Accepted Manuscript Title: Targetable fluorescent sensors for advanced cell function analysis Author: Shin Mizukami PII: DOI: Reference:
S1389-5567(16)30065-X http://dx.doi.org/doi:10.1016/j.jphotochemrev.2017.01.003 JPR 257
To appear in: Reviews
Journal of Photochemistry and Photobiology C: Photochemistry
Received date: Revised date: Accepted date:
14-9-2016 17-1-2017 20-1-2017
Please cite this article as: Shin Mizukami, Targetable fluorescent sensors for advanced cell function analysis, Journal of Photochemistry and Photobiology C:Photochemistry Reviews http://dx.doi.org/10.1016/j.jphotochemrev.2017.01.003 This is a PDF file of an unedited manuscript that has been accepted for publication. As a service to our customers we are providing this early version of the manuscript. The manuscript will undergo copyediting, typesetting, and review of the resulting proof before it is published in its final form. Please note that during the production process errors may be discovered which could affect the content, and all legal disclaimers that apply to the journal pertain.
Targetable fluorescent sensors for advanced cell function analysis Shin Mizukami* †Institute
of Multidisciplinary Research for Advanced Materials, Tohoku University,
2-1-1 Katahira, Aoba-ku, Sendai, Miyagi 980-8577, Japan * Corresponding author at: Institute of Multidisciplinary Research for Advanced Materials, Tohoku University, 980-8577, Japan. E-mail address: [email protected]
Highlights * Two fundamental strategies for fluorescent sensor targeting were reviewed. * A small-molecule-based targeting was used for in vivo imaging of osteoclast activity. * A bacterial enzyme-based protein labeling technology was introduced.
1. Introduction When Roger Y. Tsien reported the first fluorescent sensor for free Ca2+ , such chemistry-based bioimaging techniques opened a new era of cell biology. Newer Ca2+ probes, in particular Fura-2 , have contributed enormously to the elucidation of intracellular Ca2+ signaling events . The evolution of these tools essentially created a new field of biology called “chemical biology” and in the process formed a new relationship between physiologists and synthetic chemists. In the period from the 1990s to 2000s, many fluorescent sensors were developed and have subsequently been used in live cell imaging experiments ; in some cases, near-infrared sensors have been used in in vivo studies [5,6]. Soon after fluorescent proteins (FPs) became familiar tools for molecular biologists, two FP-based sensors were developed for Ca2+ imaging [7,8]. Both sensors took advantage of the fluorescence resonance energy transfer (FRET) phenomenon, whereby the energy of an excited donor fluorophore is non-radiatively transferred to an acceptor fluorophore . These protein-based sensors are genetically encoded; therefore, they can be targeted to specific subcellular domains by linking a signal peptide sequence with the protein sensors [10–12]. In comparison, classical small-molecule sensors do not have this targeting property. However, small-molecule sensors have several advantages over protein sensors, including an established probe design strategy . For example, the rational design of a FRET-based sensor would also be effective for small-molecule sensors . Photoinduced electron transfer (PeT) and internal charge transfer (ICT) have also often been utilized as other rational design strategies for small-molecule fluorescent sensors [15,16]. In addition, various physical properties of small-molecule sensors such as a large dynamic range of the fluorescence intensity, excitation/emission wavelength, or dissociation constants to the target are tunable by the organic synthesis-based molecular design. Some organic fluorophores such as ATTO647 are much more photo-stable than FPs, and are thus applicable to strong laser-based imaging such as STED (stimulated emission depletion) microscopy . Therefore, small-molecule fluorescent sensors are still useful for a variety of biological applications, and thus it is of importance to develop advanced functional, target-specific small-molecule fluorescent sensors to address these needs (Fig 1). Strategies to develop target-specific small-molecule sensors can be classified simply into two approaches. One utilizes the intrinsic properties of fluorophores or other small molecules for targeting specific regions such as subcellular organelles. The second exploits the combination of a
genetically encoded tag and its specific ligand. In this review, I will briefly discuss these two probe targeting strategies and then introduce our recent studies in this area.
2. Probe targeting strategy based on the intrinsic properties of small molecules 2.1 Targetable fluorescent sensors based on the intrinsic properties of small molecules In the simplest case, a fluorophore may possess the property of being able to be targeted to a subcellular location. Over the last several decades, various organelle-specific fluorescent probes have been reported . Fluorescent probes such as DAPI (1), MitoTracker (2), LysoTracker (3), ER-Tracker (4), and their derivatives (Fig 2) are commercially available and have been widely used for targeting nuclei, mitochondria, lysosomes, and the endoplasmic reticulum, respectively. An examination of the structure of MitoTracker Orange revealed that this compound is a derivative of rhodamine, somewhat like tetramethylrhodamine, which is one of the most popular fluorophores in biology. Rhodamines have a carboxy group at the ortho-position of the 9-substituted phenyl group; thus, they exist as neutral intramolecular salts. When the carboxylate anion is removed or modified to an ester, the total charge on the molecule become positive. In these cases, the fluorophores tend to accumulate within mitochondria. Cyanine family or other cationic fluorophores are also known to accumulate in mitochondria [19,20]. By utilizing the nature of the fluorophores, rhodamine-based sensors for Ca2+ (5a) as well as Zn2+ (5b) and cyanine-based pH sensors (6) have been reported to be mitochondria-targetable fluorescent sensors (Fig 3) [21–24]. Hamachi et al. also developed a pyronine-based nucleotide sensor (7) for targeting mitochondria . Mitochondria have two lipid membranes, an outer membrane and an inner membrane. Unlike the outer membrane, the inner membrane is highly impermeant to molecules. Numerous transporters and ion channels are therefore required to allow molecules to move in and out of the mitochondrial matrix. As a result, the mitochondrial inner membrane has a membrane potential that is maintained at about −180 mV. As a distinguishing feature of this organelle, a “lipophilic cation” has become a common structural feature for targeting molecules to the mitochondria. Fluorophores that have a tendency to accumulate in mitochondria, such as some rhodamine derivatives, generally meet this criterion. However, a more reliable method for organelle targeting is through conjugation with an organelle-targeting small molecule, despite the fact that this method increases the molecular weight. The most popular mitochondria-targeting agents are triphenylphosphonium (TPP) salts [26,27]. This strategy has proven to be very useful in the molecular design of mitochondria-targeting sensors for H2O2 , Zn2+ (8) , Cu+ (9) , Mg2+ (10) , and H2S (11)  (Fig 4).
The lysosome is one of the most important digestive organelles in cells, where various macromolecules are degraded. The pH in lysosomes is maintained in the acidic range (pH 4.5–5.0) due to the action of a series of proton pumps. Lysosomes are surrounded by a lipid membrane; thus, lysosome-targeting compounds need to exist in part in the non-charged form at neutral pH. Tertiary amines are widely used lysosome-targeting groups, such as an N,N,-dimethylaminoethyl group and a morpholinoethyl group (Figure 5). At cytosolic neutral pH, these amino groups are partially protonated, so that their neutral portions can permeate into the lysosomes. Once entering acidic compartments, these functional groups are almost fully protonated, and thus the hydrophilicity of the total compounds increases and the compounds become trapped in the compartments. Since visualization of the metal ions in the lysosome is very important for understanding the metal-related biological processes in cellular digestion, various researchers have tried to develop lysosome-targetable fluorescent sensors. To date, some Zn2+-sensors (12, 13) have been targeted to the lysosome [33, 34] (Figure 6). By using other turn-on fluorescent sensors based on the same targeting mechanism, NO (14) , H2O2 (15a) , HOCl (15b) , and H2S (15c)  were successfully imaged in the lysosome. The plasma membrane is another subcellular compartment to which fluorescent sensors have been targeted. In particular, the local activities of plasma membrane ion channels are of interest to researchers studying the physiological role of metal ions. In order to image the ion concentration near the plasma membrane, plasma membrane-targeted sensors have proven useful. In order to target fluorescent sensors to plasma membranes, a long alkyl chain [23, 39] or cholesterol  has been conjugated to the sensor either with or without a linker (16–18, Fig 7). When the fluorescent sensor moiety is very hydrophilic, the probes anchor at the outer leaflet of the plasma membrane. However, selective targeting to the inner leaflet is difficult to achieve using this approach. For this purpose, a strategy using a genetically encoded tag and its ligand, which is discussed in the next section, is effective. The nucleus is another important organelle for the targeting of various fluorescent sensors. The nucleus is different from the other organelles discussed above, because it is segregated from the cytosol by the nuclear membrane; however, the membrane has many nuclear pores, which are composed of various proteins such as nucleoporins. Therefore, small-molecule probes can freely permeate through the nuclear membrane. The most abundant target substance of the nucleus is DNA, to which many fluorogenic probes such as DAPI and Hoechst dyes are selectively localized. Nevertheless, reports of nucleus-targetable fluorescent sensors are scarce. Only a few sensors that can detect reactive oxygen species (ROS) (19)  or H2O2 (20)  in the nucleus have been
reported to date (Figure 8). In these cases, a pyrrole polyamide or a polycationic peptide, which shows high affinity to DNA, was exploited. Surprisingly, there are almost no reports about metal ion fluorescent sensors with specific nuclear targetability based on this strategy, which is probably because metal ion chelators generally possess multiple anions, which show electrostatic repulsion to DNA and may promote the export from the nucleus after the cleavage of protective groups such as the acetoxymethyl group. Despite all this, this approach has practical limitations; namely, the actual number of small molecule-based methods for the subcellular targeting of fluorescent sensors is low and only a few target organelles such as the mitochondria, lysosome, and plasma membrane have been reported to be successfully targeted. One of the reasons for these limitations is thought to be that the subcellular localization property is affected by both the targeting moiety as well as the fluorescent sensor moiety as discussed in the above paragraph related to nucleus-targeting sensors. It is therefore difficult to target the sensors to specific organelles when the localization signal from the sensing dye is strong. In targeting mitochondria or the plasma membrane, the hydrophobicity of general fluorophores assists to achieve a desirable probe localization. Another reason for the low number of targeted fluorophores based on intrinsic properties is that an alternate strategy exists whereby identical fluorescent sensors can be targeted to various organelles through the combination of a genetically encoded tag and its specific ligand.
2.2 Fluorescent pH sensors targeted to the bone surface in living animals Given that in vivo genetic manipulation is considerably more complex and time consuming than that in vitro, small molecule probes have promising in vivo applications. Although a hybrid approach, whereby fluorescent sensing probes and drug delivery systems are combined, is expected to be achieved in the near future, few examples have been reported recently [43, 44]. In our own studies, we have begun to develop new targetable probes to investigate specific cellular functions in vivo. We have chosen to focus on the targeting of a pH sensor to the surface of bone in living animals in order to image osteoclast activity. Changes in osteoclast activity are related to a variety of bone diseases such as osteoporosis and bone metastasis. However, intravital imaging of osteoclast activity has not yet been achieved, although the attainment of this novel technology could lead to progress in the study of the pathology of bone diseases and hopefully result in the discovery of new drugs to treat them. To visualize osteoclast activity, we chose to focus on the mechanism by which bone resorption by osteoclasts occurs. Bones are composed of inorganic materials such as Ca3(PO4)2 and proteins such
as type I collagen; activated osteoclasts degrade bone tissues by extruding both protons (i.e. increased acidification) and proteases in order to dissolve inorganic materials and proteins, respectively. Therefore, osteoclast activity can be monitored by observing changes in acidity between the osteoclast and bone tissue (Fig 9). Although a large number of fluorescent pH sensors have been reported [45, 46], BODIPY-based probes were chosen because BODIPY dyes have excellent photo-physical properties such as strong fluorescence intensity, i.e. large molar extinction coefficients, and high fluorescence quantum yields . In addition, BODIPY dyes have been used in numerous biological applications . Urano et al. demonstrated that BODIPY-based fluorescent sensors were applicable to in vivo imaging experiments . In order to target the fluorescent probe to bone tissue, drugs used in the clinical treatment of osteoporosis, which include a bisphosphonate structure, were chosen. Some bisphosphonate drugs such as pamidronic acid and alendronic acid (Fig 10a) contain a primary amine group, which can be conjugated with various other functional groups. Frangioni et al. reported that a pamidronate-conjugated near-infrared fluorophore was specifically distributed to bone tissues in a living animal . Therefore, we designed bone-targetable fluorescent pH sensors shown in Fig 10b. Since the pH of the resorption pit formed by an osteoclast decreases to around 4.5 , the pKa values of the pH sensors were targeted to be between 4.5 and 6.5. Two pH sensors that were synthesized, BAp-M (21a) and BAp-E (21b), had pKa values of 4.5 and 6.2, respectively, with the difference being attributed to the structures of the substituted alkyl groups . The bone targetability of the sensors was then assessed using imaging of bone tissues in vivo. To achieve this, two-photon excitation microscopy (TPEM) was adopted, because this technique can image relatively deep regions due to the use of near-infrared excitation. Real-time imaging of osteoclasts was performed in the thin parietal bone using TPEM. One advantage of using TPEM is that two-photon excitation of bone tissues induces the second harmonic generation (SHG), which enables the observation of bone tissues without staining. In this study, we used a transgenic mouse, in which mature osteoclasts were labeled red with the fluorescent protein tdTomato. In order to assess the bone-targeting property of the sensors, the control, “always-ON” probe without the pH sensing property, BAp-A (21c), was synthesized. When BAp-A was injected subcutaneously, the probe’s green fluorescence was successfully distributed over the bone surface (Fig. 11). Injection of BAp-E resulted in the appearance of green signals due to BAp-E, whereas in the case of BAp-M, the fluorescence intensity was weaker. This indicated that the pKa value is important to be able to detect acidic compartments with a high signal-to noise ratio. As a simple estimate, the brightness of BAp-E
was 1.2–7.5 times as intense as that of BAp-M between pH 4 and pH 6. This bone-targeted bisphosphonate-based pH sensor provided a first in vivo demonstration of osteoclast activation and thus provided a novel technology in bone biology. More recently, this new pH sensor BAp-E has been able to provide a new insight about the relationship between the motility of mature osteoclasts and their proton secreting activity . Despite this early success, BAp-E was found to have a severe limitation in terms of its photo-stability, thereby limiting its use in more versatile applications such as long-term imaging. Of particular importance, the photo-stability of a dye is an essential factor for TPEM or other laser-based imaging technologies such as super-resolution microscopy or single-molecule fluorescence microscopy. To address this problem, a second-generation of targetable pH sensors, the pHocas series (22a–c, Fig. 10c), was developed . The most promising probe was pHocas-3, with a pKa of 6.1 that offered the advantages of being more photostable and was also more resistant to ROS than BAp-E. Improvement of photo-stability enabled its use in various advanced measurements such as 3D imaging and long time-lapse imaging (Fig. 12). Although the binding of bisphosphonate to bone tissues is non-covalent, the strong coordination binding is effective for long-period intravital imaging. Long time-lapse imaging revealed the appearance and disappearance of the fluorescent signal beneath the osteoclast membrane, which indicated pH lowering and neutralization, respectively. To summarize, the studies have provided new and important information relating to osteoclast activity in vivo as well details about the relationship between osteoclast dynamics and proton-secreting activity.
3. Probe targeting strategy based on genetically encoded tags In this section, another strategy for targeting chemical sensors based on genetically encoded tags and their ligands (Fig 13) will be reviewed. Protein labeling approaches that use genetically encoded tags are currently used in a wide variety of applications [54,55]. Two types of genetically encoded tag systems are currently in use: covalent and noncovalent labeling systems. Bond stability is a very important parameter for targeted chemical sensors, especially in the case of long time-lapse imaging experiments. For this reason, the majority of reports concerning targeted chemical sensors using a genetically encoded tag report the use of a covalent tag.
3.1 Targetable fluorescent sensors based on genetically encoded tags Tsien et al. pioneered genetically encoded tag technology by developing the pairing of a
tetra-cysteine tag along with its covalent ligand FlAsH-EDT2 (23) (Figure 14a) . One of the advantages of this system is the small size of the tag, which is a small peptide containing only a 6-amino acid sequence. Another advantage is that FlAsH-EDT2 is almost non-fluorescent until it is labeled to the tag. As other covalent peptide tags, enzymatic conjugation using enzymes such as biotin ligase, transglutaminase, and lipoic acid ligase have also been reported [57–59]. In regard to peptide tag-based sensor targeting, the first targetable sensor was named Calcium Green FlAsH (24) (Figure 14b), which was targeted to the intercellular or intracellular specific domains in order to monitor local Ca2+ concentration change . As a similar example, a FlAsH-based Mg2+ sensor (25) was developed to target actin filaments and mitochondria . In this study, a tetracysteine tag was tandemly expressed with an FP, and this strategy enabled intramolecular calibration of the fluorescence intensity of non-ratio-metric sensors using FP fluorescence. Another proposed approach was based on the use of protein tags. Johnsson et al. developed a pioneering protein-tag, SNAP-tag . This tag system is based on a 20-kDa DNA-repair enzyme, human O6-alkylguanine-DNA alkyltransferase (hAGT), which specifically reacts with O6-alkylated guanine by irreversibly transferring the alkyl group to the reactive cysteine residue (Cys145). SNAP-tag is a mutant protein developed by directed evolution of hAGT to have specificity to benzylguanine with high reaction kinetics. Therefore, SNAP-tag quickly reacts with benzyl guanine derivatives and forms a stable covalent adduct (Figure 15a) . Another popular protein tag is HaloTag , which was developed by researchers at Promega Corporation. HaloTag is a non-catalytic mutant of a 34-kDa bacterial enzyme, haloalkane dehalogenase. The wild-type enzyme hydrolyzes haloalkane by the nucleophilic substitution of aspartate, the catalytic center, and the subsequent hydrolysis by water activated by a histidine residue. HaloTag is a noncatalytic mutant that has lost the second hydrolysis process, and rapidly forms a covalent adduct with the chloroalkane substrate (Figure 15b). Compared to a peptide tag-based approach, protein tag systems such as SNAP-tag and HaloTag have proven to be more popular for a wide variety of practical applications, probably because the small-molecule ligand quickly forms a covalent bond with the protein tag under physiological conditions. Lippard et al. utilized SNAP-tag for targeting a fluorescent Zn2+ sensor (26) to subcellular organelles . They synthesized a genetically targeted chemical sensor (ZP1BG) by coupling the sensor with O-(4-aminomethylbenzyl)guanine, a SNAP-tag ligand (Fig 16). ZP1BG targeted the mitochondria and the Golgi apparatus. Johnsson, the inventor of SNAP-tag system, also reported Ca2+ sensors capable of targeting the nucleus, cytosol, or mitochondria (27, 28) [66, 67]. Targetable
O-benzylcytosine derivatives, CLIP-tag , have been developed for detecting H2O2 (29, 30) [69,70] or gaseous biomolecules such as NO (31)  and H2S (32, 33) [72,73]. One of the advantages of using a genetically encoded tag is that a ligand-conjugated sensor can be localized at various organelles without modification of the sensor chemical structure. Summarizing the above, reported intracellular sites targeted by SNAP/CLIP-tag-based sensors include the mitochondria, plasma membrane, nucleus, Golgi apparatus, endoplasmic reticulum, lysosome, phagosome, and cytosol. HaloTag-based targeted sensors have also been reported (Fig. 17). Li et al. reported localization of a fluorescent Zn2+ sensor (34)  capable of monitoring Zn2+ release as a surrogate for insulin release from intracellular secretory granules . In addition, a Ca2+ sensor (35)  and a K+ sensor (36)  conjugated with the HaloTag ligand were targeted to the cell surface and nucleus, respectively. The basic chemical characteristic of the HaloTag system is that the ligand, i.e. the 6-chlorohexyl group, is neutral and relatively small. This is an advantage for both probe synthesis and for use in live cell experiments, in which membrane permeability is very important. One rare example of the development of a sensor target based on non-covalent labeling is based on use of the leucine zipper . In this example, a basic leucine zipper peptide (ZIP) was chemically conjugated with a fluorescent pH sensor, and a ZIP-binding peptide (ZBP) was genetically fused with an extracellular domain from the glutamate receptor. Following glutamate receptor labeling with the pH sensor-ZIP conjugate, pH changes during both endocytosis and intracellular trafficking of the glutamate receptor could be visualized.
3.2 Molecular design and application of fluorogenic labeling probes In the above subsection, I have reviewed a second sensor-targeting strategy based on a genetically encoded tag technology and showed that the melding together of chemical probe design and genetic engineering can produce powerful bioanalytical tools. In addition to SNAP-tag or HaloTag, other protein labeling technologies are available, and so I expect new bioanalytical tools exploiting the advantages of these technologies will be developed. As a part of this new direction, our group has explored a protein labeling system using a bacterial tag. To develop a practical protein labeling technology that is useful in biological experiments with mammalian cells, the system has to meet the following criteria: (1) a bio-orthogonal tag, meaning that fusion of the tag with a protein of interest (POI) does not affect the function of the POI or other biomolecules; (2) a bio-orthogonal ligand; and (3) the ligand displays rapid binding kinetics under physiological condition. To achieve bioorthogonality in mammalian cells, plant or bacterial
peptides/proteins are good candidates as tags. In addition, a small molecular size is preferable for both tags and ligands. For these reasons, we chose to focus on a mutant of β-lactamase called TEM-1. TEM-1 is a small bacterial enzyme that hydrolyzes β-lactam compounds such as penicillin and cephalosporin. TEM-1 degrades the β-lactam substrate in two enzymatic steps, acylation (nucleophilic attack) and deacylation (hydrolysis). Since the amino acid Glu166 of TEM-1 is important in the second step, a mutation (E166N) produces an enzyme that is unable to proceed further than production of the enzyme-substrate adduct . As a result of this, the mutant protein can be used as a protein tag (Fig 18). Using this approach, we have synthesized a variety of fluorophore-conjugated β-lactam compounds, and verified their use as protein labels. This labeling system is called the BL-tag system and utilizes a variety of β-lactam antibiotic compounds that have already been identified and for which some synthetic precursors are commercially available. For our purposes, we have used three kinds of β-lactam antibiotics, penicillin [79–81], bacampicillin , and cephalosporin [79,83,84], as ligands (Fig 18). The labeling rates of synthesized ligands were fast enough for practical use in live cell experiments as well as in commercial labeling studies. Since an overview of this approach has been described in another review , I focus here on the use of cephalosporin ligands. In general, a commercial labeling kit requires the washing out of unlabeled probes prior to imaging experiments. However, this process is sometimes unsuccessful due to the nonspecific adsorption of hydrophobic ligands, leading to a high background. One way to overcome this problem is the use of fluorogenic probes, which do not emit fluorescence until after tag labeling. We developed a fluorogenic labeling assay by exploiting the property of cephalosporin, which is the substituent elimination following β-lactam ring opening. The fluorogenic labeling probe design was based on the substituent elimination induced by hydrolysis of cephalosporin lactam (Fig. 19), as well as some fluorescent sensors detecting β-lactamase activity [86–88]. The first-generation fluorogenic labeling probe CCD (37, Fig 20) contained three components, coumarin, cephalosporin, and DABCYL, in which the latter quenches coumarin fluorescence as a FRET acceptor . As a result of this, the fluorescence of CCD is very weak (Φ = 0.005). As expected, fluorescence was increased after labeling of the BL-tag. However, practical improvements were needed mainly because of two drawbacks: first, the slow rate of quencher elimination (first-order kinetic parameter of auto-elimination reaction was 7.2 × 10−5 s−1), and second, low fluorescence intensity after the labeling. Subsequently, an improved labeling probe, FCAPO2 (38), was developed via substitution of the fluorophore from coumarin to fluorescein and by changing the quencher from DABCYL to
2-(4-dimethylaminophenylazo) pyridinium . FCAPO2 had a large fluorescence enhancement factor and a quick elimination after tag labeling (Fig. 21a). The elimination rate was so fast that the first-order elimination constant was not determined. The second-order labeling rate was estimated to be 7.8 × 104 M−1 s−1, which is large enough for practical use in live cell experiments at probe concentrations in the hundreds of nanomolar to micromolar range. This probe enabled no-wash protein labeling within a few minutes under fluorescence microscopic imaging (Fig. 21b). As an application of this approach, the real-time translocation of a BL-tag-fused EGF receptor protein from the intracellular region to the cell surface was viewed by imaging (Fig. 21c). Another fluorogenic labeling strategy uses static quenching of the fluorophore and the quencher. In our studies, the static quenching-based fluorogenic labeling probe FCDNB (39) uses dinitrobenzene as the quencher [83,89]. As a third strategy, we also reported a photo-induced electron transfer-based fluorogenic probe CC3DNB (40) . In this case, the distance between the fluorophore and quencher is required to be very close and intramolecular aggregation of the two components is sterically prohibited. One of the advantages of non-FRET-type strategies is that some quencher compounds become universal quenchers for a wide range of fluorescent dyes. In the case of a static quenching strategy, various fluorophores such as coumarin, fluorescein, and rhodamine are quenched by the same quencher. Because of such potential versatility, the non-FRET type of fluorogenic labeling strategy is attracting increasing attention [91–93].
4. Summary and perspective In this review, I briefly introduced recent studies describing the targeting of small-molecule fluorescent sensors to specific sites using a molecular design strategy. Current targetable sensors are classified into two categories. The first category is based on the intrinsic properties of small molecules such as the fluorescent sensors used for localization of specific organelles, including mitochondria, as reviewed in Section 2.1. As an example of challenging in vivo applications based on this strategy, recent our studies about bone-targeting pH sensors were briefly introduced in Section 2.2. By using local targeting of pH sensors, new insights concerning the properties of active osteoclasts were obtained. Small molecule-based sensor targeting will therefore be an important approach for such studies in which exogenous gene expression is technically challenging or impossible. The second category is based on genetically encoded tags produced with gene engineering techniques. This approach has the advantage that once the specific ligand-conjugated probe is
developed, this compound can be targeted to various organelles using a genetically encoded tag. Recent studies using commercially available tags for targeting various sensors, as reviewed in Section 3.1, exploit the reliability of the protein labeling methods, show generality, and have universal protocols. One of the technical advancements of protein labeling systems that make sensor targeting more purposive is the fluorogenic labeling approach. In Section 3.2, our studies on the development of a fluorogenic protein labeling method were introduced. In the future, sensor targeting will become more and more important, because this technique is essential for revealing the environmental diversity in living cells and animals. In addition, recent progress in optical microscopy has enabled nanoscopic observation below the diffraction limit. With such advanced technology, probe diffusion is undesirable because it lowers spatial resolution. Although protein-based sensors are intrinsically targetable, some small-molecule fluorophores are much more photo-stable. Therefore, targetable small-molecule sensors will remain to be standard probes for analyzing the subcellular microenvironment.
Profile Shin Mizukami graduated from the University of Tokyo (1997) and received the Ph. D. in pharmaceutical sciences from the University of Tokyo (2002). During 2002–2004, he was an AIST Postdoctoral Fellow at National Institute of Advanced Industrial Science and Technology, Japan. During 2004-2005, he was a JSPS Postdoctoral Fellow for Research Abroad at Stanford University, USA. In 2005, he became an Assistant Professor at Graduate School of Engineering, Osaka University, and was promoted to an Associate Professor in 2009. Since 2016 he is a Professor at Institute of Multidisciplinary Research for Advanced Materials, Tohoku University. He received the Pharmaceutical Society Award for Young Scientists in 2011 and the Commendation for Science and Technology by the Minister of Education, Culture, Sports, Science and Technology (the Young Scientists’ Prize) in 2012. His research interests are in the development of new technologies for molecular imaging and chemical biology.
Acknowledgements The studies described in this review were the result of collaborative research conducted at the Graduate School of Engineering, Osaka University, with many collaborators whose names are listed in the cited papers in references, under supervision by Prof. Kazuya Kikuchi. The studies were supported in part by the Ministry of Education, Culture, Sports, Science, and Technology, Japan, and
by the Japanese Society for the Promotion of Science (JSPS).
References  R. Y. Tsien, Biochemistry 19 (1980) 2396–2404.  G. Grynkiewicz, M. Poenie, R. Y. Tsien, J. Biol. Chem. 260 (1985) 3440–3450.  R. Rudolf, M. Mongillo, R. Rizzuto, T. Pozan, Nat. Rev. Mol. Cell Biol. 4 (2003) 579–586.  I. Johnson, Histochem. J. 30 (1998) 123–140.  C. Bremer, C. H. Tung, R. Weissleder, Nat. Med. 7 (2001) 743–748.  C. H. Tung, R. E. Gerszten, F. A. Jaffer, ChemBioChem 3 (2002) 207–211.  A. Miyawaki, J. Llopis, R. Heim, J. M. McCaffery, J. A. Adams, M. Ikura, R. Y. Tsien, Nature 388 (1997) 882–887.  V. A. Romoser, P. M. Hinkle, A. Persechini, J. Biol. Chem. 272 (1997) 13270–13274.  K. Truong, M. Ikura, Curr. Opin. Struct. Biol. 11 (2001) 573–578.  N. Demaurex, M. Frieden, Cell Calcium 34 (2003) 109–119.  J. L. Vinkenborg, T. J. Nicolson, E. A. Bellomo, M. S. Koay, G. A. Rutter, M. Merkx, Nat. Methods 6 (2009) 737–740.  P. J. Dittmer, J. G. Miranda, J. A. Gorski, A. E. Palmer, J. Biol. Chem. 284 (2009) 16289– 16297.  X. Li, X. Gao, W. Shi, H. Ma, Chem. Rev. 114 (2014) 590–659.  J. Fan, M. Hu, P. Zhan, X. Peng, Chem. Soc. Rev. 42 (2013) 29–43.  J. F. Callan, A. P. de Silva, D. C. Magri, Tetrahedron 61 (2005) 8551–8588.  A. P. de Silva, T.S. Moody, G. D. Wright, Analyst 134 (2009) 2385–2393.  R. Kasper, B. Harke, C. Forthmann, P. Tinnefeld, S. W. Hell, M. Sauer, Small 6 (2010) 1379– 1384.  R. P. Haugland, Handbook of Fluorescent Probes and Research Products 9th ed.  A. T. Hoye, J. E. Davoren, P. Wipf, M. P. Fink, V. E. Kagan, Acc. Chem. Res. 41 (2008) 87–97.  S. Zhang, T. Wu, J. Fan, Z. Li, N. Jiang, J. Wang, B. Dou, S. Sun, F. Song, X. Peng, Org. Biomol. Chem. 11 (2013) 555–558.  A. Minta, J. P. Kao, R. Y. Tsien, J. Biol. Chem. 264 (1989) 8171–8178.  D. R. Trollinger, W. E. Cascio, J. J. Lemasters, Biochem. Biophys. Res. Commun. 236 (1997) 738–742.  S. L. Sensi, D. Ton-That, J. H. Weiss, A. Roth, K. R. Gee, Cell Calcium 34 (2003) 281–284.  Y. Chen, C. Zhu, J. Cen, Y. Bai, W. He, Z. Guo, Chem. Sci. 6 (2015) 31187–3194.  Y. Kurishita, T. Kohira, A. Ojida, I. Hamachi, J. Am. Chem. Soc. 134 (2012) 18779–18789.  M. P. Murphy, Trends Biotechnol. 15 (1997) 326–330.
 A. Muratovska, R. N. Lightowlers, R. W. Taylor, J. A. Wlice, M. P. Murphy, Adv. Drug Delivery Rev. 49 (2001) 189–198.  B. C. Dickinson, C. J. Chang, J. Am. Chem. Soc. 130 (2008) 9638–9639.  G. Masanta, C. S. Lim, H. J. Kim, J. H. Han, H. M. Kim, B. R. Cho, J. Am. Chem. Soc. 133 (2011) 5698–5700.  S. C. Dodani, S. C. Leary, P. A. Cobine, D. R. Winge, C. J. Chang, J. Am. Chem. Soc. 133 (2011) 8606–8616.  G. Zhang, J. J. Gruskos, M. S. Afzal, D. Buccella, Chem. Sci. 6 (2015) 6841–6846.  S. K. Bae, C. H. Heo, D. J. Choi, D. Sen, E. –H. Joe, B. R. Cho, H. M. Kim, J. Am. Chem. Soc. 135 (2013) 9915–9923.  H. Zhu, J. Fan, S. Zhang, J. Cao, K. Song, D. Ge, H. Dong, J. Wang, X. Peng, Biomater. Sci. 2 (2014) 2, 89–97.  K. Sreenath, Z. Yuan, JH. R. Allen, M. W. Davidson, L. Zhu, Chem. Eur. J. 21 (2015) 867–874.  H. Yu, Y. Xiao, L. Jin, J. Am. Chem. Soc. 134 (2012) 17486–17489.  D. Kim, G. Kim, S. J. Nam, J. Yin, J. Yoon, Sci. Rep. 5 (2015) 8488.  Z. J. Qu, J. X. Ding, M. Y. Zhao, P. Li, J. Photochem. Photobiol. A 299 (2015) 1–8.  Q. L. Qiao, M. Zhao, H. J. Lang, D. Q. Mao, J. N. Cui, Z. C. Xu, RSC Adv. 4 (2014) 25790– 25794.  D. Li, S. Chen, E. A. Bellomo, A. I. Tarasov, C. Kaut, G. A. Rutter, W.-H. Li, Proc. Natl. Acad. Sci. USA 108 (2011) 21063–21068.  S. Iyoshi, M. Taki, Y. Yamamoto, Org. Lett. 13 (2011) 4558–4561.  M. Ikeda, H. Nakagawa, S. Ban, H. Tsumoto, T. Suzuki, N. Miyata, Free Radical Biol. Med. 49 (2010) 1792–1797.  Y. Wen, K. Liu, H. Yang, Y. Li, H. Lan, Y. Liu, X. Zhang, T. Yi, Anal. Chem. 86 (2014) 9970– 9976.  Y. Urano, D. Asanuma, Y. Hama, Y. Koyama, T. Barrett, M. Kamiya, T. Nagano, T. Watanabe, A. Hasegawa, P. L. Choyke, H. Kobayashi, Nat. Med. 15 (2009) 104–109.  Y. Wang, K. Zhou, G. Huang, C. Hensley, X. Huang, X. Ma, T. Zhao, B. D. Sumer, R. J. DeBerardinis, J. Gao, Nat. Mater. 13 (2014) 204–212.  J. Han, K. Burgess, Chem. Rev. 110 (2010) 2709–2728.  R. Wang, C. Yu, F. Yu, Trends Anal. Chem. 29 (2010) 1004–1013.  A. Loudet, K. Burgess, Chem. Rev. 107 (2007) 4891–4932.  T. Kowada, H. Maeda, K. Kikuchi, Chem. Soc. Rev. 44 (2015) 4953–4972.
 A. Zaheer, R. E. Lenkinski, A. Mahmood, A. G. Jones, L. C. Cantley, J. V. Frangioni, Nat. Biotechnol. 19 (2001) 1148–1154.  S. L. Teitelbaum, Science 289(2000) 1504–1508.  T. Kowada, J. Kikuta, A. Kubo, M. Ishii, H. Maeda, S. Mizukami, K. Kikuchi, J. Am. Chem. Soc. 133 (2011) 17772–17776.  J. Kikuta, Y. Wada, T. Kowada, Z. Wang, G. –H. Sun-Wada, I. Nishiyama, S. Mizukami, N. Maiya, H. Yasuda, A. Kumanogoh, K. Kikuchi, R. N. Germain, M. Ishii, J. Clin. Invest. 123 (2013) 866–873.  H. Maeda, T. Kowada, J. Kikuta, M. Furuya, M. Shirazaki, S. Mizukami, M. Ishii, K. Kikuchi, Nat. Chem. Biol. 12 (2016) 579–585.  H. M. O’Hare, K. Johnsson, A. Gautier, Curr. Opin. Struct. Biol 17 (2007) 488–494.  D. Rabuka, Curr. Opin. Chem. Biol. 14 (2010) 790–796.  B. A. Griffin, S. R. Adams, R. Y. Tsien, Science 281 (1998) 269–272.  I. Chen, M. Howarth, W. Lin, A. Y. Ting, Nat. Methods 2 (2005) 99–104.  C. -W Lin, A. Y. Ting, J. Am. Chem. Soc. 128 (2006) 4542–4543.  M. Fernádez-Suárez, H. Baruah, L. Martínez-Hernández, K. T. Xie, J. M. Baskin, C. R. Bertozzi, A. Y. Ting, Nat. Biotechnol. 25 (2007) 1483–1487.  O. Tour, S. R. Adams, R. A. Kerr, R. M. Meijer, T. J. Sejnowski, R. W. Tsien, R. Y. Tsien, Nat. Chem. Biol. 3 (2007) 423–431.  T. Fujii, Y. Shindo, K. Horra, D. Citterio, S. Nishiyama, K. Suzuki, K. Oka, J. Am. Chem. Soc. 136 (2014) 2374–2381.  A. Keppler, S. Gendreizig, T. Gronemeyer, H. Pick, H. Vogel, K. Johnsson, Nat. Biotechnol. 21 (2003) 86−89.  A. Juillerat, T. Gronemeyer, A. Keppler, S. Gendreizig, H. Pick, H. Vogel, K. Johnsson, Chem. Biol. 10 (2003) 313−317.  G. V. Los, L. P. Encell, M. G. McDougall, D. D. Hartzell, N. Karassina, C. Zimprich, M. G. Wood, R. Learish, R. F. Ohana, M. Urh, D. Simpson, J. Mendez, K. Zimmerman, P. Otto, G. Vidugiris, J. Zhu, A. Darzins, D. H. Klaubert, R. F. Bulleit, K. V. Wood, ACS Chem. Biol. 3 (2008) 373–382.  E. Tomat, E. M. Nolan, J. Jaworski, S. J. Lippard, J. Am. Chem. Soc. 130 (2008) 15776–15777.  M. Bannwarth, I. R. Corrêa Jr., M. Sztretye, S. Pouvreau, C. Fellay, A. Aebischer, L. Royer, E. Ríos, K. Johnsson, ACS Chem. Biol. 4 (2009) 179–190.  M. Kamiya, K. Johnsson, Anal. Chem. 82 (2010) 6472–6479.
 A. Gautier, A. Juillerat, C. Heinis, I. R. Correa Jr., M. Kindermann, F. Beaufils, K. Johnsson, Chem. Biol. 15 (2008) 128–136.  D. Srikun, A. E. Albers, C. I. Nam, A. T. Iavarone, C. J. Chang, J. Am. Chem. Soc. 132 (2010) 4455–4465.  M. Abo, R. Murakami, K. Miyano, M. Kamiya, T. Nagano, Y. Urano, H. Sumimoto, Anal. Chem. 86 (2014) 5983–5990.  C. Wang, X. Song, Z. Han, X. Li, Y. Xu, Y. Xiao, ACS Chem. Biol. 11 (2016) 2033–2040.  J. Chen, M. Zhao, X. Jiang, A. Sizovs, M. C. Wang, C. R. Provost, J. Huang, J. Wang, Analyst 141 (2016) 1209–1213.  L. A. Montoya, M. D. Pluth, Anal. Chem. 88 (2016) 5769–5774.  D. Lin L. Liu, W. –H. Li, ACS Chem. Biol. 10 (2015) 1054–1063.  T. Hirata, T. Terai, H. Yamamura, M. Shimonishi, T. Komatsu, K. Hanaoka, T. Ueno, Y. Imaizumi, T. Nagano, Y. Urano, Anal. Chem. 88 (2016) 2693–2700.  M. Best, I. Porth, S. Hauke, F. Braun, D.-P. Herten, R. Wombacher, Org. Biomol. Chem. 14 (2016) 5606–5611.  A. Hayashi, D. Asanuma, M. Kamiya, Y. Urano, S. Okabe, Neuropharmacology 100 (2016) 66– 75.  H. Adachi, T. Ohta, H. Matsuzawa, J. Biol. Chem. 266 (1991) 3186–3191.  S. Mizukami, S. Watanabe, Y. Hori, K. Kikuchi, J. Am. Chem. Soc. 131 (2009) 5016–5017.  A. Yoshimura, S. Mizukami, Y. Hori, S. Watanabe, K. Kikuchi, ChemBioChem 12 (2011) 1031– 1034.  S. Mizukami, T. Yamamoto, A. Yoshimura, S. Watanabe, K. Kikuchi, Angew. Chem. Int. Ed. 50 (2011) 8750–8752.  S. Watanabe, S. Mizukami, Y. Akimoto, Y. Hori, K. Kikuchi, Chem. Eur. J. 17 (2011) 8342– 8349.  K. K. Sadhu, S. Mizukami, S. Watanabe, K. Kikuchi, Chem. Commun. 46 (2010) 7403–7405.  S. Mizukami, S. Watanabe, Y. Akimoto, K. Kikuchi, J. Am. Chem. Soc. 134 (2012) 1623–1629.  S. Mizukami, Y. Hori, K. Kikuchi, Acc. Chem. Res. 47 (2014) 247–256.  J. T. Moore, S. T. Davis, I. K. Dev, Anal. Biochem. 247 (1997) 203–209.  G. Zlokarnik, P. A. Negulescu, T. E. Knapp, L. Mere, N. Burres, L. Feng, M. Whitney, K. Roemer, R. Y. Tsien, Science 279 (1998) 84–88.  W. Gao, B. Xing, R. Y. Tsien, J. Rao, J. Am. Chem. Soc. 125 (2003) 11146–11147.  K. K. Sadhu, S. Mizukami, S. Watanabe, K. Kikuchi, Mol. BioSyst. 7 (2011) 1766–1772.
 K. K. Sadhu, S. Mizukami, C. R. Lanam, K. Kikuchi, Chem. Asian J. 7 (2012) 272–276.  Y. Hori, H. Ueno, S. Mizukami, K. Kikuchi, J. Am. Chem. Soc. 131 (2009) 16610–16611.  Y. Hori, K. Nakaki, M. Sato, S. Mizukami, K. Kikuchi, Angew. Chem. Int. Ed. 51 (2012) 5611– 5614.  Y. Hori, T. Norinobu, M. Sato, K. Arita, M. Shirakawa, K. Kikuchi, J. Am. Chem. Soc. 135 (2013) 12360–12365.
Figure captions Fig. 1 Subcellular distribution of conventional fluorescent sensors and organelle targetable fluorescent sensors. Fig. 2 Organelle-specific fluorescent probes. Fig. 3 Mitochondria-targeting fluorescent sensors. AM: acetoxymethyl group. Fig. 4 Mitochondria-targeting fluorescent sensors with a triphenylphosphonium group. Fig. 5 Lysosome-targeting mechanism using a tertiary amino group. Fig. 6 Lysosome-targeting fluorescent sensors with a tertiary amino group. Fig. 7 Plasma membrane-targeting fluorescent sensors with alkyl groups. Fig. 8 Nucleus-targeting fluorescent sensors. Fig. 9 Strategy for detection of bone-resorbing osteoclast with targetable pH sensors. Fig. 10 Structures of clinically used bisphosphonate with an amino group (a) and bone-targeting pH sensors (b and c). Fig. 11 Two-photon excitation microscopy images of in vivo osteoclasts using BAps. (green) Fluorescence of Baps (green) and tdTomato (red) and second harmonic generation (SHG) from collagen in bone matrix (blue) were observed. Scale bars: 40 μm. Reprinted with permission from Ref. 34. Copyright© 2011 American Chemical Society. Fig. 12 (a) Three-dimensional fluorescence images of bone tissue reconstructed from z-stack images after injections of pHocas-3 (left) or pHocas-AL (right). Scale bars: 20 μm. (b) Time-lapse fluorescence imaging of bone tissues for 8 h after injection of pHocas-3. Scale bars: 20 μm. Reprinted with permission from Ref. 36. Copyright © 2016, Rights Managed by Nature Publishing Group. Fig. 13 Overview of a protein-labeling system based on a genetically encoded tag. Fig. 14 Labeling scheme of a tetracysteine tag with the targetable fluorophore (17) (a) and tetracysteine tag-based targetable fluorescent sensors (18, 19) (b). Fig. 15 Overview of the SNAP-tag system (a) and HaloTag system (b). Fig. 16 SNAP-tag-based targetable fluorescent sensors. Fig. 17 HaloTag-based targetable fluorescent sensors. Fig. 18 Overview of noncatalytic β-lactamase-based protein labeling system. Fig. 19 Fluorogenic labeling system based on FRET quencher elimination induced by nucleophilic attack by BL-tag. Fig. 20 Fluorogenic labeling probes based on BL-tag system.
Fig. 21 (a) Time-dependent emission spectra of 500 nM FCAPO2 (λex = 490 nm), obtained by incubation with 1 μM BL-tag. Inset, time-dependent fluorescence intensity of 500 nM FCAPO2 (λex = 490 nm, λem = 518 nm) with (circle) or without (triangle) BL-tag. (b) Real-time fluorogenic labeling of cell surface BL-EGFR with 10 nM FCAPO2 at 37 °C. (c) Real-time fluorescence imaging of BL-EGFR trafficking by using fluorogenic probe. Cell-surface BL-EGFR at 0 min was labeled with a red fluorescent probe (RA). Confocal fluorescence microscopic images for RA (top, λex = 559 nm) and FCAPO2 (middle, λex = 473 nm), and merged images of fluorescence and bright field images (bottom). Scale bar: 20 μm. Reprinted with permission from Ref. 64. Copyright© 2012 American Chemical Society.
Fig. 1 Subcellular distribution of conventional fluorescent sensors and organelle targetable fluorescent sensors.
Fig. 2 Organelle-specific fluorescent probes.
Fig. 3 Mitochondria-targeting fluorescent sensors. AM: acetoxymethyl group.
Fig. 4 Mitochondria-targeting fluorescent sensors with a triphenylphosphonium group.
Fig. 5 Lysosome-targeting mechanism using a tertiary amino group.
Fig. 6 Lysosome-targeting fluorescent sensors with a tertiary amino group.
Fig. 7 Plasma membrane-targeting fluorescent sensors with alkyl groups.
Fig. 8 Nucleus-targeting fluorescent sensors.
Fig. 9 Strategy for detection of bone-resorbing osteoclast with targetable pH sensors.
Fig. 10 Structures of clinically used bisphosphonate with an amino group (a) and bone-targeting pH sensors (b and c).
Fig. 11 Two-photon excitation microscopy images of in vivo osteoclasts using BAps. (green) Fluorescence of Baps (green) and tdTomato (red) and second harmonic generation (SHG) from collagen in bone matrix (blue) were observed. Scale bars: 40 μm. Reprinted with permission from Ref. 51. Copyright© 2011 American Chemical Society.
Fig. 12 (a) Three-dimensional fluorescence images of bone tissue reconstructed from z-stack images after injections of pHocas-3 (left) or pHocas-AL (right). Scale bars: 20 μm. (b) Time-lapse fluorescence imaging of bone tissues for 8 h after injection of pHocas-3. Scale bars: 20 μm. Reprinted with permission from Ref. 53. Copyright © 2016, Rights Managed by Nature Publishing
Fig. 13 Overview of a protein-labeling system based on a genetically encoded tag.
Fig. 14 (a) Labeling scheme of a tetracysteine tag with the targetable fluorophore (a) and tetracysteine-tag-based targetable fluorescent sensors (b).
Fig. 15 Overview of the SNAP-tag system (a) and HaloTag system (b).
Fig. 16 SNAP-tag-based targetable fluorescent sensors.
Fig. 17 HaloTag-based targetable fluorescent sensors.
Fig. 18 Overview of noncatalytic β-lactamase-based protein labeling system.
Fig. 19 Fluorogenic labeling system based on FRET quencher elimination induced by nucleophilic attack by BL-tag.
Fig. 20 Fluorogenic labeling probes based on BL-tag system.
Fig. 21 (a) Time-dependent emission spectra of 500 nM FCAPO2 (λex = 490 nm), obtained by incubation with 1 μM BL-tag. Inset, time-dependent fluorescence intensity of 500 nM FCAPO2 (λex = 490 nm, λem = 518 nm) with (circle) or without (triangle) BL-tag. (b) Real-time fluorogenic
labeling of cell surface BL-EGFR with 10 nM FCAPO2 at 37 °C. (c) Real-time fluorescence imaging of BL-EGFR trafficking by using fluorogenic probe. Cell-surface BL-EGFR at 0 min was labeled with a red fluorescent probe (RA). Confocal fluorescence microscopic images for RA (top, λex = 559 nm) and FCAPO2 (middle, λex = 473 nm), and merged images of fluorescence and bright field images (bottom). Scale bar: 20 μm. Reprinted with permission from Ref. 83. Copyright© 2012 American Chemical Society.