Core remodeling leads to long wavelength fluoro-coumarins†

Low molecular weight, uncharged far-red and NIR dyes would be enabling for a range of imaging applications. Rational redesign of the coumarin scaffold leads to Fluoro-Coumarins (FCs), the lowest molecular weight dyes with emission maxima beyond 700, 800, and 900 nm. FCs display large Stokes shifts and high environmental sensitivity, with a 40-fold increase in emission intensity in hydrophobic solvents. Untargeted variants exhibit selective lipid droplet and nuclear staining in live cells. Furthermore, sulfo-lipid derivatization enables active targeting to the plasma membrane. Overall, these studies report a promising platform for the development of biocompatible, context-responsive imaging agents.

a separate flask, a solution of Selectfluor (1.38 g, 3.90 mmol) was dissolved in DMF and added to the first mixture dropwise over 20 min at -78 ºC under argon. The reaction mixture was allowed to warm to 0 ºC and stirred for 1 h. The progress of the reaction was monitored using LC/MS and TLC (every aliquot collected for analysis was quenched with a sat. NH4Cl). Upon completion (1 h), the reaction mixture was quenched using sat. NH4Cl (50 mL) and extracted using EtOAc (200 mL). The organic fraction was washed with H2O (2 × 50 mL) and brine (2 × 30 mL) then dried (Na2SO4) and concentrated in vacuo. The crude mixture was purified using normal-phase column chromatography (120 g silica; 0®15% EtOAc: Hexane) to result in FC2 (311 mg, 38%) as a thick viscous dark red liquid.
Step 2: To a solution of the previous intermediate (115 mg, 0.45 mmol) in MeCN (5 mL), Selectfluor (158 mg, 0.45 mmol) was added in small batches. The reaction was allowed to stir at r.t. for 1 h and the progress of the reaction was monitored using LC/MS. Upon completion (1 h), the reaction was quenched by the addition of sat. NH4Cl (20 mL) and the organics were extracted using EtOAc (50 mL). The organic fraction was washed with H2O (2 × 25 mL) and brine (25 mL) then dried (Na2SO4) and concentrated in vacuo. The crude mixture was purified using normalphase column chromatography (
The reaction mixture was allowed to warm to r.t. and left to stir for 24 h. The progress of the reaction was monitored by LC/MS and TLC (every aliquot collected for analysis was quenched with a sat. NH4Cl and washed with sat. NaHCO3). 24 h were found to be result in an optimum conversion to the desired product and upon completion, the reaction was quenched with aq. sat.
NH4Cl (25 mL) and transferred into a conical flask and cooled to 0 o C. To this cold biphasic mixture, aq. sat. NaHCO3 (50 mL) was slowly added and the organics were extracted into EtOAc Scheme S1: Synthesis of FC9-SL

Step-2 (Conjugating the sulfo lipid 8) c :
The boc-deprotected analog ( c NMR spectroscopy was not useful due complexity by the amide rotamers and C-F splitting.

S3.1 Hammett Analysis d
We hypothesized that introducing electron withdrawing functionalities would lead to red-shifted FC dyes. Surprisingly, a modest red-shift was observed in the case of FC3 (22 nm) in both hydrophilic and hydrophobic solvents. FC4 also showed a small red shift (~32 nm) in its spectra which could be attributed to the relatively weak inductive property of the bromo substituent.
While, expectedly, the introduction of a -Ph group (FC5) did not result in any significant changes compared to FC2, FC6 with -CN group at position-4 resulted in a significantly red-shifted absorption (~70 nm) and emission (~70 nm) spectra in all the solvents investigated. Interestingly, our Hammett analysis ( Figure S2) shows that this observation was partially correlatable with Hammett constants (except FC3) demonstrating the importance of electron-withdrawing groups at position-4. 5 The anomaly found in the case of FC3 suggests that the modifications at position-3 is not crucial for red-shifting the FC dyes.

S3.2 Discussion of optical properties of FC dyes
In the best interest of the reader, we have limited our discussion on the optical properties of FC dyes in the main manuscript to only toluene (mimicking hydrophobic cellular compartments) and 10% FBS (biologically relevant media). Below, we summarize other notable observations that were made.
3. In the case of emission, positive solvatochromism was seen for all FC dyes going from toluene ® MeCN ( Figure S6). However, negative solvatochromism between MeCN ® 10% FBS was recorded in all cases (except FC1). This trend has been reported to be significant in the case of Nile red 8 and, less dramatically, the case of COUPY dyes. 9

S31
Section S4: Photostability studies: The photostability of the FC dyes was investigated in toluene and MeCN upon irradiation with an appropriate LED source set to an intensity of 10 mW/cm 2 . The FC dyes were diluted to 25 µM, from a principal stock solution of 10 mM (in DMSO), in a HPLC vial. In general, all the FC dyes (except FC1 and FC7) were photostable in toluene and MeCN (~ 30 -50% degradation over 60 min; Figure   S14). Dinitrile-substituted dyes (FC8 -FC10) were highly photostable upto 24 h of irradiation. Particularly, FC10 upon irradiation with a 660 nm LED at an intensity of 50 mW/cm 2 for 8 h was found to be > 80% intact.
Dark stability of these dyes was recorded, in parallel, to find that all the FC dyes were stable for the duration of the study in the absence of irradiation.
b During the geometry optimization of the S2 state (π → π*, starting from S0 optimized geometry and using small maximum step size of 0.05 Bohr) its energy dropped below the n → π* excited singlet state in the very first step, thus becoming the first excited singlet state S1. Subsequent geometry optimization of this S1 π → π * state led to it transforming to n → π * character beginning at step 33, and thus quickly acquiring an essentially zero oscillator strength. Therefore, the S1 * → S0 * emission energy given is for step 32 which is the last point that true π → π* character is retained and where the oscillator strength of 0.48 is still high. See Figure S18 for more details. a These Mulliken populations are penalized to be negative since each R-group introduces a node to the LUMO that passes between the ring-bound R-group atom (not in parentheses) and neighboring carbon atoms, i.e., in these systems the ring-bound R-group atom has a negative bond order with neighboring carbon atoms in the LUMO. For the other molecules, the ring-bound R-group atom has a positive bond order with neighboring carbon atoms in the LUMO. See Table S5 for further details.   Figure S18: Absolute energies (A) and oscillator strengths (B) of R=SiMe2 excited states S1 and S2 during geometry optimization of the π → π * singlet excited state which starts out as S2 (at S0 minimum geometry) but on the first geometry step becomes S1. Subsequent geometry optimization of this S1 π → π * state led to it transforming to n → π * character beginning at step 33, and thus quickly acquiring an essentially zero oscillator strength. As such, emission data are reported for geometry step 32. Note that the geometry optimization had a small maximum step size of 0.05 Bohr.                             -10  0  10  20  30  40  50  60  70  80  90  100  110  120  130  140  150  160  170  180  190  200  210 f1 ( -10  0  10  20  30  40  50  60  70  80  90  100  110  120  130  140  150  160  170  180  190  200  210 f1 (ppm)   -10  0  10  20  30  40  50  60  70  80  90  100  110  120  130  140  150  160  170  180  190  200  210 f1 (