Thiosulfonate‐Derived BODIPY “Stick and Glue” Strategy for Fluorescent Thiol Labeling

Abstract

Fluorescent thiol labeling is a powerful technique for investigating peptide and protein functions, interactions, and cellular localization, offering minimal structural disruption due to the low natural abundance of cysteine residues. The high nucleophilicity of cysteine thiol groups further facilitates efficient and selective labeling. However, existing thiol‐labeling strategies, most commonly involving iodoacetamides, maleimides, or methanethiosulfonates (MTS), often suffer from limited selectivity, undesirable side reactions, and the instability of the resulting conjugates. To overcome these limitations, we developed three spectrally orthogonal MTS‐derived fluorescent labeling agents based on coumarin and BODIPY photoremovable protecting groups and a BODIPY thiosulfate water‐soluble analogue. Using conventional MTS chemistry, these agents form disulfide‐linked conjugates with thiols, which are subsequently converted into stable thioethers upon light‐induced sulfur extrusion. This two‐step approach significantly improves the stability of labeled conjugates, providing a robust and permanent method for fluorescent labeling of small molecules, peptides, and proteins. Our findings offer a promising strategy for precise fluorescent labeling in biological applications.

We present a powerful photochemical strategy for converting disulfide bonds into stable thioether linkages via a [1,2]‐sigmatropic rearrangement. By integrating this reaction into traditional methanethiosulfonate‐based thiol labeling, we enable light‐triggered stabilization of peptide and protein conjugates with enhanced robustness and versatility.

Introduction

Fluorescent labeling is an excellent method allowing visualization of targeted species in real time, including in live cells. Whether genetically encoded or chemically conjugated, fluorescent labeling can be particularly useful for mapping peptide and protein functions, interactions, and cellular localization.^{[ 1 ]} But chemically conjugated methods offer a key advantage, i.e., minimal disruption of protein structure and function. Among these methods, cysteine‐specific labeling targets protein thiol groups.^{[ 2} , ³ , ^{4 ]}

Cysteine thiol groups enable specific chemical labeling thanks to their low abundance (∼2.3% in mammals)^{[ 5 ]} and excellent nucleophilicity^{[ 6 ]} in comparison to other nucleophiles abundant in proteins (amines, amides, alcohols, etc.). Furthermore, peptides can be readily functionalized during solid‐phase peptide synthesis by introducing exogenous cysteine residues, which further serve as additional reactive sites.

Current commercially available chemical thiol‐labeling agents benefit from the reactivity of iodoacetamides (Figure 1a), maleimides (Figure 1b), or methanethiosulfonates (MTS, Figure 1c).^{[ 7 ]}

Figure 1.

Comparison of different labeling methods: a) iodoacetamide, b) maleimide, and c) MTS and our approach.

Notwithstanding their advantages, these methods suffer from major limitations. Iodoacetamides show low selectivity and react with various functional groups.^{[ 8} , ^{9 ]} Even the most commonly used labeling agents, maleimides, are prone to undesirable retro‐Michael addition, which reverses labeling conjugation.^{[ 10} , ^{11 ]} In addition, maleimides introduce an additional stereogenic center to the molecule and undergo a thiazine rearrangement when labeling N‐terminal cysteine. This rearrangement decreases the overall labeling yield.^{[ 12 ]} Methanethiosulfonates (MTS) are commonly used for transient thiol labeling. However, their permanent labeling and application in cells or in‐vivo systems is limited by the high propensity for reduction of their disulfide conjugates by cellular cysteine, glutathione (GSH), enzymes, and other reductants.^{[ 13 ]} Their reduction precludes permanent MTS labeling and applications in cell cultures and in vivo models.

In this study, we improved this highly selective and practical labeling technique by an additional photochemical step (Figure 1c, our approach) to increase the cellular stability of small molecules and peptides labeled by MTS‐based agents. In addition, we developed a water‐soluble BODIPY analogue bearing a thiosulfate group. Inspired by our prior research on meso‐methyl BODIPYs,^{[ 14} , ^{15 ]} we developed three chromatically orthogonal, MTS‐derived labeling agents, namely Coumarin‐MTS (Figure 2a), BDP‐MTS (Figure 2b), and ext‐BDP‐MTS (Figure 2c). When labeling butanethiol, cysteine, and peptides with these agents, we found that their disulfide conjugates displayed a novel photoreactive pathway. Photochemical rearrangement of these disulfides yielded thioethers as the products of sulfur extrusion (Figure 2d–e). The sulfur‐extrusion step enhanced the clean, fast, and selective MTS labeling by the light‐induced formation of thioethers, exhibiting superior stability compared to their parent disulfides. This process may be described as a ’stick‐and‐glue‘ strategy, where the initial conjugation step acts as a temporary “sticker” that forms a disulfide bond for transient labeling capable of reductively restoring the thiol residue to its native form. Upon exposure to light, the system undergoes a “gluing” step that stabilizes the conjugate by transforming the disulfide into a stable thioether bond, ensuring permanent labeling. This dual approach provides researchers the ability to control the stability of labeled conjugates, providing versatility for applications that require either transient or permanent labeling.

Figure 2.

Synthesis of the labeling agents: a) Coumarin‐MTS, b) BDP‐MTS, c) ext‐BDP‐MTS. Labeling of thiol groups and subsequent photochemical sulfur extrusion: d) BDP‐SS‐Bu/BDP‐S‐Bu, BDP‐SS‐Cys/BDP‐S‐Cys, ext‐BDP‐SS‐Bu/ext‐BDP‐S‐Bu, and e) Coumarin‐SS‐Bu/Coumarin‐S‐Bu. Photolysis of f) BDP‐OAc, and ext‐BDP‐OAc. Synthetic details in Supporting Information, Chapter 3.

Results and Discussion

Synthesis

We synthesized three labeling agents, namely Coumarin‐MTS, BDP‐MTS, and ext‐BDP‐MTS, as shown in Figure 2a–c. These agents were specifically designed to have distinctly different absorption spectra that would allow chromatically orthogonal photochemical activation of individual components in a mixture. Coumarin‐MTS was synthesized from the commercially available 4‐(bromomethyl)‐7‐methoxy‐2H‐chromen‐2‐one (Coumarin‐Br) by nucleophilic substitution with sodium methanethiosulfonate (Figure 2a). BDP‐MTS was synthesized by a straightforward, two‐step synthesis, starting from the condensation of 2,4‐dimethylpyrrol with chloroacetyl chloride and further complexation with BF₃ · OEt₂, yielding the BODIPY derivative BDP‐Cl. The final compound BDP‐MTS was prepared from BDP‐Cl by a nucleophilic substitution of chloride with sodium methanethiosulfonate (Figure 2b).

ext‐BDP‐MTS was prepared by a five‐step synthesis, starting from the condensation of 2,4‐dimethylpyrrol with acetoxyacetyl chloride and further complexation with BF₃ · OEt₂ to prepare the BODIPY derivative BDP‐OAc. Then, ext‐BDP‐OAc was prepared via piperidine‐catalyzed Knoevenagel condensation with anisaldehyde, further extending the aromatic system. This was followed by hydrolysis of the acetate group with sodium hydroxide, forming an alcohol ext‐BDP‐OH. Via a modified Appel reaction with a triphenylphosphine‐iodine complex and imidazole, ext‐BDP‐OH was converted into ext‐BDP‐I. This electrophilic iodo‐derivative served as the key intermediate for the synthesis of the target compound ext‐BDP‐MTS. Moreover, ext‐BDP‐I stands out as an important milestone in the preparation of substituted extended BODIPY derivatives featuring various leaving groups at the meso‐position. Other tested methods for the installation of a specific leaving group at the meso‐methyl position, including tosylation, mesylation, Mitsunobu reaction, and Appel reaction turned out to be unsuccessful. The final stage of the synthesis involved a nucleophilic substitution of iodide by sodium methanethiosulfonate, yielding the target product ext‐BDP‐MTS (Figure 2c).

These three thiol labeling agents were subsequently mixed with butanethiol to prepare a series of mixed disulfides as model compounds (BDP‐SS‐Bu, ext‐BDP‐SS‐Bu, Figure 2d) and Coumarin‐SS‐Bu (Figure 2e). Additionally, BDP‐MTS was mixed with cysteine, forming BDP‐SS‐Cys (Figure 2d), to demonstrate the system's compatibility with various functional groups.

All disulfide derivatives (Coumarin‐SS‐Bu, BDP‐SS‐Bu, BDP‐SS‐Cys, and ext‐BDP‐SS‐Bu) showed unprecedented photoreactivity resulting in their transformation from disulfides to thioethers (Coumarin‐S‐Bu, BDP‐S‐Bu, BDP‐S‐Cys, ext‐BDP‐S‐Bu, respectively, Figure 2d–e). HPLC‐MS analysis (Figure S16–S18) indicated that this photoreaction is rapid and clean for all BODIPY derivatives, yielding more stable thioethers. For Coumarin‐SS‐Bu, the photoreaction also proceeds toward the thioether formation (Coumarin‐S‐Bu), but with further decomposition to Coumarin‐SH and other by‐products caused by the use of high‐energy UV light (Figure S19).

Photophysical and Photochemical Properties

Table 1 outlines the photophysical and photochemical properties of the prepared compounds. The BODIPY derivatives possess excellent absorption properties, exhibiting high molar absorption coefficients of up to ∼8.9 × 10⁵ M⁻¹ cm⁻¹. The MTS derivatives were selected to cover the entire range from UV to near‐IR spectrum, with low spectral overlap between chromophores. Coumarin‐MTS absorbs UV light with a maximum (λ _abs,max) at 328 nm; BDP‐MTS, green light (λ _abs,max = 519 nm); and ext‐BDP‐MTS, red light (λ_abs,max = 662 nm), facilitating enhanced light penetration into tissues. The absorption properties of the chromophores are not affected by variations in substituents at the meso‐position.

Table 1.

Photophysical and photochemical properties of the synthesized compounds.

compound name		λ abs,max a) nm	λ em b) nm	Δṽ c) cm−1	ε max d) M−1cm−1	Φfluo e) , i) %	Φr f) %
Coumarin‐	MTS	328	397	5171	10 700	0.38	n.d. g)
	SS‐Bu	326	388	4835	13 200	0.84	n.d. g)
BDP‐	MTS	519	540	749	66 100	21	n.d. g)
	SS‐Bu	514	530	587	52 300	5.1	1.71 ± 0.03 i)
	S‐Bu	510	526	596	37 300	43	n.r. h)
	SS‐Cys	514	535	764	35 900	18	8.6 ± 1.3
	S‐Cys	510	533	846	37 200	48	n.r. h)
ext‐BDP‐	MTS	662	684	486	88 800	34	n.d. g)
	SS‐Bu	654	676	498	24 400	30	0.052 ± 0.001 i)
	S‐Bu	654	678	541	16 100	25	n.r. h)

All data were measured in methanol unless specified;

^a) absorption maximum,

^b) emission maximum,

^c) Stokes shift,

^d) molar extinction coefficient,

^e) fluorescence quantum yield,

^f) photochemical sulfur extrusion quantum yield, data were obtained from the three independent measurements,

^g) n.d.: not determined,

^h) n.r.: no reaction, and

ⁱ⁾ measured in MeCN.

All prepared derivatives exhibit fluorescence in the visible region. Coumarin derivatives show large Stokes shifts (Δṽ) of up to 5171 cm⁻¹, whereas all BODIPY derivatives exhibit small Δṽ of up to 846 cm⁻¹. In parent BODIPY derivatives, sulfur extrusion was associated with an increase in fluorescence quantum yields from 5.1% to 43% for BDP‐SS‐Bu/BDP‐S‐Bu and from 18% to 48% for BDP‐SS‐Cys/BDP‐S‐Cys, which enables the detection of sulfur extrusion by monitoring the fluorescence buildup.

This effect is caused by strong intramolecular quenching of the excited singlet state of BODIPY^{[ 16 ]} by the electron‐rich disulfide moiety, which lowers the fluorescence quantum yield of the corresponding disulfides. The strong emission of sulfide photoproducts suggests that intramolecular quenching does not occur between the ‐S‐ moiety and the excited BODIPY. In extended BODIPYs, the fluorescence quantum yield remained similar upon sulfur extrusion, indicating a minimal effect of the substitution on the low‐lying singlet excited state emission.

To quantify the photoreaction efficiency of sulfur extrusion, we determined the photoreaction quantum yield for BDP‐SS‐Bu (Φ _r = 1.71 ± 0.03%), BDP‐SS‐Cys (Φ _r = 8.6 ± 1.3%), and ext‐BDP‐SS‐Bu (Φ _r = 0.052 ± 0.001%). All values were one order of magnitude higher than those of the corresponding photocages with an acetate leaving group (BDP‐OAc (Φ _r = 0.14%, Figure 2f) and ext‐BDP‐OAc (Φ _r = 0.0041%, Figure 2f). The uncaging cross‐sections Φ_rε (BDP‐SS‐Bu ∼800, BDP‐SS‐Cys ∼3000, ext‐BDP‐SS‐Bu ∼13) are well‐tuned to allow for handling the compounds under the ambient light, while still high enough to reach full conversion after short irradiation with low‐power LEDs (on a scale of ∼ 100 µmol).

The photoreaction quantum yield of Coumarin‐SS‐Bu could not be accurately determined due to the gradual photodecomposition of the primary thioether photoproduct. However, relative photoreaction rates are proportional to the photon energy in the order of ext‐BDP‐SS‐Bu < BDP‐SS‐Bu < Coumarin‐SS‐Bu (Table 1.). This trend is explained by i) the Energy Gap Law ^{[ 17 ]} —the increasing rate of non‐radiative decay with decreasing S₀‐S₁ separation, and ii) the relative energies of the excited species to the transition state energy as per the Bell–Evans–Polanyi principle, determined by quantum chemical calculations (Figure 3d). The photoreaction rate (ext‐BDP‐SS‐Bu < BDP‐SS‐Bu < Coumarin‐SS‐Bu) and fluorescence quantum yield (Coumarin‐SS‐Bu < BDP‐SS‐Bu < ext‐BDP‐SS‐Bu) trends may be attributed to variations in the activation barriers for sulfur extrusion in the excited state. While the ground‐state activation barrier is approximately the same (∼70 kcal mol⁻¹) across all compounds, the singlet excited state energy differs, being the highest for Coumarin‐SS‐Bu and lowest for ext‐BDP‐SS‐Bu. As a result, the activation barrier for the sulfur extrusion in the excited state follows the order Coumarin‐SS‐Bu < BDP‐SS‐Bu < ext‐BDP‐SS‐Bu. As the activation barrier increases, the excited state becomes less reactive, favoring relaxation through fluorescence. This, in turn, results in a higher fluorescence quantum yield. These findings highlight the competitive relationship between photoreaction and fluorescence pathways.

Figure 3.

Illustration and justification of the sulfur extrusion mechanism. a) and b) Proposed mechanistic pathways of light‐induced sulfur extrusion of the BODIPY fluorophore. Disulfide undergoes [1,2]‐sigmatropic rearrangement to thiosulfoxide, which rapidly decomposes to thioether and elemental sulfur. c) Schematic representation of the thiosulfoxide [1,2]‐sigmatropic rearrangement using Lewis‐like NBO orbitals. The thermally‐forbidden/photochemically‐allowed dichotomy is visualized by imposing symmetry restriction on the p‐orbitals, reflecting the ground‐ and excited‐state electron configuration, as per 4,3‐CASSCF (SI, Chapter 18), which prevents their symmetric combination in the ground state and, in turn, allows their symmetric combination in the excited state. Sulfur atoms are depicted in violet. d) Simplified diagram of the calculated ground‐ and excited‐state isomerization energies (Table S1, Figures S82–S84) of Coumarin‐SS‐Bu, BDP‐SS‐Bu, and ext‐BDP‐SS‐Bu, illustrating that the excited‐state reaction barrier increases with the decrease in vertical excitation energy.

Reaction Mechanism

Through a combined experimental/quantum‐chemical approach, we attempted to narrow down the reaction mechanism of sulfur extrusion. To this end, we first conducted a crossover experiment using BDP‐SS‐Cys and ext‐BDP‐SS‐Bu (Figures S22–S25). Through this experiment, we aimed to assess whether the thiols remained attached to their respective BODIPY counterparts or if they scrambled, thereby determining if the molecule is cleaved into two individual fragments during the photoinduced sulfur extrusion. HPLC‐MS analysis did not reveal any crossover product traces, indicating that the sulfur extrusion reaction is a unimolecular process.

Next, we performed a control experiment to determine which electronic states facilitate the sulfur extrusion. In the dark‐control experiment, we did not observe any changes in the starting material, indicating that the reaction is a photoinitiated process (Figure S15). Moreover, under irradiation, the kinetics of the reaction was not affected by a high concentration of triplet quencher (cyclooctatetraene, 2600 eq.), so we ruled out the involvement of long‐lived triplet excited states in the sulfur extrusion reaction^{[ 14 ]} (Figures S30–S33). In line with the fluorescence quantum yield trend (vide supra), we hence suggest that sulfur‐extrusion occurs from the first singlet excited state.

Based on the experimental evidence and on the literature,^{[ 18} , ¹⁹ , ²⁰ , ^{21 ]} we devised that sulfur extrusion proceeds in two steps. In the first photoinitiated step, the disulfide functional group isomerizes to the corresponding thiosulfoxide (Figure 3a). In the second, ground‐state step, the thiosulfoxide spontaneously and rapidly decomposes to the corresponding thioether and elemental sulfur.^{[ 19} , ^{22 ]}

Disulfide rearrangement to thiosulfoxides is a known phenomenon, observed in elemental sulfur, persulfides (RSSH), polysulfides (R‐S _n ‐R), and polythionates (‐SO₃‐S _n ‐SO₃) where n ≥ 3.^{[ 23 ]} These thiosulfoxide‐forming compounds are important sulfane sulfur precursors for biosynthetic processes, such as cyanide detoxification.^{[ 24 ]} However, disulfides do not usually undergo thiosulfoxide rearrangement, unless the C─S bond is polarized.^{[ 23 ]} So why does photochemical activation of coumarin and BODIPY derivatives promote this reactivity?

To answer this question, we performed a quantum‐chemical study in which we modeled the isomerization reaction in both ground and excited states (Supporting Information, Chapter 18). Using the nudged elastic band method^{[ 25 ]} in all compounds, we successfully converged the “alkyl migration” reaction in which the S‐alkyl sulfur atom isomerizes to an S^II− group while the alkyl group migrates to the adjacent sulfur atom through a cyclic transition state (Figure 3a, pathway A). Conversely, the contrasting “aryl‐methyl migration” (Figure 3b, pathway B) did not converge, instead leading to various chemically unfeasible geometries.

In the converged pathway A, we observed that the reaction is virtually thermally forbidden due to its high activation energy (ΔE ^TS > 70 kcal mol⁻¹) and that the thiosulfoxide products are thermodynamically unfavored (ΔE > 15 kcal mol⁻¹) in the ground state. On the other hand, in the excited state, the barrier is lowered in dye‐bearing molecules and virtually nonexistent in an unperturbed model molecule—dimethyldisulfide (Figures S68–S73). Furthermore, in dye‐bearing molecules, the reaction can also proceed through a nonadiabatic pathway facilitated by low‐lying S₁‐S₀ conical intersections (Table S1, Figures S76–S84). Thus, based on these results, we can conclude that the reaction is thermally forbidden, but photochemically allowed.

The contrasting ground‐ and excited‐state reactivities and the course of the reaction through a 4‐electron‐delocalized, cyclic transition state (Figures S70–S73) indicate that disulfide isomerization to thiosulfoxide can be classified as a pericyclic reaction. By natural bond orbital (NBO) analysis, we constructed a Lewis‐like model showing isomerization through a [1,2]‐sigmatropic rearrangement (Figure 3c). This theoretical model is in line with the observed reactivity and calculated ground‐ and excited‐state energetics, providing intuitive insights into the rearrangement mechanism.

To conclude our mechanistic study, we aimed to determine which sulfur species were released from the thiosulfoxide intermediate. We performed methylene blue assay for the detection of sulfane^{[ 26 ]} and fluorescent detection of dihydrogen disulfide, both of which yielded negative results (Figures S35–S37).^{[ 27 ]} Thus, H₂S nor H₂S₂ are not products of sulfur extrusion. We suspected that the reaction mixture could contain higher polysulfur compounds (elemental sulfur or S _n ). Because elemental sulfur can be reduced to sulfane by GSH, we repeated the methylene blue assay after reduction of the irradiated BDP‐SS‐Bu mixture by GSH. In this experiment, we detected sulfane as a product of S _n reduction by GSH (Figure S35).

Based on the sulfane presence in the irradiated reaction mixture reduced by GSH, and the previously suggested mechanism,^{[ 28 ]} we conclude that the sulfur extrusion reaction yields elemental sulfur. Formation of elemental sulfur was further supported by a recent study by Kawaguchi et al.,^{[ 29 ]} who investigated a structurally distinct coumarin derivative containing a disulfide bond also undergoing such photochemical rearrangement. Using an S⁰‐specific detection assay, they confirmed that the photoproduct indeed corresponds to elemental sulfur. Moreover, upon computational investigation, we found the extrusion of elemental sulfur highly exergonic (Figure S85), thereby explaining its spontaneity and ruling out the involvement of other reagents or coupled chemical transformations.

Thiol Labeling

As previously mentioned, the main advantage of our methodology lies in the enhanced stability of sulfide (thioether) conjugates in a reducing environment. To test this hypothesis, we conducted a stability study of BDP‐SS‐Bu and BDP‐S‐Bu under reducing conditions (2 eq. of TCEP in PBS). While BDP‐SS‐Bu was fully reduced within 1 h, BDP‐S‐Bu remained completely stable for at least 48 h under the same conditions (Figures S26–S30). We then compared the labeling performance of commercially available labeling agents, namely MTS‐AMCA (Biotium), MTS‐CR110 (Biotium), and BDP‐FL‐maleimide (Lumiprobe), with that of our labeling agent BDP‐MTS (Figure 4a). We followed a standardized protocol for thiol labeling^{[ 7 ]} using cysteine methyl ester as a thiol source. The commercially available labeling agents provided complex mixtures difficult to separate. By contrast, our labeling agent quantitatively yielded a single labeling product (Figure 4b).

Figure 4.

This panel illustrates the comparison of the commercially available thiol‐labels with our labeling agent and the scope of tested peptides used in this study. a) Structures of used labeling agents: BDP‐MTS, BDP‐FL‐ maleimide, MTS‐CR110 and MTS‐AMCA. b) HPLC chromatograms of the reaction mixtures after thermal labeling of L‐cysteine methyl ester (no photochemical step involved). Commercially available labeling agents (BDP‐FL‐maleimide, MTS‐CR110, and MTS‐AMCA) yield complex reaction mixtures, whereas BDP‐MTS yields a single labeled product. c) Scheme of Peptidisc, glutathione, and CysTAT peptide labeling. All peptides were labeled with BDP‐MTS, and Peptidisc and CysTAT were also labeled with ext‐BDP‐MTS.

To demonstrate the ability of our method to label larger systems, we labeled Peptidisc (over 4000 Da), which contains an N‐terminal cysteine in the peptide chain (Figure 4c) prone to thiazine rearrangement upon maleimide labeling. First, we labeled Peptidisc with ext‐BDP‐MTS using the standard procedure^{[ 7 ]} (a mixture of DMSO and PBS buffer), yielding ext‐BDP‐SS‐Peptidisc. When irradiating the sample with 625‐nm LED, we obtained ext‐BDP‐S‐Peptidisc (Supporting Information, Chapter 5). Subsequently, we labeled Peptidisc with BDP‐MTS (Supporting Information, Chapter 5) using an optimized procedure in a mixture of acetonitrile and ammonium bicarbonate buffer (1.7% DMF, 65% ammonium bicarbonate, 33.3% MeCN) to prevent unwanted oxidation of the peptide^{[ 30 ]} and to improve compatibility with the LC‐MS system. Because Peptidisc contains phenylalanines, its water solubility was low, requiring a higher amount of organic solvent (DMSO or MeCN, 33–50%) to solubilize the reagents.

Our hydrophobic labeling agents also labeled other hydrophilic peptides, such as glutathione and CysTAT (Figure 4c). These model targets were labeled by BDP‐MTS to BDP‐SS‐GSH and BDP‐SS‐CysTAT and, upon irradiation, converted into BDP‐S‐GSH and BDP‐S‐CysTAT, respectively (Figure 4c). Completing the scope, we labeled CysTAT with ext‐BDP‐MTS (SI, Chapter 5).

Additionally, we selected glutathione labeling to illustrate that the irradiation step can be carried out with the same results either immediately after the conjugation (one‐pot procedure) or subsequently after the purification (Figures S1–S5).

Despite the availability of standardized labeling protocols, it is essential to note that each peptide requires specific handling due to its limited solubility in different solvents and pH. Therefore, a careful selection of an appropriate buffer and solvent is necessary for every instance to achieve the best labeling conditions. However, once the peptide and the labeling agent are fully solubilized in a solution with the appropriate pH, the labeling process occurs efficiently. By this, our versatile method enables highly selective MTS‐based labeling enhanced by a photochemical step that further stabilizes the peptide‐label conjugate.

For practical application, a general protocol utilizing our labeling strategy is described in the Supporting Information (SI, Chapter 3).

Extension to Aqueous media and Protein Substrates

The aforementioned modifications were performed using a high amount of organic co‐solvent (DMSO, MeCN, up to 50%), but when targeting proteins or functional peptides, where the tertiary structure is important, the use of organic solvents is limited. Although BDP‐MTS is intrinsically hydrophobic and insoluble in pure aqueous media, we determined its solubility limit under protein‐compatible conditions (Tris‐buffered saline, TBS: 20 mM Tris, 150 mM NaCl, 1% DMSO, pH 8.0, Supporting Information, Chapter 20). The reagent remained fully soluble up to 10 µM, while precipitation occurred at concentrations above 50 µM, suggesting a solubility threshold between 10 and 50 µM. Importantly, the safe working concentration of 10 µM falls within the range typically employed for protein conjugations.

To demonstrate that BDP‐MTS labeling and subsequent photochemical stabilization can be achieved under biocompatible aqueous conditions, we selected bovine serum albumin (BSA, Figure 5a, Supporting Information, Chapter 21) as a relevant thiol‐containing target. Conjugation of BSA with BDP‐MTS, followed by SDS‐PAGE analysis, confirmed the covalent attachment of the label to BSA (Figure S121). DTT treatment partially released the label from the nonirradiated conjugate. Irradiation of the conjugate prior to reduction markedly reduced the dye loss, indicating its enhanced stability. Overall, these results show that protein labeling with BDP‐MTS is feasible in TBS (20 mM Tris, 150 mM NaCl, 1% DMSO) and that light activation can stabilize the conjugate in reducing environment, highlighting its potential for biologically relevant applications.

Figure 5.

Scheme of a) BSA labeling with BDP‐MTS and b) CysMe labeling with BDP‐TS.

To fully overcome the solubility limitation, we attempted to replace the MTS group with a thiosulfate (TS) moiety (BDP‐TS, Figure 5b). This structural change enhances the aqueous solubility while maintaining the chemo‐selectivity of the reagent. BDP‐TS can be easily prepared from the BDP‐Cl intermediate in a single step. Importantly, this structural change was introduced on the leaving group moiety; thus, all products formed upon the reaction with thiols and subsequent irradiation are identical to those obtained with the original BDP‐MTS reagent.

To test whether BDP‐TS is capable of labeling in aqueous media, we selected cysteine methyl ester as a simple water‐soluble substrate (Figure 5b, Supporting Information, Chapter 20). The first step proceeded efficiently in PBS, confirming its suitability for aqueous conditions. The subsequent photochemical step, however, was complicated by the low solubility of the product (BDP‐SS‐CysMe) in PBS, which led to partial precipitation and thus incomplete conversion. Upon the addition of acetonitrile co‐solvent, the photoreaction proceeded quantitatively. Since the photochemical step is solvent‐independent, as established in previous experiments, we regard this study as a successful proof of concept that the labeling procedure is fully water‐compatible.

To assess whether BDP‐TS could also label more complex substrates, we selected the CysTAT peptide as a model system and applied the same conditions previously used for BDP‐MTS. While BDP‐MTS efficiently formed the desired conjugate, no conversion was observed with BDP‐TS, either in acetonitrile‐containing mixtures or under purely aqueous conditions. In the latter case, the reaction resulted in pronounced precipitation of the mixture, despite both BDP‐TS and the peptide being individually soluble under these conditions. The absence of product formation together with the observed precipitation may indicate nonspecific interactions between BDP‐TS and the peptide, likely arising from the amphiphilic character of the labeling agent.

Cell Experiments

We assessed the stability and localization of BDP‐MTS, BDP‐SS‐Bu, BDP‐S‐Bu, and BDP‐S‐CysTAT in living cells under confocal fluorescence microscopy. HeLa cells were treated with a 4 µM solution of the corresponding BODIPY derivative. The results indicated that BDP‐MTS, BDP‐SS‐Bu, and BDP‐S‐Bu were predominantly localized in lipophilic parts of cells (e.g., in internal and plasma membranes, Figure 6a–c, respectively), which corresponds to the typical localization behavior of lipophilic BODIPY dyes.^{[ 31 ]} On the other hand, BDP‐S‐CysTAT (Figure 6d,f) retained the ability of the TAT peptide to penetrate membranes via endocytosis (Figure 6g), as previously described in the literature for its fluorescein‐labeled analog, Flu‐TAT (Figure 6e,f),^{[ 32 ]} highlighting the intracellular stability of our BODIPY conjugate.

Figure 6.

The panel illustrates the intracellular behavior of labeled thiols used in this study. a) BDP‐MTS (4 µM), b) BDP‐SS‐Bu (4 µM), c) BDP‐S‐Bu (4 µM), and d) BDP‐S‐CysTAT (4 µM), in HeLa cells (λ _ex = 514 nm), emission collection at 532–545 nm. The scale bar is 50 µm. e) Image of punctate fluorescence, indicating endocytic uptake for Flu‐TAT, reprinted with permission from.^{[ 32 ]} f) Structures of BDP‐S‐CysTAT (this study) and Flu‐TAT (literature).^{[ 32 ]} g) Schematic representation of Flu‐TAT endocytosis.^{[ 32 ]}

Conclusion

While classical methanethiosulfonate (MTS) labeling yields disulfides prone to reduction, our thiol‐selective labeling approach introduces a novel photochemical step, leveraging the unique photoreactivity of disulfides attached to an arylmethyl chromophore (BODIPY, coumarin). This approach significantly improves the stability of labeled conjugates, yielding stable thioethers via a [1,2]‐sigmatropic rearrangement followed by sulfur extrusion. The resulting fluorescent thioethers are resistant to reduction for at least 48 h in the presence of TCEP, whereas the parent disulfides are fully reduced within one hour under the same conditions. Butanethiol, cysteine, glutathione, Peptidisc, TAT peptide, and BSA labeling showcase the versatility of the method. In cells, the labeled TAT peptide retains its stability and native properties. Our strategy offers a robust and versatile alternative for thiol‐selective modifications, overcoming the instability of conventional MTS labeling products and the limitations of gold‐standard maleimide labeling methods, such as thiazine rearrangements on terminal cysteines and the retro‐Michael reaction.

Supporting Information

The authors have cited additional references within the Supporting Information.^{[ 7} , ¹⁴ , ¹⁹ , ²⁰ , ²⁶ , ²⁷ , ²⁸ , ³³ , ³⁴ , ³⁵ , ³⁶ , ³⁷ , ³⁸ , ³⁹ , ⁴⁰ , ⁴¹ , ⁴² , ⁴³ , ⁴⁴ , ⁴⁵ , ⁴⁶ , ⁴⁷ , ⁴⁸ , ⁴⁹ , ⁵⁰ , ⁵¹ , ⁵² , ⁵³ , ⁵⁴ , ⁵⁵ , ^{56 ]}

Conflict of Interests

The authors declare no conflict of interest.

Supporting information

Supporting Information

Šálková L., Dunlop D., Tütüncü B. B., Knejzlík Z., Slanina T., Angew. Chem. Int. Ed. 2026, 65, e15338. 10.1002/anie.202515338

Data Availability Statement

The data that support the findings of this study are available in the Supporting Information of this article.