Photoswitchable Spiropyran Dyads for Biological Imaging

The synthesis of a small-molecule dyad consisting of a far-red-emitting silicon rhodamine dye that is covalently linked to a photochromic spironaphthothiopyran unit, which serves as a photoswitchable quencher, is reported. This system can be switched reversibly between the fluorescent and nonfluorescent states using visible light at wavelengths of 405 and 630 nm, respectively, and it works effectively in aqueous solution. Live-cell imaging demonstrates that this dyad has several desirable features, including excellent membrane permeability, fast and reversible modulation of fluorescence by visible light, and good contrast between the bright and dark states.


(b) Determination of the Power Intensities
The power intensities of the Prizmatix mic-LED-405 and mic-LED-630 light sources were measured with a Coherent® power meter at various points of indicated intensity percentage. These intensity measurements were carried out at the end of the liquid light guide, immediately before the cuvette. The power of the mic-LED-405 was also confirmed by chemical actinometry, as described below.
Ferric oxalate actinometry: 3,4 A solution of ferric oxalate (0.0060 M) was prepared by dissolving K 3 [Fe(C 2 O 4 ) 3 ]·3H 2 O (300 mg, 0.60 mmol) in aqueous sulfuric acid (100 mL of 0.05 M H 2 SO 4 ). 2.00 mL of this solution was placed in a quartz cuvette, and while stirring, irradiated with mic-LED-405 at the power of the indicated output percentage for 10 s, while an identical control sample was kept in the dark. A sample of the irradiated ferric oxalate solution (0.2 mL) and a control sample (0.2 mL) were transferred to vials containing a mixture of 0.1% phenanthroline solution (0.9 mL, prepared by dissolving 13.56 g of CH 3 CO 2 Na and 0.100 g of phenanthroline in 100 mL of 0.5 M H 2 SO 4 ) and H 2 O (0.9 mL). Samples were kept in the dark for 1 h to allow the complexation to occur and absorption spectra were acquired in the 250-850 nm range. The LED power intensities and irradiation lengths were chosen to ensure that the photoconversion of ferric oxalate does not exceed 5% and that the absorbance of the Fe(phen) 3 2+ complex is within the range 0.5-1.0.
Upon exposure to light, K 3 [Fe(C 2 O 4 ) 3 ] is converted to Fe 2+ , and the quantum yield !R of formation of Fe 2+ is well-known for a range of wavelengths (222-500 nm; !R = 1.14 at 405 nm 5 ). The photon flux ƒ irradiating the solution can be calculated using Equation 1, where !R is the quantum yield of the ferric oxalate conversion at the irradiation wavelength, t (s) is the irradiation time, (1−10 − ! ) is the fraction of light absorbed by the ferric oxalate solution while A is the absorbance of 0.006 M solution at the irradiation wavelength. Formation of the colored Fe(phen) 3 2+ complex (ε 510 = 11,100 M -1 cm -1 ) allows quantification of Fe 2+ generated during the irradiation by measuring the absorption of irradiated and non-irradiated samples (Equation 2).
The calculated value of the photon flux was expressed in E·s -1 and was converted to mW using Equation 3.
A linear correlation between the power intensity (mW) and indicated output percentage (%) was derived as in Figure S6. The power intensity obtained from the power meter matches remarkably well with that from the actinometry measurement.
As the output beam was fixed in close contact with the cuvette, the irradiation area was estimated to be the surface area of the end of the liquid light guide, A = 0.071 cm 2 . This area was used to calculate the applied power density (W/cm 2 ).

(c) Acquisition of the Photostationary State (PSS) Absorption Spectra
A methanol solution of compound 6 (13.8 μM, 2.0 mL) was irradiated with 80 mW of mic-LED-405 for 5.5 s, and immediately afterwards, the spectrum was recorded while the sample was given pulses of the same LED intensity with 100 ms irradiation and 200 ms interval lengths. The reconstructed PSS spectrum from Matlab was used to calculate the FRET efficiency of compound 8. 6 The following equations were used:

(d) Calculation of the FRET efficiency (E)
where r is the distance between the donor and acceptor, R 0 is the Förster radius (distance at which the energy transfer efficiency is 50%), κ 2 is the dipole orientation factor (assumed to be 2/3 given free rotation between the donor and acceptor), ! D is the fluorescence quantum yield of the donor in the absence of the acceptor (0.39), 1 n is the refractive index of the solvent (1.33 for water), J(λ) is the spectral overlap integral based on the data of Figure S8, f D is the normalized donor emission spectrum, ε A is the acceptor molar absorption coefficient, and λ is the wavelength. The spectra shown in Figure S8 give J(λ) = 2.97 × 10 14 M -1 cm -1 nm 4 , which corresponds to a Förster radius of R 0 = 36 Å.
A conformational search of the MC form of dye 8 using DFT calculations at the PM6 level of theory with Gaussian 09 revealed two predominant conformers with distances r of 19.0 Å and 9.3 (see Figure  S9). The equations above indicate the FRET efficiency is E > 97% in both conformations.
where m is the slope of the linear fit from the time-dependent change in absorbance at 645 nm during irradiation at 405 nm (m = 0.00277 s -1 , Figure S7F), V is the sample volume (2.0 mL), N A hc are standard constants, P is the power intensity (0.0455 W), λ is the excitation wavelength (405 nm), A is the absorbance at excitation wavelength (0.134), ε prod is the molar absorption coefficient of the MC form at 645 nm, d is the path length (1.00 cm). It is difficult to determine ε prod because it is difficult to determine the composition of the PSS due to the speed of the MC→SP back reaction. Values of the quantum yield φ o were calculated from equation (5) using four different approaches to estimate ε prod , as listed below. All these approaches are crude approximations, but they provide a rough estimate of φ o : (i) In this method, we assume that the MC form of 6 has negligible absorption at 285 nm. The change in absorbance at 285 nm is attributed to conversion of the SP to MC form, and indicates 13% of the SP form was photoswitched to the MC form in the PSS ( Figure 1). Therefore, ε prod = A 645ofPSS /(0.13 [SP] 0 ) = 1.5 × 10 4 M -1 cm -1 . This approach gives φ o = 0.88%. In reality, the MC form of 6 must have some absorption at 285 nm, implying that φ o > 0.88%.
(ii) In this method, we assume that the PSS consists entirely of the MC form, i.e. photoisomerization is completed from the SP to MC form.
(iii) In this method, we assume that the ε prod of the MC form for compound 6 is the same as that of compound 5, ε = 3.4 × 10 4 M -1 cm -1 . The molar absorption coefficient of compound 5 in its MC form can easily be measured because this isomer predominates in CH 3 OH, as shown by NMR spectroscopy and supported by the fact that the spectrum shows no change before and after irradiation at 405 nm. This approach gives φ o ≈ 0.40%.
(iv) In this method, we attempted to determine the composition of the PSS by 1 H NMR spectroscopy. Compound 6 (0.53 mL of a 18 mM solution in d 6 -DMSO) was transferred to a 5-mm NMR tube. The sample was locked and shimmed, and a single-scan 1 H NMR spectrum was recorded at 25 °C. The sample was ejected, and the solution was evenly irradiated for 2 min with a laser pointer (405 nm, 1 mW). The tube was quickly inserted into the probe, the sample was locked, and a single-scan spectrum was recorded at 25 °C. The total time between the end of irradiation and acquisition of the single-scan spectrum was 25 s. The resulting spectrum revealed that after this time only 1.5% of 6 exists in the MC form (estimated by integration of two different pairs of signals and two independent measurements). A kinetic trace was obtained by measuring the decrease in absorbance at 645 nm as a function of time after irradiation. For this experiment, compound 6 (2 mL of a 0.48 mM solution in DMSO) was transferred to a 1-cm cuvette. The sample was irradiated for 12 s with a laser pointer (405 nm, 1 mW), and the decrease in absorbance was recorded as a function of time at 25 °C. After 25 s, the absorbance at 645 nm was 0.57. Knowing that at 25 s after irradiation the fraction of MC is 1.5%, the concentration of MC in the cuvette is approximately 7.3 µM, and therefore the molar absorption coefficient of the MC form is approximately ε prod = 7.9 × 10 4 M -1 cm -1 , which gives φ o ≈ 0.17%.

(f) Testing the Kinetics of Thermal Ring Closure in Cuvettes
Ring-closing kinetic traces of compound 5 were measured using solutions (12.9 µM, 2.0 mL) at different temperatures ( Figure S10). To induce ring opening, the samples were irradiated for 5.5 s using 80 mW of mic-LED-405 from the side of a squared quartz cuvette (10 mm) immediately before measuring the decay of absorbance at 570 nm as a function of time. Figure S11 shows a similar set of ring-closing kinetic traces for compound 6 (13.8 µM, 2.0 mL) by monitoring the absorbance change at 645 nm at different temperatures. Activation enthalpy (Δ! ‡ ) and activation entropy (∆! ‡ ) were calculated based on the Eyring equation, where k is the rate constant (s -1 ), T is the absolute temperature (K), R is the universal gas constant (8.3145 J mol -1 K -1 ), k B is the Boltzmann constant (1.381 × 10 -23 J K -1 ), h is the Plank constant (6.626 × 10 -34 J s). The free energy of activation (ΔG ‡ ) was calculated from ΔG ‡ = ΔH ‡ -TΔS ‡ .
Based on the Eyring equation, the Gibbs energies of activation (ΔG ‡ ) are almost the same for switches 5 and 6 at 25 °C (75 and 78 kJ mol -1 respectively), which are at the same magnitude of previously reported analogues. 31 There is a small enthalpy difference (ΔH ‡ = 77.8 and 78.0 kJ mol -1 ), whereas an entropic gain (ΔS ‡ = 9.2 J mol -1 K -1 ) was calculated for compound 5, and an entropic loss (ΔS ‡ = -0.5 J mol -1 K -1 ) for compound 6. The opposite sign for the entropy values is probably attributed to the difference steric hindrance associated with the two chalcogen atoms. With a bigger sulfur atom next to the spiro-carbon center, the transition from the MC to SP state presumably involves generation of restraints and loss of disorder.
Dependence of ring-closing kinetics of compound 6 on the solvent viscosity ( Figure S12) was measured using solutions (12 µM, 2.0 mL) in a mixtures of ethylene glycol and CH 3 OH to vary the viscosity of the medium without significant variation in the polarity. 8 To induce ring opening, the samples were irradiated for 30 s with a Tesoar hand-held laser pointer (405 nm, 1 mW) from the top of a squared quartz cuvette (10 mm). During the measurements, the solutions were stirred at the constant rate.

(g) Testing Accelerated Ring Closure Under Irradiation with Red Light
A methanol solution of compound 6 (13.8 μM, 2.0 mL) was placed in the sample holder of a UV-vis spectrometer and stirred at 25 °C. Immediately after irradiation with 80 mW of mic-LED-405 for 5.5 s, a kinetic trace of absorption intensity at 645 nm was recorded over 120 s to give the thermal ring closure kinetics ( Figure S14A). The photochemically accelerated ring closing was performed under the same conditions, except that the sample was given pulses of 147 mW mic-LED-630 with 200 ms irradiation length and 500 ms interval over the 120 s recording time ( Figure S14B).

(h) Testing Photochemical Fatigue Resistance in Cuvettes
A methanol solution of compound 6 (13.8 µM, 2.0 mL) was placed in the sample holder of a UV-vis spectrometer and stirred at 25 °C. The samples were irradiated using mic-LED-405 (80 mW) for 5.5 s at the beginning of each cycle. Immediately after irradiation, a kinetic trace of absorption intensity at 645 nm was recorded over 90 s. This procedure was repeated using the same sample for 30 cycles. For each cycle, absorbance at t = 0 and 90 s is normalized to the maximum absorption intensity that happens at t = 0 s of the first cycle. The normalized intensities at t = 0 s of each cycle is referred to as S7 '+' ('on') data and t = 90 s as '-' ('off') data (Figures 3 and S17). This experiment was also carried out under argon using a distilled and degassed methanol solution in an airtight cuvette.

(i) Cell culture and staining procedures
Human dermal lymphatic endothelial cells (HDLEC) and Chinese hamster ovarian (CHO) cells were cultured in Dulbecco's modified Eagle medium (DMEM), supplemented with 10% fetal calf serum (FCS). The cells were grown to confluence at 37 °C with 5% CO 2 in the presence of a poly-D-lysinecoated microscope coverslip 24 h before imaging. For staining, the dye was added to the growth medium as DMSO stock solution to reach a final concentration of 5 µM and the cells were incubated for 15 min. The coverslip was taken out of the growth medium, rinsed with dye-free medium, and used for imaging at 37 °C.

(j) Optical Microscopy
Imaging was carried out using a Zeiss 780 confocal microscope, operated with Zen software. The excitation light sources were a diode laser (405 nm, 30 mW) and a helium-neon laser (633 nm, 5 mW). The optimization of excitation light intensities used and pixel dwell times are given in Figures S19-S23, and the fatigue resistance in Figures S24-25.

1-Mercapto-4-nitro-2-naphthaldehyde (3)
A solution of S-(2-formyl-4-nitronaphthalen-1-yl) dimethylcarbamothioate (90 mg, 0.30 mmol) in CH 3 OH (8 mL) was degassed in an ultrasonic bath under positive pressure of nitrogen for 15 min. An aqueous solution of NaOH (1.86 mL, 1.0 M) was degassed for 10 min and added to the first solution. The reaction was stirred at 20 °C for 2 h. The red solution was cooled in an ice bath, and the mixture was acidified with aqueous hydrochloric acid (3 mL, 1.0 M). The suspension was diluted with saturated aqueous NaCl (20 mL) and extracted with EtOAc. The combined organic fractions were dried over MgSO 4 , filtered, and evaporated to yield a yellow solid, which was used immediately in the next step without further characterization. LRMS (ESI) m/z: calcd for [C 11 H 6 NO 3 S]ˉ: 232.0, found 232.0. TLC (98:2 CH 2 Cl 2 /CH 3 OH) indicated a single spot.