Organic dyes that emit fluorescence in the near infrared (NIR) window (650-900 nm) are needed for biological imaging as this wavelength region has relatively deep tissue penetration, minimal tissue auto fluorescence, and lower Rayleigh scattering of the light [1][2][3]. At present, many NIR fluorescent probes are based on cyanine dyes because of their large molar absorption coefficients, moderate to high fluorescence quantum yields, and broad wavelength tunability [4]. Although the practical value and utility of cyanine dyes in biological studies is without doubt, they can exhibit some non-optimal molecular properties such as poor chemical and photostability, limited water solubility, propensity to self-quench upon dye aggregation, and difficult purifications due to the ionic molecular structures [5]. Thus, an ongoing task for dye chemists is to prepare new classes of NIR fluorescent dyes for biological imaging applications.
Squaraine dyes are a well-known family of NIR fluorescent dyes with intense and narrow absorption bands, high fluorescence quantum yields, and excellent photostabilities [6]. Because of these attractive optical properties, squaraines have been investigated over the years for many potential applications [7][8][9][10][11]. In the specific case of biological imaging, the development of squaraine dyes as NIR fluorescent probes has been limited by technical challenges such as poor chemical stability in biological media and a propensity for dye self-aggregation [12]. Our research group has discovered that both problems can be abrogated by encapsulation of the dye inside a protective macrocycle [13]. The equilibrium in Scheme 1 shows complexation of an bis(aminothiophene)squaraine dye by a tetralactam macrocycle that has two anthracene sidewalls. The threaded complex is stabilized by a synergistic combination of aromatic stacking and hydrogen bonding interactions between the two supramolecular components, and complexation produces a red-shift of the squaraine absorption/emission maxima [14]. Moreover, the surrounding macrocycle protects the encapsulated squaraine dye from nucleophilic attack, attenuates the quenching effects caused by dye self-aggregation, and enhances the squaraine quantum yield in water [14][15][16][17]. The favorable fluorescence properties of these threaded squaraine macrocycle complexes have been exploited for different applications in biological imaging [18,19], diagnostics [20] and liposome surface functionalization [21]. Most squaraine dyes and squaraine/macrocycle complexes (including the system in Scheme 1) have fluorescence emission bands in a narrow region between 650-700 nm, which makes them useful substitutes for the common cyanine dye, Cy-5, in many types of microscopy and in vivo imaging techniques [3,18]. Looking to the future, there is little doubt that the value of squaraine dyes for biological imaging would be enhanced if additional dye structures could be produced with absorption/emission wavelengths that cover a broader window and extend up to 850 nm. Even more utility would be gained if the squaraine structures allowed macrocycle threading to produce high stability complexes. Within the literature on squaraine dyes is a small and scattered collection of structures with squaraine absorption/emission bands that are close to 800 nm [22][23][24][25][26]. None of these were considered suitable for our needs because the structures were either too hard to prepare as water-soluble molecules or they contained sterically large terminal groups that would prevent macrocycle threading. We decided to develop a new set of squaraine chromophores with incrementally extended absorption/emission wavelengths. We knew that squaraine structures containing aminophenyl or aminothiophene units would be useful for operation in the region of 600-700 nm [27]. The real challenge was to develop a stable squaraine chromophore with the extended π-conjugation needed to reach 800 nm and also allow macrocycle threading. After some preliminary experimentation, we discovered that squaraine dyes could be prepared with attached aminothienothiophene units, and that these dyes exhibited the desired optical and supramolecular properties. Herein, in Scheme 2 we showcase this discovery by describing three organic-soluble dyes (S1-S3) that were easily prepared from commercially available materials in a few steps, and water-soluble analogues (S1PEG-S3PEG) that were obtained by appending long polyethylene glycol (PEG) chains. We find that the dyes form supramolecular complexes with the tetralactam macrocycles M1 or M2 to produce a suite of 12 fluorescent compounds that span the absorption window of 600-800 nm, making them a very attractive set of dyes for a wide range of future biological imaging applications.
The unsymmetrical squaraines, S1 and S2, were synthesized by similar procedures using N,N-dibenzylaminophenyl semisquaraine 2 as a common building block (Schemes 3 and 4). In the first case, condensation of 2 with 2-aminothiophene derivative 1 furnished organic-soluble squaraine S1 in 77% yield, which was subsequently converted into water-soluble squaraine S1PEG by conducting a copper-catalyzed alkyne-azide cycloaddition reaction with 3. In the second case, the appropriate 2-aminothienothiophene 6 was produced in three steps [28] and then condensed with 2 to give organic-soluble squaraine S2 in 58% yield. A subsequent cycloaddition reaction with 3 provided the water-soluble squaraine S2PEG.
The symmetrical squaraine S3 was prepared (Scheme 5) in 92% yield by condensing squaric acid with two molar equivalents of 2-aminothienothiophene 8. The reaction was complete after 2 hours and the air-stable product appeared as a golden brown precipitate that could be isolated by simple filtration. A subsequent cycloaddition reaction with two molar equivalents of 3 provided the water-soluble squaraine S3PEG.
As a series, squaraines S1, S2 and S3 have increasing π-conjugation, and they are soluble in organic solvents such as CHCl 3 . A key molecular design feature is the small N-alkyl group (either N-methyl or N-ethyl) at one or both ends of the central chromophore, which ensures that macrocycle threading is facile [17]. Threading studies in CHCl 3 were performed by mixing separate samples of S1, S2 and S3 with one molar equivalent of organic-soluble macrocycle M1. In each case there was an immediate 20-35 nm red-shift in the squaraine absorption and emission maxima wavelength (Figure 1) which is diagnostic of squaraine encapsulation by M1. In the case of S3, the macrocycle threading process was characterized using 1 H-NMR. In Figure 2 is a comparison of partial 1 H-NMR spectra for separate solutions of S3, M1 and M1 ⊃ S3 in CDCl 3 . Squaraine S3 can adopt two low-energy conformations that differ in the relative orientation of the thienothiophene units, which can have a cis or a trans relationship. The 1 H-NMR spectrum for free S3 shows single peaks for protons 1 and 2 suggesting fast exchange between cis and trans conformations. A sample containing a 1:1 mixture of S3 and M1 instantly formed M1 ⊃ S3 in a quantitative yield. The chemical shifts for M1 ⊃ S3 indicated encapsulation of the squaraine inside the macrocycle as illustrated by the molecular model in Figure 3. Notably, there are large downfield changes in chemical shift for the macrocycle NH residues and protons B, and large upfield changes in the chemical shift for macrocycle protons D and E. Moreover, the spectral patterns for the macrocycle unambiguously indicate that the encapsulated squaraine predominantly adopts a C 2 -symmetric trans conformation, with a small fraction of the encapsulated squaraine (<10%) in a cis conformation. For example, the major signal for the macrocycle B protons is a singlet at 9.33 ppm, which can only arise if the encapsulated squaraine is trans. In contrast, the minor signal for the macrocycle B protons is split into two singlets around 9.22 ppm, which implies that the encapsulated squaraine is cis. The peaks for squaraine protons 1 and 2 in encapsulated S3 are sharp due to hindered rotation of the squaraine single bonds when S3 is inside the macrocycle. Moreover, the large upfield change in chemical shift for squaraine proton 1 and the relatively negligible change in chemical shift for proton 2 are both consistent with the anisotropic NMR shielding zones predicted by the molecular model in Figure 3.
Preliminary threading studies in water were performed by mixing separate solutions of water-soluble squaraines S1PEG, S2PEG and S3PEG with one molar equivalent of water-soluble macrocycle M2. Shown in Figure 4 The association constants for threading of M2 by S1PEG, S2PEG or S3PEG in water were measured by fluorescence titration (Figure 5, Figures S1 andS2) and the values of K a are listed in Table 1. The symmetrical thienothiophene-based squaraine S3PEG has essentially the same nanomolar affinity for M2 as analogous but shorter symmetrical thiophene-based squaraines [14]. Although the extended chromophore in S3PEG has a larger hydrophobic surface area, it does not translate into higher affinity for M2, presumably because there is no increase in the amount of hydrophobic surface area that is buried by complexation (see the molecular model in Figure 3). The K a values for unsymmetrical S1PEG and S2PEG are 7.8 and 5.3 times higher than S3PEG because these dyes have terminal N,N-dibenzyl groups which are known to stabilize the threaded complexes by stacking with peripheral surfaces of the surrounding macrocycle [29]. For each water-soluble squaraine, the second order rate constant for macrocycle threading, k on , was measured using a stopped flow device that monitored the increase in fluorescence after mixing the squaraine and macrocycle. As shown in Figure 5, threading of M2 by S3PEG was complete in less than a minute when the concentration of each binding partner was 250 nM. Compared to S3PEG, the threading of M2 by S1PEG and S2PEG (Figures S1 andS2) was 25 and 100 times faster, respectively, which was expected based on the known dependence of k on on the steric size of N-alkyl groups at the end of the squaraine chromophore [17]. Squaraines S1PEG and S2PEG have N-methyl groups whereas squaraine S3PEG has larger N-ethyl groups which were needed to increase the dye solubility.
In Table 2 is a summary of the photophysical properties of the six different dye systems in water. The photostability of each squaraine and its macrocycle complex in water was monitored by continuous NIR light excitation over 15 h and in each case there was no or little loss in sample fluorescence intensity (Figures S3-S5). The chemical stabilities of the dyes and their macrocycle complexes were tested by monitoring the fluorescence intensity over time in the presence of strongly nucleophilic Na 2 S in water (Figures S6-S8). All three dyes were quickly bleached by the Na 2 S, whereas the corresponding macrocycle complexes were quite stable, which is compelling evidence that the encapsulated squaraines were inside the protective macrocycle M2. This combination of high resistance to chemical and photobleaching exhibited by the complexes is quite notable for such highly π-extended chromophores [3]. Quantum yields of S1PEG and M2 ⊃ S1PEG were measured relative to bis(N,N-dimethylaminophenyl)squaraine, quantum yields of S2PEG and M2 ⊃ S2PEG were measured relative to cryptocyanine, and quantum yields of S3PEG and M2 ⊃ S3PEG were measured relative to ICG [30].
Presently, the only NIR dye that is approved for use in humans is indocyanine green (ICG). Although ICG is employed extensively, it is also known to have nonoptimal performance properties and there is an active community effort to find replacement dyes with very similar absorption/emission wavelengths. Shown in Figure 6 is a comparison of the absorption and emission maxima for ICG, S3PEG and M2 ⊃ S3PEG in water. The absorption spectra for ICG and S3PEG both exhibited blue-shifted aggregation bands whereas the maxima band for M2 ⊃ S3PEG was relatively sharp and narrow (Figure 6). To compare the fluorescence quantum yields, the concentrations of each sample were adjusted to produce the same absorbance value (0.08) at 750 nm, and the fluorescent emission spectrum was collected with excitation at 750 nm. Integration of the emission bands in Figure 6 indicated that the fluorescence quantum yields for S3PEG and M2 ⊃ S3PEG were 1.3 and 2 times higher than ICG (Table S1). A unique photophysical feature of these threaded squaraine macrocycle complexes is the option of squaraine excitation by internal energy transfer [14]. That is, 375 nm excitation of the anthracene sidewalls of the surrounding macrocycle (M2) is followed by efficient internal energy transfer within the complex and excitation of the encapsulated squaraine. This capability raises the possibility of multiplex excitation of multiple dyes in the same sample. Shown in Figure 7 is the two-color emission spectrum that was observed when 1:1 mixture of M2 ⊃ S1PEG and M2 ⊃ S3PEG in water was excited at 375 nm. The two emission bands nicely match the popular Cy5 and Cy7 emission channels that are commonly used for biological imaging. Single wavelength excitation and multiplex detection is a valuable tool in orthogonal fluorescence imaging, and is likely to be very helpful in technologies such as fluorescence-guided surgery and cancer tissue histology, where there is a desire to simultaneously identify the location of different molecular probes that label distinct biological targets [31,32].
1 H and 13 C-NMR spectra were recorded on Bruker AVANCE III HD 400 (Billerica, MA, USA)
and 500 MHz spectrometer (Billerica, MA, USA) Chemical shift was presented in ppm and referenced by residual solvent peak. Mass spectrometry (MS) was either performed using a Bruker microTOF II spectrometer (Billerica, MA, USA) with electron spray ionization (ESI) or Bruker Autoflex III (Billerica, MA, USA) with matrix-assisted laser desorption/ionization (MALDI). Commercially available solvents and chemicals were used without further purification unless otherwise stated. Water was de-ionized and micro filtered. Flash column chromatography was performed using Biotage flash column chromatography purification system with SNAP Ultra cartridges (Charlotte, NC, USA). All cartridges used silica gel as stationary phase unless otherwise stated.
Compound 1 [17], 2 [33] and 3 [34] were prepared using previously reported literature methods.
S1: Compound 1 (80 mg, 0.410 mmol) and compound 2 (150 mg, 0.406 mmol) were dissolved in a mixture of 1-butanol (15 mL) and benzene (45 mL) and the reaction mixture was heated to reflux for 2 h with Dean-Stark distillation. The solvent was removed and the residue was purified by column chromatography using 0-10% MeOH/CHCl 3 to produce pure S1 as a blue solid (172 mg, 77% yield).
Compound S1 (12 mg, 0.021 mmol), azido-mPEG 45 3 (40 mg, 0.020 mmol), triethylamine (2 drops) and TBTACu(I)Br (2 mg) were dissolved in CHCl 3 (5 mL) and the reaction mixture was stirred at room temperature for 12 h. After removing the solvent, the residue was purified by column chromatography using 0-10% MeOH/CHCl 3 (containing 0.2% NH 4 OH) to obtain pure S1PEG as a blue solid (48 mg, 92% yield). 13 mmol) was dissolved in acetic acid (10 mL). N-bromosuccinimide (1.27 g, 7.13 mmol) was added and the reaction mixture was stirred at room temperature for 2 h. Solvent was removed and the residue was dissolved in diethyl ether (80 mL) and washed with NaOH (1 M, 100 mL × 3), H 2 O (100 mL × 3) and brine (100 mL). The resulting solution was dried over Na 2 SO 4 , and the solvent was removed by rotary evaporation to obtain pure 4 as a light yellow liquid (1.3 g, 81% yield). (15 mL). TBA HSO 4 (200 mg) was added to the mixture as a phase transfer reagent and the resulting solution was stirred at room temperature for 5 h. Toluene was removed by rotary evaporation and H 2 O (100 mL) was added to the residue. The resulting solution was extracted with diethyl ether (100 mL × 3). The organic phase was combined and dried over Na 2 SO 4 , the solvent was removed and the residue was purified by column chromatography (neutral Al 2 O 3 as stationary phase) using 5-20% EtOAc/hexane to obtain pure 6 as a yellow liquid (650 mg, 97% yield). N-ethyl-N-(2-(prop-2-yn-1-yloxy)ethyl)thieno [3,2-b]thiophen-2-amine (8): Propargyl bromide (900 µL, 8.0 mmol) was added to a mixture of compound 7 (660 mg, 2.9 mmol) in toluene (15 mL) and NaOH solution (15 mL, 50% w%). TBA HSO 4 (200 mg) as phase transfer reagent was added to the mixture and the resulting solution was stirred at room temperature for 5 h. Toluene was removed by rotary evaporation and H 2 O (100 mL) was added to the residue. The resulting solution was extracted with diethyl ether (100 mL × 3). Organic phase was combined and dried over Na 2 SO 4 , the solvent was removed and the residue was purified by column chromatography (neutral Al 2 O 3 as stationary phase) using 5-20% EtOAc/hexane to obtain pure 8 as a yellow liquid (520 mg, 84% yield). S3: Compound 8 (150 mg, 0.57 mmol), 3,4-dihydroxycyclobut-3-ene-1,2-dione (32 mg, 0.28 mmol) was dissolved in a mixture of 1-butanol (15 mL) and benzene (45 mL). The reaction was heated to reflux with Dean-Stark distillation for 2 h. After cooling to room temperature, the solvent was removed and the residue was purified by recrystallization in MeOH to obtain pure S3 as a brown solid (158 mg, 92% yield).
Compound S3 (7.0 mg, 0.012 mmol), azido-mPEG 45 3 (50 mg, 0.024 mmol), triethylamine (2 drops) and TBTACu(I)Br (2 mg) were mixed in CHCl 3 (5 mL) and the reaction mixture was sonicated at 40 • C for 4 h. After removing the solvent, the residue was purified by column chromatography (silica gel as stationary phase) using 0-10% MeOH/CHCl 3 to obtain the crude product. The crude product was further purified by a send column chromatography (neutral Al 2 O 3 as stationary phase) using 0-5% MeOH/CHCl 3 to obtain pure S3PEG as a brown solid (24 mg, 42% yield).