Selective fluorescent labeling of cellular proteins and its biological applications

Abstract

Proteins, which are ubiquitous in cells and critical to almost all cellular functions, are indispensable for life. Fluorescence imaging of proteins is key to understanding their functions within their native milieu, as it provides insights into protein localization, dynamics, and trafficking in living systems. Consequently, the selective labeling of target proteins with fluorophores has emerged as a highly active research area, encompassing bioorganic chemistry, chemical biology, and cell biology. Various methods for selectively labeling proteins with fluorophores in cells and tissues have been established and are continually being developed to visualize and characterize proteins. This review highlights research findings reported since 2018, with a focus on the selective labeling of cellular proteins with small organic fluorophores and their biological applications in studying protein-associated biological events. We also discuss the strengths and weaknesses of each labeling approach for their utility in living systems.

---

Joo Hee Choi

Joohee Choi was born in Seoul, Korea, in 1996. She earned her BS degree in Chemistry from Sungshin Women's University in 2021. She is currently pursuing her PhD under the guidance of Professor Injae Shin. Her doctoral research focuses on the synthesis and development of multi-targeting fluorescent probes for selective cancer detection as well as organelle-specific ionophores.

Sooin Kim

Sooin Kim was born in Seoul, Korea, in 1996. He received his BS degree in Chemistry from Sogang University in 2021. He completed his MS degree in 2023 under the guidance of Professor Hyun Soo Lee in the Department of Chemistry at Sogang University. During his master's studies, he focused on protein engineering using genetic code expansion technology. Since 2023, he has been working as a master's researcher with Professor Hyun Soo Lee on synthesizing antibody–drug conjugates using a novel protein labeling technique.

On-Yu Kang

On-Yu Kang was born in Gyeongsangbuk-do, Korea, in 1991. She earned her BS in 2014 and MS degree in 2016, both in Chemistry from Chungbuk National University. From 2017 to 2023, she pursued her PhD at Sungkyunkwan University under the guidance of Professors Do Hyun Ryu and Hwan Jung Lim, focusing on the development of new organic reactions and their applications in discovering therapeutic agents. She is currently a postdoctoral associate in the Therapeutics & Biotechnology Division of Korea Research Institute of Chemical Technology (KRICT). Her research interests include the development of small molecule-based anticancer agents and drug conjugates.

Seong Yun Choi

Seong Yun Choi was born in Busan, Korea, in 1998. He received his BS degree in Chemistry from Pukyong National University in 2022. Since 2023, he has been pursuing his MS research under the guidance of Professor Ji Young Hyun at Korea National University of Science and Technology (UST) and Korea Research Institute of Chemical Technology (KRICT). His research focuses on the synthesis of pharmacologically active compounds.

Ji Young Hyun

Ji Young Hyun, born in Seoul, Korea in 1990, earned her BS degree in Chemistry from Yonsei University in 2013. She completed her PhD research at Yonsei University under the guidance of Professor Injae Shin from 2013 to 2019, focusing on the synthesis and biological activity of various glycoconjugates for functional studies of glycans. After her postdoctoral studies at Yonsei University with Professor Injae Shin, she began her independent career in 2020 as a senior researcher at the Therapeutics & Biotechnology Division of Korea Research Institute of Chemical Technology (KRICT). Her research interests include the development of carbohydrate-based therapeutic agents.

Hyun Soo Lee

Hyun Soo Lee earned his BS and MS degrees in Chemistry from POSTECH in 2001 and 2003, respectively. He completed his PhD in Chemistry at the Scripps Research Institute in 2009, working under the mentorship of Dr Peter Schultz. In 2010, he joined Sogang University, where he achieved the rank of full professor in 2019. His research focuses on protein engineering, particularly leveraging genetic code expansion technology. His work encompasses the development of innovative site-specific protein modification techniques, the design of protein sensors, and the synthesis of antibody–drug conjugates using cutting-edge site-specific antibody modification strategies.

Injae Shin

Injae Shin received his BS (1985) and MS degrees (1987) in Chemistry from Seoul National University in Korea. His PhD research was conducted at University of Minnesota with Professor Hung-wen Liu (1991–1995). After postdoctoral studies at University of California at Berkeley (1995–1998), he began his independent career in 1998 at Yonsei University in Korea. He is a Fellow of the Korean Academy of Science and Technology and a Fellow of the Royal Society of Chemistry. His research interests include the discovery of bioactive molecules that modulate biological processes and functional studies of glycans using chemical tools including glycan microarrays.

---

1. Introduction

Proteins are fundamental to nearly every cellular function that plays vital roles in cell survival and operation. Given their critical importance to the structure, function, and regulation of living organisms, researchers have devoted significant efforts to unveil the mysteries surrounding their diverse functions. Fluorescence imaging has emerged as a powerful technique, allowing for non-invasive, direct detection of target molecules and analysis of biological events within living systems.^1–9 The beneficial features for fluorescence imaging have made it a leading tool for understanding biological processes involving proteins and their molecular dynamics in real-time, with exceptional temporal and spatial precision. Progress in fluorescence imaging has significantly deepened our understanding of protein functions.

For fluorescence imaging of cellular proteins using conventional fluorescence microscopy and advanced super-resolution microscopy, it is essential to selectively label proteins of interest in complex biological systems with small organic fluorophores or fluorescent proteins (FPs).¹⁰ Advances in biological imaging are intrinsically connected to advances in labeling techniques. FPs have been widely utilized in fused forms to visualize target proteins in live cells and organisms.^11–13 Despite their clear benefits for protein imaging, FPs have inherent drawbacks. Their large size (>25 kDa) often perturbs the structure, localization, and function of target proteins, and they also suffer from issues such as photobleaching and restricted spectral properties.^14,15 In addition, fusion of FPs to the N- or C-terminus of proteins of interest can further influence their cellular function, limiting their application in studying protein-associated cellular events. Furthermore, in most cases, FPs lack the necessary brightness and photostability for super-resolution imaging at resolutions below 20 nm.

In contrast, small organic fluorophores, which offer sensitive, specific, and multiplexing capabilities in biological research, have become indispensable tools for imaging cells, tissues, and animals.^1–9 Also, their advantages, such as easy synthesis, high biocompatibility, favorable excretion, and beneficial pharmacokinetics, position them as promising candidates for biomedical applications. Consequently, fluorescent labeling of proteins with organic dyes in living systems has become a vibrant research field in bioorganic chemistry, chemical biology, and cell biology. A variety of labeling strategies have been established and continue to evolve for the visualization and characterization of specific cellular proteins.

Protein labeling strategies with organic dyes developed thus far include the site-specific incorporation of noncanonical amino acids (ncAA) into cellular proteins by genetic code expansion (GCE) technology. In this approach, fluorescent ncAAs can be directly inserted into target proteins, or functionalized non-fluorescent ncAAs are introduced into proteins followed by labeling with organic dyes via bioorthogonal ligation. Additionally, self-labeling enzyme (SLE) tag or peptide-tag based methods are also popularly utilized for the selective fluorescent labeling of cellular proteins. Although less popular, affinity-based fluorescent labeling approaches have advantages in labeling proteins in their intact form. The primary application of fluorescence-labeled proteins is to visualize proteins of interest in cells and tissues. They are also employed in studies aimed at determining protein–protein interactions (PPIs) and exploring protein translocation. Furthermore, fluorescence-labeled proteins are utilized to investigate oligomeric status, internalization, conformational changes, and membrane potential of cell–surface proteins. Other applications include studying protein degradation and processing, as well as detecting protein-specific glycans.

This review highlights recent advances (2018–2023) made in the development of labeling strategies for specific cellular proteins with small organic fluorophores. The advantages and limitations of each protein labeling strategy are also discussed. Moreover, it summarizes the biological applications of fluorescence-labeled proteins. Finally, we provide suggestions for possible directions in the development of new, efficient protein labeling methods and their biological applications.

2. Fluorescent labeling methods

2.1 Incorporation of noncanonical amino acids into cellular proteins

The site-specific incorporation of ncAAs into proteins using GCE technology has garnered significant attention as an effective method for generating modified proteins in cells and organisms. Since its initial report in 2001,¹⁶ this technology has continuously evolved and thus now enables the incorporation of over 200 different ncAAs into proteins.¹⁷ Many of these ncAAs possess fluorescent groups or reactive functional groups that allow for the selective labeling of target proteins. This section discusses the fluorescent labeling of cellular proteins utilizing GCE technology.

2.1.1 Incorporation of fluorescent amino acids into cellular proteins. GCE technology requires the expression of a modified aminoacyl-tRNA (aatRNA) and aminoacyl-tRNA synthetase (aaRS) pair within the host cell. The aaRS charges the ncAA onto its cognate tRNA (Fig. 1a). It is crucial that this aatRNA/aaRS pair functions orthogonally to the host cell's native pairs, with the aaRS's active site engineered to specifically recognize the ncAA. Typically, the amber stop codon (TAG) is used to incorporate the ncAA into the target protein.^18–20 The insertion of fluorescent amino acids into target proteins in cells using GCE technology is a highly intriguing and valuable tool for fluorescently labeling specific proteins. In this approach, the aaRS is engineered to recognize and incorporate the fluorescent amino acid, and the target protein's codon at the desired position is mutated to TAG.

Fig. 1 Incorporation of fluorescent amino acids into cellular proteins using GCE technology. (a) Schematic for the insertion process of a fluorescent amino acid into a target protein. An orthogonal aaRS/tRNA pair is expressed in host cells. The aaRS acylates the tRNA with the ncAA, which is incorporated site-specifically into the protein at an amber codon. (b) Chemical structures of fluorescent amino acids incorporated into proteins.

In 2006, the Schultz group incorporated two fluorescent amino acids, L-(7-hydroxycoumarin-4-yl) ethylglycine (CouA) and 2-amino-3-(5-(dimethylamino)naphthalene-1-sulfonamide)propanoic acid (DanA), into proteins via GCE (Fig. 1b and Table 1).^21,22 CouA and DanA were initially employed for fluorescent labeling of cellular proteins due to their intriguing fluorescent properties and small fluorophores. A coumarin moiety in CouA exhibits dramatic changes in fluorescence depending on the protonation state of the hydroxyl group, making it highly sensitive to pH and environmental conditions.²³ Similarly, the fluorescence intensity of the dansyl group in DanA is highly sensitive to the surrounding polarity, making it useful for analyzing biochemical environmental changes.²² Despite these useful fluorescent properties,^24–26 their application for fluorescence imaging of cellular proteins is limited because the Methanococcus janaschii (Mj) TyrRS/tRNA pair used for CouA incorporation is not orthogonal in mammalian cells,²⁷ and the fluorescence intensity of DanA is too weak in aqueous solutions.²⁸

Table 1 Spectroscopic properties of fluorescent amino acids

	CouA	DanA	HCK	ACK	ANAP	Acd
a Data were obtained from an anionic form,^24 and ΦF and ε were measured at 360 nm.^21 b Data were obtained in water, and ΦF was measured at 363 nm.^25 c Data were obtained with dansylglycine, and ε was measured at 325 nm.^26 d Data were obtained in PBS buffer solutions at pH 7.4.^29 e Data were obtained with 7-diethylamino-4-methylcoumarin in a 1 : 1 mixture of EtOH and water.^30 f Data were obtained in water.^31 g Data were obtained in EtOH at 360 nm.^31 h Data were obtained in 100 mM phosphate buffer at pH 7.^32 i Data were obtained in water at 386 nm.^33
λ abs (nm)	320^a	324^c	326^d	348^d	360^f	386^h
λ em (nm)	450^a	559^c	483^d	477^d	490^f	446^h
Quantum yield (ΦF)	0.63^a	0.10^b	0.21^d	0.40^d	0.48^g	0.93^b
Extinction coefficient (ε, M^−1 s^−1)	17 000^a	3900^c	10 000^d	14 000^d	17 500^g	5700^i
Lifetime (τ, ns)	1.64^b	3.30^b	1.64^b	1.4^e	2.25^b	14.04^b

---

Later, the Deiters group devised another fluorescent amino acid, 7-hydroxycoumarin-lysine (HCK), for its insertion into proteins in mammalian cells utilizing the Methanosarcina barkeri (Mb) PylRS/tRNA system (Fig. 1b and Table 1).³⁴ However, the hydroxycoumarin moiety of HCK displayed low fluorescence efficiency under physiological pH conditions, limiting its utility for fluorescently visualizing cellular proteins.²⁹ To overcome this limitation, the same group subsequently developed 7-aminocoumarin-lysine (ACK), which is pH-insensitive (Fig. 1b and Table 1).^29,30,35 This fluorescent amino acid could also be introduced into proteins in mammalian cells using the same MbPylRS/tRNA pair through GCE.

Acridon-2-ylalanine (Acd) is a fluorescent amino acid with a high quantum yield in water (ϕ = 0.93), excellent photostability, and prolonged fluorescence lifetime (τ ∼ 14 ns) (Fig. 1b and Table 1).^32,33 This fluorescent amino acid was incorporated into proteins using the MjTyrRS/tRNA pair via GCE.³⁶ However, this pair is only orthogonal in prokaryotic systems, limiting its applicability in mammalian cells.³⁷ To address this issue, the MbPylRS/tRNA pair, which is orthogonal in mammalian cells, was engineered to insert Acd into proteins in mammalian cells.²⁵ Specifically, a variant of MbPylRS capable of recognizing and incorporating Acd was constructed through mutations at five amino acid positions (N311S, C313G, V366A, W382T, and L155V) within the Acd binding site of MbPylRS. This variant successfully facilitated the introduction of Acd into the insulin receptor expressed in HEK293T cells for fluorescence imaging of live cells. Furthermore, fluorescence lifetime imaging (FLIM) was employed to observe the localization of the insulin receptor at different stages of maturation within the endoplasmic reticulum (ER) and Golgi apparatus.

Another fluorescent amino acid employed for protein labeling in cells through GCE is 3-((6-acetylnaphthalen-2-yl)amino)-2-aminopropanoic acid (ANAP), which contains a prodan derivative as a fluorophore (Fig. 1b and Table 1).³⁸ The prodan fluorophore is notable for its sensitivity to environmental polarity.³⁹ ANAP is efficiently incorporated into proteins in mammalian cells using the E. coli LeuRS/tRNA pair.^22,40 Importantly, ANAP exhibits low background fluorescence after washing, making it suitable for protein imaging in cells.^31,41–43 The brightness and clarity of ANAP-incorporated proteins are comparable to those achieved with traditional FPs. Given its sensitivity to environmental polarity, ANAP is particularly valuable for investigating the structural dynamics of cellular proteins. Notably, as the polarity surrounding ANAP increases, its emission wavelength shifts from 430 nm to 490 nm.³⁸ ANAP has been utilized to study conformational changes in cellular proteins and to investigate PPIs in cells. Detailed discussions on its biological applications are provided in Section 3.

2.1.2 Two-step fluorescent labeling of cellular protein. The site-specific incorporation of fluorescent amino acids into cellular proteins through GCE is technically straightforward. However, currently available fluorescent amino acids possess smaller fluorophores with lower brightness (quantum yield × extinction coefficient) compared to commonly used organic dyes. Consequently, they require high-energy light for excitation, which can potentially lead to phototoxicity.^44,45 Moreover, the options for fluorescent amino acids suitable for protein labeling are limited.

To overcome these limitations, a two-step approach has been developed and proven more effective for fluorescently labeling proteins. This method involves using GCE technology to insert reactive ncAAs into proteins, followed by labeling the modified proteins with fluorophores ranging from UV to NIR emission through bioorthogonal ligation (Fig. 2a). Recent advancements in bioorthogonal chemistry have significantly advanced the field of cellular protein labeling in conjunction with GCE technology. Bioorthogonal reactions coupled with the genetic incorporation of reactive ncAAs include copper-catalyzed azide–alkyne cycloaddition (CuAAC), strain-promoted azide–alkyne cycloaddition (SPAAC), and inverse electron-demand Diels–Alder (IEDDA) reaction (Fig. 2b).^46,47 In this section, we discuss the two-step process of fluorescently labeling cellular proteins by combining GCE technology with bioorthogonal ligation.

Fig. 2 Two-step fluorescent labeling of cellular proteins. (a) Reactive ncAAs are incorporated into cellular proteins via GCE and are subsequently labeled with functionalized fluorophores through bioorthogonal ligation. (b) Examples of bioorthogonal reactions used for labeling reactive ncAA-incorporated proteins with functionalized fluorophores. (c) Chemical structures of reactive ncAAs used for bioorthogonal ligation.

CuAAC utilizes a copper(I) catalyst to promote the reaction between an alkyne and an azide to form a stable triazole ring.^48,49 Traditional Huisgen 1,3-dipolar cycloadditions require high temperatures, organic solvents, and long reaction times. However, CuAAC allows for rapid and regioselective reactions in aqueous solutions at room temperature due to the catalytic property of Cu(I). Given the scarcity of alkyne and azide functional groups in cellular environments, CuAAC has become an invaluable tool for selectively modifying biomolecules in cells. ncAAs containing azide or alkyne groups have been designed and incorporated into cellular proteins using GCE technology.⁵⁰ For example, PrK (Fig. 2c), which includes an alkyne group in a carbamate form at the epsilon amine of Lys, was incorporated into proteins with high efficiency using a PylRS/tRNA pair.⁵¹ Subsequently, CuAAC was employed to label the modified proteins with azide-conjugated fluorophores. This method was applied to achieve super-resolution imaging of cellular proteins such as Rab protein, β-actin, and α-synuclein in mammalian cells.⁵²

Recently, PrK was also introduced into the M2 muscarinic acetylcholine receptor (M2R), a member of the G protein-coupled receptor (GPCR) family, on the cell surface and then labeled with an azide-appended fluorophore by CuAAC for imaging M2R (Fig. 3a).⁵³ Also, this approach enabled MINIFLUX super-resolution imaging of β-actin within filopodia of U2OS cells (Fig. 3b), allowing precise measurements of interfilament separation (12 nm).

Fig. 3 (a) (Left) Schematic for the fluorescent labeling of M2R and (right) confocal fluorescence microscopy image of PrK-containing M2R labeled with azide-linked Alexa Fluor-647 in cells. (b) Super-resolution image of PrK-containing β-actin labeled with azide-linked Alexa Fluor-647 in filopodia of cells (reproduced from ref. 53 with permission from the National Academy of Sciences, copyright 2022).

Another ncAA used for CuAAC is ACPK (Fig. 2c), a pyrrolysine (Pyl) analogue equipped with an azido group. This ncAA was inserted into acid chaperones using an engineered PylRS and subsequently labeled with alkyne-linked fluorophores in cells via CuAAC.^54–56 In this study, a tris(triazolylmethyl)amine-based ligand was employed to reduce the cytotoxic effects of Cu(I), facilitating the fluorescent labeling of acid chaperones in both E. coli and BHK-21 cells.

CuAAC is highly useful for labeling biomolecules due to its rapid reaction rate and bioorthogonality. However, the Cu(I) catalyst can generate reactive oxygen species (ROS) and cause cytotoxicity, which restricts the application of CuAAC primarily to labeling cell–surface proteins.^57–59 To address these concerns, a Cu(I)-free cycloaddition reaction was developed. It was discovered that cyclooctyne and phenyl azide could form a triazole without a catalyst, leading to the development of the SPAAC reaction (Fig. 2b).^60,61 SPAAC has successfully been used for labeling various biomaterials in cells, effectively resolving the cytotoxicity issues associated with CuAAC.^62,63 Moreover, the development of more reactive cyclooctyne derivatives, such as dibenzocyclooctyne (DBCO) (0.36 M⁻¹ s⁻¹) and bicyclo[6.1.0]nonyne (BCN) (0.14 M⁻¹ s⁻¹), has partially addressed the issue of the slow reaction rate (2 × 10⁻³ M⁻¹ s⁻¹) in cyclooctyne.^64–67 SPAAC combined with GCE technology has been applied for the fluorescent labeling of cellular proteins.

For instance, amino acids containing azide groups, such as ACPK and p-azidophenylalanine (AzF, Fig. 2c), have been introduced into intracellular proteins for fluorescence imaging through SPAAC with cyclooctyne-derivatized fluorophores.^68,69 Recently, AzF was incorporated into SifA, an effector protein involved in Salmonella infection (Fig. 4).⁷⁰ HeLa cells infected with Salmonella expressing AzF-incorporated SifA were treated with a DBCO-dye for fluorescence imaging. The results revealed the cytoplasmic distribution of SifA and its colocalization with the lysosome marker LAMP1 (Fig. 4b). Because fusing a FP to an effector protein could disrupt its secretion,⁷¹ chemical labeling of SifA using this two-step procedure proved useful for studying its biological process through fluorescence imaging.

Fig. 4 (a) Schematic for the fluorescent labeling of effector proteins. A ncAA is incorporated into an effector protein in Salmonella, which is then secreted into host cells via the Type III secretion system (T3SS). Infected host cells are subjected to SPAAC with a DBCO-dye or SiR-tetrazine. (b) Confocal fluorescence microscopy images of HeLa cells infected with SifA (LAMP1: a lysosome marker). (c) Temporal localization of another effector protein, SsaP, modified with TCO*K and labeled with SiR-tetrazine via IEDDA reaction (reproduced from ref. 70 with permission from eLife Sciences Publications, copyright 2021).

Conversely, amino acids bearing strained alkyne groups, such as CoK and BCNK (Fig. 2c), have also been inserted into cellular proteins using a PylT/PylRS pair and subsequently labeled with azide-appended fluorophores.^72–74 One example of this approach involved introducing BCNK into the Cdc42 binding domain (CBD), followed by labeling with an environment-sensitive merocyanine-azide to visualize the localization of the CBD-Cdc42 complex (Fig. 5).⁷⁵ The results showed that the complex is located along the cell periphery and within cellular extensions (Fig. 5b).

Fig. 5 (a) CBD-BCNK-EGFP is labeled with an environment-sensitive merocyanine-azide through SPAAC to generate the non-fluorescent fusion protein. When CBD binds to Cdc42, the fluorescence of the labeled protein increases, enabling the detection of the active complex. (b) Fluorescence images of migrating HeLa cells expressing CBD-BCNK-EGFP after labeling with merocyanine-azide. The CBD–Cdc42 complex was seen at the cell periphery and within cellular extensions (left). In contrast, cells expressing a mutant CBD (mCBD-BCNK-EGFP) with reduced affinity for Cdc42 did not display this pattern (right). White arrows indicate protrusions leading to cell movement (reproduced from ref. 75 with permission from the American Chemical Society, copyright 2019).

Despite improvements in labeling efficacy by the development of reactive cycloalkynes, the slow reaction rate of SPAAC remained a significant issue for cellular protein labeling. Researchers sought bioorthogonal reactions that could overcome this limitation and turned their attention to the IEDDA reaction with tetrazine (Fig. 2b). In 2008, IEDDA reactions of tetrazines with cycloalkenes were introduced.^76,77 Since then, many strained alkenes and alkynes have been developed to react with tetrazines at very rapid reaction rates (1–10⁶ M⁻¹ s⁻¹).⁷⁸ These reactions produce only molecular nitrogen as a byproduct and do not require any additional catalyst, making them a promising bioorthogonal reaction for labeling biomolecules in cells.

The IEDDA reaction, combined with GCE technology, has been effectively utilized to label proteins in cells. Specifically, BCNK and TCOK (or TCO*K with higher incorporation efficiency than TCOK) (Fig. 2c), which contain strained alkyne and alkene groups respectively, are incorporated into proteins using engineered PylRS variants applicable to both prokaryotic and eukaryotic cells.^{73,74,79–81} The proteins bearing the strained groups are then labeled with dye-conjugated tetrazines. The combination of the genetic incorporation of BCNK or TCO*K with the rapid IEDDA reaction has proven to be a highly effective technique for labeling cellular proteins, and it has been extensively applied to study various cellular events.^81–86

Tetrazine is known to quench the fluorescence of organic dyes through energy transfer.⁸⁷ As a result, tetrazine-conjugated dyes are non-fluorescent until they undergo an IEDDA reaction with strained alkynes and alkenes, at which point they become fluorescent. This feature allows for turn-on fluorescence labeling, significantly reducing background fluorescence. In addition, Me-tetrazine displays better stability than H-tetrazine, leading to improved labeling efficiency when TCO is used as a dienophile.⁸⁸ The quenching effect of tetrazine on the fluorescence of conjugated dyes also increases as the distance between the tetrazine and the fluorophore decreases. Using this approach, TCO*K was inserted into actin, tubulin, and membrane receptors in mammalian cells, followed by labeling with fluorophore-linked tetrazines (Fig. 6a).⁸⁸ Owing to the non-fluorescent nature of fluorophore-linked tetrazines, fluorescence images could be obtained without the need to wash out the remaining dye-tetrazines (Fig. 6b and c).

Fig. 6 (a) The intact tetrazine-dye is non-fluorescent due to the fluorescence quenching effect of tetrazine. However, when it conjugates with TCO*K-incorporated proteins, the tetrazine-dye becomes fluorescent. (b) Wash-free confocal fluorescence microscopy images of actin in fixed cells labeled with indicated dye-conjugated tetrazines. (c) Confocal fluorescence microscopy images of the EMTB-TCO*K-GFP after labeling with the membrane-permeable H-Tet-SiR in live cells (reproduced from ref. 88 with permission from the Springer Nature, copyright 2019).

Dye-tetrazines with enhanced quenching effects was developed for imaging proteins in cells. While commercially available SiR-tetrazine, which has a relatively longer distance between the fluorophore and the tetrazine group, displayed a lower quenching effect, the newly devised HD653 dye with a shorter distance significantly enhanced the quenching effect and the fluorescence turn-on ratio (Fig. 7a).⁸⁹ This HD653 dye was utilized for wash-free imaging of cellular proteins. Specifically, BCNK was inserted into the nucleoporin 153 protein and subsequently labeled with SiR-tetrazine or HD653 dye. Analysis of confocal fluorescence microscopy images of cells without washing showed that while SiR-tetrazine labeled cells display noticeable background, cells labeled with HD653 exhibit minimal background (Fig. 7b).

Fig. 7 Wash-free fluorescence imaging of proteins using a tetrazine-dye with an enhanced fluorogenic effect. (a) SiR-tetrazine has a relatively longer distance between the fluorophore and the tetrazine group, resulting in a lower quenching effect. In contrast, an engineered HD653 dye features a much shorter distance between these groups, significantly enhancing both the quenching effect and the fluorescence turn-on ratio. (b) Wash-free confocal fluorescence microscopy images of cells expressing EGFP-nucleoporin (EGFP-Nup153) containing BCNK after labeling with HD653 or SiR-tetrazine (reproduced from ref. 89 with permission from the American Chemical Society, copyright 2021).

For general protein labeling using BCNK and tetrazine probes, the Elia group developed a GCE-tag that consisted of 14-amino acid sequence composed of a hemagglutinin (HA) tag, linker, and BCNK (Fig. 8).⁹⁰ Several proteins (α-tubulin, GFP, and organelle marker proteins such as LAMP1 and CD63) containing the GCE-tag at the N-terminus were expressed in cells and labeled with SiR-tetrazine via the IEDDA reaction for fluorescence imaging. Also, a TCO*K and SiR-tetrazine pair was employed to analyze host cell infection by Salmonella.⁷⁰ In this study, TCO*K was incorporated into the effector protein SsaP in Salmonella, which is crucial for the infection of host cells, and fluorescent labeling was achieved using SiR-tetrazine through the IEDDA reaction. Infection of HeLa cells by SsaP during the infection process was fluorescently visualized (Fig. 4c). SsaP was localized at the pole of Salmonella at 8 h post-infection, but it entered the host cell at 16 h post-infection. It should be noted that fusion of the FP to an effector protein disrupts its secretion, which is critical when studying the infection of mammalian cells by Salmonella.⁷¹

Fig. 8 Fluorescent labeling of cellular proteins using a GCE-tag. The GCE-tag, which consists of a hemagglutinin (HA) tag, linker, and BCNK, is fused to target proteins at the N-terminus in cells. The subsequent labeling of GCE-tagged proteins with SiR-tetrazine via the IEDDA reaction enables the fluorescent visualization of these proteins.

Traditional optical microscopy cannot resolve structures smaller than approximately half the wavelength of the emitted light (typically no less than ∼200 nm). However, super-resolution techniques achieve single nanometer resolution, providing more detailed information about biological structures and allowing visualization of smaller features.⁹¹ Super-resolution imaging relies on fluorescent dyes with specific photophysical or photochemical properties. Antibody-based immunofluorescence imaging is unsuitable for super-resolution owing to the large size of antibodies (>10 nm).⁹² In addition, FPs are not ideal for this purpose because of their photophysical properties.⁹³ The two-step fluorescent labeling approach, which combines GCE technology with bioorthogonal ligation and has no restrictions on fluorophore selection, is particularly advantageous for super-resolution imaging applications.^81,83

For super-resolution imaging of cell–surface proteins, TCO*K was introduced into the NR1 subunit of the extracellular domain of the N-methyl-D-aspartate receptor (NMDAR), that is a glutamate receptor and predominant Ca²⁺ ion channel found in neurons, using GCE technology. The modified receptor was then labeled with dye-appended tetrazine through the IEDDA reaction (Fig. 9a). For comparison, the NMDAR was also immunostained with its specific antibody. Analysis of super-resolution images (dSTORM) showed that while the basal membrane structure of NR1 appears more discontinuous when visualized by immunolabeling, the two-step labeling method provides a more homogeneous and densely packed NR1 distribution at the plasma membrane (Fig. 9b).⁹⁴

Fig. 9 (a) Schematic for fluorescent labeling of the TCO*K-incorporated NR1 subunit of the NMDAR with dye-tetrazine via IEDDA in cells. (b) Super-resolution images of the NMDAR after click labeling (top) and immunostaining with NR1 antibody (bottom) (reproduced from ref. 94 with permission from the Wiley-VCH, copyright 2018).

Recently, the neurofilament light chain (NFL) in primary mouse cortical neurons was labeled with two distinct dyes at different incubation times.⁹⁵ In this study, TCO*K was incorporated into NFL in primary neurons, which was then sequentially labeled with BODIPY-tetrazine and SiR-tetrazine via the IEDDA reaction (Fig. 10). Super-resolution image analysis indicated that while initially synthesized NFL extends distally along the axon, later synthesized NFL remains in the cell body (Fig. 10b). Moreover, using CRISPR/Cas9 genome editing technology, endogenous NFL containing TCO*K in neurons was fluorescently labeled with SiR-tetrazine without the need for external plasmid introduction.

Fig. 10 Dual-color fluorescent labeling of neurofilament light chain (NFL) in primary mouse cortical neurons. (a) Schematic for the dual-color labeling process of NFL with two distinct dye-tetrazines at different time points. Neurons containing a PylRS/tRNA pair are transfected with NFL-TAG in the presence of TCO*K. After a few days of incubation, neurons are labeled with the first dye-tetrazine. Following additional incubation with TCO*K, neurons are labeled with the second dye-tetrazine. (b) Super-resolution image of neurons expressing NFL that were labeled with the two different dyes (reproduced from ref. 95 with permission from the Springer Nature, copyright 2022).

Multiple proteins in cells were also labeled with different dyes for simultaneous imaging using a two-step labeling procedure.⁹⁶ In this approach, PrK was inserted into one protein in a cell, while TCO*K was incorporated into another protein in a different cell (Fig. 11a). The two cells were then fused to construct a single cell containing two different proteins, each with a distinct ncAA. The PrK-incorporated protein was labeled with an azide-appended dye via CuAAC, and the TCO*K-inserted protein was labeled with dye-tetrazine using the IEDDA reaction, generating two different proteins labeled with distinct dyes within a single cell (Fig. 11b). This labeling method was applied to study the spatial distribution and interactions of the proteins of interest in cells.

Fig. 11 Fluorescent labeling of two different proteins with distinctive dyes in cells. (a) Schematic for the process where two sets of cells are separately transfected with genes containing amber codons for two distinct proteins of interest (POI). In the first set, PrK is incorporated into POI1 using the PylRS^WT/tRNA system. Concurrently, in the second set, TCO*K is inserted into POI2 employing the PylRS^AF/tRNA system. After separate incubation, two cell types are fused to allow for co-localization of POI1 and POI2 within the same cells. Subsequent click chemistry generates two different proteins that are labeled with distinct fluorophores, allowing for dual fluorescence imaging of the proteins. (b) Two-color fluorescence images of cells containing two different proteins labeled with distinctive dyes (reproduced from ref. 96 with permission from the American Chemical Society, copyright 2018).

A single protein in cells was also labeled with two different dyes. In this case, two mutually orthogonal PylT/PylRS pairs, derived from different archaeal strains,⁹⁷ were employed to suppress the ochre (UAA) and amber (UAG) codons for incorporating PrK and TCO*K into the protein (Fig. 12a).⁹⁸ The modified protein, containing the two different ncAAs, was then labeled with corresponding dyes using IEDDA and CuAAC reactions. This enabled imaging of a single protein labeled with two different dyes in cells (Fig. 12b).

Fig. 12 Dual fluorescent labeling of a single protein using two orthogonal stop codons. (a) Schematic for the suppression of two different stop codons (UAA and UAG) using two orthogonal PylRS/tRNA pairs, each tailored to a specific ncAA. After incorporation of the ncAAs into a protein at their respective stop codons, the protein is subsequently labeled with different dyes via IEDDA and CuAAC, resulting in dual fluorescence. (b) Fluorescence images of corticotropin-releasing factor type 1 receptor (CRFR1) labeled with Cy3-tetrazine and sulfo-Cy5-picolyl azide in cells (reproduced from ref. 98 with permission from the Elsevier, copyright 2020).

2.2 Peptide tag-based fluorescent labeling

Peptide tags are frequently utilized for the selective fluorescent labeling of cellular proteins. This strategy involves genetically fusing target proteins in cells with short peptide tags ranging from 0.6 to 6 kDa that are subsequently labeled with fluorophore-appended substances through either covalent or noncovalent bonds. This method offers benefits such as straightforward genetic manipulation and the minimal impact of the small peptide tags on the structure and function of the target proteins. Because peptide tags are inherently non-fluorescent, they need to be conjugated with fluorophores prior to cell imaging. This section begins with a brief discussion of peptide tag-based fluorescent labeling strategies. Detailed methods for peptide tag-based fluorescent labeling have been extensively reviewed in other articles.^99,100

Since the 1990s, small molecule-binding peptide tags have been discovered for labeling peptide-tagged proteins with fluorophores. The Tsien group pioneered this field by designing a peptide tag with a tetracysteine core (CCxxCC, commonly CCGPCC) that was labeled with the non-fluorescent fluorescein–arsenical–helix binder (FlAsH), which becomes fluorescent upon binding to Cys residues of the peptide tag (Fig. 13a).¹⁰¹ Other derivatives, such as ReAsH, CrAsH and Cy3As, were later devised to provide diverse fluorescence emissions.^102–104 Despite its usefulness for detecting cellular proteins, nonspecific protein labeling can occur due to interactions between FlAsH or its derivatives and other Cys-rich proteins. In addition, Cys-rich sequences fused to proteins can sometimes lead to incorrect disulfide bond formation within proteins, thereby perturbing protein structures.

Fig. 13 Fluorescent labeling of (a) a tetracysteine-tagged protein with FlAsH, (b) a fluorette-tagged protein with an a-chloroacetylated fluorophore, (c) dC10α-tagged protein with a dimaleimide fluorogen, (d) a His₆-tagged protein with a fluorophore-conjugated trisNTA (Fl-trisNTA), and (e) a CA₆D₄-tagged protein with an α-chloroacetylated Zn²⁺-DpaTyr appended by a fluorophore (Fl-Zn²⁺-DpaTyr).

Another small molecule-binding peptide tag is the fluorophore-binding peptide (fluorette) tag that was identified from a phage display peptide library (Fig. 13b).^105,106 Binding of α-chloroactylated fluorophores to fluorette-tagged proteins promotes proximity-induced ligation to generate fluorescence-labeled proteins via a thioether linkage.¹⁰⁶ The dC10α tag (LSAAECAAREAACREAAARAGGK), which contains two Cys residues separated by two α-helix turns, is another example of a small molecule-binding peptide tag (Fig. 13c).¹⁰⁷ One of the Cys residues in this tag reacts with a maleimide moiety of a non-fluorescent dimaleimide fluorogen through thiol addition to form a non-fluorescent conjugate. The remaining maleimide in the conjugate further reacts with the other Cys residue of the dC10α tag to produce a fluorescently labeled protein. Because both the dimaleimide fluorogen and the mono-conjugated adduct are non-fluorescent, washing procedures are unnecessary prior to cell imaging.

Metal ion-mediated fluorescent labeling exploits interactions between electron-deficient metal ions and electron-rich ligands. A representative example of this class of peptide tags is the His₆-tag, which binds to Ni²⁺–nitrilotriacetic acid (Ni²⁺–NTA) and was originally discovered as an affinity tag for protein purification (Fig. 13d).¹⁰⁸ Because of the non-covalent nature of interactions between metal ions and ligands, high binding affinities are crucial for achieving stable protein labeling. To enhance binding affinity and stability, multivalent complexes possessing two to four NTA moieties were developed.¹⁰⁹

The oligo-Asp tag (D₄ tag), which interacts with the dinuclear Zn²⁺–bis((dipicolylamino)methyl)tyrosine (Zn²⁺–DpaTyr) complex, was devised for protein labeling (Fig. 13e).¹¹⁰ Although the dinuclear Zn²⁺–DpaTyr complex displayed moderate binding affinity for the D₄ tag, elongating the peptide tag to the D₄ × 3 peptide tag and utilizing dimeric Zn²⁺–DpaTyr complexes resulted in strongly bound dimeric complexes.¹¹¹ This improvement enabled the fluorescence imaging of proteins in live cells. To further improve the stability of the label, a nucleophilic Cys residue was added to the D₄ tag, and the electrophilic α-chloroacetyl group was attached to DpaTyr, forming a covalent linkage between the D₄ tag and Zn²⁺–DpaTyr complex (Fig. 13e).¹¹⁰

A peptide-templated labeling method, that relies on the high affinity (nanomolar K_D) for the formation of heterodimeric coiled coils between two α-helical peptides, was created to fluorescently label target proteins in cells.^112,113 In this strategy, two α-helical peptides noncovalently interact with each other, and thus the fluorescent label can be lost during prolonged imaging. To address this issue, the peptides are modified to form a covalent linkage when they come into close proximity through coiled coil interactions. One example of this covalent bond formation is proximity-induced S-alkylation which involves the reaction between a Cys-containing P1-tagged protein and an α-chloroacetylated P2, where P1 and P2 form the heterodimeric coiled coil (Fig. 14a).¹¹⁴ However, this S-alkylation process requires more than 2 h for complete ligation.

Fig. 14 Fluorescent labeling of a protein via (a) peptide-templated S-alkylation, (b) peptide-templated acyl transfer, and (c) peptide-templated acyl transfer followed by hybridization of PNA with complementary DNA appended with a fluorophore.

To enhance the reaction kinetics for covalent protein labeling, the Seitz group developed a peptide-templated acyl transfer reaction (Fig. 14b).¹¹⁵ In this approach, an N-terminal Cys residue was introduced to the P3 tag, while a thioester-conjugated fluorophore was attached to the N-terminus of the P4 tag. Upon the heterodimeric coiled coil interaction between P3 and P4 tags, peptide-templated acyl transfer takes place, leading to the generation of fluorescence-labeled proteins. The transfer of the fluorophore from the P4 probe to the Cys-P3-tagged protein occurs rapidly within a few minutes.

Later, peptide nucleic acid (PNA)-conjugated proteins were generated through peptide-templated labeling for erasable fluorescent labeling of cell–surface proteins (Fig. 14c).¹¹⁶ This process involves the association of the Cys-P3-protein with the P4 peptide containing thioester-conjugated PNA, resulting in the formation of PNA-conjugated protein. Subsequently, the PNA-conjugated protein is labeled with a fluorophore via PNA–DNA hybridization with complementary DNA appended with a fluorophore. An advantage of this labeling method is the easy removal of the fluorophore from the labeled protein using displacement DNA. In addition, the unlabeled protein can be relabeled with complementary DNA linked to different fluorophores (see Fig. 20 for detailed explanation and Fig. 59 for its biological application).

Enzymes possess unique features such as fast reaction rates, high substrate specificity, and the ability to catalyze reactions under mild conditions. This led to the development of enzyme-catalyzed labeling for peptide-tagged proteins to construct fluorophore-labeled proteins in cells. Biotin ligase (BirA) derived from E. coli, for instance, promotes the ATP-dependent conjugation of biotin to a Lys residue within the AP-tag (GLNDIFEAQKIEWHE) (Fig. 15a).¹¹⁷ Interestingly, BirA enzyme from different organisms can accept biotin derivatives containing functional groups such as ketone, alkyne or azide group (keto-B, cis-PB or DTB-Az, respectively) as substrates.^118,119 Reporter groups within biotin derivatives that are attached to the peptide tag are subsequently conjugated to a fluorophore through bioorthogonal ligation, enabling the fluorescent labeling of target proteins.

Fig. 15 Fluorescent labeling of proteins tagged with (a) AP and (b) LAP using BirA and LpIA, respectively, followed by bioorthogonal ligation with a fluorophore.

Lipoic acid ligase (LplA) from E. coli, which is compatible with BirA, catalyzes the ATP-dependent attachment of lipoic acid to a Lys residue within the LAP tag (DEVLVEIETDKAVLEVPGGEEE or GFEIDKVWYDLDA) (Fig. 15b).^120,121 Moreover, this enzyme can accept lipoic acid derivatives with diverse reporter units such as alkyne, azide, trans-cyclooctene, aryl aldehyde and aryl hydrazine groups.^121–124 Labeling of LAP tagged proteins with lipoic acid derivatives and subsequent bioorthogonal ligation with fluorophores result in the generation of fluorescence-labeled proteins. Notably, unlike BirA, LplA can utilize fluorophore-bearing substances as substrates, enabling one-step fluorescent labeling of proteins. Besides these two enzymes, other enzymes including sortase A,^125–127 formylglycine-generating enzyme (FGE),^128,129 and transglutaminase (TGase)¹³⁰ have been discovered and utilized for fluorescently labeling target proteins in cells.

Since 2018, several peptide tags, including the FlAsH-binding peptide tag, Ni²⁺–NTA binding peptide tag, dC10α peptide tag, and α-helical peptide tag, have been employed for visualizing proteins of interest in cells. For example, a Ni²⁺–NTA chelator with high affinity for the His-tag was developed for imaging His-tagged proteins in fixed cells (Fig. 16a).¹³¹ Among cyclic, linear, and dendritic trivalent Ni²⁺–NTAs, the cyclic trivalent Ni²⁺–NTA (cyclic trisNTA–Ni²⁺) exhibited the strongest binding affinity (K_D = 9.5 nM) to His₆-tagged proteins as determined by surface plasmon resonance (SPR) spectroscopy. Using a picomolar concentration of the Cy5-appended cyclic trisNTA–Ni²⁺ chelator, imaging of His₁₀-EGFP-lamin A, a fibrous protein in type V intermediate filaments responsible for structural support and transcriptional regulation in the nucleus, was achievable in fixed cells (Fig. 16b).

Fig. 16 (a) Chemical structure of cyclic trisNTA–Ni²⁺ conjugated with a fluorophore and (b) fluorescence images of His₁₀-mEGFP-lamin A labeled with Alexa 647-conjugated cyclic trisNTA–Ni²⁺ in fixed cells (reproduced from ref. 131 with permission from the Wiley VCH, copyright 2018).

In an endeavor to further advance the Ni²⁺–NTA chelator for tighter binding to the His-tag, two cyclic trisNTA–Ni²⁺ moieties were linked via a short tether to create hexaNTA–Ni²⁺ (Fig. 17a).¹³² This chelator displayed picomolar binding affinity toward the His₁₂-tagged protein and remarkable stability even in the presence of 30 mM imidazole. Utilizing the fluorophore-conjugated HexaNTA-Ni²⁺ chelator in conjunction with confocal and super-resolution microscopy, His₁₂-EGFP-lamin A was visualized with high signal-to-background in live cells expressing the mechanosensitive channel MscL, which facilitated rapid delivery of the chelator inside cells (Fig. 17b and c).

Fig. 17 (a) Chemical structure of fluorophore-conjugated hexaNTA-Ni²⁺. (b) Confocal and (c) super-resolution (dSTORM) microscopy images of His₁₂-tagged EGFP-lamin A labeled with Alexa 647-conjugated hexaNTA-Ni²⁺ in live cells expressing the mechanosensitive channel MscL for fast delivery of the chelator inside cells (reproduced from ref. 132 with permission from Wiley-VCH, copyright 2018).

A BODIPY-based dimaleimide probe, YC23, was designed and prepared for highly fluorogenic, no-wash labeling of the dC10α-tagged protein in live cells (Fig. 18a).¹³³ This probe remained non-fluorescent because of the quenching effect of two maleimide groups. However, reaction of YC23 with two thiols of the dC10α-tag led to a remarkable increase in fluorescence with a maximum emission of 525 nm. YC23 was effectively employed to visualize the dC10α-tagged histone H2B in the nucleus of live cells (Fig. 18b).

Fig. 18 (a) Schematic for labeling of the dC10α-tagged protein with a BODIPY-based dimaleimide fluorogen YC23. (b) Fluorescence images of HeLa cells expressing dC10 α-tagged histone H2B labeled with YC23 (reproduced from ref. 133 with permission from the Wiley-VCH, copyright 2018).

Previously developed FlAsH, with the interarsenic distance of approximately 6 Å, covalently bound to the four Cys residues of CCXXCC sequences found abundantly in cellular proteins, leading to nonspecific protein labeling (Fig. 13a). To address the nonspecific labeling issue, a cell-permeable Cy3-based biarsenical probe, AsCy3, was developed with an increased interarsenic distance of ca. 15 Å to specifically image peptide-tagged proteins in cells (Fig. 19a).¹³⁴ Probe AsCy3 bound to CCKAEAACC (Cy3TAG) and CCKAEAAKAEAAKCC (Cy3TAG+6) sequences, that were individually inserted into EGFP, with submicromolar K_D (Fig. 19b). The effectiveness of this probe was evaluated by determining FRET signals from EGFP to Cy3 in E. coli expressing EGFP tagged with either Cy3TAG or Cy3TAG+6. While a significant and highly reproducible FRET signal was observed in E. coli expressing EGFP–Cy3TAG labeled with AsCy3, cells expressing EGFP–Cy3TAG+6 labeled with AsCy3 exhibited a smaller and more variable FRET signal. In another investigation, FlAsH was utilized for live cell imaging of spiral-shaped bacteria, spirochetes, by labeling inner or outer membrane proteins tagged with tetracysteine motifs.¹³⁵

Fig. 19 (a) Chemical structure of AsCy3 and (b) labeling of two peptides (Cy3TAG and Cy3TAG+6) with AsCy3.

Peptide-templated acyl transfer (Fig. 14c) was employed for erasable and exchangeable fluorescence imaging of proteins on the surface of live cells (Fig. 20a)¹¹⁶ In this approach, the Cys-P1-tagged protein on the cell surface was initially labeled with PNA-linked P2, where P1 and P2 form a heterodimeric coiled coil. The resultant PNA-conjugated protein was subsequently hybridized with fluorophore-appended complementary DNA (Complex III) for fluorescent labeling (Fig. 20b left). In addition, a noncovalently bound fluorophore on the cell–surface protein was removed by treatment with displacement DNA-23mer (Fig. 20b middle). Furthermore, the resultant non-fluorescent protein was relabeled with a fluorophore by treatment with fluorophore-appended complementary DNA-23mer (Fig. 20b right). This developed method of erasable fluorescence labeling was applied to clearly visualize the internalization of cell–surface proteins into cells (see Fig. 59).

Fig. 20 (a) A PNA-conjugated cell–surface protein, prepared through peptide-template acyl transfer, is hybridized with fluorophore-conjugated complementary DNA (complex III) for fluorescent labeling of the protein. Treatment with displacement DNA converts the fluorophore-labeled protein into a non-fluorescent form, which can be subsequently relabeled with fluorophore-linked complementary DNA. (b) Reversible fluorescent labeling of PNA-tagged EGFR-YFP on the cell surface. Atto565-labeled EGFR-YFP (left) was treated with displacement DNA to generate Atto565-free EGFR-YFP (middle). This form was relabeled with Cy7-DNA (right) (reproduced from ref. 119 with permission from the Springer Nature, copyright 2021).

A multi-color fluorescence labeling strategy using orthogonal peptide-templated acyl transfer was devised to concurrently visualize different receptors on the cell surface. In this method, the orthogonality of the peptide tag system is critical for selective labeling of distinct target proteins with specific fluorophores without any cross-reactivity. For example, Cys-P1-tag and Cys-P3-tag were fused to the N-termini of two different GPCRs (endothelin receptors A and B (ET_AR and ET_BR), angiotensin II receptor type 1 (AT₁R) or apelin receptor (APJ)), that play a crucial role in cardiovascular regulation (Fig. 21a).¹³⁶ As peptides P1 and P3 selectively bind to their respective peptides P2 and P4 through coiled coil interactions, the Cys-P1 and Cys-P3-tagged GPCRs were simultaneously and selectively labeled with Atto488-P2 and Atto565 (or TMR)-P4, respectively, via orthogonal peptide-templated acyl transfer (Fig. 21b). This method was employed to assess the oligomeric state of these receptors, which were separately labeled with a FRET donor and acceptor, by measuring FRET signals (see Fig. 56).

Fig. 21 (a) Schematic for fluorescent labeling of two different GPCRs via orthogonal peptide-templated acyl transfer. (b) Fluorescence images of cells expressing CysP1/CysP3-tagged receptors labeled with different fluorophores (reproduced from ref. 136 with permission from the Wiley-VCH, copyright 2022).

In 2020, a versatile interacting peptide (VIP) tag named MiniVIPER, comprising a MiniE–MiniR heterodimer, was constructed to visualize cell–surface proteins in live cells or intracellular proteins in fixed cells (Fig. 22a).¹³⁷ Besides this peptide tag, two additional peptide tags, TinyVIPER and PunyVIPER, consisting of TREER–TinyERRE and PuRRRE–PunyEEER, respectively, were identified based on the MiniVIPER tag for multi-color protein imaging in cells (Fig. 22a).¹³⁸ These three orthogonal coiled–coil tags were utilized to selectively image actin (TREER-Actin) in the cytoskeleton, histone H2B (H2B-MiniR) in the nucleus and TOMM20 (TOMM20-PuRRRE) in mitochondria after concurrent labeling with three different fluorophores (TinyERRE-Cy3, MiniE-AF488 and PunyEEER-Cy5) in fixed cells (Fig. 22b). Consequently, MiniVIPER, TinyVIPER and PunyVIPER tags enable efficient and simultaneous labeling of three distinctive proteins in cells.

Fig. 22 (a) Peptide sequences used in orthogonal peptide-templated protein labeling. (b) (Left) Schematic for selective fluorescent labeling of three distinct proteins via three orthogonal coiled–coil interactions. (Right) Fluorescence images of fixed cells expressing three peptide-tagged proteins labeled with distinct fluorophore-conjugated peptides (reproduced from ref. 138 with permission from the American Chemical Society, copyright 2019).

2.3 Self-labeling enzyme (SLE)-based fluorescent labeling

Initially, we briefly introduce self-labeling enzymes (SLEs) that allow specific and covalent conjugation of SLE-tagged proteins with fluorophore-appended ligands (or substrates) in cells. For detailed information on fluorescent labeling of proteins using SLEs, readers are encouraged to refer to other review articles.^139,140 Attachment of functionalized ligands to SLEs typically occurs with high specificity and rapid kinetics under physiological conditions. Among SLEs, SNAP-tag and HaloTag have emerged as the most popular choices for imaging proteins of interest in cells owing to their excellent specificity and simple manipulation, as well as the commercial availability of fluorophore-appended substrates (Fig. 23).¹⁴¹ These tags have been extensively applied to study a broad range of biological processes using confocal microscopy and super-resolution microscopy.¹³⁹ Additionally, other SLEs, such as E. coli dihydrofolate reductase (eDHFR), the photoactive yellow protein (PYP) and the β-lactamase mutant (E166N), have been utilized for selective imaging of target proteins in cells (Fig. 24).

Fig. 23 Schematic for fluorescent labeling of (a) SNAP-tagged, (b) CLIP-tagged, and (c) halo-tagged proteins with their corresponding dye-conjugated ligands.

Fig. 24 Schematic for fluorescent labeling of (a) eDHFR-tagged, (b) PYP-tagged, and (c) BL-tagged proteins with their corresponding dye-conjugated ligands.

The SNAP-tag, approximately 19 kDa in size and engineered from human O⁶-alkylguanine–DNA alkyltransferase, promotes the transfer of an O⁶-alkyl moiety from substrates containing O⁶-benzylguanine (BG) to its Cys side chain (Fig. 23a).^142–145 Additionally, its variant known as the CLIP-tag selectively reacts with O²-benzylcytosine (BC) derivatives as substrates instead of BG (Fig. 23b).¹⁴⁶ Because of their orthogonal enzymatic reactions, SNAP-tag and CLIP-tag can be simultaneously employed for selective labeling of two different target proteins with distinctive fluorophores in cells.¹⁴⁷ While the SNAP-tag displayed slower labeling kinetics (with BG-tetramethylrhodamine (TMR) at 2.71 × 10⁵ M⁻¹ s⁻¹) compared to the HaloTag (with TMR-conjugated substrate at 1.88 × 10⁷ M⁻¹ s⁻¹),¹⁴⁸ its smaller size minimizes interference with the dynamics and function of fused proteins.

The HaloTag (33 kDa), evolved from a bacterial haloalkane dehalogenase, forms a covalent ester linkage with synthetic substrates consisting of a chloroalkane linker attached to various moieties (e.g., fluorophores) (Fig. 23c).¹⁴⁹ Specifically, the aspartate (Asp) side chain of the HaloTag attacks chloroalkylated substrates to form the covalent bond. The HaloTag efficiently labels fusion proteins with rapid kinetics,¹⁴⁸ high specificity, and stability.¹⁰⁴ However, its relatively large size compared to other SLEs may potentially interfere with the function of target proteins.¹⁵⁰

The eDHFR-tag (approximately 18 kDa) interacts noncovalently with a folate analog trimethoprim (TMP), a classical antibiotic with a nanomolar K_D for eDHFR but lacking binding affinity for mammalian DHFR. Consequently, eDHFR-tagged proteins are noncovalently labeled with fluorophore-conjugated TMP analogues.¹⁵¹ To enhance labeling stability, proximity-induced reactivity-based TMP derivatives containing an acrylamide electrophile and a fluorophore were developed for covalent labeling (Fig. 24a).^152–154 In this scenario, the Cys side chain of eDHFR attacks acrylamide within TMP derivatives, resulting in covalent labeling of the enzyme. Presently, the eDHFR-tag is employed to image target proteins in live cells, even employing sophisticated microscopy techniques.

The PYP-tag (14 kDa), derived from purple bacteria, covalently binds to the CoA thioester of 4-hydroxycinnamic acid as a natural cofactor via transthioesterification with the Cys residue. Interestingly, this enzyme also covalently interacts with thioester derivatives of 7-hydroxycoumarin-3-carboxylic acid through transthioesterification.¹⁵⁵ Because 7-hydroxycomarin and 7-dimethylaminocoumarin (DMAC) derivatives serve as labeling ligands for PYP (Fig. 24b),^156,157 this enzyme can be utilized as a SLE for fluorescently labeling target proteins in cells.

TEM-1 β-lactamase (29 kDa) catalyzes the hydrolysis of β-lactam antibiotics (e.g., penicillin, ampicillin, and cephalosporin) by initially forming an acyl–enzyme intermediate followed by its breakdown (the deacylation step). By mutating glutamic acid (Glu)-166, which is responsible for deacylation of an acyl–enzyme intermediate, to asparagine (Asn), the acyl–enzyme intermediate remains unbroken and accumulates within the mutant enzyme.¹⁵⁸ Consequently, the mutant β-lactamase-E166N (BL-tag) has been utilized as a SLE to fluorescently label fusion proteins in cells (Fig. 24c).

These SLEs have found widespread application in studying protein localization, dynamics, trafficking, PPIs, and various other protein-associated biological events.^139,140 In many cases, these enzyme tags are labeled with ‘always-on’ fluorophores spanning from the visible to near-infrared (NIR) spectrum, which often require a washing step prior to cell imaging.¹⁴¹ To circumvent this washing step, ‘turn-on’ fluorogenic labeling probes for SLEs have been developed for wash-free protein labeling in live cells and organisms.¹⁴¹ Among SLEs mentioned above, SNAP-tag, HaloTag, eDHFR-tag, PYP-tag and BL-tag have been employed to fluorescently monitor target proteins in cells since 2018.

As mentioned earlier, BG derivatives are commonly employed as substrates for the SNAP-tag. Interestingly, O⁶-(5-pyridylmethyl)guanine (5PG) derivatives, where the benzyl group of BG is replaced with the 5-pyridylmethyl group, were able to serve as ligands for the SNAP-tag (Fig. 25).¹⁴⁰ As a consequence, a 5PG derivative containing a coumarin dye was used to visualize SNAP-tagged proteins within subcellular organelles. Notably, the fluorescence intensity of SNAP-tagged proteins labeled with the 5PG derivative in cells was enhanced in the presence of Zn²⁺ due to binding of this ion to the pyridyl group of 5PG.

Fig. 25 Chemical structure of the 5PG derivative as the SNAP-tag ligand.

Recently, the application of the SNAP-tag has expanded to include the fluorescent imaging of specific proteins in plant cells, which possess distinctive cell wall compositions compared to animal cells.¹⁵⁹ Specifically, the SNAP-tagged α-tubulin (SNAP-TUA5) was expressed in tobacco BY-2 cells and subsequently labeled individually with three commercially available BG derivatives, SNAP-Cell 430, SNAP-Cell TMR-Star and SNAP-Cell 647SiR (Fig. 26a). The microtubules were clearly visualized in cells labeled with these ligands at different emission wavelengths, as depicted in Fig. 26b. Moreover, cytoskeletal structures such as the spindle and phragmoplast in Arabidopsis roots expressing SNAP-TUA5 were also imaged using the SNAP-tag/BG labeling system. The results underscore the utility of the SNAP-tag/BG labeling system in studying plant protein-associated biological processes.

Fig. 26 (a) Chemical structures of SNAP-Cell 430, SNAP-Cell TMR-Star, and SNAP-Cell 647SiR. (b) Fluorescence images of BY-2 cells expressing SNAP-TUA5 labeled with indicated ligands (reproduced from ref. 159 with permission from the Oxford Academic, copyright 2020).

The SNAP-tag was also utilized for fluorescence imaging of endogenously expressed proteins in tissues composed of heterogeneous cell states.¹⁶⁰ Employing CRISPR/Cas9 genome editing technology, a SNAP_f-tag, a variant of a SNAP-tag with faster labeling kinetics, was inserted between the signal peptide and the ectodomain of the glucagon-like peptide-1 receptor (GLP1R, a class B GPCR), which plays a role in the modulation of glucose homeostasis and food intake, in mice (Fig. 27a). Islets isolated from GLP1R^SNAP/SNAP mice were treated with several BG-ligands. As illustrated in Fig. 27b, sulfonated ligands such as SBG-TMR, BG-Sulfo549 and BG-SulfoCy5 selectively and efficiently labeled the cell–surface SNAP-tagged GLP1R to a comparable extent. This labeling approach was also used to investigate ligand-induced internalization of endogenous GPCRs in tissues (see Fig. 60 in Section 3.4).

Fig. 27 (a) Schematic for fluorescent labeling of SNAP_f-tagged GLP1R with sulfonated BG derivatives in mouse islets. Included are chemical structures of SBG-TMR, BG-SulfoCy5, and BG-Sulfo549. (b) Fluorescence images of SNAP_f-tagged GLP1R in islets after labeling with indicated ligands (reproduced from ref. 160 with permission from the Springer Nature, copyright 2023).

The SNAP-tag mimic (SmFP485) of the fluorescent protein was generated by enzymatic reaction of the SNAP_f-tag with BG-F485, enabling the study of several cellular events (Fig. 28a).¹⁶¹ BG-F485 itself was non-fluorescent in solution owing to its intramolecular motion (Fig. 28b). However, BG-F485 fluorescence increased over 350-fold upon reaction with SNAP_f, with 17 000 M⁻¹ s⁻¹ of the second-order rate constant, by restricting its intramolecular motion within the enzyme. Because free BG-F485 in cells lacks fluorescence, there is no need for a washing step to remove remaining ligand within cells prior to cell imaging. Moreover, additional BG-F485 derivatives were developed to create a range of SmFPs spanning from the visible to infrared spectrum. SmFP485 was utilized to fluorescently monitor several proteins in cells (Fig. 28c), as well as for real-time tacking of protein expression, degradation, and interaction (see Fig. 50, 52 and 67).

Fig. 28 (a) A non-fluorescent ligand BG-F485 becomes fluorescent upon forming a covalent complex with SNAP_f to generate SmFP485. (b) Absorption (dashed) and emission (solid) spectra of SmFP485 (cyan), along with the emission spectrum of BG-F485 (red). (c) Fluorescence images of SmFP485-tagged proteins in subcellular organelles (reproduced from ref. 161 with permission from the Springer Nature, copyright 2023).

The nuclear pore complex (NPC), the largest protein assembly composed of approximately 30 nucleoporins (NUPs) in eukaryotic cells, facilitates selective transport of various cargoes between the nucleus and the cytoplasm.¹⁶² Utilizing DNA-based point accumulation for imaging in nanoscale topography (DNA-PAINT), super-resolution imaging of single copies of NPC proteins was conducted with SNAP-tag and HaloTag.¹⁶³ In this investigation, NUP proteins (NUP96 or NUP107) tagged with either SNAP-tag or HaloTag were expressed in mammalian cells through CRISPR/Cas9 technology (Fig. 29a). Following cell fixation and permeabilization, SNAP-tag or HaloTag fused NUP proteins were conjugated with the respective BG- or chloroalkane (CA)-modified oligonucleotide, followed by hybridization with fluorophore-appended complementary oligonucleotides (Fig. 29b). Cells expressing NUP96-HaloTag labeled with a fluorophore displayed optimal 2D DNA-PAINT images with the smallest distribution width (Fig. 29c). Additionally, distinctive pairs of close-by localization clouds, possibly attributed to single NUP96 proteins, were visualized in 3D DNA-PAINT images of cells expressing NUP96-HaloTag (Fig. 29d). Thus, the combination of the SLE tag with DNA-PAINT enabled the direct visualization of single NPC proteins.

Fig. 29 (a) NPCs contain 16 copies of NUP96 and NUP107 in the cytoplasmic and the nuclear ring. (Top right) C-terminally labeled (blue, marked by green arrows) NUP96 structure (orange). (Bottom right) N-terminally labeled (blue, marked by green arrows) NUP107 structure (orange). (b) Schematic for SNAP-tag and HaloTag labeling with BG- and CA-oligonucleotides, respectively, followed by hybridization with fluorophore-appended complementary oligonucleotides. (c) Individual NPC and sum images (n = 398) obtained by means of 2D DNA-PAINT. (d) 3D DNA-PAINT images of single NPCs. Arrows indicate two copies of NUP96 proteins (reproduced from ref. 163 with permission from Wiley-VCH, copyright 2019).

The labeling approach based on SLE tags primarily relies on the irreversible covalent coupling of fluorophore-linked ligands to enzymes. However, fluorophores that are irreversibly attached to enzymes often undergo photobleaching during imaging, thereby limiting their application for super-resolution imaging in live cells. To overcome this limitation, reversible labeling techniques that allow continuous exchange of photobleached fluorophores with new ones were very recently developed for long-term time-lapse imaging with super-resolution microscopy. One such example of reversible labeling of SLE tags is the exchangeable fluorogenic HaloTag ligand (xHTL) (Fig. 30).¹⁶⁴ Si–rhodamine ligands (SiR-S5 and SiR-T5) containing methylsulfonamide (S5) and trifluoromethylsulfonamide (T5) interacted noncovalently with HaloTag7 with submicromolar K_D values (Fig. 30a and b). These ligands displayed fast exchange rates crucial for efficient probe exchange to overcome photobleaching. Also, fluorescence emitted by SiR-T5 bound to HalTag7 increased 13-fold compared to the free ligand. This reversible labeling technique was applied for super-resolution imaging of Halo-tagged proteins in cells using PAINT and MINFLUX microscopy. Furthermore, another reversible labeling system with different ligand specificity was devised for dual-color super-resolution imaging. Specifically, Asp106 in HaloTag7 was mutated to Ala to generate dHaloTag7, which weakly interacted with Fl-S5 and Fl-T5 but strongly with primary hydroxy derivatives with varying alkyl chain lengths (Fl-Hy4 and Fl-Hy5) (Fig. 30c and 30d). The HaloTag7/Fl-S5 and dHaloTag7/Fl-Hy4 labeling systems were utilized for dual-color super-resolution imaging of target proteins with PAINT and stimulated emission depletion (STED) microscopy.

Fig. 30 (a) Schematic for noncovalent labeling of HaloTag7 with an exchangeable fluorogenic HaloTag ligand (xHTL) containing sulfonamide. (b) Chemical structures of Fl-S5 and Fl-T5 that contain methylsulfonamide (S5) and trifluoromethylsulfonamide (T5), respectively. (c) Schematic for noncovalent labeling of dHaloTag7 with the xHTL possessing the hydroxyl group. (d) Chemical structures of Fl-Hy4 and Fl-Hy5.

In an effort to develop a reversible labeling method, HaloTag7 was mutated to restore its dehalogenase activity.¹⁶⁵ Two HaloTag7 variants, reHaloTagS (HaloTag7–N272H) and reHaloTagF (HaloTag7–N272H/L273Y), were generated by introducing mutations at Asn272 to His or at Asn272/Leu273 to His/Tyr, respectively. Both reHaloTagS and reHaloTagF exhibited reversible interactions with fluorophore-linked ligands with turnover kinetics ranging from seconds to minutes (Fig. 31). Fluorophores attached to reHaloTag variants demonstrated enhanced photostability during imaging with confocal and STED microscopy. Furthermore, reHaloTag variants labeled with fluorophores were suitable for prolonged time-lapse imaging using advanced and super-resolution microscopy techniques.

Fig. 31 Schematic for reversible labeling of retro-engineered HaloTag variants (reHaloTagS and reHaloTagF) with the fluorophore-conjugated ligand.

A peptide tag-assisted split HaloTag complementation (TA-splitHalo) system was constructed to visualize target proteins and detect PPIs in cells.¹⁶⁶ In this setup, HaloTag was split into two inactive fragments, nHalo (1–155) and cHalo (156–257) (Fig. 32a). The resulting fragments were fused individually to two distinct binders, nanobody NbALFA and SpyCatcher002 (SpyC), which interact with the ALFA tag and SpyTag002 (SpyT), respectively. The two HaloTag fragments alone lacked self-complementarity and thus remained inactive for labeling with fluorophore-linked ligands. However, when the peptide tags (ALFA tag and SpyT) were in close proximity, the binders (NbALFA and SpyC) brought together the HaloTag fragments, restoring HaloTag activity and enabling enzyme labeling with chloroalkane substrates. The TA-splitHalo/JF646 system was applied to image the nuclear lamina protein of lamin A/C (LMNA) (Fig. 32b) and fluorescently detect interactions between TUBA1B and TUBB4B in cells (Fig. 32c).

Fig. 32 The TA-splitHalo system. (a) The nHalo and cHalo fragments are separately fused to each binder. When two tags come into proximity, the binders facilitate assembly of the splitHalo fragments, restoring HaloTag activity and enabling labeling of the enzyme with JF646. Table shows two peptide tags and their corresponding binders. Detection of (b) target proteins and (c) protein–protein interactions using the TA-splitHalo system.

The PYP-tag is normally labeled with 7-hydroxycoumarin or 7-dimethylaminocoumarin (DMAC) derivatives through a thioester linkage (Fig. 24b). However, this thioester linkage is susceptible to nucleophilic cleavage by endogenous thiols such as glutathione, leading to a decrease in protein fluorescence in cells over time. To address this instability issue, a ligand capable of forming a stable thioether bond with the PYP-tag was devised and evaluated for its utility in cell imaging.¹⁶⁷ Among several tested ligands (Fig. 33a), a chloromethylketone-containing DMAC-CMK ligand displayed optimal labeling properties, hydrolytic stability, and fluorescence characteristics. DMAC-CMK itself was only weakly fluorescent in aqueous solutions owing to fluorescence quenching by twisted intramolecular charge transfer (Fig. 33b). However, when attached to the hydrophobic binding pocket of the PYP-tag, DMAC-CMK experienced a significant increase in fluorescence intensity. Consequently, there was no need for a washing step to remove residual DMAC-CMK within cells prior to imaging. The DMAC-CMK/PYP-tag labeling system was utilized to visualize telomeric repeat binding factor 2 (TRF2) that is crucial for maintaining telomere integrity and DNA repair (Fig. 33c).

Fig. 33 (a) Chemical structures of electrophilic DMAC derivatives that react with the Cys69 side chain of the PYP-tag. (b) Fluorescence spectrum of DMAC-CMK in the absence (cyan) and presence (red) of the PYP-tag. (c) Fluorescence images of cells expressing HA-PYP-hTRF2 labeled with DMAC-CMK. White and yellow triangles indicate the nucleoli and the nucleoplasm, respectively (reproduced from ref. 167 with permission from the Royal Society of Chemistry, copyright 2020).

The second generation TMP ligand (TMP2) for the eDHFR-tag, developed in 2012,¹⁵³ was covalently attached to the enzyme through Michael addition of the Cys side chain to the acrylamide of TMP2. However, due to its relatively low cell permeability, TMP2 was used for protein imaging in fixed cells. To improve synthetic procedures, labeling kinetics and cell permeability, the third generation TMP ligand (TMP3-MaP555 or TMP3-SiR) was designed and synthesized. Its labeling performance with the eDHFR-tag was compared to that of the HaloTag and SNAP-tag with their cognate ligands (Fig. 34a).¹⁵⁴ The labeling rate of the eDHFR-tag with TMP3-TMR ligand was comparable to that of the HaloTag with Halo-TMR. In addition, the lipophilic chloroalkylated ligand displayed the best cell permeability, with TMP3-SiR and TMP3-MaP555 demonstrating practical permeability (Fig. 34b and c), while BG ligands exhibited the poorest cell permeability. Although the eDHFR-tag, SNAP-tag and HaloTag were inserted at the loop of GFP without affecting GFP fluorescence, the eDHFR-tagged protein was expressed at the highest level in cells. Owing to their orthogonal labeling properties, these three SLE tags were utilized to simultaneously label fusion proteins with three distinct fluorophores.

Fig. 34 (a) Schematic for labeling the eDHFR-tagged protein with TMP3-MaP555 or TMP3-SiR. (b) Confocal fluorescence microscopy images of nucleus-localized histone H2B, ER-localized Sec61b, and F-actin in cells expressing eDHFR-tagged proteins after labeling with TMP3-MaP555. (c) Confocal fluorescence microscopy and super-resolution (STED) images of microtubule-associated protein 7 (MAP7), H2B, F-actin, and outer mitochondrial membrane-localized Tomm20 in cells expressing eDHFR-tagged proteins after labeling with TMP3-SiR (reproduced from ref. 154 with permission from the Wiley-VCH, copyright 2022).

The BL-tag has limitations for protein labeling due to the chemical instability of the β-lactam ring in its ligand under acidic and alkaline conditions, as well as its susceptibility to nucleophiles. In addition, the acyl–enzyme intermediate formed by the reaction of the BL-tag with β-lactam-derived ligands undergoes deacylation through intramolecular rearrangement, limiting long-term protein imaging in cells. To overcome these limitations, non-β-lactam ligands based on the β-lactamase inhibitor MK-7655, containing a diazabicyclo[3.2.1]octane (DBO) scaffold, were developed for prolonged imaging of BL-tagged proteins (Fig. 35a).¹⁶⁸ In these ligands, the Ser side chain of the BL-tag attacks the carbonyl carbon of cyclic urea in DBO to form a stable carbamate linkage (Fig. 35b). The initially developed ligand RD was able to label cell–surface BL-tagged proteins but not intracellular proteins owing to its poor cell permeability. The second ligand, RD2c, was efficiently attached to intracellular fusion proteins. However, the fluorescence signal gradually diminished presumably owing to ligand dissociation from the BL-tag via deacylation through ring closure. To form a more stable complex with the BL-tag by increasing binding affinity, an alkyl spacer was inserted between the TMR dye and the bicyclic ring to generate the RD3 ligand. RD3 was covalently conjugated to the BL-tag with enhanced affinity, enabling the imaging of intracellular proteins over extended incubation periods in live cells.

Fig. 35 (a) Chemical structures of MK-7655 and BL-tag ligands. (b) Schematic for labeling the BL-tagged protein with MK-7655 based ligands.

2.4 Affinity-based fluorescent labeling

The aforementioned methods, including GCE technology, peptide tags, and SLEs, have been widely utilized for fluorescence imaging of cellular proteins. However, these approaches necessitate genetic manipulations to incorporate ncAAs into target proteins or to fuse peptide tags and SLEs to proteins of interest. Such genetic alterations can complicate the analysis of the native state of proteins in cells. To address this, an affinity-based fluorescent labeling strategy has been developed. This method involves modifications of protein-selective ligands to include reactive groups that can interact with and label binding proteins, enabling the fluorescent labeling of native proteins in cells without the need for genetic manipulation.^169,170

2.4.1 Photoaffinity labeling. An initial affinity-based fluorescent labeling method involves covalent labeling of proteins with photoaffinity labels under light irradiation (Fig. 36a). Photoaffinity labels are typically composed of target protein ligands, photoactivatable moieties (e.g., diazirine, diazo, benzophenone, and phenylazide),^171–173 and reporter groups. When these labels bind to target proteins, light irradiation induces covalent labeling of proteins. This process allows for spatiotemporal control by modulating the timing and location of light exposure,¹⁷⁴ making photoaffinity labeling suitable for fluorescence imaging of cellular proteins.

Fig. 36 Affinity-based fluorescent labeling of cellular proteins. (a) Photoaffinity labeling. The target protein recognizes a corresponding ligand that possesses both a fluorophore and a photoreactive group such as diazirine, diazo, benzophenone, or phenylazide. Upon light irradiation, the photoreactive ligand covalently binds to the target protein for fluorescent labeling. (b) Ligand-directed chemistry-based labeling. When a ligand containing a fluorophore linked through a reactive moiety binds to the target protein, the reactive group covalently modifies a nucleophilic residue within the protein for fluorescent labeling. Chemical structures of the reactive moieties are shown, with the leaving groups highlighted in red. (c) Affinity-guided organic catalyst-based labeling. Organic catalysts, such as DMAP and PyOx, are used to activate acyl donors such as thioesters and NASA. These activated donors form reactive intermediates that react with nucleophilic residues within target proteins. The leaving groups in these moieties are also highlighted in red.

For instance, to fluorescently label the AXL protein, which plays a role in tumorigenesis, a high-affinity binding ligand (9im) for AXL was modified with diazirine and alkyne to create AX-1 (Fig. 37a and b).¹⁷⁵ After UV irradiation of cells treated with AX-1, cells were labeled with TMR-N₃via click chemistry, revealing that TMR fluorescence in cells co-localized well with immunofluorescence (Fig. 37c). However, photoaffinity labeling can generate highly reactive intermediates that nonspecifically react with cellular proteins and can cause cytotoxicity due to short wavelength light irradiation.¹⁷⁶ To address nonspecific labeling and reduce background fluorescence, malachite green (MG), which fluoresces only when bound to a fluorogen-activating protein (FAP), was utilized for fluorescent labeling of proteins.¹⁷⁷ Specifically, FAP fused to mCerulean3 was expressed in cells and treated with MG-diazirine composed of MG, Arg, and diazirine (Fig. 38a and b). UV irradiation resulted in the fluorescent labeling of the fused protein with a significantly reduced background (Fig. 38c).

Fig. 37 (a) Schematic for photoaffinity labeling of AXL with AX-1. (b) Chemical structure of AX-1. (c) Cells were treated with AX-1, irradiated, and labeled with TMR-N₃via click chemistry. The fluorescence of TMR was co-localized with immunofluorescence (IF) (reproduced from ref. 175 with permission from the Wiley-VCH, copyright 2018).

Fig. 38 (a) Schematic for photoaffinity labeling of fluorogen-activating protein (FAP) with MG-diazirine. (b) Chemical structure of MG-diazirine. (c) HeLa cells expressing mCerulean3-FAP were exposed to MG-diazirine followed by UV irradiation (reproduced from ref. 177 with permission from the American Chemical Society, copyright 2019).

Also, to alleviate the cytotoxicity caused by UV irradiation, a photoaffinity label containing a diazo group activated by visible blue light was developed for fluorescent labeling of endogenous carbonic anhydrase II (CA-II) in cells.¹⁷⁸ This photoaffinity label, SA-Cou-N₂, consisted of a sulfonamide ligand for CA-II and a non-fluorescent diazocoumarin that becomes fluorescent only when the generated carbene undergoes the insertion reaction with the protein, thereby reducing background fluorescence (Fig. 39a and b). This label was employed to fluorescently image endogenous CA-II in live HCT116 cells (Fig. 39c). Furthermore, in situ two-photon activation using longer wavelength light (800 nm) was utilized to detect endogenous CA-II, reducing cytotoxic effects associated with UV light.

Fig. 39 (a) Schematic for photoaffinity labeling of CA-II with SA-Cou-N₂. (b) Chemical structure of SA-Cou-N₂. (c) HCT116 cells were treated with SA-Cou-N₂ in the absence or presence of a CA-II inhibitor, and then irradiated with blue LED light (reproduced from ref. 178 with permission from the American Chemical Society, copyright 2020).

2.4.2 Ligand-directed chemistry-based labeling. Photoaffinity labeling has significantly contributed to establishing the concept of affinity-based protein labeling. However, challenges such as cytotoxicity from light irradiation and the unintended attachment of photoaffinity labels, which can render target proteins nonfunctional, persisted.^179,180 To mitigate these issues, the Hamachi group introduced ligand-directed (LD) chemistry for the selective labeling of native proteins. Probes used in LD chemistry typically consist of three components: a ligand for the target protein, a reactive moiety that reacts with a nucleophilic amino acid residue within the target protein, and a reporter group (Fig. 36b). When the probe binds to the target protein, the nearby amino acid residue reacts with the reactive moiety, resulting in the transfer of the probe to the protein. Since the ligand dissociates from the protein after labeling, the protein activity remains intact.^169,170 This method has been used to analyze endogenous proteins in live cells because it is technically straightforward and avoids the side effects associated with photoaffinity labeling.

In designing probes for LD chemistry-based labeling, selecting an appropriate reactive moiety is crucial for ensuring labeling efficiency and specificity. If the reactivity of the moiety is too high, it may react with other cellular components or undergo hydrolysis before reaching the target protein. Conversely, if the reactivity is too low, labeling may be inefficient. An early reactive moiety developed for this purpose is the LD tosyl (LDT) probe (Fig. 36b).^181,182 This probe was initially utilized to label endogenous CA and FK506-binding protein 12 (FKBP12) in cells by employing a sulfonamide ligand and a synthetic ligand, respectively.¹⁸³ However, the reactivity of LDT probes is low, thereby leading to slow labeling reactions and low labeling efficiency.

To enhance labeling kinetics, LD acyl imidazole (LDAI) chemistry using alkyloxyacyl imidazole was developed (Fig. 36b).¹⁸⁴ The LDAI probe containing methotrexate, which serves as a ligand for DHFR, was used to label this protein in vitro. The results revealed that the LDAI probe achieves a labeling efficiency of 84%, compared to only 9% with the LDT probe. LDAI chemistry was further applied to image endogenous opioid receptors in rat and mouse brain slices (Fig. 40).¹⁸⁵ The LDAI probe, NAI-A594, used in this study, was comprised of β-naltrexamine (an antagonist for opioid receptors), an acyl imidazole moiety, and a fluorophore (Fig. 40b). Cell studies showed that endogenous opioid receptors in live cells are fluorescently labeled with NAI-A594 even after washing with the high-affinity antagonist naloxone (Fig. 40c). Notably, since the ligand portion of the probe dissociated from the receptor after labeling, the function of the labeled receptor was preserved. Very recently, LDAI chemistry was employed to label various neurotransmitter receptors, such as AMPA receptors (AMPARs), mGlu1, NMDAR, and GABA_AR, in mouse brains.¹⁸⁶ The LDAI probes for this study consisted of a ligand for each neurotransmitter receptor, an acyl imidazole moiety, and a fluorophore (Fig. 41a). These probes were employed for three-dimensional spatial mapping of endogenous receptors (Fig. 41b), multi-color receptor imaging, and pulse-chase analysis of receptor dynamics.

Fig. 40 (a) When NAI-A594 binds to the opioid receptor, a nucleophile residue within the receptor attacks the carbonyl carbon of acyl imidazole, facilitating the transfer of the fluorophore to the protein. (b) Chemical structure of NAI-A594. (c) Fluorescence images of the endogenous opioid receptor labeled with NAI-A594 in locus coeruleus neurons. The labeling of neurons with NAI-A594 remains intact even after washing with the high-affinity antagonist naloxone (reproduced from ref. 185 with permission from eLife Sciences Publications copyright 2019).

Fig. 41 (a) Chemical structures of LDAI probes, designed for AMPAR, mGlu1, NMDAR, and GABA_AR, that contain an acyl imidazole moiety. (b) Confocal fluorescence microscopy images of brain slices labeled with CAM2, CmGlu1M, CNR1M, or CGABAaRM (reproduced from ref. 186 with permission from the National Academy of Sciences, copyright 2024).

LD chemistry is effective for the fluorescent labeling of endogenous receptors in live cells but is inadequate for the quantitative analysis of receptor distribution and population changes when labeled receptors undergo internalization. To tackle this limitation, a two-step labeling method combining LD chemistry with IEDDA reactions was developed.¹⁸⁷ For example, AMPAR was labeled with CAM2(TCO) which consisted of 6-pyrrolyl-7-trifluoromethyl-quinoxaline-2,3-dione (PFQX) serving as the AMPAR ligand, an acyl imidazole moiety, and TCO* (Fig. 42a and b). The labeled protein was then treated with dye-tetrazine for its fluorescent labeling through the IEDDA reaction. Owing to the rapid labeling kinetics with dye-tetrazine (Fig. 42c), AMPAR internalization after fluorescent labeling was successfully monitored by fluorescently imaging the protein.

Fig. 42 (a) CAM2(TCO) binds to AMPAR, where a nucleophilic residue attacks the carbonyl carbon of acyl imidazole. This process facilitates the transfer of TCO to the protein. Subsequently, the modified protein is subjected to the IEDDA reaction with dye-tetrazine for fluorescent labeling. (b) Chemical structure of CAM2(TCO). (c) Confocal fluorescence microscopy images of HEK293T cells treated with CAM2(TCO), followed by time-dependent fluorescent labeling with dye-tetrazine (reproduced from ref. 187 with permission from the Springer Nature, copyright 2021).

Phenyl ester derivatives have been also utilized as reactive groups in LD chemistry. The Hamachi group, for example, developed LD dibromophenyl benzoate (LDBB) probes (Fig. 36b). These probes, designed to impose steric hindrance, are resistant to cellular esterases, offering enhanced stability and reactivity, which facilitates efficient protein labeling.¹⁸⁸ Later, a fluorophenyl ester-based LD probe was devised for fluorescent labeling of the adenosine A_2A receptor (A_2AR) on the surface of cells. Fluorophenyl esters have the advantages of minimal interference with ligand binding due to their small size and high reactivity.¹⁸⁹ This LD probe was composed of ZM241385 (an A_2AR antagonist), o-fluorophenyl ester, and a fluorophore (Fig. 43b). The developed LD probe specifically labeled A_2AR without detecting other subtypes (A₁R, A_2BR, and A₃R) in live HEK293 cells (Fig. 43c). Also, difluorophenyl ester-based LD probes with sulfonamide were designed and utilized to detect hCA isoforms within cells (R = Ac) and on the cell surface (R = H) under hypoxic conditions (Fig. 44).¹⁹⁰ A cell-permeable version of this probe (R = Ac) was further employed for fluorescent detection of tumor in xenografted mice.

Fig. 43 (a) Schematic for fluorescent labeling of A_2AR with a fluorophenyl ester-based LD probe. (b) Chemical structure of a fluorophenyl ester-based LD probe. (c) (Upper) Fluorescence images of HEK293 cells following treatment with a fluorophenyl ester-based LD probe and (lower) fluorescence images of HEK293 cells expressing SNAP-tagged receptors labeled with the SNAP ligand (reproduced from ref. 189 with permission from the Springer Nature, copyright 2020).

Fig. 44 (a) Chemical structure of difluorophenyl ester-based LD probes (R = H; a cell-impermeable form, R = Ac; a cell-permeable form). (b) Fluorescence images of endogenous CAs treated with Probe-H (left) or Probe-Ac (right) under conditions with and without a CA inhibitor in MCF-7 cells (reproduced from ref. 190 with permission from the Wiley-VCH, copyright 2021).

The LD N-sulfonyl pyridone (LDSP) probe was designed for labeling endogenous proteins in cells (Fig. 36b).¹⁹¹ This probe exhibited a low rate of hydrolysis and labeled proteins through a sulfonylation reaction, which is orthogonal to typical cellular enzymes, thereby maintaining its stability within cells. The LDSP probe was used to fluorescently visualize endogenous CAs in live MCF7 cells as well as to determine the binding affinity for CA inhibitors by monitoring FRET signals between the FRET donor-labeled CA and FRET acceptor-linked ligands.

Another type of LD probes containing N-acyl-N-alkyl sulfonamide (NASA) as the reactive moiety was developed for fluorescent labeling of cellular proteins (Fig. 36b).¹⁹² An electron-withdrawing cyanomethyl group was inserted into LDNASA probes to enhance the electrophilicity of sulfonamide. Consequently, LDNASA probes exhibited a reaction rate comparable to the IEDDA reaction between tetrazine and TCO (2.9 × 10⁴ M⁻¹ s⁻¹), with excellent stability against hydrolysis and enzymatic degradation. These features enabled LDNASA probes to achieve quantitative labeling of endogenous FKBP12 within 2 h, whereas LDT probes labeled only 30% of the protein after 48 h. Notably, LDNASA probes selectively labeled Lys residues of target proteins, unlike other forms of LD probes that labeled various amino acid residues. Very recently, a universal platform of LDNASA probes was developed for cellular protein labeling.¹⁹³ These probes were composed of the receptor-binding ligand (bait), a fluorophore (cargo), and a linker possessing a NASA moiety (platform reagent) that reacts with the Lys residue of the protein (Fig. 45a). The linker facilitates the transfer of the cargo to the target protein as well as enables easy modifications with a ligand and a cargo for various applications. The LDNASA probe bearing a HU-308 derivative (Fig. 45b) serving as a cannabinoid receptor 2 (CB2R) agonist was employed to fluorescently label the CB2R in various types of cells (Fig. 45c).

Fig. 45 (a) Structure of a universal platform of LDNASA probes consisting of a receptor-binding ligand (bait), a fluorophore (cargo), and a linker bearing a NASA moiety (platform reagent). (b) Chemical structure of a HU-308 derivative used as a CB2R agonist. (c) Fluorescence images of hCR2B in live AtT-20 cells following treatment with a LDNASA probe in the absence and presence of its inhibitor (reproduced from ref. 193 with permission from the American Chemical Society, copyright 2023).

An LDNASA probe containing an aptamer instead of a small molecule ligand was constructed for fluorescence imaging of the endogenous target protein in cells.¹⁹⁴ This probe comprised the sgc8c aptamer, which binds to protein tyrosine kinase 7 (PTK7), a NASA moiety, and biotin (Fig. 46). Upon recognition of PTK7 by the aptamer, the Lys residue attacks the NASA moiety in the probe, resulting in the labeling of the protein with biotin, which can be detected using dye-labeled streptavidin. This probe was utilized to monitor the internalization and localization of PTK7 in NIH3T3 cells.

Fig. 46 Fluorescent labeling of protein tyrosine kinase 7 (PTK7) with a LDNASA probe equipped with a PTK7-targeting aptamer. Upon binding of the sgc8c aptamer of the probe to PTK7, the Lys residue undergoes modification with biotin, which is subsequently detected using dye-labeled streptavidin.

2.4.3 Affinity-guided organic catalyst-based labeling. An affinity-guided organic catalyst-based labeling approach has also been developed to selectively label target proteins (Fig. 36c). This method utilizes both ligand-conjugated organic catalysts and dye-linked acyl donors. Initially, dimethylaminopyridine (DMAP) was employed as an acyl transfer organic catalyst that facilitates the labeling of target proteins with a dye-conjugated thioester acyl donor.¹⁹⁵ When a DMAP-linked ligand binds to the target protein, the DMAP moiety attacks the thioester of the supplemented acyl donor, followed by an attack of nucleophilic residues of target proteins to the activated carbonyl group, thereby leading to protein labeling. Using a DMAP-conjugated sugar ligand and a dye-linked thioester acyl donor, target lectins were labeled with a dye up to 62% efficiency in vitro after 3 h incubation.

Photoaffinity and LD chemistry-based labeling strategies typically require small molecule ligands for target proteins, which can be challenging to discover. Alternatively, small proteins that selectively bind to target proteins can serve as surrogates for these small molecule ligands. In this approach, protein ligands are first modified with organic catalysts. When organic catalyst-conjugated protein ligands bind to target proteins and are subsequently treated with dye-linked acyl donors, the fluorophores transfer from the acyl donors to the target proteins. For instance, HER2 in cells was fluorescently labeled using this method.¹⁹⁶ Specifically, maleimide-appended DMAP was attached to the Cys residues of a single-chain variable fragment (scFv) of an anti-HER2 antibody. When HER2-expressing N87 cells were incubated with the DMAP-conjugated scFv and then treated with a dye-linked thioester acyl donor, HER2 was fluorescently labeled in cells.

Despite its capability to label cellular proteins with dyes, DMAP-based protein labeling requires a relatively high pH (above pH 8) to function efficiently. In addition, thioester acyl donors are prone to enzymatic hydrolysis and can cause nonspecific labeling due to their high reactivity. To overcome these limitations, affinity-guided oxime chemistry was developed using pyridinium aldoxime (PyOx) as an acyl transfer catalyst and a NASA acyl donor.¹⁹⁷ NASA is orthogonal to cellular enzymes, resistant to enzymatic degradation, and inherently less reactive than thioesters, thereby minimizing nonspecific labeling.

To enhance labeling efficiency, a pre-formed complex of the oxime catalyst and the acyl donor was attached to the protein ligand (Fig. 47).¹⁹⁸ The proximity of the oxime catalyst and the acyl donor significantly increased the efficiency for target protein labeling. Specifically, PyOx was fused to the His₆-tag containing anti-HER2 scFv, and a dye-linked acyl imidazole donor conjugated with NTA was then attached to the His₆-tag via Ni²⁺ ions (Fig. 47a). Upon binding of the modified anti-HER2 scFv to HER2, the oxime attacked the acyl imidazole donor to form a reactive intermediate, which reacted with nucleophiles within HER2 for fluorescent labeling. This strategy achieved nearly 20 times greater labeling compared to the acyl donor lacking the NTA complex. Using this method, HER2 in N87 cells was fluorescently detected, while HER2-negative HEK293T cells showed no detectable fluorescence (Fig. 47b). Despite the advantages of affinity-guided organic catalyst-based labeling for protein labeling, the low cell permeability of NASA acyl donors limits its application for labeling intracellular proteins. Therefore, acyl donors with improved cell permeability need to be developed to expand this strategy for labeling a wide range of cellular proteins.

Fig. 47 (a) Schematic for HER2 labeling with the anti-HER2 scFv containing a His₆-tag conjugated with PyOx and an acyl imidazole donor. (b) Confocal fluorescence microscopy images of N87 and HEK293T cells. N87 and HEK293T cells were treated with the modified anti-HER2 scFv bearing the PyOx and the acyl imidazole donor. HEK293T cells were biotinylated using biotin-NHS and subsequently labeled with an Alexa647-streptavidin (SAv-647) to distinguish them from N87 cells (reproduced from ref. 198 with permission from the American Chemical Society, copyright 2023).

2.5 Comparison of labeling methods

Over the last two decades, numerous methods have been developed to fluorescently label target proteins with small organic dyes, enabling their visualization in cells and tissues. Despite significant progress in these techniques, there remains substantial room for improvement. The optimal system for fluorescence imaging of proteins in live cells should have several critical attributes: specific targeting of proteins, compact tag sizes, high efficiency in labeling, rapid kinetics for labeling, stable linkages between proteins and fluorophores, versatility for labeling both cell–surface and intracellular proteins, and the capability for multi-color labeling (Table 2). Because there is no ideal labeling method and each comes with its own strengths and limitations, researchers must carefully consider various practical factors when implementing fluorescently labeled proteins in cellular studies.

Table 2 Comparison of labeling methods

Labeling method		Tag size	Genetic modification	Labeling selectivity	Labeling efficiency	Versatility	Labeling kinetics	Conjugate stability	Multi-color labeling
Genetic code expansion	Fluorescent amino acid	Small	Required	High	High	High	—	High	Difficult
	Non-fluorescent amino acid	Small	Required	High	Moderate to high	High	Moderate to fast	High	Possible
Peptide tag	Small molecule binder	Medium	Required	Moderate to high	High	High	Moderate to fast	Moderate to high	Possible
	Metal complex binder	Medium	Required	Moderate to high	High	Moderate	Moderate to fast	Moderate	Difficult
	Peptide assembling	Medium	Required	High	High	Low	Fast	Moderate to high	Possible
	Enzyme-catalyzed	Medium	Required	High	High	High	Fast	High	Possible
Self-labeling enzyme tag		Large	Required	High	High	Moderate to high	Fast	High	Possible
Affinity-based labeling		Small	Not required	Moderate to high	Moderate to high	Moderate	Moderate to fast	High	Possible

---

Generally, tags are much smaller than the target proteins. However, larger tags, particularly enzyme tags, can sometimes interfere with the structure and function of proteins within biological systems. The smallest tags currently available are ncAAs, which can be incorporated site-specifically into target proteins using GCE technology.^17,18 This approach offers remarkable versatility, allowing the labeling of proteins at any position on the cell surface or inside cells. Fluorescent amino acids are particularly advantageous for protein labeling compared to non-fluorescent amino acids that require an additional ligation step with fluorophores via bioorthogonal handles. Importantly, background fluorescence arising from free fluorescent amino acids and their nonspecific incorporation into endogenous amber codons is minimal.⁴¹

Despite these advantages, the range of fluorescent amino acids available for protein labeling is limited in terms of fluorescence emission characteristics. In contrast, non-fluorescent amino acids with various bioorthogonal handles enable labeling with fluorophores emitting from UV-VIS to NIR wavelengths. While GCE technology eliminates the need for additional peptide tags, it necessitates genetic engineering of host cells and supplementation with ncAAs, which can be costly or unavailable commercially. Moreover, inserting multiple ncAAs into target proteins at different positions for multi-color fluorescent labeling, if feasible, is a challenging task.^98,199,200 Overall, GCE technology using ncAAs offers substantial flexibility in protein labeling, but researchers must navigate trade-offs between fluorescent and nonfluorescent amino acids based on specific experimental requirements and practical considerations.

Peptide tag-based fluorescent labeling of cellular proteins provides an alternative approach to using ncAAs via GCE technology.^99,100 Genetic fusion of peptide tags to target proteins is straightforward, and peptide tags are relatively small in size (0.6–6 kDa). Various bioorthogonal ligation methods have been developed to facilitate the fluorescent labeling of proteins tagged with peptides.²⁰¹ In this approach, small molecule or metal ion-dependent labeling of peptide tags can be employed to label both cell–surface and intracellular proteins, with labeling kinetics varying from moderate to fast depending on the specific labeling sites. The linkages formed between small molecules or metal complex binders and peptide tags are relatively stable. However, these substances may also interact with other endogenous proteins that share sequences similar to the peptide tags. In addition, the incorporation of Cys-rich sequences into proteins can sometimes lead to improper disulfide bond formation, potentially affecting protein structures. An additional issue of peptide tag-based labeling is the limited number of peptide tags recognized by small molecules or metal complex binders, which restricts the ability to perform multi-color protein labeling effectively.

Coiled–coil peptide-based fluorescent labeling of cellular proteins offers rapid labeling kinetics, particularly suitable for cell–surface proteins.^99,100 However, this method is less suitable for labeling intracellular proteins owing to the low cell permeability of binding peptides. Although multi-color protein labeling has been achieved using coiled–coil peptides, further development of additional coiled–coil peptide pairs is needed to broaden its applicability. Furthermore, the noncovalent nature of coiled–coil peptide-based labeling restricts prolonged cell imaging. Peptide-templated acyl transfer-based labeling has been developed to address this issue. This approach requires the insertion of peptide tags at the exposed terminus of cell–surface proteins to facilitate peptide-templated acyl transfer.

Enzyme-catalyzed peptide tag labeling allows rapid labeling of target proteins within cells through internal enzyme expression or rapid labeling of cell–surface proteins by treating with external enzymes. This method also enables multi-color fluorescence imaging of proteins by utilizing different peptide tags and corresponding enzymes. Despite its precision in protein labeling, enzyme-catalyzed peptide tag labeling necessitates sufficient enzyme expression to ensure complete ligation.

SLE-based fluorescent labeling of proteins provides high specificity and rapid labeling reactions, making it particularly well-suited for studying dynamic biological processes in time-lapse experiments.^139,140 This method also allows multi-color protein labeling by using two different SLEs, and it is effective for labeling both cell–surface and intracellular proteins. Another advantage for this method is the commercial availability of ligands containing fluorophores with diverse emission wavelengths. However, the fusion of large enzymes (18–33 kDa) to target proteins can potentially perturb their structure and function, thereby limiting applications such as studying protein trafficking and PPIs. Moreover, bulky tags are not suitable to develop sensitive sensors that can detect subtle changes in protein conformation, which are crucial for understanding the complexity of cell–surface receptor signaling. To mitigate any potential interference with protein localization and function, efforts should focus on minimizing the size of SLE tags.

Affinity-directed labeling represents a distinct advantage by enabling the direct labeling of native proteins in cells. This approach is especially valuable for labeling endogenous proteins in their natural states directly in host cells, a capability not achievable with other previously discussed methods.^169,170 This approach involves creating a probe that integrates a linker and a fluorophore with a small molecule ligand specifically targeting the desirable protein. This design allows efficient labeling of the target protein simply by exposing host cells to the probe. However, the effectiveness of affinity-directed labeling hinges on the availability of a small molecule ligand that exhibits specificity for the target protein. In addition, the design of labeling probes must be meticulous since the labeling efficiency depends greatly on both the binding affinity of the ligand for the target protein and the reactivity of the linker.

3. Applications

Fluorescently labeled proteins have been widely utilized in a broad range of protein-associated biological investigations. The most common application is visualizing proteins of interest in cells and tissues. Also, they are valuable for assessing PPIs and exploring protein translocation. Fluorescence-labeled proteins are also instrumental in studying the oligomeric status, internalization, conformational changes, and the membrane potential of proteins embedded in cell plasma membranes. Furthermore, they are employed to investigate protein degradation and processing, as well as to detect protein-specific glycans. Owing to their wide range of applications, proteins labeled with fluorophores are essential tools in biological research. This section discusses the various biological applications of fluorescence-labeled proteins in cells.

3.1 Protein–protein interactions (PPIs)

Proteins in cells modulate numerous biological processes through protein complexes formed by PPIs.^202,203 Abnormal PPIs can lead to various human diseases, particularly cancer and neurodegenerative disorders, making them significant targets for therapeutic agent discovery.^204–209 Fluorophore-conjugated proteins have been applied to study PPIs in cells. For instance, we analyzed the Bax–Hsp70 interaction in a single live cell using a FRET system consisting of Hsp70-YFP fusion and ANAP-incorporated Bax (Bax-ANAP) by GCE (Fig. 48a).^41,42 Because the emission spectrum of ANAP overlaps well with the absorption spectrum of YFP, ANAP serves as a FRET donor and YFP as a FRET acceptor. Bax is a key proapoptotic protein involved in mitochondria-associated apoptotic pathways.²¹⁰ Upon activation by various apoptotic stimuli, Bax undergoes oligomerization in mitochondria, inducing mitochondrial outer membrane permeabilization (MOMP) and leading to apoptosis.²¹¹ Hsp70 interacts directly with Bax to suppress apoptosis and thus acts as an anti-apoptotic protein.^212,213 When apoptosis is evoked under certain conditions, Bax dissociates from the Hsp70 complex and is translocated to mitochondria to promote mitochondria-associated apoptosis.²¹³ The FRET system composed of Bax-ANAP and Hsp70-YFP was used to evaluate the effects of various substances on the Bax–Hsp70 interaction. Analysis of real-time FRET images of treated cells revealed that a Bax activator Bam7, which binds to Bax trigger sites, disrupts the Bax–Hsp70 interaction, whereas another Bax activator SMBA1, which blocks phosphorylation of S184, does not influence this interaction (Fig. 48b). In addition, a p53 activator (nutlin 3a), an ER stress inducer (brefeldin A), a lysosomal membrane destabilizer (apoptozole),^{212,214–218} and a death ligand (TRAIL) promoted the release of Bax from the Bax/Hsp70 complex. However, an inhibitor of lysosomal acidification (bafilomycin A1) and a mitochondria-targeted apoptosis inducer (raptinal) had no influence on this PPI. The FRET system consisting of ANAP-incorporated protein and YFP fusion proteins proved to be a powerful method for determining PPIs in cells.

Fig. 48 (a) FRET system composed of Hsp70-YFP and ANAP-incorporated Bax (Bax-ANAP) for study of the Bax–Hsp70 interaction. When Bax interacts with Hsp70, FRET takes place from ANAP to YFP with excitation of ANAP. However, upon dissociation of Bax from Hsp70 induced by substances, FRET signals are reduced with excitation of ANAP. (b) Real-time FRET images of cells co-expressing Bax-ANAP and Hsp70-YFP after incubation with indicated substances (reproduced from ref. 41 and 42 with permission from the American Chemical Society, copyright 2019 and the Wiley-VCH, copyright 2020, respectively).

The combination of GCE technology with the SNAP-tag was utilized to study interactions between AMPARs and transmembrane AMPAR regulatory proteins (TARPs) in cells.⁸⁵ AMPARs are glutamate receptors involved in excitatory synaptic transmission, while TARPs are crucial regulators of AMPAR-mediated synaptic transmission and plasticity. In this effort, TCO*K was incorporated into the Ex1 loop of two TARP family members, γ2 and γ8, followed by labeling with Cy3-linked tetrazine (H-Tet-Cy3) as a FRET acceptor (Fig. 49a). To fluorescently label the AMPAR subunit GluA1, a SNAP-tag was fused either to its N-terminus (SNAP-GluA1) or within the linker between the amino-terminal domain (ATD) and ligand-binding domain (LBD) of GluA1at position 396 (GluA1–SNAP396). These constructs were subsequently labeled with BG-AF488 as a FRET donor. AF488 fluorescence lifetime did not decrease in cells co-expressing SNAP–GluA1 and either γ2 or γ8, presumably owing to a long distance between the two fluorophores (Fig. 49b left). In contrast, its lifetime was robustly attenuated in cells expressing GluA1–SNAP396 and either γ2 or γ8 because of FRET from AF488 to Cy3 (Fig. 49b right). These findings demonstrated that the distance between a FRET donor and acceptor is crucial for studies aimed at determining PPIs.

Fig. 49 (a) Schematic for fluorescent labeling of GluA1 and TARPs (γ2 or γ8) used in FRET analysis. (b) Spinning disk (AF488 and Cy3) and widefield illumination FLIM (AF488 lifetime) images of cells co-expressing either (left) Cy3-labeled γ2 and AF488-labeled SNAP-GluA1 or (right) Cy3-labeled γ2 and AF488-labeled GluA1-SNAP396 (reproduced from ref. 85 with permission from the Springer Nature, copyright 2021).

Split SLEs (split HaloTag and split SNAP-tag), whose reconstitution induced by PPIs restores enzyme activity for fluorescent labeling of proteins, were also utilized to probe PPIs.^161,166 For example, the SNAP_f-tag was divided into two parts, SNAP_f (1–74, SmFP485N) and SNAP_f (75–182, SmFP485C), which were fused to two different interacting proteins (Fig. 50a).¹⁶¹ In this system, PPIs induce the association of the two SNAP_f parts, consequently restoring its activity for labeling with BG-F485 to generate fluorescent split SmFP485 (Fig. 28). Using split SmFP485, the rapamycin-induced interaction between the rapamycin binding domain of mTOR (FRB) and FK506 binding protein (FKBP) was evaluated in cells. The results revealed that fluorescence in cells co-expressing SmFP485N–FRB and FKBP–SmFP485C after labeling with BG-F485 rapidly increases in the presence of rapamycin, which promotes the FRB–FKBP interaction (Fig. 50b).

Fig. 50 (a) Schematic for bimolecular fluorescence complementation of SmFP485 to detect the rapamycin-induced interaction between FRB and FKBP. (b) Real-time fluorescence images of rapamycin-induced fluorescence of SmFP485 (reproduced from ref. 161 with permission from the Springer Nature, copyright 2023).

3.2 Protein translocation

Protein translocation to appropriate destinations within cells is critical for their proper function, often relying on specific targeting signals or sequences.^219,220 Conventional methods, such as Western blotting and immunocytochemistry, are not suitable for real-time analysis of protein translocation in live cells. In contrast, fluorescence imaging is an excellent technique for studying protein translocation in real time within live cells.

We used fluorescently labeled apoptosis-inducing factor (AIF) to assess its real-time translocation from mitochondria to the nucleus via the cytosol in cells (Fig. 51a).⁴³ The N-terminal portion of AIF is attached to the inner mitochondrial membrane, with the C-terminus exposed to the mitochondrial intermembrane space. Upon apoptotic stimulation, the N-terminal sequence of AIF is cleaved by proteases to produce truncated AIF, which is translocated to the cytosol through MOMP. In the cytosol, AIF is further translocated to the nucleus, where it binds directly to DNA and recruits proteases and nucleases that degrade chromatin and DNA, inducing AIF-mediated apoptosis. For real-time spatial and temporal analysis of AIF translocation induced by various apoptosis inducers, ANAP-incorporated AIF was constructed via GCE. As expected, AIF was mostly localized in the mitochondria of untreated cells (Fig. 51b). However, treatment with a topoisomerase inhibitor (doxorubicin and camptothecin), a microtubule stabilizer (paclitaxel) and an Akt activator (silibinin) led to AIF translocation to the nucleus in a time-dependent manner, indicating that these substances act as AIF-mediated apoptosis inducers. In contrast, a Hsp70 inhibitor (apoptozole), a Bax activator (Bam7), a p53 activator (RITA), an ER stress inducer (brefeldin A) and a mitochondrial membrane disruptor (FCCP) enhanced the translocation of mitochondrial AIF to only the cytosol without its translocation to the nucleus (Fig. 51b). As a consequence, these substances serve as AIF-independent apoptosis inducers.

Fig. 51 (a) Schematic for translocation of ANAP-incorporated AIF to the cytosol and then the nucleus. (b) HeLa cells expressing ANAP-AIF (green) were treated with indicated substances in the presence of MitoTracker Red (red). Shown are merged images of cells in green and red (reproduced from ref. 43 with permission from the American Chemical Society, copyright 2021).

The SNAP-tag mimic of the fluorescent protein, SmFP, was employed for real-time visualization of protein translocation (see Fig. 28).¹⁶¹ Pulse-chase studies using conventional fluorophores normally require a washing step to remove unlabeled probes during the chase period. However, SmFP does not require washout, allowing real-time monitoring of newly synthesized proteins in the ER of live cells. In this effort, a trans-Golgi targeting signal peptide was fused to SNAP_f to construct Golgi-SNAP_f. Trafficking of Golgi-SNAP_f in live cells was monitored after sequential labeling with BG-F643, BG-F555 and BG-F485 for 4 h each to generate SmFP643, SmFP555 and SmFP485, respectively (Fig. 52a and b). Golgi-SNAP_f expressed at different times displayed distinct localizations (Fig. 52c). Specifically, the newly synthesized Golgi-SNAP_f labeled with SmFP485 accumulated in Golgi-like structures and reticular structures dispersed throughout the cytoplasm. Older Golgi-SNAP_f labeled with BG-F643 appeared as puncta that colocalized with the lysosome marker Lamp1-RFP fusion. Furthermore, the fluorescence arising from Golgi-SmFP643 gradually decreased over time, indicating protein degradation in lysosomes. Thus, the multi-color pulse-chase approach using SmFP proved to be an efficient tool for real-time monitoring of protein translocation in live cells.

Fig. 52 (a) Structures of BG-fluorophores used in the multi-color pulse-chase study. (b) Schematic for Golgi-SNAP_f translocation. (c) Multi-color pulse-chase images of Golgi-SNAP_f trafficking. Cells co-expressing Golgi-SNAP_f and Lamp1-RFP were sequentially incubated with BG-F643, BG-F555 and BG-F485 each for 4 h (reproduced from ref. 161 with permission from the Springer Nature, copyright 2023).

To visualize lysosomal translocation of proteins during autophagy induced by starvation, the BL-tag (see Fig. 35) was fused to an intracellular protein and labeled with a pH-activatable protein labeling probe pH-RD4 (Fig. 53a).¹⁶⁸ The pH-RD4 probe, containing the pentyl ester group to enhance cell permeability, is hydrolyzed by cytosolic esterases inside cells to generate pH-RD4–CO₂⁻. When recognized by the BL-tagged protein, pH-RD4–CO₂⁻ forms a covalently-labeled, non-fluorescent BL-tagged protein (Fig. 53b). Under acidic conditions such as inside lysosomes, the non-fluorescent pH-RD4–CO₂⁻ is converted to its fluorescent form, allowing detection of the target protein. Cells expressing the BL-tagged EGFP labeled with pH-RD4 in starved media displayed red fluorescence puncta from pH-RD4, which colocalized with LysoTracker fluorescence (Fig. 53c). These findings indicate the translocation of BL-tagged EGFP to acidic lysosomal compartments during starvation-induced autophagy. In contrast, red fluorescence puncta from pH-RD4 were not observed in cells under nutrient-rich conditions.

Fig. 53 (a) Chemical structure of a cell-permeable pH-activatable pH-RD4 and its fluorescence activation at the acidic pH. (b) Schematic for the fluorescent detection of the BL-tagged protein after labeling with pH-RD4–CO₂⁻ at the acidic pH. (c) Fluorescence images of cells expressing the BL-tagged EGFP cultured in nutrient and starved media after labeling with pH-RD4 (reproduced from ref. 168 with permission from the Wiley-VCH, copyright 2023).

Most Gram-negative pathogens possess protein translocation systems that facilitate the delivery of pathogenic effector proteins into target host cells.²²¹ One such system is the Type III secretion system (T3SS), which is crucial for translocating virulence effector proteins to host cells.²²² The HaloTag, SNAP-tag, and CLIP-tag were utilized to investigate the translocation and subcellular localization of T3SS effector proteins within host cells.²²³ In addition, these SLEs were employed to assess the single-molecule localization and dynamics of effector proteins in live host cells infected by bacterial pathogens.

3.3 Oligomerization of G protein-coupled receptors (GPCRs)

G protein-coupled receptors (GPCRs), also known as 7-transmembrane receptors, represent the largest family of signaling receptors. GPCRs mediate numerous cellular responses to hormones and neurotransmitters, and they play a critical role in the sense of vision, smell and taste.^224,225 Due to their involvement in various diseases such as diabetes, obesity, depression, cancer, and Alzheimer's disease, GPCRs have attracted significant attention for their potential as targets in the discovery of therapeutic agents for a wide range of diseases.^226,227 GPCRs are categorized into four subfamilies based on amino acid sequences: class A (rhodopsin), class B (secretin and adhesion), class C (glutamate), and class F (frizzled).²²⁷ Originally, these receptors were thought to function as simple monomers, but numerous studies have shown that they also form dimers and higher-order oligomers for their cellular functions.^228–231 Fluorescence-labeled GPCRs have been employed to investigate their oligomeric status on the cell surface.

One example of studying GPCR oligomerization involves determining the oligomerization state of the metabotropic glutamate receptor (mGluR) in neurons.²³² mGluRs, which are predominantly expressed in the central nervous system, are responsible for the regulation of synaptic transmission and neuronal excitability. Malfunctions in these receptors are closely linked to diverse neurological disorders.²³³ Previous studies have shown that mGluRs exist as homodimers connected by an extracellular disulfide bridge, which is essential for G protein activation induced by agonists.²³⁴ To characterize the oligomeric status of mGluR in rat hippocampal neurons, the SNAP-tag was fused to the N-terminus of mGluR-2 to generate ST-mGluR-2 and subsequently labeled with a cell-impermeable dye Alexa488 (Fig. 54a). Fluorescence images of neurons were analyzed using fluorescence fluctuation microscopy and scanning Number and Brightness (sN&B) techniques (Fig. 54b and c). The analysis revealed that mGluR-2 mainly exists as dimers at physiological expression levels, but higher-order oligomers form as expression levels increase. Moreover, upon activation (with an agonist) or inhibition (with an antagonist) of mGluR-2, larger oligomers formed even at expression levels where mGluR-2 remained dimeric under basal conditions.

Fig. 54 (a) Fluorescent labeling of SNAP-tagged homodimeric mGluR-2 (ST-mGluR-2) with Alexa488. (b) Two-photon fluorescence microscopy image of primary neurons expressing ST-mGluR-2 labeled with Alexa488. (c) Molecular brightness map of the dashed square in (b) (reproduced from ref. 232 with permission from the Springer Nature, copyright 2018).

The CXC chemokine receptor 4 (CXCR4), a class A GPCR, is another GPCR member that regulates physiological processes such as cell migration and development.²³⁵ Dysregulated CXCR4 function causes cancer progression, and viral and immune diseases.²³⁵ To study the organization of CXCR4 in live cells under different expression levels and in the presence of CXCR4 ligands, cells expressing N-terminally SNAP-tagged CXCR4 (SNAP–CXCR4) were labeled with SNAP-surface 549.²³⁶ Analyses of cell images using single molecule microscopy and fluorescence fluctuation spectroscopy showed that SNAP–CXCR4 predominantly exists as monomers at low expression levels. However, as CXCR4 expression increased, homodimers became more prevalent. Furthermore, agonist-induced stimulation of CXCR4 promoted the formation of receptor dimers that are conformationally distinct from the basal dimers.

The μ-opioid receptor (μOR, a class A GPCR) is a member of the GPCR family with high affinity for enkephalins and β-endorphin and is involved in mediating the analgesic effects of opioids. In a study aimed at evaluating the oligomeric status of the μOR, cells expressing N-terminally SNAP-tagged μOR (SNAP-μOR) were labeled with SNAP-Surface 549.²³⁷ Analysis of cell images utilizing single-molecule total internal reflection fluorescence (TIRF) microscopy combined with super-resolution techniques revealed that under basal conditions, μOR displays a distinctively monomeric behavior. Subsequently, the effects of agonists and antagonists on receptor dimerization were explored. It was found that the full agonist DAMGO ([D-Ala², N-MePhe⁴, Gly-ol]-enkephalin) exclusively promotes the formation of dimeric μOR. In contrast, other ligands tested, including the reversible antagonist naloxone, the irreversible antagonist β-funaltrexamine, and the agonist morphine (which does not induce receptor internalization), do not alter the monomeric state of the μOR. These findings shed light on the dynamic behavior of μOR oligomerization in response to specific ligands.

Bioorthogonal ligation between TCO and tetrazine was utilized to explore the oligomerization state of the corticotropin releasing factor receptor type 1 (CRF₁R, a class B GPCR) (Fig. 55).²³⁸ In this investigation, TCO*K was introduced into EGFP-fused CRF₁R at position 263 in cells through GCE. The resulting TCO*K–CRF₁R–EGFP in cells was labeled with varying ratios of Cy3- and Cy5-tetrazine. This dual-color competitive TCO labeling approach obviated the necessity for a monomeric protein reference to assess the receptor's oligomerization state. The apparent molecular brightness was evaluated across three fluorescence channels (Cy3, Cy5, and EGFP) and plotted against the percentage of Cy3 in the labeling solution. The apparent brightness of EGFP remained consistent regardless of the labeling mixture (Cy3 and Cy5). In contrast, analysis of the measured brightness of Cy3 and Cy5 suggested the formation of dimeric receptors on the cell surface. However, researchers should be mindful that the substantial variance in brightness values is observed due to the significantly heterogeneous expression of the receptor in cells, necessitating meticulous data analysis to accurately determine the receptor's oligomerization state using this method.

Fig. 55 Schematic for fluorescent labeling of EGFR-fused and TCO*K-incorporated CRF₁R with a mixture of two dyes (Cy3- and Cy5-tetrazine) to evaluate the oligomerization state of the CRF₁R.

An orthogonal peptide-templated labeling approach was employed to examine the lateral proximity of two different GPCRs on plasma membranes of cells (Fig. 21).¹³⁶ In this study, the Cys-P1-tag and Cys-P3-tag were individually fused to the N-termini of two different GPCRs involved in cardiovascular regulation (Fig. 56). Leveraging the selective coiled–coil interactions of P1/P2 or P3/P4, Cys-P1 and Cys-P3-tagged GPCRs expressed on the cell surface were simultaneously and selectively labeled with respective Atto488-P2 (a FRET donor) and Atto565-P4 (a FRET acceptor) through peptide-templated acyl transfer. Analysis of FRET images of cells co-expressing endothelin receptors A and B (ET_AR and ET_BR) revealed their close proximity, indicating the formation of a heterodimer.

Fig. 56 Fluorescent labeling of two different GPCRs with the FRET donor (Atto488) and acceptor (Atto565) through orthogonal peptide-templated acyl transfer. The interaction between two proteins leads to an increase in the FRET signal.

3.4 Cell–surface receptor internalization

Receptors residing in plasma membranes of cells transmit extracellular signals to the cytoplasm through interactions with specific ligands outside the cell. These receptors are involved in a wide range of fundamental cellular processes, including the cell cycle, proliferation, and communication.^227,239 Transmembrane receptors are typically comprised of three components: an extracellular ligand-binding region, a transmembrane region, and an intracellular region. Under basal conditions, most receptors exist as freely diffusing monomers or discrete oligomers that do not function as signaling molecules. However, upon ligand binding, receptors become activated through ligand-induced receptor oligomerization.^240,241 In addition, activated receptors undergo internalization into the cell, which can have both positive and negative effects on signaling.²⁴² Thus, membrane receptor internalization is an integral part of cell signaling. Fluorescent labeling techniques have been utilized to monitor ligand-induced internalization of cell–surface receptors.

An example illustrating this research involves the investigation of epidermal growth factor receptor (EGFR) endocytosis in real time. This was achieved through the combination of GCE technology and bioorthorgonal ligation to fluorescently label EGFR (Fig. 57a).²⁴³ EGFR, a single transmembrane glycoprotein, regulates vital cellular processes such as proliferation, differentiation, and survival.^244,245 Its overactivation is implicated in various human tumors, making it a promising target for anticancer drug discovery.^246,247 Upon binding to epidermal growth factor (EGF), EGFR undergoes conformational changes that promote its dimerization and activation. Also, EGF-stimulated EGFR enters cells via endocytosis to evoke signaling pathways. The fluorescent labeling of EGFR on the cell surface involved the insertion of PrK using GCE technology, followed by click chemistry with a cell-impermeable, azide-appended fluorophore (AZDye 488) (Fig. 57a). The results revealed that EGF induces the rapid internalization of EGFR into cells, with an EGFR inhibitor (erlotinib) and antibody effectively blocking EGF-induced EGFR internalization. Additionally, glycans linked to the EGFR influenced EGF-promoted receptor endocytosis. Specifically, EGFR variants with desialylated or defucosylated glycans, achieved through treatment with respective sialidase or fucosidase, exhibited enhanced entry into cells compared to wild-type EGFR (Fig. 57b), These findings underscore the inhibitory role of sialic acid and fucose residues of EGFR-associated glycans in EGF-induced endocytosis.

Fig. 57 Fluorescent labeling of cell–surface EGFR for real-time monitoring of EGFR internalization induced by EGF. (a) Schematic for PrK incorporation into EGFR via GCE technology. (b) HeLa cells expressing the fluorescently labeled EGFR (green) on the cell surface were incubated with either fucosidase or sialidase followed by exposure to Lysotracker Deep Red (purple). Cells were then treated with EGF before cell imaging. A graph shows percentage of time-dependent colocalization of green with purple fluorescence (reproduced from ref. 243 with permission from the Royal Society of Chemistry, copyright 2024).

A dual-color pulse-chase labeling method based on peptide-templated acyl transfer reaction was utilized to track the internalization of the human neuropeptide Y2 receptor (hY2R, a class A GPCR), which is involved in central and peripheral diseases such as epilepsy (Fig. 58a).²⁴⁸ In this effort, the N-terminally Cys-E3-tagged hY2R on the cell surface was pulse-labeled with the TMR-K3 peptide via peptide-templated acyl transfer reaction (Fig. 58b). The first stimulation with the endogenous ligand porcine neuropeptide Y (pNPY) induced receptor internalization (Fig. 58c). By adjusting the agonist concentration and stimulation time, a subset of hY2R remained on the cell surface after this initial stimulation. The remaining unlabeled hY2R was then conjugated to the ATTO488 fluorophore using the ATTO488-K3 peptide for the chase label (Fig. 58c). Upon the second stimulation with pNPY, the chase-labeled hY2R underwent internalization into cells. In this case, the pulse- and chase-labeled hY2Rs in the early endosomal compartment were distinguishable shortly after the second stimulation (Fig. 58d). However, 10–12 min post-second stimulation, the pulse- and chase-labeled hY2Rs were observed to colocalize, indicating vesicular fusion within cells (Fig. 58e). This peptide-templated labeling strategy was also applied to monitor the constitutive internalization of ET_AR, ET_BR, the angiotensin II receptor type 1 (AT₁R) responsible for vasoconstriction, and the apelin receptor APJ serving as a vasodilator.¹³⁶ The results showed that while ET_AR, AT₁R, and APJ do not undergo internalization into the cytoplasm in the absence of an agonist, ET_BR quickly enters cells even without agonist-mediated activation.

Fig. 58 (a) Schematic for the dual-color pulse-chase study to track hY2R internalization. (b) Fluorescence images of unstimulated cells labeled with TMR-K3. (c) Fluorescence images of cells initially stimulated with pNPY for 30 min followed by chase-labeling with ATTO488-K3. (d) Fluorescence images of cells secondly stimulated with pNPY for 5 and (e) 28 min (reproduced from ref. 248 with permission from the American Chemical Society, copyright 2018).

An erasable fluorescence imaging method was employed to facilitate the clear monitoring of membrane protein internalization into cells.¹¹⁶ Initially, the N-terminally Cys-P1-tagged EGFR, in which the C-terminus was fused to YFP, was linked via peptide-templated acyl transfer to PNA using the P2–PNA conjugate (see Fig. 20a). The resulting PNA-conjugated receptor was then labeled through PNA–DNA hybridization with Complex III, composed of an adaptor DNA-105mer carrying five Atto565-labeled DNA-23mers (Fig. 59a). Activation of EGFR by EGF prompted the internalization of Atto565-labeled EGFR into cells. However, the Atto565 fluorescence from non-internalized EGFR was comparable in intensity to that from internalized EGFR, posing challenges for the analysis of internalized EGFR (Fig. 59b). To enhance the identification of EGFR internalized into transport vesicles, Atto565-23mers present in the labeled EGFR were selectively removed from non-internalized EGFR remaining on the cell surface. This removal was achieved by treating the cells with a displacement DNA-23mer in the presence of EGF. As a result, EGF-stimulated EGFR internalization could be readily assessed by erasing the fluorescent label from non-internalized EGFR (Fig. 59c).

Fig. 59 (a) Schematic for erasable fluorescence imaging of PNA-conjugated EGFR-YFP on the cell surface. (b) PNA-conjugated EGFR-YFP that was hybridized with Complex III was stimulated with EGF for 15 min. (c) Fluorophore-conjugated DNA in (b) was removed from non-internalized EGFR-YFP on the cell surface by treatment with displacement DNA in the presence of EGF (reproduced from ref. 116 with permission from the Springer Nature, copyright 2021).

To explore the ligand-induced internalization of endogenous GLP1R in complex tissues (Fig. 27a), a SNAP_f-tag was introduced between the signal peptide and the ectodomain of the GLP1R in mice using the CRISPR/Cas9 technology.¹⁶⁰ Islets isolated from GLP1R^SNAP/SNAP mice was then labeled with a cell-impermeable, sulfonated ligand SBG-TMR to confine labeling to the cell–surface receptor. After this, the labeled islets were treated with either a GLP1R agonist (exendin4, semaglutide and tirzepatide) or its antagonist (exendin4(9–39)). The results revealed that an antagonist exendin4(9–39) reduces GLP1R internalization, whereas the agonist exendin4 induces widespread GLP1R internalization (Fig. 60). Moreover, semaglutide and tirzepatide induced GLP1R internalization with lower efficiency compared to exendin4.

Fig. 60 Fluorescence images of TMR-labeled GLP1R in mouse islets after stimulation with indicated substances (reproduced from ref. 160 with permission from the Springer Nature, copyright 2023).

3.5 Conformational change of cell–surface receptors

Conformational changes of proteins, especially those embedded in cell plasma membranes, are pivotal in numerous signaling processes.²⁴⁹ In general, cell–surface proteins serve as molecular transducers crucial for communication and molecular transport between the cell and its external environment. Membrane proteins undergo structural rearrangements triggered by a multitude of stimuli, including the binding of extracellular and intracellular ligands, covalent modifications (e.g., phosphorylation), and alterations in membrane potential. The utilization of fluorescently labeled proteins has proven invaluable in tracking these conformational changes within live cells. In particular, the insertion of ANAP into the target protein within cells is a highly effective approach for this purpose, as its small size minimizes any impact on the protein's conformation and function.

To analyze ligand-induced conformational changes in membrane proteins, ANAP serving as a FRET donor was inserted into maltose-binding protein (MBP) at position either 295 or 322 through GCE (Fig. 61).²⁵⁰ Cu²⁺–TETAC acting as a FRET quencher was conjugated to the Cys residue of MBP at position 237 or 309 via a mixed disulfide linkage. In addition, the CAAX domain was fused to MBP to facilitate its expression on the cell plasma membrane. The ANAP fluorescence of these two FRET systems in cells was measured using a fluorometer to monitor maltose-induced conformational changes in MBP. The results indicated that upon addition of maltose, ANAP fluorescence arising from MBP-ANAP295 is more effectively quenched by labeled Cu²⁺-TETAC compared to MBP-ANAP322. Furthermore, these FRET systems were employed to determine the distances between FRET donors and acceptors based on the observed FRET efficiency.

Fig. 61 Labeling of MBP with ANAP (a FRET donor) and Cu²⁺-TETAC (a FRET acceptor) at positions of (upper) 322/309 or (lower) 295/237. The conformational change of MBP leads to changes in the FRET signal.

The Mg²⁺- and nucleotide-dependent conformational changes of the regulatory subunit sulfonylurea receptor-1 (SUR1), which interacts with the pore-forming subunit Kir6.2 in the ATP-sensitive K⁺ (K_ATP) channel, were explored in cells using ANAP as a FRET donor and the ATP analog, TNP-ATP, as a FRET acceptor (Fig. 62).²⁵¹ K_ATP channels are comprised of octameric complexes containing four SUR1 subunits and four Kir6.2 subunits.²⁵² Both Mg²⁺-ATP and Mg²⁺-ADP interact with the nucleotide binding domain 1 and 2 (NBD1 and NBD2) of the SUR1, but the binding of Mg²⁺-ADP to the NBD2 triggers the dimerization of NBD1 and NBD2 as well as promotes K_ATP channel opening.²⁵² ANAP was incorporated into NBS2 of SUR1 at position of either 1353 or 1397 via GCE, resulting in the generation of SUR1-ANAP1353 or SUR1-ANAP1397, respectively. FRET analysis of cells using a fluorometer revealed that while both SUR1-ANAP1353/Kir6.2 and SUR1-ANAP1397/Kir6.2 channels undergo conformational changes dependent on Mg²⁺ ions and nucleotides, only SUR1-ANAP1397/Kir6.2 channels are activated by Mg²⁺-nucleotides. Furthermore, both ADP and ATP bound to NBS2 in the absence of Mg²⁺, but only Mg²⁺-bound nucleotides promoted the dimerization of the NBDs and initiated the conformational change of SUR1 necessary for channel opening.

Fig. 62 (a) The topology of the K_ATP channel complex, featuring two Kir6.2 subunits and one SUR1 subunits for clarity. (b) Chemical structures of TNP-ATP and ANAP. (c) Structure of the ligand-bound NBDs of SUR1 (reproduced from ref. 251 with permission from eLife Sciences Publications, copyright 2019).

Owing to its property of red-shifting in polar environments and blue-shifting in hydrophobic environments, ANAP was utilized as an environmentally sensitive fluorophore to probe the conformational changes of membrane-bound ion channels. Specifically, the conformational dynamics of the human voltage-gated sodium channel, Na_v1.5, was investigated in live cells by incorporating ANAP into the Na_v1.5 inactivation gate at position 1475 using GCE technology (Fig. 63a).²⁵³ Na_v1.5 plays a critical role in cardiac excitability and conduction, with its inactivation being crucial for the proper propagation of the cardiac action potential.²⁵⁴ Genetic mutations in this channel can disrupt inactivation, leading to the development of cardiac arrhythmia disorders. Na_v1.5 consists of four homologous domains I–IV interconnected by intracellular linkers (Fig. 63a). The intracellular loop between DIII-S6 and DIV-S1 acts as an inactivation gate that folds into the channel pore during depolarization.²⁵⁵ Spectral imaging analysis revealed a blue-shift in the fluorescence of Nav1.5-ANAP1475 in cells, indicating that ANAP localizes within the hydrophobic environment of the channel. To evaluate the effect of K⁺ channel activity on the conformational change in the Na_v1.5 inactivation gate, full-length or C-terminally truncated (ΔCT) Na_v1.5-ANAP1475, along with the K⁺ channel Kir4.1, were expressed in cells (Fig. 63b). K⁺ depolarization induced a minor yet significant red shift in the ANAP fluorescence of full-length Na_v1.5-ANAP1475. However, K⁺ depolarization did not affect the ANAP fluorescence of ΔCT Na_v1.5-ANAP1475. These findings suggest that the shifts in ANAP fluorescence induced by the conformational changes of the Na_v1.5 inactivation gate are dependent on the external K⁺ concentration and the presence of the distal C-terminal region of Na_v1.5.

Fig. 63 (a) The topology of the human voltage-gated sodium channel α subunits (reproduced from ref. 255 with permission from the American Chemical Society, copyright 2015). (b) Schematic for K⁺ depolarization in cells expressing Nav1.5-ANAP1475 and GFP-Kir4.1 (reproduced from ref. 253 with permission from Elsevier, copyright 2019).

ANAP was also utilized to investigate the channel opening and movement of the S4 segment of the voltage-gated proton channel (H_V1) in response to voltage and the pH gradient.²⁵⁶ The H_V1 channel comprises an intracellular N-terminal region containing a phosphorylation site, four transmembrane helices (S1–S4), and the C-terminal α-helix (Fig. 64a). This channel forms a functional dimer through C-terminal coiled–coil interactions, with each monomer possessing its own proton channel activity (Fig. 64b).²⁵⁷ ANAP was inserted into the S4 segment of human H_V1 at position 197 through GCE. The patch-clamp fluorometry technique was employed to record both currents and fluorescence signals. Depolarization led to an increase in ANAP fluorescence alongside proton currents. The authors proposed that the depolarization-induced rise in ANAP fluorescence is not attributable to a change in environmental polarity but rather to the conformation-dependent quenching effect by Phe150 in the S2 segment of the H_v1 channel, as indicated by mutagenesis studies. Specifically, Phe150 is in close proximity to ANAP197 and thus quenches ANAP fluorescence in the inactive state of the H_v1 channel. However, during channel activation, ANAP197 moves away from Phe150 as the S4 segment moves outward, diminishing the quenching effect of the Phe150 and leading to an increase in ANAP fluorescence. Furthermore, it was suggested that the absolute pH value has an influence on the channel opening step.

Fig. 64 (a) The topology of the voltage-gated proton channel H_V1 incorporated with ANAP at position 197. (b) The dimeric structure of H_v1 formed by the C-terminal coiled–coil interaction. Phosphorylation of the N-terminal intracellular region converts the resting state to the channel opening state.

3.6 Membrane potential

The membrane potential arises from the disparity in electric potential between the interior and exterior of a cell. In electrically excitable cells such as neurons and muscle cells, the opening or closing of ion channels embedded in the cell plasma membrane results in localized changes in membrane potential, generating propagating signals. Particularly, significant depolarization in excitable cells triggers the generation of action potentials via the activation of voltage-gated ion channels. During this process, the membrane potential undergoes rapid and substantial changes for a short time.

In neurons, rapid alterations in membrane potential prompt a transient elevation in intracellular Ca²⁺ levels, leading to the release of neurotransmitters from synaptic vesicles in presynaptic neurons. To directly observe changes in membrane potential in living systems, the transmembrane domain (TMD) of the platelet-derived growth factor receptor (PDGFR) fused to the HaloTag was labeled with RhoVR1-PEG₂₅-Halo, a cell-impermeable, photoinduced electron transfer-triggered voltage-sensitive ligand, creating a chemigenetic voltage indicator (Fig. 65a).²⁵⁸ The RhoVR1 dye attached to the TMD-HaloTag on the surface of transfected cells displayed high voltage sensitivity and rapid response times. Also, action potentials were detected in rat hippocampal neurons expressing the TMD-HaloTag labeled with the RhoVR1 dye (Fig. 65b and c). Furthermore, both the RhoVR1-Halo dye and the genetically encoded calcium sensor GCaMP6s were employed to simultaneously monitor membrane potential and Ca²⁺ levels in rat hippocampal neurons. Finally, action potentials were recorded in single imaging trials from cortical neurons in a rat brain slice expressing the RhoVR1-labeled TMD-HaloTag.

Fig. 65 (a) Labeling of the Halo-tagged TMD of PDGFR with voltage-sensitive RhoVR1-PEG₂₅-Halo to detect membrane potential in neurons. Included is the chemical structure of RhoVR1-PEG₂₅-Halo. (b) Confocal fluorescence microscopy images of cultured rat hippocampal neurons expressing the TMD-HaloTag labeled with RhoVR1-PEG₂₅-Halo. (c) Plots of the fractional change in fluorescence (ΔF/F) over time in hippocampal neurons labeled with RhoVR1-PEG₂₅-Halo (reproduced from ref. 258 with permission from the American Chemical Society, copyright 2020).

The HaloTag was also utilized to develop chemigenetic indicators for calcium ions and membrane potential.²⁵⁹ Previous studies showed that whereas rhodamine dyes mainly exist in a fluorescent zwitterionic form in aqueous solutions, far-red Si–rhodamines (e.g., JF635–HaloTag ligand) are exclusively present in a non-fluorescent lactone form (Fig. 66a).²⁶⁰ Initially, the HaloTag-based Ca²⁺ indicator (HaloCaMP₆₃₅) was constructed by fusing the Ca²⁺-sensor protein calmodulin (CaM) and the CaM-binding peptide to the C- and N-termini of a circularly permuted HaloTag (cpHaloTag), respectively, followed by labeling with the JF635-HaloTag ligand (Fig. 66b). The resultant sensor, HaloCaMP₆₅₃, exhibited substantial fluorescence increases (4–9 fold) in the presence of Ca²⁺, indicating a transition of the JF₆₃₅ dye from a non-fluorescent to fluorescent form. Fluorescence responses were then determined in rat hippocampal neurons expressing HaloCaMP₆₅₃ following electrical stimulation to induce action potentials, indicating a significant increase in fluorescence upon field electrode stimulation. Next, the voltage-sensitive indicator HASAP₆₃₅ was developed by inserting a cpHaloTag into the loop between the third and fourth transmembrane α-helices of a voltage-sensitive domain (VSD) from Gallus gallus, followed by labeling with the JF₆₃₅-HaloTag ligand (Fig. 66c). Additionally, the voltage indicator HArcLight₆₃₅ with a different topology was constructed by attaching the HaloTag to the C-terminus of a VSD from Ciona intestinalis, followed by labeling with the same HaloTag ligand (Fig. 66d). Because the change in membrane potential led to a change in the fluorescence of HASAP₆₃₅ and HArcLight₆₃₅ indicators, alterations in membrane potential in single imaging trials from rat hippocampal neurons were assessed. The results revealed that while fluorescence from HASAP₆₃₅ increases, that from HArcLight₆₃₅ decreases, in response to a voltage step from −70 mV to +30 mV.

Fig. 66 (a) Equilibrium between non-fluorescent and fluorescent states of rhodamine and Si–rhodamine dyes. (b) Detection of Ca²⁺ ions using the chemigenetic calcium indicator HaloCaMP₆₃₅ that consists of the CaM-binding peptide, cpHaloTag, and CaM. Detection of changes in membrane potential using (c) the chemigenetic voltage-sensitive indicator HASAP₆₃₅ composed of VSD and cpHaloTag and (d) the chemigenetic voltage-sensitive indicator HArcLight₆₃₅ comprising VSD and HaloTag.

3.7 Protein degradation and proteolytic processing

Protein homeostasis is essential for sustaining normal cellular metabolism through processes such as protein biosynthesis, trafficking and degradation.²⁶¹ As cells age, the accumulation of damaged, misfolded or aggregated proteins can contribute to various age-related diseases, including neurodegenerative disorders and cancer.²⁶² To maintain protein homeostasis, cells have evolved intricate quality control mechanisms such as the ubiquitin-proteasome system and autophagy-lysosomal system. Thus, protein turnover, which encompasses the continuous cycle of protein synthesis and degradation, is a key determinant for cellular homeostasis.

Fluorescence-labeled proteins are valuable to monitor protein degradation within cells. For instance, BCNK was incorporated into members (IL-12α and IL-23α) of the human interleukin 12 (IL-12) family, crucial signaling molecules in the immune system, and then labeled with TMR-tetrazine or SiR-tetrazine.⁸⁴ Time-dependent degradation studies conducted through in-gel fluorescence measurements revealed that the half-life of IL-12α degradation is approximately 2 h, consistent with previous findings. Although cell imaging study was not conducted, this method holds promise for the detection of protein degradation in live cells.

To visualize auxin-based protein degradation, SmFP485, prepared according to the strategy depicted in Fig. 28 was employed.¹⁶¹ It is known that the interaction of the complex of auxin and the auxin receptor TIR1 with the auxin-inducible degron (AID) triggers AID ubiquitination and subsequent proteosomal degradation.²⁶³ The SNAP_f-AID mCherry-H2B histone fusion protein expressed in the nucleus of cells was labeled with BG-F485 (Fig. 67a). Treatment with auxin led to a decrease in fluorescence from both SmFP485 and mCherry in the nucleus, with a half-life of approximately 40 min (Fig. 67b).

Fig. 67 Real-time detection of auxin-based protein degradation using SmFP485. (a) Schematic for the SNAP_f-mCherry fusion protein degradation under the control of the AID system. (b) Real-time fluorescence images of the SNAP_f-AID-mCherry-histone fusion protein in cells treated with auxin (reproduced from ref. 161 with permission from the Springer Nature, copyright 2023).

The PYP-tag, in conjunction with the OFF–ON–OFF labeling probe F5-DNB2 (Fig. 68a), was utilized to fluorescently detect protein degradation in live cells.²⁶⁴ F5-DNB2 consisted of a 7-hydroxycoumarin ligand, a fluorescein fluorophore and a dinitrobenzene (DNB) quencher. To facilitate faster reaction with the anionic ligand F5-DNB2 and to enhance expression levels in cells, three negatively charged residues (Asp71, Glu74 and Asp97) located on the surface of PYP-tag were mutated to neutral amino acids, Asn, Gln and Asn, respectively, resulting in the construction of the PYP^NQN-tag. F5-DNB2 itself exhibited weak fluorescein fluorescence due to the quenching effect of the DNB moiety (Fig. 68a and b). However, upon covalent binding of F5-DNB2 to the hydrophobic pocket of the PYP^NQN-tag, there was a noticeable increase in fluorescein fluorescence (an ‘OFF–ON’ switch). The thioester linkage formed through the transthioesterification reaction of the thiol group of the Cys69 residue of the PYP-tag with F5-DNB2 was resistant to glutathione. As anticipated, degradation of the PYP-tag labeled with F5-DNB2 by trypsin led to a decrease in fluorescein fluorescence due to the intramolecular quenching effect of the coumarin moiety (an ‘ON–OFF’ switch). Because F5-DNB2 had limited cell permeability, it was converted to its cell-permeable form, Ac₂F5-DNB2, for cellular studies, wherein cytosolic esterases hydrolyzed acetyl groups to generate F5-DNB2. Through the utilization of the PYP^NQN/Ac₂F5-DNB2 system, both the expression (indicated by fluorescence ‘ON’) and proteolytic degradation (indicated by fluorescence ‘OFF’) of the short-lived protein, mouse ornithine decarboxylase (MODC), were visualized in the nucleus of cells in real time (Fig. 68c).

Fig. 68 (a) Structure and mechanism of F5-DNB2 for the fluorescent detection of protein degradation. (b) Fluorescence spectrum of F5-DNB2 in the absence (red) and presence of the PYP-tag (blue). (c) Real-time fluorescence images of cells expressing HA-PYP^NQN-NLS-MODC^422–461 after labeling with Ac₂F5-DNB2 (reproduced from ref. 264 with permission from the Royal Society of Chemistry, copyright 2022).

Certain cellular proteins undergo proteolytic processing to generate functionally active protein isoforms (proteoforms) that serve as hormones, neurotransmitters, and disease pathogens.²⁶⁵ Amyloid-β (Aβ) peptides with 37–49 amino acid residues are typical proteoforms involved in the pathogenesis of Alzheimer's disease. These peptides are produced through the proteolytic cleavage of amyloid precursor protein (APP) by the sequential action of membrane-bound β- and γ-secretases that catalyze cleavage of the N- and C-termini of the Aβ domain, respectively (Fig. 69).²⁶⁶ Also, the C-terminal fragment of APP generated by β-secretase can be internalized into cells and processed by γ-secretase to produce Aβ peptides within endocytic compartments. A combination of genetic code expansion and bioorthogonal ligation techniques were employed to selectively visualize Aβ peptides generated from APP in live cells.⁸⁶ BCNK was incorporated into the Aβ domain of the Halo-tagged APP via GCE, followed by labeling with Cy5-tetrazine as a FRET donor and the HaloTag ligand QSY21-Cl as a FRET quencher (Fig. 69). In this dually labeled protein, Cy5 fluorescence would be quenched by QSY21 conjugated to the HaloTag. Furthermore, Cy5 fluorescence of Halo-tagged APP cleaved only by β-secretase would also be quenched by QSY21. However, Aβ fully liberated from APP by both β- and γ-secretases would exhibit Cy5 fluorescence. Cell studies revealed that Cy5 fluorescence, which rarely overlaps with that of anti-myc stained APP, is observed in cells expressing the dually labeled APP-HaloTag, indicating the generation of Aβ peptides cleaved by both secretases.

Fig. 69 Schematic for FRET-based imaging of Aβ generated from APP by actions of β- and γ-secretases.

3.8 Detection of protein-specific glycans

Receptors on the mammalian cell surface are generated in a form of glycoproteins heavily modified with diverse glycans.²⁶⁷ These glycans greatly affect their physicochemical properties of receptors, such as solubility, stability and folding, as well as cellular functions, including ligand binding, oligomerization and signaling pathways.^268–272 Owing to the significant impact of glycans on receptor biology, methods for the detection of cell surface-receptor specific glycans are in great demand. However, imaging of fluorescence-labeled receptors was ineffective for monitoring glycans conjugated to receptors.

To address this challenge in detecting cell–surface receptor-specific glycans, we employed two complementary techniques: GCE technology for fluorescence labeling of proteins and metabolic glycan incorporation for fluorescence labeling of glycans.²⁴³ Specifically, PrK and an azidosugar were introduced into EGFR and cellular glycans, respectively, through GCE and metabolic glycan incorporation methods (Fig. 70a). The azide-containing cellular glycans and the alkyne group of PrK in EGFR were then sequentially labeled with rhodamine-ADIBO (Rh-ADIBO) as a FRET acceptor and AZDye 488-N₃ as a FRET donor using SPAAC and CuAAC, respectively. Detection of cell–surface EGFR-specific glycans was achieved by observing FRET from AZDye 488 to rhodamine upon excitation of AZDye 488. FRET signals were detected exclusively on the surface of cells transfected with the EGFR-N128TAG gene in the presence of both PrK and Ac₄ManNAz, but not in the presence of PrK or Ac₄ManNAz alone (Fig. 70b). Furthermore, treatment with sialidase resulted in the abolition of FRET-induced fluorescence, indicating the critical role of sialic acid residues on EGFR in the observed FRET phenomenon.

Fig. 70 (a) Schematic for FRET-based detection of cell–surface EGFR-specific glycosylation. (b) Fluorescence images of cells expressing PrK-incorporated EGFR in the presence Ac₄ManNAz that were sequentially treated with Rh-ADIBO and AZDye 488-N₃ (reproduced from ref. 243 with permission from the Royal Society of Chemistry, copyright 2024).

4. Conclusions and perspective

Fluorescence imaging of specific proteins in cells is a rapidly growing field that has greatly advanced our understanding of biological processes involving proteins. Numerous strategies have been devised to label cellular proteins, encompassing methods such as incorporating fluorescent amino acids into proteins, utilizing functionalized non-fluorescent amino acids followed by bioorthogonal ligation with fluorophores, peptide tag-based fluorescent labeling, SLE tag-based fluorescent labeling, and affinity-based fluorescent labeling. As discussed in Section 2.5, each strategy for protein labeling presents unique strengths and weaknesses based on factors such as selectivity, tag size, labeling efficiency, kinetics, stability of the linkage between the protein and fluorophore, and versatility. These considerations are crucial for selecting the most suitable method depending on the specific requirements of the experiment or study.

The site-specific insertion of fluorescent amino acids into proteins is highly effective for fluorescent labeling, but its application is constrained by the availability of suitable fluorescent amino acids. Discovering more orthogonal tRNA/tRNA synthetase pairs is crucial to enable the incorporation of a wide range of fluorescent amino acids, particularly those with long-wavelength emission, into cellular proteins. Conversely, non-fluorescent amino acids require additional bioorthogonal chemistry for protein labeling with fluorophores, but they offer greater versatility compared to their fluorescent counterparts.²⁰¹ Labeling cell–surface proteins with functionalized fluorophores via bioorthogonal chemistry is straightforward. However, labeling intracellular proteins presents greater challenges owing to slower kinetics and lower specificity caused by the complex intracellular environment and nonselective nucleophilic addition of biothiols to the bioorthogonal handles.

In this context, bioorthogonal chemistry that allows rapid protein labeling with low concentrations of fluorophores would be invaluable for efficient labeling of proteins inside live cells and tissues.⁴⁶ Moreover, residual free fluorophores after bioorthogonal reaction should be removed through extensive washing to improve signal-to-background ratios, especially for detecting low-abundance intracellular targets. Similarly, fluorescent ligands used in enzyme-catalyzed labeling should be washed out before cell imaging. Fluorogenic probes or ligands that become fluorescent only upon protein labeling offer significant advantages by eliminating the need for a washing step.⁴⁷ This advancement enhances detection sensitivity by suppressing nonspecific interactions or side reactions that can reduce signal-to-noise ratios, thereby allowing the live-cell imaging of endogenous proteins with low abundance.

Another challenge lies in multi-color labeling of proteins to visualize multiple proteins simultaneously in cells. This has been achieved through combinations of distinct labeling methods or by integrating multiple compatible bioorthogonal reactions. Nevertheless, the optimization of bioorthogonal handles and reagents for existing reactions and the development of new bioorthogonal chemistries are essential steps.

While chemists have primarily focused on developing selective protein labeling strategies with fluorophores, there has been relatively less emphasis on applying these methods to elucidate detailed protein-associated biological events. Overcoming these challenges requires robust interdisciplinary collaborations between chemists and biologists, enabling the exploration of previously inaccessible questions at the forefront of biology. Advancements in protein imaging within cells are expected to deepen our understanding of protein functions and dynamics. We anticipate that this review will contribute to the design and development of more effective and selective protein labeling methods, ultimately advancing our comprehension of protein roles in cellular processes.

Data availability

No primary research results have been included and no new data were generated or analysed as part of this review. We have added data citations as bibliographic references within the main text as they are mentioned.

Conflicts of interest

There are no conflicts to declare.

Acknowledgements

This study was supported financially by the National Research Foundation of Korea (grant no. 2020R1A2C3003462 to I. S., 2022M3E5F2017857 to H. S. L., and RS-2023-00211730 to J. Y. H) and the KRICT (grant no. KK2431-20 to J. Y. H).

References

1. L. Wang, W. Du, Z. Hu, K. Uvdal, L. Li and W. Huang, Angew. Chem., Int. Ed., 2019, 58, 14026–14043.
2. M. Gao, F. Yu, C. Lv, J. Choo and L. Chen, Chem. Soc. Rev., 2017, 46, 2237–2271.
3. Y. Kim, H. Li, J. Choi, J. Boo, H. Jo, J. Y. Hyun and I. Shin, Chem. Soc. Rev., 2023, 52, 7036–7070.
4. X. Chen, F. Wang, J. Y. Hyun, T. Wei, J. Qiang, X. Ren, I. Shin and J. Yoon, Chem. Soc. Rev., 2016, 45, 2976–3016.
5. X. Chen, X. Tian, I. Shin and J. Yoon, Chem. Soc. Rev., 2011, 40, 4783–4804.
6. S. K. Ko, X. Chen, J. Yoon and I. Shin, Chem. Soc. Rev., 2011, 40, 2120–2130.
7. A. L. Vahrmeijer, M. Hutteman, J. R. van der Vorst, C. J. van de Velde and J. V. Frangioni, Nat. Rev. Clin. Oncol., 2013, 10, 507–518.
8. S. H. Park, N. Kwon, J. H. Lee, J. Yoon and I. Shin, Chem. Soc. Rev., 2020, 49, 143–179.
9. H. Li, Y. Kim, H. Jung, J. Y. Hyun and I. Shin, Chem. Soc. Rev., 2022, 51, 8957–9008.
10. J. Liu and Z. Cui, Bioconjugate Chem., 2020, 31, 1587–1595.
11. E. A. Rodriguez, R. E. Campbell, J. Y. Lin, M. Z. Lin, A. Miyawaki, A. E. Palmer, X. Shu, J. Zhang and R. Y. Tsien, Trends Biochem. Sci., 2017, 42, 111–129.
12. M. Wang, Y. Da and Y. Tian, Chem. Soc. Rev., 2023, 52, 1189–1214.
13. D. M. Chudakov, M. V. Matz, S. Lukyanov and K. A. Lukyanov, Physiol. Rev., 2010, 90, 1103–1163.
14. W. Margolin, J. Bacteriol., 2012, 194, 6369–6371.
15. D. W. Piston and G. J. Kremers, Trends Biochem. Sci., 2007, 32, 407–414.
16. L. Wang, A. Brock, B. Herberich and P. G. Schultz, Science, 2001, 292, 498–500.
17. H. B. Yi, S. Lee, K. Seo, H. Kim, M. Kim and H. S. Lee, Chem. Rev., 2024, 124, 7465–7530.
18. C. C. Liu and P. G. Schultz, Annu. Rev. Biochem., 2010, 79, 413–444.
19. L. Wang, J. Xie and P. G. Schultz, Annu. Rev. Biophys., 2006, 35, 225–249.
20. L. Wang and P. G. Schultz, Angew. Chem., Int. Ed., 2005, 44, 34–66.
21. J. Wang, J. Xie and P. G. Schultz, J. Am. Chem. Soc., 2006, 128, 8738–8739.
22. D. Summerer, S. Chen, N. Wu, A. Deiters, J. W. Chin and P. G. Schultz, Proc. Natl. Acad. Sci. U. S. A., 2006, 103, 9785–9789.
23. P. E. Zinsli, J. Photochem., 1974, 3, 55–69.
24. M. Amaro, J. Brezovský, S. Kováčová, J. Sýkora, D. Bednář, V. Němec, V. Lišková, N. P. Kurumbang, K. Beerens, R. Chaloupková, K. Paruch, M. Hof and J. Damborský, J. Am. Chem. Soc., 2015, 137, 4988–4992.
25. C. M. Jones, D. M. Robkis, R. J. Blizzard, M. Munari, Y. Venkatesh, T. S. Mihaila, A. J. Eddins, R. A. Mehl, W. N. Zagotta, S. E. Gordon and E. J. Petersson, Chem. Sci., 2021, 12, 11955–11964.
26. T. Takeuchi, J. Chromatogr. Libr., 2005, 70, 229–241.
27. G. Charbon, E. Brustad, K. A. Scott, J. Wang, A. Løbner-Olesen, P. G. Schultz, C. Jacobs-Wagner and E. Chapman, ChemBioChem, 2011, 12, 1818–1821.
28. R. F. Chen, Arch. Biochem. Biophys., 1967, 120, 609–620.
29. B. Cürten, P. H. M. Kullmann, M. E. Bier, K. Kandler and B. F. Schmidt, Photochem. Photobiol., 2005, 81, 481–710.
30. G. Jones II, W. R. Jackson and A. M. Halpern, Chem. Phys. Lett., 1980, 72, 391–395.
31. A. Chatterjee, J. Guo, H. S. Lee and P. G. Schultz, J. Am. Chem. Soc., 2013, 135, 12540–12543.
32. J. M. Goldberg, L. C. Speight, M. W. Fegley and E. J. Petersson, J. Am. Chem. Soc., 2012, 134, 6088–6091.
33. A. Szymańska, K. Wegner and L. Łankiewicz, Helv. Chim. Acta, 2003, 86, 3326–3331.
34. J. Luo, R. Uprety, Y. Naro, C. Chou, D. P. Nguyen, J. W. Chin and A. Deiters, J. Am. Chem. Soc., 2014, 136, 15551–15558.
35. W. Brown, J. Wesalo, S. Samanta, J. Luo, S. E. Caldwell, M. Tsang and A. Deiters, ACS Chem. Biol., 2023, 18, 1305–1314.
36. L. C. Speight, A. K. Muthusamy, J. M. Goldberg, J. B. Warner, R. F. Wissner, T. S. Willi, B. F. Woodman, R. A. Mehl and E. J. Petersson, J. Am. Chem. Soc., 2013, 135, 18806–18814.
37. K. T. Grasso, M. J. R. Yeo, C. M. Hillenbrand, E. D. Ficaretta, J. S. Italia, R. L. Huang and A. Chatterjee, Biochemistry, 2021, 60, 489–493.
38. H. S. Lee, J. Guo, E. A. Lemke, R. D. Dimla and P. G. Schultz, J. Am. Chem. Soc., 2009, 131, 12921–12923.
39. G. Weber and F. J. Farris, Biochemistry, 1979, 18, 3075–3078.
40. N. Wu, A. Deiters, T. A. Cropp, D. King and P. G. Schultz, J. Am. Chem. Soc., 2004, 126, 14306–14307.
41. S. H. Park, W. Ko, H. S. Lee and I. Shin, J. Am. Chem. Soc., 2019, 141, 4273–4281.
42. S. H. Park, W. Ko, S. H. Park, H. S. Lee and I. Shin, ChemBioChem, 2020, 21, 59–63.
43. S. H. Park, S. Kim, H. S. Lee and I. Shin, ACS Chem. Biol., 2021, 16, 2462–2471.
44. C. P. Toseland, J. Chem. Biol., 2013, 6, 85–95.
45. S. Douthwright and G. Sluder, J. Cell. Physiol., 2017, 232, 2461–2468.
46. V. Rigolot, C. Biot and C. Lion, Angew. Chem., Int. Ed., 2021, 60, 23084–23105.
47. P. Shieh and C. R. Bertozzi, Org. Biomol. Chem., 2014, 12, 9307–9320.
48. V. V. Rostovtsev, L. G. Green, V. V. Fokin and K. B. Sharpless, Angew. Chem., Int. Ed., 2002, 41, 2596–2599.
49. C. W. Tornøe, C. Christensen and M. Meldal, J. Org. Chem., 2002, 67, 3057–3064.
50. K. Lang and J. W. Chin, Chem. Rev., 2014, 114, 4764–4806.
51. D. P. Nguyen, H. Lusic, H. Neumann, P. B. Kapadnis, A. Deiters and J. W. Chin, J. Am. Chem. Soc., 2009, 131, 8720–8721.
52. I. C. Vreja, I. Nikić, F. Göttfert, M. Bates, K. Kröhnert, T. F. Outeiro, S. W. Hell, E. A. Lemke and S. O. Rizzoli, ACS Nano, 2015, 9, 11034–11041.
53. T. S. Mihaila, C. Bate, L. M. Ostersehlt, J. K. Pape, J. Keller-Findeisen, S. J. Sahl and S. W. Hell, Proc. Natl. Acad. Sci. U. S. A., 2022, 119, e2201861119.
54. Z. Hao, Y. Song, S. Lin, M. Yang, Y. Liang, J. Wang and P. R. Chen, Chem. Commun., 2011, 47, 4502–4504.
55. M. Yang, A. S. Jalloh, W. Wei, J. Zhao, P. Wu and P. R. Chen, Nat. Commun., 2014, 5, 4981.
56. M. Yang, Y. Song, M. Zhang, S. Lin, Z. Hao, Y. Liang, D. Zhang and P. R. Chen, Angew. Chem., Int. Ed., 2012, 51, 7674–7679.
57. B. Meineke, J. Heimgärtner, A. J. Craig, M. Landreh, L. W. K. Moodie and S. J. Elsässer, Front. Chem., 2021, 9, 768535.
58. A. Kumar, K. Li and C. Cai, Chem. Commun., 2011, 47, 3186–3188.
59. S. Li, L. Wang, F. Yu, Z. Zhu, D. Shobaki, H. Chen, M. Wang, J. Wang, G. Qin, U. J. Erasquin, L. Ren, Y. Wang and C. Cai, Chem. Sci., 2017, 8, 2107–2114.
60. G. Wittig and A. Krebs, Chem. Ber., 1961, 94, 3260–3275.
61. N. J. Agard, J. A. Prescher and C. R. Bertozzi, J. Am. Chem. Soc., 2004, 126, 15046–15047.
62. N. J. Agard, J. M. Baskin, J. A. Prescher, A. Lo and C. R. Bertozzi, ACS Chem. Biol., 2006, 1, 644–648.
63. J. M. Baskin, J. A. Prescher, S. T. Laughlin, N. J. Agard, P. V. Chang, I. A. Miller, A. Lo, J. A. Codelli and C. R. Bertozzi, Proc. Natl. Acad. Sci. U. S. A., 2007, 104, 16793–16797.
64. J. Dommerholt, F. P. J. T. Rutjes and F. L. van Delft, Top. Curr. Chem., 2016, 374, 1–20.
65. J. Dommerholt, S. Schmidt, R. Temming, L. J. A. Hendriks, F. P. J. T. Rutjes, J. C. M. van Hest, D. J. Lefeber, P. Friedl and F. L. van Delft, Angew. Chem., Int. Ed., 2010, 49, 9422–9425.
66. M. F. Debets, S. S. van Berkel, S. Schoffelen, F. P. J. T. Rutjes, J. C. M. van Hest and F. L. van Delft, Chem. Commun., 2010, 46, 97–99.
67. A. Kuzmin, A. Poloukhtine, M. A. Wolfert and V. V. Popik, Bioconjugate Chem., 2010, 21, 2076–2085.
68. M. Cheng, W. Zhang, J. Yuan, W. Luo, N. Li, S. Lin, Y. Yang, X. Fang and P. R. Chen, Chem. Commun., 2014, 50, 14724–14727.
69. S. H. Alamudi, R. Satapathy, J. Kim, D. Su, H. Ren, R. Das, L. Hu, E. Alvarado-Martínez, J. Y. Lee, C. Hoppmann, E. Peña-Cabrera, H. H. Ha, H. S. Park, L. Wang and Y. T. Chang, Nat. Commun., 2016, 7, 11964.
70. M. K. Singh, P. Zangoui, Y. Yamanaka and L. J. Kenney, eLife, 2021, 10, e67789.
71. Y. Akeda and J. Galán, Nature, 2005, 437, 911–915.
72. T. Plass, S. Milles, C. Koehler, C. Schultz and E. A. Lemke, Angew. Chem., Int. Ed., 2011, 50, 3878–3881.
73. A. Borrmann, S. Milles, T. Plass, J. Dommerholt, J. M. M. Verkade, M. Wiessler, C. Schultz, J. C. M. van Hest, F. L. van Delft and E. A. Lemke, ChemBioChem, 2012, 13, 2094–2099.
74. K. Lang, L. Davis, S. Wallace, M. Mahesh, D. J. Cox, M. L. Blackman, J. M. Fox and J. W. Chin, J. Am. Chem. Soc., 2012, 134, 10317–10320.
75. C. J. MacNevin, T. Watanabe, M. Weitzman, A. Gulyani, S. Fuehrer, N. K. Pinkin, X. Tian, F. Liu, J. Jin and K. M. Hahn, J. Am. Chem. Soc., 2019, 141, 7275–7282.
76. M. L. Blackman, M. Royzen and J. M. Fox, J. Am. Chem. Soc., 2008, 130, 13518–13519.
77. N. K. Devaraj, R. Weissleder and S. A. Hilderbrand, Bioconjugate Chem., 2008, 19, 2297–2299.
78. B. L. Oliveira, Z. Guo and G. J. L. Bernardes, Chem. Soc. Rev., 2017, 46, 4895–4950.
79. K. Lang, L. Davis, J. Torres-Kolbus, C. Chou, A. Deiters and J. W. Chin, Nat. Chem., 2012, 4, 298–304.
80. T. Plass, S. Milles, C. Koehler, J. Szymański, R. Mueller, M. Wiessler, C. Schultz and E. A. Lemke, Angew. Chem., Int. Ed., 2012, 51, 4166–4170.
81. I. Nikić, T. Plass, O. Schraidt, J. Szymański, J. A. G. Briggs, C. Schultz and E. A. Lemke, Angew. Chem., Int. Ed., 2014, 53, 2245–2249.
82. G. Lukinavičius, K. Umezawa, N. Olivier, A. Honigmann, G. Yang, T. Plass, V. Mueller, L. Reymond, I. R. Corrêa, Z.-G. Luo, C. Schultz, E. A. Lemke, P. Heppenstall, C. Eggeling, S. Manley and K. Johnsson, Nat. Chem., 2013, 5, 132–139.
83. C. Uttamapinant, J. D. Howe, K. Lang, V. Beránek, L. Davis, M. Mahesh, N. P. Barry and J. W. Chin, J. Am. Chem. Soc., 2015, 137, 4602–4605.
84. Y. G. Mideksa, M. Fottner, S. Braus, C. A. M. Weiss, T. A. Nguyen, S. Meier, K. Lang and M. J. Feige, ChemBioChem, 2020, 21, 1861–1867.
85. D. Bessa-Neto, G. Beliu, A. Kuhlemann, V. Pecoraro, S. Doose, N. Retailleau, N. Chevrier, D. Perrais, M. Sauer and D. Choquet, Nat. Commun., 2021, 12, 6715.
86. K. Sappakhaw, K. Jantarug, S. A. Slavoff, N. Israsena and C. Uttamapinant, Angew. Chem., Int. Ed., 2021, 60, 3934–3939.
87. W. Chi, L. Huang, C. Wang, D. Tan, Z. Xu and X. Liu, Mater. Chem. Front., 2021, 5, 7012–7021.
88. G. Beliu, A. J. Kurz, A. C. Kuhlemann, L. Behringer-Pliess, M. Meub, N. Wolf, J. Seibel, Z. Shi, M. Schnermann, J. B. Grimm, L. D. Lavis, S. Doose and M. Sauer, Commun. Biol., 2019, 2, 261.
89. P. Werther, K. Yserentant, F. Braun, K. Grußmayer, V. Navikas, M. Yu, Z. Zhang, M. J. Ziegler, C. Mayer, A. J. Gralak, M. Busch, W. Chi, F. Rominger, A. Radenovic, X. Liu, E. A. Lemke, T. Buckup, D. Herten and R. Wombacher, ACS Cent. Sci., 2021, 7, 1561–1571.
90. I. Segal, D. Nachmias, A. Konig, A. Alon, E. Arbely and N. Elia, BMC Biol., 2020, 18, 5.
91. S. Liu, P. Hoess and J. Ries, Ann. Rev. Biophys., 2022, 51, 301–326.
92. C. G. Galbraith and J. A. Galbraith, J. Cell Sci., 2011, 124, 1607–1611.
93. S. Wang, J. R. Moffitt, G. T. Dempsey, X. S. Xie and X. Zhuang, Proc. Natl. Acad. Sci. U. S. A., 2014, 111, 8452–8457.
94. F. Neubert, G. Beliu, U. Terpitz, C. Werner, C. Geis, M. Sauer and S. Doose, Angew. Chem., Int. Ed., 2018, 57, 16364–16369.
95. A. Arsic, C. Hagemann, N. Stajkovic, T. Schubert and I. Nikic-Spiegel, Nat. Commun., 2022, 13, 314.
96. K. Saal, F. Richter, P. Rehling and S. O. Rizzoli, ACS Nano, 2018, 12, 12247–12254.
97. J. C. W. Willis and J. W. Chin, Nat. Chem., 2018, 10, 831–837.
98. B. Meineke, J. Heimgärtner, J. Eirich, M. Landreh and S. J. Elsässer, Cell Rep., 2020, 31, 107811.
99. J. Lotze, U. Reinhardt, O. Seitz and A. G. Beck-Sickinger, Mol. BioSyst., 2016, 12, 1731–1745.
100. Y. Zhang, K.-Y. Park, K. F. Suazo and M. D. Distefano, Chem. Soc. Rev., 2018, 47, 9106–9136.
101. B. A. Griffin, S. R. Adams and R. Y. Tsien, Science, 1998, 281, 269–272.
102. S. R. Adams, R. E. Campbell, L. A. Gross, B. R. Martin, G. K. Walkup, Y. Yao, J. Llopis and R. Y. Tsien, J. Am. Chem. Soc., 2002, 124, 6063–6076.
103. H. Cao, B. Chen, T. C. Squier and M. U. Mayer, Chem. Commun., 2006, 2601–2603.
104. H. Cao, Y. Xiong, T. Wang, B. Chen, T. C. Squier and M. U. Mayer, J. Am. Chem. Soc., 2007, 129, 8672–8673.
105. M. N. Rozinov and G. P. Nolan, Chem. Biol., 1998, 5, 713–728.
106. M. Sunbul, L. Nacheva and A. Jäschke, Bioconjugate Chem., 2015, 26, 1466–1469.
107. Y. Chen, C. M. Clouthier, K. Tsao, M. Strmiskova, H. Lachance and J. W. Keillor, Angew. Chem., Int. Ed., 2014, 53, 13785–13788.
108. E. Hochuli, W. Bannwarth, H. Döbeli, R. Gentz and D. Stüber, Nat. Biotechnol., 1988, 6, 1321–1325.
109. A. Tinazli, J. Tang, R. Valiokas, S. Picuric, S. Lata, J. Piehler, B. Liedberg and R. Tampë, Chem. – Eur. J., 2005, 11, 5249–5259.
110. H. Nonaka, S. Tsukiji, A. Ojida and I. Hamachi, J. Am. Chem. Soc., 2007, 129, 15777–15779.
111. A. Ojida, K. Honda, D. Shinmi, S. Kiyonaka, Y. Mori and I. Hamachi, J. Am. Chem. Soc., 2006, 128, 10452–10459.
112. J. R. Litowski and R. S. Hodges, J. Biol. Chem., 2002, 277, 37272–37279.
113. Y. Yano, A. Yano, S. Oishi, Y. Sugimoto, G. Tsujimoto, N. Fujii and K. Matsuzaki, ACS Chem. Biol., 2008, 3, 341–345.
114. J. Wang, Y. Yu and J. Xia, Bioconjugate Chem., 2014, 25, 178–187.
115. U. Reinhardt, J. Lotze, S. Zernia, K. Mörl, A. G. Beck-Sickinger and O. Seitz, Angew. Chem., Int. Ed., 2014, 53, 10237–10241.
116. G. C. Gavins, K. Gröger, M. D. Bartoschek, P. Wolf, A. G. Beck-Sickinger, S. Bultmann and O. Seitz, Nat. Chem., 2021, 13, 15–23.
117. D. Beckett, E. Kovaleva and P. J. Schatz, Protein Sci., 1999, 8, 921–929.
118. I. Chen, M. Howarth, W. Lin and A. Y. Ting, Nat. Methods, 2005, 2, 99–104.
119. S. A. Slavoff, I. Chen, Y.-A. Choi and A. Y. Ting, J. Am. Chem. Soc., 2008, 130, 1160–1162.
120. D. E. Green, T. W. Morris, J. Green, J. E. Cronan and J. R. Guest, Biochem. J., 1995, 309, 853–862.
121. M. Fernández-Suárez, H. Baruah, L. Martínez-Hernández, K. T. Xie, J. M. Baskin, C. R. Bertozzi and A. Y. Ting, Nat. Biotechnol., 2007, 25, 1483–1487.
122. J. D. Cohen, P. Zou and A. Y. Ting, ChemBioChem, 2012, 13, 888–894.
123. D. S. Liu, A. Tangpeerachaikul, R. Selvaraj, M. T. Taylor, J. M. Fox and A. Y. Ting, J. Am. Chem. Soc., 2012, 134, 792–795.
124. J. Z. Yao, C. Uttamapinant, A. Poloukhtine, J. M. Baskin, J. A. Codelli, E. M. Sletten, C. R. Bertozzi, V. V. Popik and A. Y. Ting, J. Am. Chem. Soc., 2012, 134, 3720–3728.
125. H. Mao, S. A. Hart, A. Schink and B. A. Pollok, J. Am. Chem. Soc., 2004, 126, 2670–2671.
126. J. M. Antos, G. M. Miller, G. M. Grotenbreg and H. L. Ploegh, J. Am. Chem. Soc., 2008, 130, 16338–16343.
127. M. W. Popp, J. M. Antos, G. M. Grotenbreg, E. Spooner and H. L. Ploegh, Nat. Chem. Biol., 2007, 3, 707–708.
128. I. S. Carrico, B. L. Carlson and C. R. Bertozzi, Nat. Chem. Biol., 2007, 3, 321–322.
129. E. L. Smith, J. P. Giddens, A. T. Iavarone, K. Godula, L.-X. Wang and C. R. Bertozzi, Bioconjugate Chem., 2014, 25, 788–795.
130. C.-W. Lin and A. Y. Ting, J. Am. Chem. Soc., 2006, 128, 4542–4543.
131. K. Gatterdam, E. F. Joest, V. Gatterdam and R. Tampe, Angew. Chem., Int. Ed., 2018, 57, 12395–12399.
132. K. Gatterdam, E. F. Joest, M. S. Dietz, M. Heilemann and R. Tampe, Angew. Chem., Int. Ed., 2018, 57, 5620–5625.
133. Y. Chen, K. Tsao, S. L. Acton and J. W. Keillor, Angew. Chem., Int. Ed., 2018, 57, 12390–12394.
134. K. A. Hecht, Y. Xiong, D. A. Barrack, N. R. Ford, G. Roesijadi and T. C. Squier, ACS Omega, 2018, 3, 2104–2110.
135. C. Hillman, P. E. Stewart, M. Strnad, H. Stone, T. Starr, A. Carmody, T. J. Evans, V. Carracoi, J. Wachter and P. A. Rosa, Front. Cell. Infect. Microbiol., 2019, 9, 287.
136. P. Wolf, A. Mohr, G. Gavins, V. Behr, K. Morl, O. Seitz and A. G. Beck-Sickinger, ChemBioChem, 2022, 23, e202100340.
137. J. K. Doh, S. J. Tobin and K. E. Beatty, Biochemistry, 2020, 59, 3051–3059.
138. A. Suyama, K. L. Devlin, M. Macias-Contreras, J. K. Doh, U. Shinde and K. E. Beatty, Biochemistry, 2023, 62, 1735–1743.
139. D. Saimi and Z. Chen, Smart Mol., 2023, 1, e20230002.
140. M. Macias-Contreras, H. He and L. Zhu, Bioconjugate Chem., 2018, 29, 4104–4109.
141. C. A. Hoelzel and X. Zhang, ChemBioChem, 2020, 21, 1935–1946.
142. R. Damoiseaux, P. G. Schultz and K. Johnsson, ChemBioChem, 2002, 3, 573–575.
143. A. Keppler, S. Gendreizig, T. Gronemeyer, H. Pick, H. Vogel and K. Johnsson, Nat. Biotechnol., 2003, 21, 86–89.
144. A. Juillerat, C. Heinis, I. Sielaff, J. Barnikow, H. Jaccard, B. Kunz, A. Terskikh and K. Johnsson, ChemBioChem, 2005, 6, 1263–1269.
145. C. Heinis, S. Schmitt, M. Kindermann, G. Godin and K. Johnsson, ACS Chem. Biol., 2006, 1, 575–584.
146. A. Gautier, A. Juillerat, C. Heinis, I. R. Correa, Jr., M. Kindermann, F. Beaufils and K. Johnsson, Chem. Biol., 2008, 15, 128–136.
147. A. Gautier, E. Nakata, G. Lukinavicius, K. T. Tan and K. Johnsson, J. Am. Chem. Soc., 2009, 131, 17954–17962.
148. J. Wilhelm, S. Kuhn, M. Tarnawski, G. Gotthard, J. Tunnermann, T. Tanzer, J. Karpenko, N. Mertes, L. Xue, U. Uhrig, J. Reinstein, J. Hiblot and K. Johnsson, Biochemistry, 2021, 60, 2560.
149. 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 and K. V. Wood, ACS Chem. Biol., 2008, 3, 373–382.
150. R. F. Ohana, L. P. Encell, K. Zhao, D. Simpson, M. R. Slater, M. Urh and K. V. Wood, Protein Expression Purif., 2009, 68, 110–120.
151. L. W. Miller, Y. Cai, M. P. Sheetz and V. W. Cornish, Nat. Methods, 2005, 2, 255–257.
152. S. S. Gallagher, J. E. Sable, M. P. Sheetz and V. W. Cornish, ACS Chem. Biol., 2009, 4, 547–556.
153. Z. Chen, C. Jing, S. S. Gallagher, M. P. Sheetz and V. W. Cornish, J. Am. Chem. Soc., 2012, 134, 13692–13699.
154. J. Mo, J. Chen, Y. Shi, J. Sun, Y. Wu, T. Liu, J. Zhang, Y. Zheng, Y. Li and Z. Chen, Angew. Chem., Int. Ed., 2022, 61, e202207905.
155. M. A. van der Horst, J. C. Arents, R. Kort and K. J. Hellingwerf, Photochem. Photobiol. Sci., 2007, 6, 571–579.
156. Y. Hori, H. Ueno, S. Mizukami and K. Kikuchi, J. Am. Chem. Soc., 2009, 131, 16610–16611.
157. Y. Hori, T. Norinobu, M. Sato, K. Arita, M. Shirakawa and K. Kikuchi, J. Am. Chem. Soc., 2013, 135, 12360–12365.
158. S. Mizukami, S. Watanabe, Y. Hori and K. Kikuchi, J. Am. Chem. Soc., 2009, 131, 5016–5017.
159. R. J. Iwatate, A. Yoshinari, N. Yagi, M. Grzybowski, H. Ogasawara, M. Kamiya, T. Komatsu, M. Taki, S. Yamaguchi, W. B. Frommer and M. Nakamura, Plant Cell, 2020, 32, 3081–3094.
160. J. Ast, D. Nasteska, N. H. F. Fine, D. J. Nieves, Z. Koszegi, Y. Lanoiselee, F. Cuozzo, K. Viloria, A. Bacon, N. T. Luu, P. N. Newsome, D. Calebiro, D. M. Owen, J. Broichhagen and D. J. Hodson, Nat. Commun., 2023, 14, 301.
161. D. Zhang, Z. Chen, Z. Du, B. Bao, N. Su, X. Chen, Y. Ge, Q. Lin, L. Yang, Y. Hua, S. Wang, X. Hua, F. Zuo, N. Li, R. Liu, L. Jiang, C. Bao, Y. Zhao, J. Loscalzo, Y. Yang and L. Zhu, Cell Discovery, 2023, 9, 56.
162. C. W. Akey, D. Singh, C. Ouch, I. Echeverria, I. Nudelman, J. M. Varberg, Z. Yu, F. Fang, Y. Shi, J. Wang, D. Salzberg, K. Song, C. Xu, J. C. Gumbart, S. Suslov, J. Unruh, S. L. Jaspersen, B. T. Chait, A. Sali, J. Fernandez-Martinez, S. J. Ludtke, E. Villa and M. P. Rout, Cell, 2022, 185, 361–378.
163. T. Schlichthaerle, M. T. Strauss, F. Schueder, A. Auer, B. Nijmeijer, M. Kueblbeck, V. Jimenez Sabinina, J. V. Thevathasan, J. Ries, J. Ellenberg and R. Jungmann, Angew. Chem., Int. Ed., 2019, 58, 13004–13008.
164. J. Kompa, J. Bruins, M. Glogger, J. Wilhelm, M. S. Frei, M. Tarnawski, E. D'Este, M. Heilemann, J. Hiblot and K. Johnsson, J. Am. Chem. Soc., 2023, 145, 3075–3083.
165. M. Holtmannspotter, E. Wienbeuker, T. Dellmann, I. Watrinet, A. J. Garcia-Saez, K. Johnsson, R. Kurre and J. Piehler, Angew. Chem., Int. Ed., 2023, 62, e202219050.
166. S. Makhija, D. Brown, R. M. Rudlaff, J. K. Doh, S. Bourke, Y. Wang, S. Zhou, R. Cheloor-Kovilakam and B. Huang, ACS Chem. Biol., 2021, 16, 671–681.
167. N. Kumar, Y. Hori, M. Nishiura and K. Kikuchi, Chem. Sci., 2020, 11, 3694–3701.
168. M. Minoshima, T. Umeno, K. Kadooka, M. Roux, N. Yamada and K. Kikuchi, Angew. Chem., Int. Ed., 2023, 62, e202301704.
169. T. Tamura and I. Hamachi, J. Am. Chem. Soc., 2019, 141, 2782–2799.
170. K. Shiraiwa, R. Cheng, H. Nonaka, T. Tamura and I. Hamachi, Cell Chem. Biol., 2020, 27, 970–985.
171. J. Das, Chem. Rev., 2011, 111, 4405–4417.
172. M. M. Hassan and O. O. Olaoye, Molecules, 2020, 25, 2285.
173. E. Smith and I. Collins, Future Med. Chem., 2015, 7, 159–183.
174. N. Ankenbruck, T. Courtney, Y. Naro and A. Deiters, Angew. Chem., Int. Ed., 2018, 57, 2768–2798.
175. B. Zheng, H. Guo, N. Ma, Y. Ni, J. Xu, L. Li, P. Hao, K. Ding and Z. Li, Chem. – Asian J., 2018, 13, 2601–2605.
176. F. Kotzyba-Hibert, I. Kapfer and M. Goeldner, Angew. Chem., Int. Ed. Engl., 1995, 34, 1296–1312.
177. T. M. Ayele, S. D. Knutson, S. Ellipilli, H. Hwang and J. M. Heemstra, Bioconjugate Chem., 2019, 30, 1309–1313.
178. D. Yang and S. Y. Dai, J. Am. Chem. Soc., 2020, 142, 17156–17166.
179. T. J. McMillan, E. Leatherman, A. Ridley, J. Shorrocks, S. E. Tobi and J. R. Whiteside, J. Pharm. Pharmacol., 2008, 60, 969–976.
180. Y. Takaoka, H. Tsutsumi, N. Kasagi, E. Nakata and I. Hamachi, J. Am. Chem. Soc., 2006, 128, 3273–3280.
181. S. Tsukiji, M. Miyagawa, Y. Takaoka, T. Tamura and I. Hamachi, Nat. Chem. Biol., 2009, 5, 341–343.
182. T. Tamura, Y. Kioi, T. Miki, S. Tsukiji and I. Hamachi, J. Am. Chem. Soc., 2013, 135, 6782–6785.
183. T. Clackson, W. Yang, L. W. Rozamus, M. Hatada, J. F. Amara, C. T. Rollins, L. F. Stevenson, S. R. Magari, S. A. Wood, N. L. Courage, X. Lu, F. Cerasoli, M. Gilman and D. A. Holt, Proc. Natl. Acad. Sci. U. S. A., 1998, 95, 10437–10442.
184. S. H. Fujishima, R. Yasui, T. Miki, A. Ojida and I. Hamachi, J. Am. Chem. Soc., 2012, 134, 3961–3964.
185. S. Arttamangkul, A. Plazek, E. J. Platt, H. Jin, T. F. Murray, W. T. Birdsong, K. C. Rice, D. L. Farrens and J. T. Williams, eLife, 2019, 8, e49319.
186. H. Nonaka, S. Sakamoto, K. Shiraiwa, M. Ishikawa, T. Tamura, K. Okuno, T. Kondo, S. Kiyonaka, E. A. Susaki, C. Shimizu, H. R. Ueda, W. Kakegawa, I. Arai, M. Yuzaki and I. Hamachi, Proc. Natl. Acad. Sci. U. S. A., 2024, 121, e2313887121.
187. K. Ojima, K. Shiraiwa, K. Soga, T. Doura, M. Takato, K. Komatsu, M. Yuzaki, I. Hamachi and S. Kiyonaka, Nat. Commun., 2021, 12, 831.
188. Y. Takaoka, Y. Nishikawa, Y. Hashimoto, K. Sasaki and I. Hamachi, Chem. Sci., 2015, 6, 3217–3224.
189. L. A. Stoddart, N. D. Kindon, O. Otun, C. R. Harwood, F. Patera, D. B. Veprintsev, J. Woolard, S. J. Briddon, H. A. Franks, S. J. Hill and B. Kellam, Commun. Biol., 2020, 3, 722.
190. K. Y. Lin, C. Hin Lam, X. H. Lin, J. I. Hsu, S. Y. Fan, N. K. Gupta, Y. C. Lin, B. K. Tee, J. P. Li, J. K. Chen and K. T. Tan, Chem. – Asian J., 2021, 16, 937–948.
191. K. Matsuo, Y. Nishikawa, M. Masuda and I. Hamachi, Angew. Chem., Int. Ed., 2018, 57, 659–662.
192. T. Tamura, T. Ueda, T. Goto, T. Tsukidate, Y. Shapira, Y. Nishikawa, A. Fujisawa and I. Hamachi, Nat. Commun., 2018, 9, 1870.
193. M. Kosar, D. A. Sykes, A. E. G. Viray, R. M. Vitale, R. C. Sarott, R. L. Ganzoni, D. Onion, J. M. Tobias, P. Leippe, C. Ullmer, E. A. Zirwes, W. Guba, U. Grether, J. A. Frank, D. B. Veprintsev and E. M. Carreira, J. Am. Chem. Soc., 2023, 145, 15094–15108.
194. S. Albright, M. Cacace, Y. Tivon and A. Deiters, J. Am. Chem. Soc., 2023, 145, 16458–16463.
195. Y. Koshi, E. Nakata, M. Miyagawa, S. Tsukiji, T. Ogawa and I. Hamachi, J. Am. Chem. Soc., 2008, 130, 245–251.
196. T. Hayashi, Y. Yasueda, T. Tamura, Y. Takaoka and I. Hamachi, J. Am. Chem. Soc., 2015, 137, 5372–5380.
197. T. Tamura, Z. Song, K. Amaike, S. Lee, S. Yin, S. Kiyonaka and I. Hamachi, J. Am. Chem. Soc., 2017, 139, 14181–14191.
198. V. Thimaradka, H. Utsunomiya, T. Tamura and I. Hamachi, Org. Lett., 2023, 25, 2118–2122.
199. K. Wang, A. Sachdeva, D. J. Cox, N. W. Wilf, K. Lang, S. Wallace, R. A. Mehl and J. W. Chin, Nat. Chem., 2014, 6, 393–403.
200. D. L. Dunkelmann, S. B. Oehm, A. T. Beattie and J. W. Chin, Nat. Chem., 2021, 13, 1110–1117.
201. S. L. Scinto, D. A. Bilodeau, R. Hincapie, W. Lee, S. S. Nguyen, M. Xu, C. W. am Ende, M. G. Finn, K. Lang, Q. Lin, J. P. Pezacki, J. A. Prescher, M. S. Robillard and J. M. Fox, Nat. Rev. Methods Primers, 2021, 1, 30.
202. U. Stelzl, U. Worm, M. Lalowski, C. Haenig, F. H. Brembeck, H. Goehler, M. Stroedicke, M. Zenkner, A. Schoenherr, S. Koeppen, J. Timm, S. Mintzlaff, C. Abraham, N. Bock, S. Kietzmann, A. Goedde, E. Toksoz, A. Droege, S. Krobitsch, B. Korn, W. Birchmeier, H. Lehrach and E. E. Wanker, Cell, 2005, 122, 957–968.
203. J. F. Rual, K. Venkatesan, T. Hao, T. Hirozane-Kishikawa, A. Dricot, N. Li, G. F. Berriz, F. D. Gibbons, M. Dreze, N. Ayivi-Guedehoussou, N. Klitgord, C. Simon, M. Boxem, S. Milstein, J. Rosenberg, D. S. Goldberg, L. V. Zhang, S. L. Wong, G. Franklin, S. Li, J. S. Albala, J. Lim, C. Fraughton, E. Llamosas, S. Cevik, C. Bex, P. Lamesch, R. S. Sikorski, J. Vandenhaute, H. Y. Zoghbi, A. Smolyar, S. Bosak, R. Sequerra, L. Doucette-Stamm, M. E. Cusick, D. E. Hill, F. P. Roth and M. Vidal, Nature, 2005, 437, 1173–1178.
204. A. W. White, A. D. Westwell and G. Brahemi, Expert Rev. Mol. Med., 2008, 10, e8.
205. L. L. Blazer and R. R. Neubig, Neuropsychopharmacology, 2009, 34, 126–141.
206. M. Rosell and J. Fernandez-Recio, Expert Opin. Drug Discovery, 2018, 13, 327–338.
207. H. Lu, Q. Zhou, J. He, Z. Jiang, C. Peng, R. Tong and J. Shi, Signal Transduction Targeted Ther., 2020, 5, 213.
208. L. G. Milroy, T. N. Grossmann, S. Hennig, L. Brunsveld and C. Ottmann, Chem. Rev., 2014, 114, 4695–4748.
209. D. E. Scott, A. R. Bayly, C. Abell and J. Skidmore, Nat. Rev. Drug Discovery, 2016, 15, 533–550.
210. P. E. Czabotar, G. Lessene, A. Strasser and J. M. Adams, Nat. Rev. Mol. Cell Biol., 2014, 15, 49–63.
211. D. R. Green and G. Kroemer, Science, 2004, 305, 626–629.
212. S. K. Ko, J. Kim, D. C. Na, S. Park, S. H. Park, J. Y. Hyun, K. H. Baek, N. D. Kim, N. K. Kim, Y. N. Park, K. Song and I. Shin, Chem. Biol., 2015, 22, 391–403.
213. T. Gotoh, K. Terada, S. Oyadomari and M. Mori, Cell Death Differ., 2004, 11, 390–402.
214. D. R. Williams, S. K. Ko, S. Park, M. R. Lee and I. Shin, Angew. Chem., Int. Ed., 2008, 47, 7466–7469.
215. H. J. Cho, H. Y. Gee, K. H. Baek, S. K. Ko, J. M. Park, H. Lee, N. D. Kim, M. G. Lee and I. Shin, J. Am. Chem. Soc., 2011, 133, 20267–20276.
216. K. H. Baek, H. Zhang, B. R. Lee, Y. G. Kwon, S. J. Ha and I. Shin, Sci. Rep., 2015, 5, 17642.
217. S. H. Park, W. J. Kim, H. Li, W. Seo, S. H. Park, H. Kim, S. C. Shin, E. R. P. Zuiderweg, E. E. Kim, T. Sim, N. K. Kim and I. Shin, Sci. Rep., 2017, 7, 3537.
218. S. H. Park, K. H. Baek, I. Shin and I. Shin, Cell Chem. Biol., 2018, 25, 1242–1254e1248.
219. W. Wickner and R. Schekman, Science, 2005, 310, 1452–1456.
220. L. Wang and Y. Ye, Front. Cell Dev. Biol., 2020, 8, 610689.
221. J. E. Galan and G. Waksman, Cell, 2018, 172, 1306–1318.
222. J. E. Galan and A. Collmer, Science, 1999, 284, 1322–1328.
223. V. Goser, C. Kommnick, V. Liss and M. Hensel, mBio, 2019, 10, e00769–19.
224. D. M. Rosenbaum, S. G. Rasmussen and B. K. Kobilka, Nature, 2009, 459, 356–363.
225. H. Tian, A. Furstenberg and T. Huber, Chem. Rev., 2017, 117, 186–245.
226. K. Sriram and P. A. Insel, Mol. Pharmacol., 2018, 93, 251–258.
227. D. Yang, Q. Zhou, V. Labroska, S. Qin, S. Darbalaei, Y. Wu, E. Yuliantie, L. Xie, H. Tao, J. Cheng, Q. Liu, S. Zhao, W. Shui, Y. Jiang and M. W. Wang, Signal Transduction Targeted Ther., 2021, 6, 7.
228. S. Ferre, V. Casado, L. A. Devi, M. Filizola, R. Jockers, M. J. Lohse, G. Milligan, J. P. Pin and X. Guitart, Pharmacol. Rev., 2014, 66, 413–434.
229. G. Milligan, D. Ramsay, G. Pascal and J. J. Carrillo, Life Sci., 2003, 74, 181–188.
230. M. Cottet, O. Faklaris, D. Maurel, P. Scholler, E. Doumazane, E. Trinquet, J. P. Pin and T. Durroux, Front. Endocrinol., 2012, 3, 92.
231. H. Guo, S. An, R. Ward, Y. Yang, Y. Liu, X. X. Guo, Q. Hao and T. R. Xu, Biosci. Rep., 2017, 37, BSR20160547.
232. T. C. Moller, J. Hottin, C. Clerte, J. M. Zwier, T. Durroux, P. Rondard, L. Prezeau, C. A. Royer, J. P. Pin, E. Margeat and J. Kniazeff, Sci. Rep., 2018, 8, 10414.
233. R. Crupi, D. Impellizzeri and S. Cuzzocrea, Front. Mol. Neurosci., 2019, 12, 20.
234. D. El Moustaine, S. Granier, E. Doumazane, P. Scholler, R. Rahmeh, P. Bron, B. Mouillac, J. L. Baneres, P. Rondard and J. P. Pin, Proc. Natl. Acad. Sci. U. S. A., 2012, 109, 16342–16347.
235. D. J. Scholten, M. Canals, D. Maussang, L. Roumen, M. J. Smit, M. Wijtmans, C. de Graaf, H. F. Vischer and R. Leurs, Br. J. Pharmacol., 2012, 165, 1617–1643.
236. A. Isbilir, J. Moller, M. Arimont, V. Bobkov, C. Perpina-Viciano, C. Hoffmann, A. Inoue, R. Heukers, C. de Graaf, M. J. Smit, P. Annibale and M. J. Lohse, Proc. Natl. Acad. Sci. U. S. A., 2020, 117, 29144–29154.
237. J. Moller, A. Isbilir, T. Sungkaworn, B. Osberg, C. Karathanasis, V. Sunkara, E. O. Grushevskyi, A. Bock, P. Annibale, M. Heilemann, C. Schutte and M. J. Lohse, Nat. Chem. Biol., 2020, 16, 946–954.
238. R. Serfling, L. Seidel, A. Bock, M. J. Lohse, P. Annibale and I. Coin, ACS Chem. Biol., 2019, 14, 1141–1149.
239. C. H. Heldin, B. Lu, R. Evans and J. S. Gutkind, Cold Spring Harbor Perspect. Biol., 2016, 8, a005900.
240. C. H. Heldin, Cell, 1995, 80, 213–223.
241. J. Li, L. Wang, J. Tian, Z. Zhou, J. Li and H. Yang, Chem. Soc. Rev., 2020, 49, 1545–1568.
242. Y. Zwang and Y. Yarden, Traffic, 2009, 10, 349–363.
243. C. H. Lee, S. Park, S. Kim, J. Y. Hyun, H. S. Lee and I. Shin, Chem. Sci., 2024, 15, 555–565.
244. D. A. Sabbah, R. Hajjo and K. Sweidan, Curr. Top. Med. Chem., 2020, 20, 815–834.
245. J. Chen, F. Zeng, S. J. Forrester, S. Eguchi, M. Z. Zhang and R. C. Harris, Physiol. Rev., 2016, 96, 1025–1069.
246. A. Ayati, S. Moghimi, S. Salarinejad, M. Safavi, B. Pouramiri and A. Foroumadi, Bioorg. Chem., 2020, 99, 103811.
247. A. Gschwind, O. M. Fischer and A. Ullrich, Nat. Rev. Cancer, 2004, 4, 361–370.
248. J. Lotze, P. Wolf, U. Reinhardt, O. Seitz, K. Morl and A. G. Beck-Sickinger, ACS Chem. Biol., 2018, 13, 618–627.
249. K. Henzler-Wildman and D. Kern, Nature, 2007, 450, 964–972.
250. S. E. Gordon, M. Munari and W. N. Zagotta, eLife, 2018, 7, e37248.
251. M. Puljung, N. Vedovato, S. Usher and F. Ashcroft, eLife, 2019, 8, e41103.
252. K. P. K. Lee, J. Chen and R. MacKinnon, eLife, 2017, 6, e32481.
253. M. A. Shandell, J. R. Quejada, M. Yazawa, V. W. Cornish and R. S. Kass, Biophys. J., 2019, 117, 1352–1363.
254. D. M. Roden, J. R. Balser, A. L. George, Jr. and M. E. Anderson, Annu. Rev. Physiol., 2002, 64, 431–475.
255. M. de Lera Ruiz and R. L. Kraus, J. Med. Chem., 2015, 58, 7093–7118.
256. E. Suarez-Delgado, M. Orozco-Contreras, G. E. Rangel-Yescas and L. D. Islas, eLife, 2023, 12, e85836.
257. T. E. DeCoursey, Physiology, 2010, 25, 27–40.
258. P. E. Deal, P. Liu, S. H. Al-Abdullatif, V. R. Muller, K. Shamardani, H. Adesnik and E. W. Miller, J. Am. Chem. Soc., 2020, 142, 614–622.
259. C. Deo, A. S. Abdelfattah, H. K. Bhargava, A. J. Berro, N. Falco, H. Farrants, B. Moeyaert, M. Chupanova, L. D. Lavis and E. R. Schreiter, Nat. Chem. Biol., 2021, 17, 718–723.
260. J. B. Grimm, A. K. Muthusamy, Y. Liang, T. A. Brown, W. C. Lemon, R. Patel, R. Lu, J. J. Macklin, P. J. Keller, N. Ji and L. D. Lavis, Nat. Methods, 2017, 14, 987–994.
261. R. I. Morimoto and A. M. Cuervo, J. Gerontol., Ser. A, 2014, 69, S33–S38.
262. X. Q. Chen, T. Shen, S. J. Fang, X. M. Sun, G. Y. Li and Y. F. Li, Front. Cell Dev. Biol., 2023, 11, 1143532.
263. K. Nishimura, T. Fukagawa, H. Takisawa, T. Kakimoto and M. Kanemaki, Nat. Methods, 2009, 6, 917–922.
264. S. I. Reja, Y. Hori, T. Kamikawa, K. Yamasaki, M. Nishiura, S. D. Bull and K. Kikuchi, Chem. Sci., 2022, 13, 1419–1427.
265. L. M. Smith and N. L. Kelleher, Nat. Methods, 2013, 10, 186–187.
266. R. J. O'Brien and P. C. Wong, Annu. Rev. Neurosci., 2011, 34, 185–204.
267. C. Reily, T. J. Stewart, M. B. Renfrow and J. Novak, Nat. Rev. Nephrol., 2019, 15, 346–366.
268. A. Varki, Glycobiology, 2017, 27, 3–49.
269. S. Park, M. R. Lee and I. Shin, Chem. Soc. Rev., 2008, 37, 1579–1591.
270. Y. Gao, X. Luan, J. Melamed and I. Brockhausen, Cells, 2021, 10, 1252.
271. I. G. Ferreira, M. Pucci, G. Venturi, N. Malagolini, M. Chiricolo and F. Dall'Olio, Int. J. Mol. Sci., 2018, 19, 580.
272. M. Takahashi, T. Tsuda, Y. Ikeda, K. Honke and N. Taniguchi, Glycoconjugate J., 2004, 20, 207–212.

---

Footnote

† Two authors equally contributed to this work.

---