Fluorophore-Assisted Click Chemistry through Copper(I) Complexation.

The copper-catalyzed alkyne-azide cycloaddition (CuAAC) is one of the most powerful chemical strategies for selective fluorescent labeling of biomolecules in in vitro or biological systems. In order to accelerate the ligation process and ensure efficient formation of conjugates under diluted conditions, external copper(I) ligands or sophisticated copper(I)-chelating azides are used. This latter strategy, however, increases the bulkiness of the triazole linkage, thus perturbing the biological function or dynamic behavior of the conjugates. In a proof-of-concept study, we investigated the use of an extremely compact fluorophore-based copper(I) chelating azide in order to accelerate the CuAAC with concomitant fluorescence labeling; in our strategy, the fluorophore is able to complex copper(I) species while retaining its photophysical properties. It is believed that this unprecedented approach which was applied for the labeling of a short peptide molecule and the fluorescent labeling of live cells, could be extended to other families of nitrogen-based fluorophores in order to tune both the reaction rate and photophysical characteristics.


Introduction
Copper-catalyzed alkyne-azide cycloaddition (CuAAC), also known as click chemistry [1], has attracted tremendous interest in recent years for the site-specific, in vitro modification of biomolecules, such as proteins, glycans, lipids, or nucleic acids, and for the bioorthogonal fluorescent labeling of cell extracts or living systems [2][3][4]. Prominent applications in this latter area include, among others, the understanding of diverse biological processes [5,6], the development of detection tools [7], and real-time live-cell imaging and therapy [8][9][10][11][12]. In contrast to other successful bioorthogonal chemistry, such as strain promoted alkyne azide cycloaddition (SPAAC) [13] and tetrazine-based inverse electron demand Diels−Alder ligation (IEDDA) [14,15], CuAAC ligation has been widely adopted due to the smallness and inertness of the alkyne and azide handles that can be incorporated into biomolecules by using the genetic code expansion, or the cellular metabolic machinery [16], and the fact that small triazole adducts also impose a minimal perturbation of resulting conjugates.
Reaction of alkynes with azides generally involves the in situ formation of copper(I) catalyst from a copper(II) source (e.g., CuSO 4 ) in the presence of reducing agent, sodium ascorbate. Although coordinating ligands are not strictly required for CuAAC [5], their combined use has shown to significantly accelerate the alkyne-azide cycloaddition, which thus ensures efficient bioconjugation under diluted conditions imposed by biological systems, while preventing both the deactivation of the copper(I) catalyst by biomolecules and copper-mediated oxidative damage. Tris(triazolylmethyl)amine derivatives, such as TBTA or its water-soluble analogues THPTA, and BTTAA, constitute an important family of such accelerating ligands (Figure 1a) [17][18][19]. However, ligands, copper sources, and a fluorescent bioorthogonal handle are generally required in excess amounts relative to the chemical reporter group for the bioconjugate reaction to proceed accordingly with high yields [20]. These ligands are mostly used for cell-surface labeling, or require covalent attachment to cell-penetrating peptides to facilitate their cellular uptake [21]. To circumvent these limitations, another strategy consists of using azides equipped with a chelating moiety which are capable of complexing the copper(I) species and thus accelerate the reaction in the absence of external ligands by both facilitating the formation of the metallacycle intermediate and increasing the electrophilicity of the azide function [22][23][24]. Highly performant azide chelating systems were designed in this context ( Figure 1b) [25,26]. Nevertheless, these chelating azides are constituted of polycyclic ligands with aromatic characters, and when linked to bulky and rigid fluorophores, water-solubilizing groups are required to counterbalance their overall hydrophobicity. As a consequence, subsequent biological, physico-chemical properties of the corresponding conjugates, in particular, for short labeled biomolecules or pharmacophores, may be dramatically altered, which questions the benefits of using small alkyne and azide handles in CuAAC, with respect to other biorthogonal reactions [27].
Following these considerations, in order to perform bioconjugate reactions with alkyne modified biomolecules, we want to report the unprecedented use of an extremely compact, fluorophore-based ligand capable of complexing copper(I) species while retaining its photophysical characteristics useful for bioimaging applications. We previously reported the Kondrat'eva ligation based on a one-pot Diels-Alder/aromatization of 5-alkoxyoazole with maleimide to furnish the corresponding azaphthalimide fluorophore, which has found different applications in chemical biology [28,29]. Interestingly, this small bicyclic azaphthalimide dye displays relatively high excitation and emission wavelengths (ex. 420 nm, em. 520 nm) suitable for live-cell imaging. Herein, we wish to combine both the fluorescent properties of the azaphthalimide dye and its underestimated copper-binding ability to design a readily-available and extremely compact fluorescent copper-chelating azide ( Figure 2).

General Information
All chemicals were used as received from commercial sources without further purification. Solvents, unless otherwise stated, were purchased at reagent grade or HPLC grade and used as received, except tetrahydrofuran, which was freshly distilled over sodium prior to use. PBS (pH 7.4, 0.1 M) and aqueous mobile phases for HPLC were prepared with water purified by means of a MilliQ system (purified to 18.2 MΩ cm). All reactions were monitored by thin layer chromatography (TLC) and/or RP-HPLC. TLC were carried out on Merck DC Kieselgel 60 F-254 aluminum sheets. Visualization of spots was performed under a UV lamp at λ = 254 or 365 nm, and/or staining with a KMnO 4 solution/K 2 CO 3 + 5% NaOH, and developed with heat. Flash column chromatography purifications were performed manually on silica gel (40-63 µM) under pressurized air flow.
High resolution mass spectrometry (HRMS) were obtained by using a Waters Micromass LCT Premier XE ® (Manchester, United Kingdom) equipped with an orthogonal acceleration time-of-flight (oa-TOF) and an electrospray source in positive mode. 1 H, 13 C, and NMR spectra were recorded on Bruker 300 machine operating at ambient temperature.
The solvent resonance was used as the internal standard for 1  Fluorescence spectroscopic studies (emission/excitation spectra and time course for kinetics monitoring) were performed on a Varian Cary Eclipse ® spectrophotometer (Le Plessis-Robinson, France) using a quartz fluorescence cell (Hellma, Jena, Germany, 104F-QS, 10 × 4 mm, lightpath: 10 mm, chamber volume 1.4 mL) and excitation/emission spectra were recorded at 20 • C. UV-Vis absorption spectra were obtained on a Varian Cary 60 UV-Vis ® (Agilent, Les Ulis, France) using a standard cell (10 × 10 mm, chamber volume 3.5 mL) at 20 • C.
Fluorescence quantum yields were measured on a HORIBA Fluorolog 3 spectrophotometer (Longjumeau, France) with a quartz fluorescence cell (Hellma, 104F-QS, lightpath: 10 × 10 mm, chamber volume: 3.5 mL, excitation and emission slit: 2 nm) at 25 • C by a relative method using Lucifer Yellow (Φ F = 21% in Water, 430 nm) [30] as a standard. The following equation was used to determine the relative fluorescence quantum yield: where A is the absorbance (in the range of 0.01-0.1 a.u.), F is the area under the emission curve, n is the refractive index of the solvents (at 25 • C) used in measurements, and the subscripts s and x represent standard and unknown, respectively. The following refractive index values were used: 1.337 for PBS 0.1 M pH 7.4.
Images (1024 × 1024 pixels) were acquired by using an upright fixed-stage Leica TCS SP8 confocal microscope (Leica Microsystems, Nanterre, France) equipped with diode laser at 405 nm to excite the fluorophores and a conventional scanner at 400Hz. Using a 25× objective (NA 0.95, water immersion) fluorescence emission was detected through a hybrid detector (HyD) in photon counting mode with a specific band from 500 to 550 nm. Mosaic image (square of 5 by 5 images at 1024 × 1024 pixels) was performed to obtain a large view, and a merging was realized after acquisition through module LAS X Navigator. For image acquisitions, focusing on cells was performed in bright-field mode before fluorescence acquisition. For the fluorescence quantification, all values are expressed as fluorescence intensity (A.U.) means of at least 100 cells ± sems. Statistical analysis was performed using the GraphPad Prism 4 software (GraphPad Software Inc., San Diego, CA, USA) and a one-way analysis of variance (ANOVA) with a Tukey-Kramer multiple comparisons tests.

Synthesis of the Fluorophore-Based Chelating Azide
The fluorescent chelating agent 1 was prepared from commercial L-aspartic acid diethyl ester hydrochloride (Scheme 1). First, the reported oxazole ester 2 obtained in two steps [33], was reduced into the corresponding alcohol derivative 3 in 46% isolated yield. The hydroxyl group was then converted into mesylate intermediate 4, which was subsequently displaced with azide ion to afford compound 5 in 74% over two steps. Finally, the oxazole 5 underwent a [4+2] cycloaddition/aromatization process with N-methylmaleimide to furnish the corresponding azaphthalimide 1 in 40% isolated yield. Importantly, this compound proved to be relatively soluble in PBS pH 7.4 as the only solvent, up to 4 mM (Scheme 1a,b). Interestingly, the LogS value of azaphthalimide 1 was found to be −1.58, which is in the range of those calculated for water-soluble tris(triazolylmethyl)amine THPTA and BTTAA (Table S8).

Comparative Study with Chelating and Non-Chelating Azide
Azaphthalimide 1's ability to accelerate 1,3-dipolar cycloadditions was compared to a reportedly effective but not fluorescent pyridine-based chelating azide, 2-(2-azidoethyl)pyridine 6, [25] and a non-chelating derivative, 2-phenylethyl azide 7. We chose these three azide-based reagents, since in all of them, the azide function is separated from the aromatic ring by a two carbon-atom alkyl chain length in order to form a 6-membered metallocycle system (for chelating azides 1 and 7 only). The fluorogenic 7-ethynyl coumarin 8 whose fluorescence increases upon reaction with azides [25], was chosen as the click partner in order to monitor the click reactions progress by fluorescence spectroscopy. For compounds 6 and 7, an excitation and emission wavelengths of 320 and 400 nm were selected, corresponding to the absorption and emission of the triazole-substituted coumarin ( Figures S2  and S3). With regard to azaphthalimide azide 1, the coumarin and azaphthalimide scaffold display complementary photophysical properties to constitute a suitable FRET pair with an excitation at the coumarin wavelength (λ ex = 320 nm), and an emission at the azaphthalimide emission wavelength (λ em = 550 nm, Figure S1). For each triazole product, a calibration curve was performed to convert fluorescence intensity signal into product conversion ( Figures S4-S6). Of note, the presence of copper(I) did not significantly modify the fluorescence emission of either triazole products 9-11.
First, the 2-(2-azidoethyl)-pyridine 6 led to the rapid formation of the triazole product (Figure 3a, curve in blue), while on the other hand, the non-chelating azide 7 gave no detectable product, even after 100 min of reaction (curve in orange). Meanwhile, the azaphthalimide enables the formation of 20% triazole. Although this conversion is significantly lower than that of 6, this result showed that the azapthalimide was able to accelerate the reaction, presumably by acting as a bidentate ligand for CuAAC reactions. This chelating ability has been further validated with reactions carried out at 175 µM, and monitored by HPLC (Figure 2b, grey curve). In fact, a plateau at ca. 60% conversion yield was quickly reached, while almost no product was obtained with the non-chelating compound (orange curve). The reaction performed with the azaphthalimide was approximatively 100 times faster than the one achieved with the phenyl scaffold. Importantly, the presence of the copper complex did not impact the absorption and emission wavelengths of either fluorophores ( Figure S1).

Applications
The potency of a small fluorophore-assisted click chemistry for minimal conjugate perturbation was illustrated through the fluorescent labeling of a heptapeptide-based short biomolecule (Scheme 2). The starting peptide was prepared according to standard solid-phase peptide synthesis protocols with the incorporation of the pentynoic acid at the N-terminus (see Materials and Methods). The reaction was monitored by HPLC and showed a clean and complete reaction within 30 min (Figure 4a). The isolated conjugate 13 obtained in a 38% isolated yield after RP-HPLC and lyophilization was subject to photophysical measurements. Absorption and emission spectra recorded either in the presence or absence of copper(I) (generated in situ from CuSO 4 ·5H 2 O and sodium ascorbate), showed no change in positions of peak maxima, which were found at λ abs~4 16 nm and λ em~5 25 nm, respectively (Figure 4b). Only a widening of the absorption peak was observed. Besides, quantum yield for the green fluorescence-labeled peptide was found to be around 8% in the presence and absence of in situ generated copper(I) determined in PBS pH 7.4 at 25 • C. This value is in line with those reported for similar azaphthalimide derivatives [29]. After establishing that small conjugates could be readily labeled in vitro using a heptapeptide as the model, then the labeling of alkyne-modified live cells with azaphthalimide azide 1 was investigated. As a proof-of-concept study, the alkyne reporter group was first chemically incorporated into live cells by using a maleimide derivative 14 (50 µM in PBS for 30 min at 37 • C), a functional group known to react rapidly and covalently with biological thiols through a Michael addition process [35]. Cells were washed twice with PBS to remove unreacted alkyne-maleimide 14 in the supernatant; then, a 50 µM solution of azaphthalimide-based chelating azide 1 in PBS pretreated with a 50 µM solution of in situ generated copper(I) (cocktail solution) was incubated for 15 min with cells at 37 • C, and then rinsed with PBS. Subsequent confocal analysis microscopy revealed strong fluorescence labeling of cells at 500-550 nm ( Figure 5)-highly statistically significant compared to control conditions (fluorescence poorly detectable in cells with only PBS incubation (ctrl) and "cocktail alone," when cells that were not incubated with the alkyne reporter group, thereby demonstrating successful CuAAC-mediated ligation of azaphthalimide azide and alkyne incorporated into cells, data not shown). Images revealed some discernible vesicles that could be identified as lipid droplets, in which the lipophilic alkyne reporter presumably accumulates ( Figure 5A,B and Figure S7) [36].

Conclusions
Herein, we disclosed an unprecedented "all-in-one" fluorophore-based chelating azide with an exceptionally low molecular weight (260 Da), which proved useful for the fluorescent labeling of small biomolecules and biological systems. In this approach, the azaphthalimide played the role of both the copper ligand and fluorophore in order to accelerate the CuAAC reaction with concomitant installation of a fluorescent tag. Importantly, the fluorescence properties of the native fluorophore were conserved upon its complexation to cooper(I). This process advantageously does not require an external ligand, a fluorophore, or water solubilizing group, which complicates the synthesis and increases the risk of impacting the mobility of azides and properties of conjugated molecules. This strategy could be extended to other families of nitrogen-containing fluorophores in order to further improve the ligation rates or to red-shift the fluorescence emission [37,38].