Figure 1. Common strategies to design opto-chemical tools. (A) Photo-induced conformational change which is usually reversible. Photo-induced isomerization is shown as an example. (B) Photo-uncaging of small molecules. (C) Photo-uncaging of oligonucleotides. Cyclic caged morpholino is shown as an example. (D) Photo-uncaging of peptides and proteins of which single amino acid residues can be caged.
Essential types of natural biological processes such as vision and phototropism rely on photo-induced conformational changes. During these processes, chromophores switch between two or more isomeric forms and subsequently alter the activities of their associated proteins. The reactions usually occur in a reversible manner to produce many rounds of inactive/active states, and for that reason, these photosensors are commonly called photoswitches
[7]. Photoswitches can be generally categorized into two groups: photoreceptor derived, and non-photoreceptor based.
2.1. Photoreceptor-Derived Photoswitches
The most convenient opto-chemical tools exist in nature and do not require further chemical modifications. These natural molecules are mostly found in photoreceptors of various organisms and typically change conformations upon light illumination. For example, photoreceptors in animals and some microbes rely on the seven-transmembrane-domain proteins called opsins
[8]. Opsins use the retinal as a chromophore which isomerizes from the
cis to
all-trans form when exposed to light (
Figure 2A), and the
all-trans retinal is subsequently converted back to the
cis retinal by a series of enzymes in the retina
[9]. Another major chromophore harnessed by a host of bacteria and plants is the flavin group which includes riboflavin, flavin mononucleotide (FMN) and flavin adenine dinucleotide (FAD). Natural photoreceptors that require flavin groups include the light-oxygen-voltage sensing domain (LOV), sensors of blue-light using FAD (BLUF), and cryptochrome (CRY)
[10][11][12]. Flavins absorb light and covalently bind to the cysteine (Cys) at the LOV active site of the photoreceptor to induce a conformational change (
Figure 2B). Similar acting mechanisms apply to other natural chromophores such as tryptophan antennas from Arabidopsis UV-B resistance receptor
[13][14][15] and bilin analogs from phytochromes
[16][17].
Figure 2. Representative examples of photoreceptor and non-photoreceptor derived photoswitches. (A) Isomerization of retinal in opsins. (B) Light-induced binding of flavin to LOV domain of naturally occurring photoreceptors. (C) Isomerization of azobenzene between the Z and E conformations. Light-sensitive covalent bonds are colored red.
2.2. Non-Photoreceptor Based Photoswitches
Some biologically active molecules that are not coupled with natural photoreceptors in particular can also alter their activities upon light illumination. For example, when irradiated by UV light, retinoic acid, a critical signaling molecule during vertebrate development, isomerizes between
9-cis,
13-cis and
all-trans forms which possess distinct biological activities
[18][19][20]. Many naturally occurring dyes such as carbocyanine
[21], rhodamine
[22], and azobenzene
[23] compounds are photosensitive and shift their fluorescence spectra after undergoing conformational changes. Some fluorescent proteins such as Dronpa and Padron can also respond to light activation in a reversible manner
[24] while other photoactivatable fluorescent proteins irreversibly convert fluorescence such as Kaede, Dendra2, and mEos
[25]. In the laboratory, synthetic chemistry has modified this natural toolkit to render photocontrol much more dynamic and versatile. A good example is azobenzene and its derivatives which are widely employed to achieve activity control via photo-induced conformational changes. Responding to both light and temperature, azobenzene and its derivatives can isomerize between the
Z and
E conformations (
Figure 2C). By choosing different substituents on the benzene ring, the physical properties of azobenzene compounds such as solubility and excitation wavelength can vary greatly
[26][27][28], facilitating their applications to diverse needs. Numerous azobenzene-based photo-switchable components have since been engineered and applied to study various biological processes both in vitro and in vivo
[29][30][31][32][33][34][35][36]. Besides azobenzene-based photoswitches, other light-sensitive ligands
[37] such as stilbene
[38][39], hemithioindigo
[40][41], spiropyran
[42][43], and diarylethene-derived
[44] molecules have also been explored as synthetic photoswitches.
3. Light-Induced Uncaging
Photoswitches based on conformational change are valuable opto-chemical tools, however, the vast majority of biological processes are not conformationally regulated by light. A more powerful and generic opto-chemical tool is light-induced uncaging in which a biomolecule of interest is covalently linked to a light-responsive protecting (caging) group. Light irradiation initiates a photolytic reaction with subsequent release of the biomolecule to function normally. In principle, photo-uncaging allows any molecules, small or large, to be chemically modified to become photoactivatable. The critical component of photo-uncaging is the caging group. The ideal caging group needs to be evaluated on several criteria: (1) it should inactivate the caged biomolecule without producing any secondary activity or toxicity; (2) the caged biomolecule should be stable outside and inside the biological system unless irradiated by light of the proper wavelength; (3) the light used should not be deleterious to the biological system and the photolysis reaction should be fast and high yielding; (4) any by-products from photo-uncaging should also be non-detrimental and preferably inert in the biological system. A variety of caging groups have been developed that satisfy these criteria fully or partially, and some of the most successful and prevalent caging groups (
Table 1) include the 2-nitrobenzyl, coumarin, 7-nitroindolinyl and BODIPY derivatives
[45][46][47][48]. Other caging groups have also been synthesized and tested in various contexts
[49].
3.1. Photocaged Small-Molecule Actuators and Probes
Small-molecule actuators and probes are essential tools for perturbing and investigating biological processes. Photo-uncaging of bioactive molecules provides additional control at the spatial and temporal level. The concept of small-molecule photo-uncaging dates back to the 1970s when cyclic adenosine monophosphate (cAMP)
[50] and adenosine triphosphate (ATP)
[51] were first photocaged. Since then, a wide range of small molecules including inhibitors, agonists, metabolites, and probes have been caged and successfully photo-activated in various biological systems
[45][46][47][49][52][53][54][55][56][57]. Neurotransmitters were among the first biomolecules to be photo-uncaged to precisely control neuronal activities. By adding an ⍺-(4,5-dimethoxy-2-nitrobenzyl) group to glutamate, the caged molecule can be locally released via UV illumination
[58]. Improved photo-uncaging of glutamate
[59][60][61] as well as other neural signaling molecules such as γ-aminobutyric acid (GABA)
[62][63] and calcium (Ca
2+)
[64][65] has also been achieved. In other cases, physiological metabolites such as inositol
[66] and retinoic acid
[67] can be photocaged to obtain spatiotemporal activation while photocaged dyes
[68][69] have also been synthesized to facilitate biological labeling and imaging. Notably, small-molecule ligands that regulate signaling pathways or gene transcription have been broadly and fruitfully caged to render photocontrol. For example, rapamycin was caged to optically control the heterodimerization of the FK506 binding protein (FKBP) and FKBP-rapamycin binding protein (FRB)
[70]. Caged doxycycline
[71][72] and tamoxifen
[73][74] were synthesized to activate gene expression with light. To improve photostability, caged 4-hydroxy-cyclofen (Cyc)
[75][76][77][78] has also been developed to optimize its use in most physiological conditions. Optical control of gene expression with caged small molecules directly against mRNA is scarce, but recently, synthetic 5′ cap analogues with photo-cleavable groups have been developed
[79]. UV irradiation efficiently releases the 5′ cap to interact with the eukaryotic translation initiation factor 4E (eIF4E), prompting the start of translation.
While photo-sensitive small molecules are mostly engineered towards their photo-activation, photo-deactivation approaches are less common. One method for photo-deactivation is chromophore-assisted light inactivation (CALI) in which engineered proteins and light-sensitive dye molecules produce reactive oxygen species (ROS) to deactivate biological systems upon light absorption
[80][81]. However, due to the many limitations of the technique such as inconvenient cellular delivery and off-target ROS toxicity, CALI has not been broadly applied, especially in vivo. Another photo-deactivation approach with caged small molecules can be achieved through protein degradation. For example, a photoactivatable auxin has been shown to act as a photoactivatable inducer of protein degradation in mammalian cells in which transposing components of the plant auxin-dependent degradation pathway were genetically introduced
[82]. Small-molecule degraders such as proteolysis-targeting chimeras (PROTACs) have also been caged to recruit the E3 ligase to the target protein upon light activation, resulting in ubiquitination and protein degradation
[83][84]. Alternatively, PROTACs have also been designed as photoswitches that can be reversibly activated at different wavelengths
[85][86][87].
Caging of small molecules is by far the most widely used photocaging technique due to the relatively straightforward yet highly effective design and synthesis. They are also ideal molecular tools for zebrafish study as most small molecules can freely diffuse in embryonic tissues and are readily accessible to light illumination.
3.2. Caged Oligonucleotides
As many biological programs are initiated at the genetic level in cells, the ability to directly control DNA or RNA provides an additional layer of precision in studying biological processes. The idea of using nucleic acids as tools to control biological activity was realized soon after the “central dogma” was first proposed by Francis Crick in the late 1950s. However, developing light-activable oligonucleotides is challenging due to the large size and structural complexity
[88]. Sub-optimal caging of oligonucleotides leads to either activity leakage while caged or insufficient activity upon photo-uncaging. Nevertheless, there have been great and promising developments in the synthesis of photocaged oligonucleotides in the past decades.
The first working example of photocaged oligonucleotide was obtained by adding several hundred 1-(4,5-dimethoxy-2-nitrophenyl) diazoethane groups to plasmids coding for luciferase and green fluorescent protein (GFP)
[89]. Similarly, photo-uncaging of mRNA has also been reported to control the expression of
GFP and
Engrailed2a, a gene involved in zebrafish eye and brain development
[90]. Efforts have been also made to cage small interfering RNA (siRNA). SiRNA is a class of double-stranded RNA of 20–24 base-pair nucleotides, and it interferes with gene expression by specifically targeting and degrading its complementary mRNA. Caged siRNAs were synthesized by adding photo-activable groups to the phosphate backbone
[91], the 5′ terminal phosphate of the antisense strand
[92] or the central region
[93]. Further optimization of photocaged siRNAs that used new photolabile groups to induce tighter control have also been reported
[94][95][96]. In zebrafish studies, the most common anti-sense knockdown tools use morpholino oligonucleotides (MOs)
[97][98]. Instead of ribose rings, MOs contain morpholino rings and are neutrally charged by replacing the phosphate backbone with phosphorodiamidates
[99]. MOs do not bind to the RNA-induced silencing complex (RISC). Consequently, MOs can effectively inhibit translation without inducing RNA degradation. Caged MOs were first developed by tethering complementary oligomers through photolabile linkers and were tested in zebrafish to silence gene expression of chordin and no tail (Ntl) in zebrafish embryos
[100][101]. To avoid the use of multiple caging groups, MOs have also been circularized with a photocleavable linker joining the 3′-amine and 5′-carboxylic acid ends of linear oligonucleotides
[102][103]. While most caged MOs (cMOs) are activated by UV light, various photolabile caging groups allow for non-UV photolysis and enable sequential control through wavelength-selective illumination
[104]. MOs and cMOs have been widely used for gene knockdown in zebrafish particularly before the advent of other generic and targeted tools for gene manipulation
[100][105][106][107]. With the increased popularity of gene editing tools such as the clustered regularly interspaced short palindromic repeats (CRIPSR)/Cas system, methods to cage guide RNA (gRNA) have also been developed, enabling precision control of gene editing and transcription
[108][109][110][111][112]. Reversely, gRNAs can also be modified to deactivate CRISPR/Cas by light. For example, a recently developed CRISPRoff method
[113] incorporates photocleavable
o-nitrobenzyl groups to gRNAs which can then go through photo-induced degradation and inactivate the editing machinery. Photo-induced uncaging of a single nucleotide in the CRISPR RNA (crRNA) with subsequent release of truncated crRNAs with 15 or fewer nucleotides of target complementarity has also been shown to be sufficient to enable light-induced deactivation
[114].
3.3. Caged Peptides and Proteins
Peptides and proteins are important players in many cellular processes and are critical in maintaining homeostasis and regulating signaling pathways. Although the activities and expression of peptides and proteins can be optically manipulated by caged small molecules and caged oligonucleotides as described above, both methods only allow for indirect control of their protein targets with concomitant uncertainty in both space and time. For example, transcription and translation can take many minutes while uncaged small molecules can diffuse and affect cellular compartments that are not directly light activated. In contrast, photo-induced uncaging of peptides and proteins is instantaneous, and its effect is largely limited to a confined space set by the light. Such direct and immediate control is especially pivotal for studying biological processes that rely on fast dynamics.
Caged polypeptides and proteins typically rely on the photo-induced cleavage of C-X (i.e., C-O, C-N and C-S, but not C-C) bonds. Therefore, the amino acid residues that are usually caged are limited to those with polar or charged side chains such as cysteine (Cys), lysine (Lys), serine (Ser), tyrosine (Tyr), glutamate (Glu), glutamine (Gln), aspartate (Asp) and asparagine (Asn)
[48]. Other non-polar residues such as glycine (Gly) can be caged only at the C- or N- terminus using a terminal backbone caging strategy
[115]. Introducing photocaging groups to polypeptides has been accomplished in both solid-phase and solution-phase synthesis. The earliest report of a photocaged peptide describes the addition of an
o-nitrobenzyl group to L-leucyl-L-leucine methyl ester
[116], a small peptide-based lysosomal damaging agent that induces apoptosis in mast cells. Longer peptides were later successfully photocaged, targeting many amino acid residues including Cys, Ser, Lys, Glu and Asp
[117][118][119][120]. Although proteins are chemically similar to peptides, caging proteins through synthetic chemistry has turned out to be much more challenging. Proteins usually possess complex structures and modifications of their amino acids could render the perturbed protein inactive or unstable. Proteins also generally require aqueous conditions, further limiting the chemical techniques that can be used for photolabile caging. Nonetheless, a few examples of synthetically caged proteins have been reported. A photoactivatable hen egg lysozyme was made by total chemical synthesis
[121] and site-selective caging of whole proteins has also been achieved
[122][123][124][125].
Compared to synthetic chemistry, a more powerful and established method to cage peptides and proteins has been engineered through genetic code expansion that allows incorporation of unnatural amino acids (UAAs)
[126][127][128][129][130]. Genetic encoding of UAAs is made possible with installation of orthogonal transfer RNA (tRNA) synthetases and their cognate tRNAs incorporating UAAs in response to the one of three stop codons (UAG, UAA and UGA). To minimize interference with the endogenous translational machinery, a stop codon that is used less frequently by endogenous tRNA synthetase/tRNA pair is preferred. As the amber codon (UAG) is the least used in
Escherichia coli (7–8%), the organism in which most proteins are expressed and purified in the laboratory, it is the most preferred one for the incorporation of UAAs
[130]. The use of the ochre (UAA) and opal (UGA) stop codon is less common but has also been implemented
[131]. Based on the genetic code expansion technology, caged amino acids critical for protein functions could be inserted in a protein of interest with little change to its overall structure. Light irradiation releases the caging groups from corresponding amino acids and restores the activity of the protein. The first photocaged protein implementing genetic incorporation was the muscle nicotinic acetylcholine receptor in which the tyrosine residue was replaced with
o-nitrobenzyl tyrosine
[132]. That study also suggested that the receptor could respond differently in
Xenopus oocytes when the tyrosine was caged at different positions, stressing the importance of rational caging design and downstream validation. Encouragingly, the genetic incorporation of photolabile UAAs to proteins has proven widely useful in the past decades, and photoswitchable proteins have also been engineered with azobenzene-based caging groups
[133]. This unique opto-chemical tool for controlling proteins has been extensively employed in the study of numerous biological processes
[134][135][136] including control of ion channels
[132][137][138][139], kinase and phosphatase activities
[140][141][142], gene expression
[143][144], epigenetic regulation
[145][146], DNA recombination
[147], protein localization
[148][149], protein self-organization
[150], intein splicing
[151][152], and O-linked-N-acetylglucosaminylation
[153].
Table 1. Some common caging groups used in photocaging. Leaving groups (X) are colored red.