to recruit FPs to target proteins. A fusion protein is constructed consisting of a POI and up to 24 repeating sequences that allow for complimentary attachment via a second FP fusion protein [22,23]. Specifically, a single-chain fragment variable (ScFv) antibody is used which recognizes GCN4; the GCN4 repeats fused to the POI are used to recruit ScFv-bound FPs to tag the POI for detection (Fig 1B). Sun-tag’s “scaffolding system” produces increased brightness because several fluorescent molecules can simultaneously attach to the POI [21]. Llama-tags are another protein tagging system allowing for faster visualization in vivo due to their use of already matured FPs [18]. Llama-tags use optimized nanobody fusion proteins to directly recruit FPs to POIs (Fig 1C). This mechanism therefore requires creation of only 1 transgene to be used and allows for transgenic addition of the tags into FP-expressing animals which are now ubiquitous [24,25].

Real-time transcriptional visualization

Rapid fluorescence detection via pooled mature fluorophores at the ready has grown in popularity not only for visualizing protein locations, but also for detecting gene expression and mRNA dynamics in vivo [19,26,27]. By far the most popular systems for real-time transcription detection are the MS2 [28,29] and PP7 [30–33] systems, whereby MS2 coat protein (MCP) and PP7 coat protein (PCP), fused to fluorophores, bind MS2 and PP7 RNA stem loops, respectively. These have been used to generate a broad array of transgenic animals and computational tools to track gene expression timing in live imaging experiments [19,34–39]. In addition to these popular mRNA visualization options, multiple other aptamer-coat protein pairs are currently available for transgenic studies including BglG [40], U1Ap [41], λN22 [42], HTLV-1 Rex [43], TAT/REV [44], and several less tested variants [40,45,46] (Fig 1D).

The two-part aptamer-binding eIF4a [47] provides an intriguing twist on the system, providing a reliable, straight-forward method to label RNA without ubiquitous background fluorescence by fusing the 2 parts to split fluorophores [48,49] (Fig 1E). Further, RNA-binding Pumilio [50,51] allows for sequence modification to bind to chosen, engineered eight-nucleotide sequences of RNA and has been implemented in the Pumby [52] labeling system to label adjacent mRNA sites to utilize split fluorophores (Fig 1F). The major limitation of most tag systems being the necessity for high background fluorescent signal from a pool of overexpressed FPs poised for binding to the tag, these split fluorophore approaches present significant advancement by eliminating this excess fluorescent noise. Similarly, DNA sequences can be targeted transgenically for visualization using dCAS9-FP fusions to target sequences in a guide RNA directed manner [53] (Fig 1G) or via traditional engineering of zinc finger-FP fusions [54] (Fig 1H).

2 Transgenic optical sensors

FRET-based sensors

Beyond simply labeling cellular components, a growing array of transgenic optical tools for measuring subcellular and tissue conditions have arisen in recent decades [55,56]. The field of transgenic biosensors has, again, bounded forward from the earliest methods, such as fluorescence recovery after photobleaching (FRAP, Fig 2A), in a mere few decades [57,58]. These remarkable nano- and micro-scale auto-biological devices present a unique opportunity in the fields of developmental and disease biology allowing for measurements where even the most modern and advanced exogenous technologies are insufficient [59].

Fig 2. Transgenic optical sensors.

(A) One of the earliest subcellular sensor system, FRAP is used to detect protein diffusion and translation rates in vivo through photobleaching of an FP, followed by imaging of fluorescent recovery. (B) Fluorophores with overlapping excitation and emission wavelengths can share energy through FRET when sufficiently close in proportion to their distance from one another. (C) Separating FRET pairs with molecular springs allows for measurement of extension of the molecular springs, and therefore, tension. (D) Ace Sensors utilize modified rhodopsin that alters configuration upon formation of membrane potential altering the FRET efficiency of an inserted FP. (E) Because flavin-binding FP fluorescence maturation is oxygen independent but GFP-based FPs are not, FluBO reports O₂ levels via O₂-dependent FRET efficiency variation. (F) Separating a FRET pair with a caspase recognition sequence allows for detection of caspase activity via SCAT sensors. (G) Steric alteration of the PS3 ε-subunit when bound to ATP allows the ATeam ATP sensor to detect ATP via FRET efficiency alteration. (H) Similar to FRET, luminescing proteins can sometimes share energy. (I) The GAP sensor utilizes shift of GFP excitation wavelength when in proximity to aequorin when bound or unbound to calcium ions. (J) The biochemical domains of some FPs can be isolated, allowing for artificially induced fragility. (K) Detection of F-actin polymerization with PriSSM is achieved by wedging a GFP and cpGFP into a Myosin II motor domain such that binding of the motor domain to F-actin produces a steric alteration which increases fluorescence and increases blue:near-IF excitation ratio. (L) In G-CaMP, a cpGFP is flanked by Calmodulin and a myosin light chain fragment which bind one another such that steric conditions disrupt fluorescence in the absence of calcium binding. (M) ADP/ATP steric alterations in GlnK fluctuate fused cpGFP optimal excitation wavelength of the Perceval sensor. Green arrows represent time in decades since the sequencing of GFP in 1990, left, to the present, right. FP, fluorescent protein; FRAP, fluorescence recovery after photobleaching; FRET, Förster resonance energy transfer.

Perhaps the most versatile of these transgenic instruments is proximity detection via Förster resonance energy transfer (FRET) [60–62]. In this method, 2 proteins are labeled with fluorophores capable of FRETing with one another and the distance between them can be calculated when they’re in sufficient proximity to one another [63] (Fig 2B). Mechanical forces are known to be significant in several developmental and disease pathways [64–71]. A particularly innovative extension of this method is the creation of several tension sensors [72–77] (Fig 2C). Because the distance between 2 FRETing fluorophores can be detected optically, fusing compatible fluorophores together separated by a molecular spring, generally spectrin [78] or flagelliform [79] repeats, allows for the measurement of forces on the FRET pairs via a simple calculation of the measured distance and the known spring-like characteristics of the separator [80–82]. These sensors have been developed in mechanical components of cells and tissues ranging from extracellular matrix and cell–cell connections [83,84], cytoskeletal connections [85,86], and all the way into the nucleus and onto the genome itself [75].

An innovative use of FRET in sensors has facilitated the detection of membrane voltage potentials via opsin-based Acetabularia opsin (Ace) sensors. In the Ace sensor systems, an FP is fused to an opsin with which it can FRET; when the intracellular side of the membrane gains a net positive charge, increased FRET reduces fluorescence of the FP (Fig 2D). Ace sensors now come in a variety of spectral flavors including red (VARNAM [87]) and green (Ace2N-mNeon [88]). Another creative use of FRET to sense subcellular conditions is FluBO [89]. While optical sensing tools have long been a field of inquiry and optimization [90], FluBO cleverly uses a flavin-binding FP (FbFP) as donor with a YFP acceptor to detect molecular oxygen (O₂) within cells. YFP is sensitive to O₂ levels while FbFP is not, allowing for detection of higher FRET efficiency in O₂ rich cells (Fig 2E). Caspase activity can further be measured using FRET pairs separated by caspase cleavage sites thereby providing a reduction in FRET as a readout for caspase activity [91–97] (see Fig 2F). ATP concentration is an important subcellular condition which has motivated the creation of a multitude of sensors [98]. The transgenic ATeam sensors [99] utilize modified ATP synthase ε-subunit transversely terminally fused to a donor and acceptor FPs such that binding to ATP [99–101] or MgATP [102] brings the FPs into proximity whereby increased FRET efficiency can be detected (Fig 2G).

Bioluminescence-based sensors

A somewhat similar method for functional and proximity detection to the FRET-based sensors is the use of bioluminescent resonant energy transfer (BRET) whereby a bioluminescent molecule is used to shift the emission fluorescence of a fluorophore in sufficient proximity via molecular transference of the bioluminescent enzyme’s emission to the fluorophore [103,104] (Fig 2H). FRET detection in FRET sensors, while precise, has proven difficult to effectively implement in vivo as autofluorescence and relatively weak signal-to-noise making alternatives like BRET attractive [105,106].

However, while a panoply of BRET sensors exist [106,107], e.g., Ca²⁺ sensing with LuCID [108], cAMP sensing with CAMYEL [109,110], cytoskeletal tension sensing [111], O₂ sensing [112], caspase activity sensing [113], and POI/POI interaction detection [114], current iterations remain shackled to the requirement of exogenous introduction of their cofactor substrates [115]. As such, these tools remain distinct from pure transgenic optical tools. Despite this, recent advances have been made to achieve autoluminescence from bioluminescence systems by introducing genes that ultimately catalyze endogenous synthesis of chemical cofactors [116–119]. Conversely, a clever employment of Aequorea victorea apo-aequorin allows for Ca²⁺ sensing without its bioluminescent cofactor requirement in the GAP (GFP and apo-aequorin) sensors [120]. In these sensors, binding of calcium ions to aequorin shifts the excitation maximum of GFP allowing for Ca²⁺ concentration quantification via ratiometric fluorescence measurement (Fig 2I).

Modified and permutated fluorophore sensors

A methodological extension of the split-fluorophore mechanism described above for in vivo tagging of RNA with low background is a class of sensors utilizing split-fluorophores [48,49]. In most of these systems, interactions of proteins labeled with complimentary subunits of the split fluorophores are detected by fluorescence via sufficient proximity of the subunits. While this method is similar to the FRET pair systems above, it has significant benefits and drawbacks in comparison: split-fluorophores confer a binary interaction measurement and not a distance, but offer higher detection resolution with less background noise and require less sophisticated imaging and analysis techniques than FRET systems. Thus, split-fluorophores have also been used to develop binary force sensors where the subunits are separated by a flexible linker allowing for subunit separation upon sufficient tension [76,121].

Fluorophores have further been modified by rearrangement and insertions in their amino acid sequences to produce various effects, termed circularly permutated fluorophores (cpFPs) [122,123] (Fig 2J). A remarkable application of one such cpFP is the strain sensor PriSSM [124]. Proximity imaging (PRIM) [125]-based strain sensor module (PriSSM) facilitates detection of F-actin-myosin II strain via ratiometric fluorescence following 490 or 390 nm laser activation. Strain on PriSSM changes the orientations of tandem, contacting GFP and cpGFP FPs making the optimal activation wavelength 390 nm and 490 nm detached from F-actin or under strain, respectively, as well as increasing fluorescence under strain (Fig 2K). A further implementation of a circularly permutated GFP (cpEGFP) was used to construct a Ca²⁺ sensor, G-Camp [17,126], which increases in fluorescence due to conformational changes induced by calcium ion binding (Fig 2L). The reduced efficiency of cpEGFP in the unbound state has the additional, key advantage of increasing the signal-to-noise ratio of G-CaMP relative to alternative strategies. The Perceval sensor [127–129] permits in vivo measurement of ATP/ADP ratios via measurement of fluorescence ratio from a circularly permuted monomeric Venus (cpmVenus) fused to Methanococcus jannaschii GlnK1 when excited by 490 or 405 nm wavelength lasers (Fig 2M). Further, use of modified FPs to be sensitive to other molecules has also been accomplished as in ClopHensor which uses a modified, chloride-ion-sensitive GFP, E²GFP, fused to DsRed-m to detect chloride levels in cells by ratiometric fluorescence [130,131].

3 Optogenetics: Optical tools for real-time subcellular manipulations

Activation/Inactivation systems

Optogenetics is a more recent technique that facilitates convenient manipulation of cellular function [132–136]. With the use of light, proteins can be utilized to affect the function of the cell. Optogenetics makes use of light-sensitive proteins (either artificial or naturally occurring) and adjusts their functionality through adjusting their secondary, tertiary, or quaternary structure. There is a myriad of mechanisms that can be affected with optogenetic tools, but all are initiated by a chromophore or light-absorbing amino acid. Optogenetics is commonly preferred over chemically inducible systems for a multitude of reasons. Specifically, optogenetic tools allow for greater specificity through avoidance of secondary chemical effects, availability of multiple isolated wavelengths allows for easy combinatorial implementations, modern laser microscopy enables precise 3D localization of effect, and physical barriers can often be overcome through the use of light rather than chemical compounds. These tools can often be adjusted within milliseconds allowing for specified cellular manipulation and can be used to study a broad range of cell types in live animals, especially during development when tissues depth is minimal and tissue transparency is greatest.

Inactivation of proteins by exposure to specific wavelengths of light is a straight-forward and reliable mechanism to modulate protein function in vivo. Degrons are naturally occurring peptide markers which facilitate degradation of the protein to which they are attached [137]. To be used as optogenetic tools, these degradation tags are fused to cryptochromes, enzymes which change conformation in response to light to expose or conceal cryptic domains within themselves, allowing for concealment and longevity of the target protein or exposure and degradation of the target protein in response to light.

While optogenetic degrons, as a class, are often referred to as photo-sensitive degrons, psd is a specific optogenetic degron developed by fusing the photosensitive domain of Arabidopsis thaliana phototropin1, light oxygen voltage 2 (LOV2), to mouse ornithine decarboxylase carboxy-terminal degron (cODC) [138,139]. When fused to a protein of interest (POI), psd exposure to blue light induces a conformational change to expose the cryptic degron, leading to protease degradation of the construct which can then be repopulated in the dark (Fig 3A). Another system developed for triggering protease degradation of fusion proteins by exposure to light is the blue light inducible degradation (B-LID) method [140]. B-LID contains a small peptide degron [141] rendered cryptic by fusion to a LOV2 domain; in this way, degradation of fusion proteins is triggered by exposure of the degron upon exposure to blue light (see Fig 3A). Photo-N-degron was recently developed to take advantage of light-dependent N-end rule-mediated protein degradation [142]. Light-induced uncoiling of the Jα helix in LOV2 exposes an N-terminal arginine amino acid and triggers N-end rule degradation [143,144] (see Fig 3A).

Fig 3. Optogenetic tools for real-time subcellular manipulations.

(A) Degrons are used with LOV2 to yield blue light triggering of fusion protein degradation. (B) Photosensitizers are FPs which produce ROS biproducts when excited and can be used to distort fused proteins. (C) LANS and LINuS trigger reversible nuclear import upon blue light exposure by exposure of a cryptic NLS in a LOV2 fusion. (D) LINX and LEXY trigger reversible nuclear export upon blue light exposure by exposure of a cryptic NES in a LOV2 fusion. (E) iLID/SspB heterodimerize reversibly when exposed to blue light. (F) CRY2/CIB1 reversibly heterodimerize when exposed to blue light. It is important to remember that CRY2 also oligomerizes under blue light exposure. (G) Reversible CRY homooligomerization occurs when exposed to blue light. (H) COP1/UVR8 reversibly heterodimerizes when exposed to ultraviolet light. In the dark, UVR8 reversibly forms homodimers. (I) Positive (pMag) and negative (nMag) domains within Magnets tools dimerize in the presence of blue light via exposure from cryptic domains within LOV2 fusions. (J) PixE/PixD heterooligomerizes in the dark and can be reversibly dissociated into PixD homodimers and PixE monomers by exposure to blue light. (K) Q-PAS1/BphP1 heterodimers are reversibly formed when exposed to far red light, whereas BphP1 homodimers and Q-PAS1 monomers are produced via red light or darkness. (L) PhotoCleavable cleaves upon exposure to violet light. (M) Opsin membrane transport pumps are used to induce transmembrane pumping of ions when exposed to appropriate wavelengths of light. Orange arrows represent time in decades since the sequencing of GFP in 1990, left, to the present, right. FP, fluorescent protein; LANS, light-activated nuclear shuttle; LOV2, light oxygen voltage 2; ROS, reactive oxygen species.

Apart from degrons, there are other strategies for deactivation of proteins. One such method is photosensitizers used in chromophore-assisted light inactivation (CALI) [145,146]. Photosensitizers are chromophores which produce reactive oxygen species (ROS) in response to light activation [147]. CALI takes advantage of this production to inactivate proteins attached to a photosensitizer. The first genetically encoded photosensitizer was KillerRed, derived from the hydrozoan chromoprotein anm2CP, which produces phototoxic affects via ROS production in response to green light exposure [148]. KillerRed is an effective optogenetic tool for selectively killing cells and tissues through green light-mediated ROS production but is also a useful tool in CALI schemes (Fig 3B). Because CALI is a more direct system than degrons, inactivation is substantially more rapid than degron-based systems but have the drawback of producing cytotoxic ROS. Since KillerRed, multiple alternate photosensitizers have been developed to counter KillerRed’s propensity to dimerize and facilitate use of other wavelengths of light [149,150]. LightsOut introduces an AsLOV2 domain into the Gal4 transcription factor (TF), widely used for conditional transgene expression [151,152], between DNA-binding and gene activation domains to suppress Gal4-mediated expression when exposed to blue light [153].

Subcellular translocation systems

AsLOV2 has further been implemented in light-activated nuclear shuttle (LANS) [154] and light-inducible nuclear localization signal (LINuS) [155] to enable conditional nuclear localization of POIs upon exposure to blue light using a cryptic NLS (see Fig 3C). Conversely, the light-inducible nuclear export systems LINX [156,157] and LEXY [158,159] employ a light-exposable cryptic NES within an AsLOV2 fusion allowing for blue light-induced nuclear export of POIs (Fig 3D).

The ability to reversibly trigger Botrytis cinerea BcLOV4 membrane binding via blue light exposure has enabled its use as a conditional fusion protein localization system whereby blue light is used to target fusion proteins to the membrane [160–163]. An interesting, naturally occurring optogenetic tool for localization to DNA is Erythrobacter litoralis EL222 which dimerizes and binds to DNA in response to blue light [164,165]. This system is currently used in multiple contexts to control transcription via light exposure [166–170].

Binding/Polymerization systems

An enormously useful optogenetic innovation has been transgenic tools which facilitate manipulation of binding dynamics of POIs [171]. It is now possible to use these tools to hold proteins in contact with one another, sequester them against membranes, or release them via exposure to laser light (Fig 3E–3K). This mechanism is not only of inherent utility in POI functional control but also has been shown extensively to be greatly advantageous as a compounding factor in conjunction with other optogenetic mechanisms [132].

iLID, an improved and modified variant of LID [141], is a blue light inducible dimerizing protein which forms a dimer with SspB in light and dissociates in the dark [172] (Fig 3E). iLID utilizes an AsLOV2 domain fused to an SsrA peptide which is then able to bind with its binding partner, SspB [173], upon exposure via reversible, light-induced LOV2 conformational transformation. A deservedly popular use of the iLID system is OptoSOS [174–178]. OptoSOS utilizes a membrane-anchored iLID to recruit an SspB-fused SOS to trigger a Ras/Erk cascade when exposed to blue light. Another light-triggered dimerization system is cryptochrome 2 (CRY2) and CIB1 [179]. In this system, blue light induces heterodimerization of Arabidopsis thaliana CRY2/CIB1 (Fig 3F). Simultaneously, however, this system also causes CRY2 homo-oligomerization creating both a potential experimental challenge and a useful mechanistic expansion beyond other hetero/homodimerization systems [180–182] (Fig 3G). Intriguingly, an optogenetic heterodimerization system exists which utilizes UV-B light [183]. In the presence of UV-B, Arabidopsis COP1 binds to UVR8 and can be used bring POIs into proximity to one another in living cells (Fig 3H).

Similarly, heterodimerization can be controlled by red and far-red light, as in the phytochrome B (PhyB)/phytochrome interacting factor (PIF) system [184]. However, these red or far-red phytochrome systems require plant chromophore 3-Z phycocyanobilin (PCB) which is not naturally synthesized in most animals, making them challenging to use in transgenic systems. Nevertheless, PCB can be synthesized in mammalian cells by transgenically introducing 4 genes required to generate it from heme [118,119]. An intriguing use of this system is SOScat which operates identically to the iLID optoSOS system via red and far-red light rather than blue, allowing for greater freedom in designing combinatorial transgenic systems [185]. Furthermore, Arabidopsis thaliana PhyA/FHY1 behave very similarly allowing for the creation of REDMAP which facilitates gene activation under red light through heterodimerization and deactivation under far-red light via dissociation [186]. Bacteriophytochrome BphP1 is used as a light-sensitive binding partner to Rhodopseudomonas palustris PpsR2 [187]. More recently, a truncated variant of PpsR2, Q-PAS1, was engineered to reduce its size and mitigate natural oligomerization of PpsR2 [188,189]. In this light-sensitive binding method, far-red light exposure causes binding of BphP1 to its binding partner which is reversed in red light or darkness conditions (Fig 3K).

A further innovation in light-dependent interaction, Magnets, has been developed to induce POIs heterodimerization in the presence of blue light [190,191]. Magnets was developed by modifying the Per-Arnt-Sim (PAS) domain of the Neurospora crassa LOV protein Vivid which forms heterodimers when exposed to blue light [192–194]. Modifications were made to the protein sequence in the binding domain to either introduce positive (pMag) or negative (nMag) amino acids, thereby creating 2 modified Vivid photoreceptors which heterodimerize but do not homodimerize in the presence of blue light (Fig 3I). Another tool derived from Vivid is LightOn which uses Gal4 fused to Vivid and a p65 transactivation domain, GAVP, to induce dimerization and activation of UAS target gene expression when exposed to blue light [195].

Contrary to the bind-in-the-light/dissociate-in-the-dark scheme of the systems discussed so far, there also exist systems which employ the contrary scheme. An example of this is the Synechocystis PixD/PixE system. In the dark, PixE and PixD form oligomers containing 5 PixE and 10 PixD proteins which dissociate into PixE monomers and PixD homodimers when exposed to blue light [196] (Fig 3J). Interesting, bacterial photoactivated adenylyl cyclase (bPAC) [197], which utilizes a BLUF (blue light receptor using FAD) light sensor domain similarly to PixD, facilitates light-dependent increase in cAMP levels. In LOV2 trap and release of protein (LOVTRAP), a small protein, Zdark (Zdk), was developed to bind to the dark state of LOV2 [198]. This facilitates the conditional tethering and release of LOV2 to Zdk under dark or blue light exposure, respectively. Somewhat orthogonally paralogous to the Zdk/LOV2 binding pair is the Erbin PDZ/LOV2 binding pair, TULIPS [199], whereby a cpePDZ [200] can bind to a short peptide sequence fused to LOV2 upon blue light exposure. A disparate mechanism of light-mediated anchoring control is through light-controlled protein cleavage. An example of this system is PhotoCleavable (PhoCl) [201] (Fig 3L). PhoCl consists of a modified circularly permutated mMaple which spontaneously dissociates upon exposure to violet light and disconnection of N- and C- terminal fused POIs.

Opsins

Perhaps the simplest examples of optogenetic tools are opsins which have gained immense popularity in neuron studies and are lurching towards translational implementation. Opsins are light-sensitive membrane proteins which regulate transmembrane ion transfer. This makes them simple, inherent optogenetic tools whereby membrane potential can be controlled by light exposure [202] (Fig 3M).

An example of an opsin optogenetic tool is pHoenex, an optical proton pump tool [203]. pHoenex uses a modified Halorubrum sodomense Archaerhodopsin-3 (Arch3) light-sensitive proton pump fused to an ATPase, vesicular protein Synaptophysin, and optical pH sensor pHluorin [204] to facilitate H⁺ shuttling to acidify the interior of synaptic vesicles via yellow light and pH detection via blue light. Similarly, several other optogenetic opsin tools are available which facilitate light-mediated transmembrane proton pumping [205], sodium ion pumping [206], and chloride ion pumping [207,208]. Furthermore, guanylyl and adenylyl cyclase rhodopsins, such as rhodopsin-guanylyl cyclase (RhGC) [209], rhodopsin cyclic nucleotide phosphodiesterase (Rh-PDE) [210], and guanylyl cyclase rhodopsin (CyclOp) [211], have demonstrated utility in mediating light-dependent cAMP and/or cGMP levels to exogenously regulate subcellular signaling.

4 Modern advances in combinatorial, real-time optical methods

Techniques involving simple, direct light-dependent messenger manipulation are quite popular and elegant; however, the most exciting implementations of transgenic optical tools are those which involve cooperative implementation of multiple systems simultaneously to enable greater control over and comprehension of the molecular biological systems being interrogated. It is feasible that experimental protocols can already be developed to visualize, measure, and manipulate one or more subcellular molecular processes in the same cell(s) in real- or near-real-time. Indeed, a growing number of such systems combining multiple transgenic optical tools and strategies are being developed to combine or improve upon individual schemes.

A complex blend of transgenic optical tools, the blue light-inducible TEV protease (BLITz) system, combines the CRY2/CIB1 system with the AsLOV2 system to precisely induce transcription upon exposure to blue light [212] (Fig 4A). BLITz consists of 2 parts, a membrane-bound fusion of CIBN (a truncated version of CIB1), the N-terminal portion of the tobacco etch virus protease (TEV), AsLOV2, TEV cleavage site, and a transcriptional activator, and a soluble portion consisting of CRY2PHR (a truncated version of CRY2) and the C-terminus of TEV (C-TEV). Upon exposure to blue light, binding of CRY2PHR to CIBN facilitates interaction of the 2 parts of TEV while simultaneously exposing the TEV cleavage site, allowing cleavage and release of the transcriptional activator.

Fig 4. Modern advances in combinatorial, real-time optical methods.

(A) The BLITZ system uses CRY2/CIBN blue light-induced dimerization to connect TEV fragments which are then able to free a membrane-anchored factor by cleaving a cryptic recognition site within a LOV2 domain. (B) LANSTRAP and CLASP improve upon LANS/LiNUS by introducing membrane sequestration of cytoplasmic fusion proteins via Zdk/LOV to more reliably and constantly maintain nuclear exclusion in the absence of blue light. (C) LINXnano improves upon the LINX/LEXY systems by fusing LINX to a minimal version of SspB, nano, which binds to membrane-tethered iLID in the presence of blue light. (D) Amg2 of the BICYCL system reversibly binds BAm Red when exposed to red light and BAm Green when exposed to green light, facilitating wavelength-directed transcriptional toggling by recruiting either activators or repressors to genomic regulatory regions. (E) The Blue-OFF system triggers transcriptional silencing and simultaneous degradation of a POI by combining an optogenetic degron fused to the POI with a repressor-bound LOV2 harboring a cryptic DNA-binding protein. (F) iLight utilizes IsPadC red-light-inducible homodimerization and isomeric steric hindrance of fused proteins in dark or far-red conditions to reversibly recruit TFs to gene loci. Red arrows represent time in decades since the sequencing of GFP in 1990, left, to the present, right. POI, protein of interest; TF, transcription factor.

Another tool in the modern genetic engineer’s toolbox is the light-activated reversible inhibition by assembled trap (LARIAT) strategy [213]. LARIAT consists of a CIB1-bound CaMKIIα and a CRY2-bound POI. Since CaMKIIα self-oligomerizes, exposure to blue light reversibly traps the POI in clusters. The PixE/PixD homo/heterodimer scheme has also been used to conditionally induce/dissociate subcellular protein droplets in the PixELLs system [219]. In PixELLs, blue light exposure diffuses phase-separated droplets formed in the dark through a combination of the liquid–liquid phase separation function of N-terminal intrinsically disordered protein region (IDR) of the human FUS protein (FUS^N) [214] and the oligomerization of PixE/PixD under dark conditions (see Fig 3J). A similar mechanism to PixELLs, developed in the same study, is optoDroplets which uses CRY2. In optoDroplets, FUSN is fused to CRY2 to form phase separated droplets when blue light is applied to promote oligomerization of CRY2 [215].

LANSTRAP and CLASP both improve upon the LANS system for LOV2-mediated nuclear import [216,217] (see Fig 4B). LANSTRAP uses a membrane-bound Zdk2 to directly bind LANS in the dark, whereas CLASP uses a membrane-bound LOV2 to trap a Zdk2:POI:LANS fusion in the dark, thereby improving nuclear exclusion of LANS pre-blue light exposure. On the other hand, LINXnano [157] fuses LINX to nano, a truncated version of SspB, which facilitates binding to a membrane-bound iLID in blue light (see Fig 4C). This improvement of the LINX system produces more complete nuclear export by tethering exported LINXnano:POIs within the cytoplasm. Similar to LANSTRAP and CLASP in function is iRIS, which sequesters the LOV2:NLS-fused POI in the cytoplasm via fusion to Q-PAS1 which facilitates binding to a membrane-anchored BphP1 when exposed to far-red light and nuclear import by NLS exposure when exposed to blue light [218].

Specific Protein Association tool giving transcriptional Readout with rapid Kinetics (SPARK) is a tool devised to conditionally report cells in which 2 POIs are interacting [219]. One POI is fused to a TEV protease while the other is fused to a modified LOV2:TEV cleavage site:TF. Upon exposure to blue light, the TEV cleavage site is exposed allowing for release of the TF and activation of reporter gene(s) in cells where the 2 POIs are interacting.

Cyanobacteriochrome-based light-inducible dimers (BICYCL) employs a modified light-induced isomerizing GAF (cGMP-specific phosphodies-terases, adenylyl cyclases and FhlA) domain derived from Acaryochloris marina AM1_C0023g2, Amg2, to shift binding between 2 binding partners when exposed to either red or green light [220]. Binding partners, binder of Amg2-red state (BAmRed) and binder of Amg2-red state (BAmGreen), were engineered such that Amg2 binds BAmRed when exposed to red light and to BAmGreen when exposed to green light, allowing for POI swapping via light exposure (Fig 4D). The developers of BICYCL showed that it can be used to conditionally recruit either a transcriptional repressor or activator to DNA-tethered Amg2 depending on laser color choice.

Blue-OFF combines LOV2-mediated transcriptional inhibition with psd-mediated POI degradation to more completely eliminate POI presence when exposed to blue light [221] (Fig 4E). LOV2 is used to create a cryptic DNA-binding domain and is further fused to a transcriptional inhibitor (KRAB) to eliminate transcription of the POI which is fused to B-LID for degron-mediated depletion. Analogously, iLight controls gene transcription via far red mediated homodimerization of the photosensory core module of Idiomarina IsPadC (IsPadC-PMC) fused to either LexA408 to block transcription or Gal4 and VP16 to trigger gene expression, which can be reversed via exposure to near-infrared light [222] (Fig 4F). In the dark or after exposure to near-infrared light, the fused TFs are sterically inhibited by isomeric transformations of IsPadC-PMC.

Discussion

While there have been many reviews on optical tagging systems, optical biosensors, optogenetics, and biologically expressed imaging tools more broadly [223], we have focused this compendium specifically on tools currently available for use in transgenic plant and animal studies. In a few decades, molecular biology and genetic engineering fields have exploded from complex transgenics being a hypothetical potentiality to thousands of transgenic organisms, hundreds of fluorescent proteins and tags, scores of subcellular sensors, and dozens of applications for several optogenetic systems.

It is important to carefully consider the limitations of each class of optical tool as well as for each individual instrument therein. For example, labeling with one of the tag systems described here comes with the disadvantage of occupying a laser and detector on the microscope merely to label, so it’s sometimes prudent to use a sensor also as a label which may come with the disadvantages of slow fluorescent maturation and low signal. Similarly, within classes of optical tools, different tools with similar functions have important advantages and limitations to consider. For example, degrons require transcription and translation of replacement fusion POIs to recover after depletion while translocation tools require mere shuttling within the cell to recover; however, degrons more quickly and completely deplete POI levels. Further, it’s necessary to contemplate the activation/inactivation times of optogenetic tools which often take a couple to tens of minutes to reach peak activation as well as their efficiencies of activity and reversibility which tend to land all over the map. In addition, when considering the use of sensors or tags in combination with optogenetic tools, it’s imperative to weigh the cost in photobleaching and phototoxicity against the reward of maintained optogenetic activation and tailor experiments accordingly. As discussed in the previous section, combinatorial applications of these tools are being developed which synergistically supplement each individual component’s limitations.

It is expected that we will soon see a surge in development of sophisticated tools and studies monitoring chromatin architecture, transcription, translation, protein localization, and cellular responses via reporters, tags, and sensors while simultaneously manipulating transcription or protein levels or function in living transgenic plants and animals. One such study has already reached preprint from the lab of Hernan Garcia manipulating transcription factors via LEXY while simultaneously monitoring transcription of their targets via MS2/MCP [224]. Another ongoing study from the lab of Sanjeevi Sivasankar seeks to combine optogenetics and BioID by fusing CRY2 or CIB1 to 2 halves of a split-TurboID biotinylation enzyme thereby allowing for precise temporal control of the enzyme’s activity [225].

Additionally, we have conceived of a few hypothetical advancements which we believe can be useful if developed. First, precise spatiotemporal control of epigenetic states has the potential to be an invaluable tool for interrogating development and disease and can feasibly be accomplished by AsLOV2-toggling of histone modifiers and genomic targeting via gRNAs and dCAS9 (see Fig 5A). Second, combining tags with other systems is sure to soon gain popularity; using SUN-tag to sequester transcription factors conditioned upon pre-cellularization conditions in the early Drosophila embryo, e.g., could be useful for specific enquiries into early development (see Fig 5B). Third, the BLITZ system can feasibly be modified to control overexpressed transgenic transcription factors which would otherwise accumulate in systems like LINXnano and confound expectations of spatiotemporal precision (see Fig 5C).

Fig 5. Future directions of transgenic optical tools.

(A) Histone modification by CRISPR (“HisPR”): LOV2 can hypothetically be used to conditionally expose a cryptic histone modifying enzyme targeted to genomic loci by dCAS9. (B) Jabba-trapped LINX via SunTag (“J-A-bba LINXsun”): Jabba-trap was developed to trap fusion proteins on lipid droplets dispersed throughout the pre-cellularization Drosophila embryo [226]. In conjunction with LINX and SunTag, it may be possible to precisely trigger nuclear export of TFs via blue light until gastrulation begins. (C) “Blue-ON”: Blue light releases a membrane-caged caspase fragment via PixE/PixD action. Caspase fragments assemble in similar fashion to BLITZ under blue light exposure via iLID/SspB action to release an NLS-fused nuclear factor via caspase recognition sequence exposure by LOV2. LANS:TF fusion is uptaken by the nucleus while exposed to blue light as in LANS. In the dark, the LANS construct is exported from the nucleus via NES, sequesters at mitochondrial-bound Zdk, and is degraded via fused degron sequence. While quite complex, this system would produce near complete silencing of overexpression models in the dark; further, once a transgenic is produced in one species, only the TF needs be replaced to produce tools for any other target. (D) Currently, it may be possible to use up to 7 transgenic optical tools simultaneously. LANS, light-activated nuclear shuttle; TF, transcription factor.

Nontoxic blue and red light are certainly most useful for in vivo studies in transgenic animals, complex systems encorporating varieties of optogenetic tools will require multiple lasers to be used in concert. It is exciting that several systems have already been developed which implement bohemian wavelengths for activation, including far-red, near-infrared, violet, and UV-B (see Fig 5D). With appropriate experimental design and controls, it is hypothetically possible to simultaneously use a handful of different lasers to affect different optical tools and achieve increasingly complex objectives.

We will undoubtedly see further exponential growth of the field of transgenic optical tool development and application in the coming years. Neurological and metabolic disorders stand to benefit greatly from future advances which will undoubtedly include translational applications with medical potential. Encouragingly, advances in ex utero [227–229] and in utero [230,231] experimental techniques and technologies promise to bring optogenetics to mammalian embryonic development, an important step toward this goal. Further, the genetically encoded devices described herein offer our best chance at supplementing our insufficient understanding of basic cell, developmental, and disease mechanisms which have hitherto remained inaccessible. This review is meant to be a brief almanac of the tools available to those who will develop and use these future models and devices.

Funding Statement

The study received support from University of Texas at Arlington, UTA STARS, to Dr TK. The funders had no role in the study design, data collection and analysis, decision to publish, or preparation of the manuscript.