Knightletin is a novel, high quantum-yield, substituted coumarin in the group of fluorescent aminocoumarins. The molecule reveals the influence of a hydroxyl group on the aminocoumarin structure through its absorbance-fluorescence properties and interaction with several solvents. The substitution on the benzene ring of the coumarin affects electronic states, the molecular dipole moment and hydrogen bonding between vicinal groups as well as hydrogen bonding with solvents. This report describes the synthesis of knightletin, its absorbance-fluorescence characteristics in three solvents of varying polarity, and insights on its conformation from density functional theory modeling. The quantum yield of 0.81 in methanol is remarkable. The modeling indicates a ground state geometry with a pyramidal NH2 group interacting with the OH in an NH-donor hydrogen bond and predicts spectroscopic characteristics consistent with the experimental observations.
Research Article
Spectral properties of a high quantum yield, novel aminocoumarin with a vicinal hydroxyl group
https://doi.org/10.21203/rs.3.rs-1608217/v1
This work is licensed under a CC BY 4.0 License
Version 1
posted
You are reading this latest preprint version
Substituted coumarin
quantum yield
Stokes shift
fluorescence lifetime
Aminocoumarins are well known to have high fluorescence quantum yields and perform well in a variety of substituted forms as commercial laser dyes [1]. This class of dyes has in common the amino group in position 7 according to the conventional coumarin numbering scheme (Scheme 1). This substituent can be unsubstituted or substituted, including cyclic structures that prevent the amine moiety from rotating. We describe for the first time a 7-aminocoumarin with a vicinal hydroxyl group: 6-hydroxy-7-amino-4-methylcoumarin, which we have given the name knightletin. This novel compound is like aesculetin, 6,7-dihydroxycoumarin, having vicinal hydrogen bonding groups on the benzene ring. However, the novel compound has a much higher quantum yield than aesculetin. Aesculetin is notable for its bioactivity and has been the subject of a recent study from our lab characterizing it as an extraordinarily strong photoacid capable of forming a complex with borate ion [2]. Besides its similarities to aesculetin, knightletin can be compared to well-studied 7-aminocoumarins that lack the hydroxyl group, such as Coumarin 120 (7-amino-4-methylcoumarin, C120) [3–5]. Interestingly, aminocoumarins are also an established class of antibiotics; in these cases, the substituted amino group is placed at position 3, with novobiocin as an example [6]. This study focuses on the photophysical properties rather than any potential bioactivity of knightletin.
The synthesis is serendipitous because the original goal of this synthetic work was to place hydroxyl groups on the coumarin structure in positions 5 and 6 so that we could study the photoacid and complex-forming behavior of vicinal dihydroxycoumarins. The target compound offers an intriguing comparison with aesculetin and with daphnetin (7,8-dihydroxycoumarin), both of which are commercially available. Knightletin––named after the mascot of Calvin University, where it was synthesized––also possesses vicinal groups that can form an intramolecular hydrogen bond, which makes it an interesting comparison to aesculetin and C120. Here we describe the synthesis of the novel aminocoumarin and characterization of its absorbance-fluorescence properties in three solvents. Future studies will examine its potential photoacid behavior and whether any species complexation occurs, such as is the case for aesculetin with borate.
The starting material for the synthesis, 4-methylgrevillone (4-methyl-6-hydroxycoumarin, compound 1a in the synthesis section), is from TCI America (Portland, OR, USA). Steady-state absorption and emission spectra were collected on a UV-Vis spectrophotometer (Hitachi U3900H) and spectrofluorometer (Horiba FluoroMax-4).
Quantum Yield Determination
A quantitative determination of the quantum yield was carried out relative to the fluorescence standard quinine sulfate (QS, 0.1 N H2SO4, 22°C) using 350-nm excitation for all solutions. The absolute fluorescence quantum yield of knightletin in MeCN (acetonitrile) is \({\varPhi }_{A}={\varPhi }_{Q}({I}_{A}/{I}_{Q})({A}_{Q}/{A}_{A}){\left({n}_{S}/{n}_{W}\right)}^{2}\), where \({\varPhi }_{Q}\) is 0.577 for QS [7],\({I}_{A}/{I}_{Q}\) is the ratio of the wavelength-integrated fluorescence intensity of knightletin to quinine sulfate, \({A}_{Q}/{A}_{A}\) is the ratio of the absorbance of QS to knightletin at 350 nm, and \({n}_{S}/{n}_{W}\) is the ratio of refractive indices of the knightletin solution solvent, in this case MeCN, and water. This analysis is performed with five absorbances over a range from 0.01 to 0.11 and yields the expected linear increase in the integrated intensity with increasing absorbance. We estimate uncertainty to be about 5% using this approach. In addition, the relative quantum yield for knightletin solutions in EtOAc (ethyl acetate) and MeOH (methanol) were compared to MeCN by determining the integrated emission intensity normalized by absorbance at the excitation wavelength. Emission spectra were recorded at the same solution concentration, excitation wavelength, excitation intensity, and emission sensitivity.
Time-resolved Fluorescence
The equipment used for time-resolved fluorescence spectroscopy has recently been described in a paper from our lab [8]. This system, used in partnership with the Blanchard group in the Department of Chemistry at Michigan State University, utilizes a picosecond laser light source. For our experiment, we used 5-ps pulses of 350-nm light and time-correlated single photon counting (TCSPC) detection. Time resolved fluorescence decays were collected for unpurged ~ 100 µM solutions of knightletin in EtOAc, MeCN, and MeOH. The effect of purging with N2(g) was tested for the knightletin/MeCN solution, purging for 10 minutes, and collecting the decay for 1.5 minutes immediately after purging; the purged decay yielded a lifetime that was about 9% longer. A study on oxygen solubility in organic solvents indicates MeCN has the highest O2 solubility of the three selected solvents with MeOH and EtOAc being 10 and 20% lower solubility, respectively [9]. For reference, a fluorescence decay for a 100 µM solution of coumarin 1 (7-diethylamino-4-methylcoumarin) in EtOH (ethanol) collected on the same day yielded a lifetime of 3.07 ± 0.01 ns (unpurged), which agrees well with the literature report of 3.1 ns (purged) [10]. Unpurged lifetimes are reported below since purging does not appear to significantly impact the lifetime for these coumarins.
Computational Methods
All quantum chemistry modeling using density functional theory (DFT) was performed with the Gaussian 16 program package [11] and the WebMO Pro interface [12]. Full geometry optimizations of knightletin (neutral and anion) in the gas phase were performed for the ground state using the B3LYP functional and cc-pVTZ basis set. For modeling in the three solvents of this study, the SMD solvation model [13] was used as implemented in Gaussian 16 to represent a diffuse dielectric medium of the solvent. Care was taken to ensure each ground state structure was at a global minimum energy, which for the neutral structure involves comparing the intramolecular hydrogen bond with either OH or NH acting as donor. To model the excited state, the time-dependent DFT/B3LYP/cc-pVTZ method was used including the SMD solvent approach where appropriate; the predicted absorbance wavelength comes from the un-optimized and optimized excited state TD-DFT computation to model the vertical excitation and the fluorescent emission, respectively.
Synthesis Of Substituted Coumarins
The original goal of this work was the preparation of 5,6-dihydroxy-4-methylcoumarin (1b, see Scheme 2) for fluorescence studies on substituted coumarins with vicinal hydroxyl groups on the benzene ring akin to the more thoroughly studied 6,7-dihydroxycoumarin (aesculetin). The synthesis was accomplished using the sequence of reactions, 1a → 1c → 1d → 1b, as described in the literature by Kaufman et al. [14] Although there are reports in the literature describing the synthesis of 1c from 1a using HNO3/H2SO4, [14, 15] no mention is made regarding its isomer 1e, which is obtained in low yield in the nitration. Reduction of the nitro group in 1c leads to 1d in very good yield using dithionite, S2O4− 2. The last step in the sequence leading to 1b was accomplished by treating 1d with Fe+ 3 in aq. HCl, which results in a very low yield. Although NMR data are available in the literature for 1c and 1e, [16] 1f was an unknown compound at the outset of this work.
5,6-Dihydroxy-4-methylcoumarin (1b). To a stirred solution of 0.634 g (3.28 mmol) of 1a in 7.5 mL of 10% HCl was added 9.40 mL of 10% aq. FeCl3 (3.48 mmol) over 5 min which produced a black paste. This paste was combined with 50 mL of 3:1 CH2Cl2 /EtOAc and the mixture was stirred with heating to dissolve 1b. The CH2Cl2 /EtOAc filtrate was chromatographed on a silica gel column and eluted with the same solvent. Removal of the solvent in the eluent gave 71 mg (11%) of 1b: mp 244–245°C (dec) (lit. mp 247–249°C) [14]; 1H NMR (500 MHz, d6-acetone) δ 8.86 (br s, 1H), 8.20 (br s, 1H), 7.10 (d, J = 8.75 Hz, 1H), 6.66 (d, J = 8.75 Hz, 1H), 6.06 (s, 1H), 2.65 (s, 3H); 13C NMR (500 MHz, d6-acetone) δ 159.9, 153.7, 148.2, 144.0, 140.3, 117.7, 113.7, 109.1, 106.4, 22.9; IR (ATR) 3131, 1644, 1609, 1596, 1564, 1473, 1364, 1307, 1269, 1197, 1060, 1036, 995, 933, 799, 637, 599, 542 cm− 1; HRMS (ESI) m/z [M + H]+ calculated for C10H9O4: 193.0501 amu; found 193.0489 amu.
6-Hydroxy-5-nitro-4-methylcoumarin (1c) and 6-hydroxy-7-nitro-4-methylcoumarin (1e). A solution of 2.92 g of 70% HNO3 (0.0328 mmol) was added over 40 min to a solution of 5.00 g (0.0285 mmol) of 1a in 48 mL of conc. H2SO4 cooled in an ice bath. The temperature was kept below 5°C during the addition. After standing in an ice-bath for 2 h, the reaction mixture was poured over 0.5 kg of ice. The yellow solid was collected by vacuum filtration and dried giving 6.28 of crude product, which produced an NMR spectrum that showed a mixture of 1c and 1e and a small amount of an unidentified product. The reaction mixture was introduced onto a silica gel column and eluted with 8:1 CH2Cl2-EtOAc to give 0.560 g (9%) of 1e as a yellow solid: mp 191°C (dec) (lit mp 185°C)[16]; 1H NMR (500 MHz, d6-DMSO) δ 11.1 (s, 1H), 7.92 (s, 1H), 7.34 (s, 1H), 6.55 (s, 1H), 2.38 (s, 3H); 13C NMR (500 MHz, d6-DMSO) δ 159.6, 151.8, 147.8, 145.0, 139.0, 125.0, 118.0, 114.4, 113.0, 18.4; IR (ATR) cm− 1 3278, 1710, 1572, 1483, 1445, 1221, 1173, 930, 883, 827, 618, 582 cm− 1.
Elution with 4:1 CH2Cl2-EtOAc gave 4.62 g (74%) of 1c as yellow crystals: mp 240–245°C (dec), (lit. mp 220–222°C)[REF RB1]; 13C NMR (500 MHz, d6-acetone) δ 11.2 (s, 1H), 7.46 (d, J = 9.15 Hz, 1H), 7.31 (d, J = 9.15 Hz, 1H), 6.52 (s, 1H), 2.22 (s, 3H); 13C NMR (500 MHz, d6-acetone) δ 158.8, 148.4, 146.2, 146.1, 135.4, 121.2, 120.3, 119.1, 112.1, 17.89; IR (ATR) 3295, 1685, 1577, 1437, 1266, 1228, 1208, 1172, 936, 857, 822, 637, 622, 574 cm− 1; HRMS (ESI) m/z [M + H]+ calculated for C10H8NO5: 222.0403 amu; found 222.0402 amu.
6-Hydroxy-5-amino-4-methylcoumarin (1d). To an ice-cold solution of 2.50 g (11.3 mmol) of 1c in 31 mL of conc. NH3 was added with stirring an ice-cold solution of 38.5 g (0.225 mmol) of 82% Na2S2O4 in 150 mL of water. The reaction mixture rapidly changed color from red to yellow and, after stirring at 25°C for 3 h, a yellow solid was collected by vac. filt. and dried to give 1.62 g (75%) of 1d: mp 257–259°C (dec) (lit. mp 253–256°C)[14]; 1H NMR (500 MHz, d6-DMSO) δ 9.56 (br s, 1H), 6.90 (d, J = 9.57 Hz, 1H), 6.46 (d, J = 9.57 Hz, 1H), 5.99 (s, 1H), 5.02 (br s, 2H), 2.62 (s, 3H); 13C NMR (500 MHz, d6-DMSO) δ 165.2, 159.8, 152.9, 146.0, 140.3, 121.7, 117.3, 112.9,108.8, 28.5; IR (ATR) 3515, 3407, 3205, 1664, 1593, 1478, 1381, 1367, 1266, 1220, 1029, 798, 679 cm− 1.
6-Hydroxy-7-amino-4-methylcoumarin (1f) (knightletin). To an ice-cold solution of 110 mg (0.496 mmol) of 1e in 1.5 mL of conc. NH3 was added with stirring an ice-cold solution of 1.69 g (7.96 mmol) of 82% Na2S2O4 in 7.5 mL of water. The reaction mixture rapidly changed color from red to yellow and, after stirring at 25°C for 3 h, the yellow solid that formed was collected by vac. filt. and dried to give 84 mg (88%) of 1f: mp > 286°C (dec); 1H NMR (500 MHz, d6-DMSO) δ 6.90 (s, 1H), 6.57 (s, 1H), 5.95 (s, 1H), 4.86 (s, 3H), 2.34 (s, 3H); 13C NMR (500 MHz, d6-DMSO) δ 163.7, 154.8, 149.6, 142.8, 141.5, 109.4, 107.4, 106.8, 99.5, 17.3; IR (ATR) 3490, 3368, 3068, 1657, 1611, 1552, 1453, 1404, 1367, 1236, 1214, 1188, 938, 857, 828, 768, 563 cm− 1; HRMS (ESI) m/z [M + H]+ calculated for C10H10NO3: 192.0661 amu; found 192.0653 amu.
Absorbance and emission spectra, Stokes shift
Figure 1 shows the absorbance and emission spectra of knightletin in three solvents that vary in polarity and ability to hydrogen bond, EtOAc, MeCN and MeOH. These spectra show behavior that is consistent with the character of many other fluorescent coumarins. Table 1 gathers the fluorescence characteristics of knightletin and includes the Lippert solvent polarity parameter (orientation polarizability, ∆f) [3, 7]. The ∆f values show that EtOAc is significantly less polar than the other two solvents, while MeCN and MeOH are distinguished by the latter’s ability to form hydrogen bonds. The fact that MeOH invokes a much larger red shift than MeCN suggests that hydrogen bonding influences the solvated ground electronic state. The significant absorbance red-shift due to MeOH has been observed in other 7-aminocoumarins [3, 17].
Table 1 Photophysical properties of knightletin in selected solvents, Lippert-Mataga bulk solvent polarity parameter, absorbance maximum wavelength, absorbance bandwidth, molar attenuation coefficient, fluorescence maximum wavelength, fluorescence emission bandwidth, Stokes shift, fluorescence quantum yield, observed fluorescence lifetime, and natural fluorescence lifetime.
The Stokes shift, which depends on both the absorbance and emission maximum wavelength, is also largest for MeOH indicating that specific hydrogen bonding interactions also play a key role in the electronic excited state. This is consistent with similar observations for many polar fluorophores [7]. Furthermore, the Stokes shift is well correlated with the solvent polarity parameter ∆f like that observed for C120 in solutions of moderate to high polarity [3]. The slope of the Stokes shift versus ∆f is like that of C120 indicating a similar sensitivity to solvent polarity. It is interesting to note that the knightletin Stokes shift of 4600 cm–1 in MeOH is considerably smaller than that for C120 (5800 cm–1) and aesculetin (6000 cm–1). The effect of the OH substitution in position 6 involves the electronic states due to its role in the molecular orbitals but will also influence the interaction with the solvent by affecting the molecular dipole moment and providing intermolecular hydrogen bonding with protic solvents.
Table 2 shows spectral properties of coumarins closely related to knightletin. The absorbance maximum wavelengths of knightletin are just outside the red end of the range of absorbance wavelengths of the other coumarins in Table 2. C120 resembles knightletin only lacking the OH in position 6; like the relationship between umbelliferone (7-hydroxycoumarin) and aesculetin (6,7-dihydroxycoumarin). Scopoletin (6-methoxy-7-hydroxycoumarin) represents a similar coumarin that is also di-substituted on the benzene ring in the same positions as knightletin. Comparing C120 to knightletin in the three solvents, the absorbance of the latter shifts 15-20 nm to the red; the absorbance red shift from umbelliferone to aesculetin in MeOH is 25 nm. The Stokes shifts for knightletin are comparable to the range of values for other coumarins in Table 2 lying just outside the low end of that range.
Table 2 Reference photophysical characteristics of similar coumarins
a Ref [2]
b Ref [1]
c Ref [8]
Quantum yield
The quantum yield for knightletin in MeCN, determined relative to quinine sulfate, is 0.74 ± 0.04, making it about 15% more efficient than the commercial laser dye, C120, in MeCN. Figure S1 shows the knightletin/MeCN experimental decays and linear fits for the relative quantum yield determination. The emission spectra in Fig. 1 visually illustrate that MeOH has the largest quantum yield of the three solvents (at just over 80%) since the spectra are normalized to absorbance at the excitation wavelength and scaled relative to the MeOH emission. Our observation that the knightletin quantum yield in both MeCN and EtOAc is about 10% lower than MeOH is the reverse of what is reported for C120 in the same set of solvents where the two aprotic solvents have ~20% larger quantum yield than for MeOH [3]. This may further point to an enhanced role for hydrogen bonding in the novel compound due to the additional hydroxy group. We also note that the knightletin quantum yield in MeOH is about four times larger than aesculetin in MeOH.
Time resolved fluorescence
The fluorescence lifetimes in Table 1 are from time resolved fluorescence decays of knightletin in the three solvents that are well represented by single exponential functions – see Figure S2 for fit details. The average fluorescence lifetime of around 4 ns is in good agreement with the range of 3 to 6 ns found in highly fluorescent coumarins. For example, 5 ns is the fluorescence lifetime of the anion in water of both scopoletin [8, 18] and umbelliferone [19]. The effect of different solvents on is consistent with the values in Table 2 by Pal et al. for C120 [3] with the least polar EtOAc having the lowest t and MeOH yielding the largest lifetime. Table 2 also shows that another study reports a C120/MeCN lifetime about 10% larger than our values for knightletin/MeCN, showing that the lifetimes are similar. The observed knightletin fluorescence lifetime and quantum yield combine to give a consistent natural lifetime, 
of 5 ns.
Computational modeling
Modeling using a straightforward, DFT approach reveals key structural features of the novel knightletin molecule. The planar symmetry of the core coumarin structure leaves only the question of how the two benzene ring substituents interact. Hydroxyl group substitutions on the benzene are normally oriented along the plane, as is found in dihydroxy coumarins, such as aesculetin [2]. Orientation of the pyramidal amine group is more complex, but we can expect the vicinal hydroxyl and amine groups will interact forming a hydrogen bond (HB) with either N or O as the acceptor atom. Free of any coupling into delocalized molecular orbitals, we expect the OH donor HB (OH···N) to be stronger than the NH donor HB (NH···O). The report by David et al., giving the average energy of these specific hydrogen bonding interactions, indicates that OH···N is more than 3 times stronger than NH···O [20]. For comparison, a non-aromatic, vicinal amino and hydroxyl intramolecular interaction was modeled in 2-aminoethanol showing the gas-phase conformer with the OH donor HB is 8 kJ/mol lower energy than the corresponding conformer with NH donor HB using the same modeling level of theory applied to knightletin below – Figure S3 is a visualization of the 2-aminoethanol model.
Conformers
Describing conformers of knightletin requires a notation scheme where + or – indicates the orientation of the hydroxyl group in the molecular plane, pointing in the direction of decreasing numbers in the coumarin numbering scheme (Scheme 1), which we designate –, or towards increasing numbers, designated +. The amine group orientation, as influenced by the expected HB formation, follows suit and is given the same sign as the OH. Thus, the conformer shown in Scheme 1 is the ‘– –’ conformer with the NH acting as donor to the HB, and the ‘+ +’ conformer has the opposite orientation with the OH acting as the HB donor. The +/– designation for the pyramidal NH2 group can in some cases be ambiguous since the two bonds can have two kinds of symmetric orientation with respect to the coumarin plane, above and below the plane or both bonds on the same side of the plane. However, when the amine group is influenced by the vicinal OH group, the +/– notation is unambiguous to indicate the arrangement of the intramolecular hydrogen bond. To fully characterize the geometry of the substituent orientation, we define three dihedral angles involving NH or OH described in Scheme 3. Furthermore, the knightletin molecular structure has a classical Lewis-structure resonance represented in Scheme 4 in which the amine group is either pyramidal, structure I, or planar, II; modeling reveals structures that are between these limits where the C–N bond is between a single and double bond. As such, the C–N bond length is a proxy for the degree to which the amino group is coupled into the frontier molecular orbitals. Polar solvents tend to stabilize the charge transfer from amine to carbonyl, represented by the resonance structure on the right, shortening the C–N bond.
Our modeling reveals the KL– – conformer in the gas-phase is 4.8 kJ/mol lower energy than the KL+ + conformer indicating the NH donor configuration is more stable than the OH donor, the reverse of our expectation based simply on average HB strength. Substituent coupling into molecular orbitals must influence the orientation of the intramolecular HB.
Ground and first electronic state geometry
Table 3 collects selected geometric parameters for conformers of knightletin in the gas phase and in the three solvents. The hydrogen bond is represented by the hydrogen to acceptor atom distance. In the ground state of the – – conformer, the NH2 group is pyramidal and has both NH bonds oriented on the same side of the coumarin plane as shown by the dihedral angles of around 20 degrees for both NH bonds. The N atom is slightly out of the coumarin plane – the dihedral N-C7-C6-C5 is 2.5° below the coumarin plane with H atoms on the other side of the plane. The amine group is twisted slightly, by less than 2°, towards the O atom possibly due to the presence of the HB. The d-OH dihedral is slightly out of plane by ~1° above the plane. Figure S4 is a visualization of the ground state knightletin – – conformer model illustrating the out-of-plane geometry. Our modeled NH2 configuration in MeCN agrees well with the modeling of C120 [5] at a similar level of theory that shows a pyramidal NH2 with corresponding dihedrals of 17.4° and –17.8° in MeCN.
Table 3 Geometric parameters and TD-DFT spectral results for conformers of knightletin at the B3LYP/cc-pVTZ level of theory and, where indicated the solvent is modeled using the SMD method; the dipole moment, dihedral angles (defined in Scheme 3), hydrogen bond length (H atom to acceptor atom distance), and NC bond length are from the optimized geometry, the wavelength and intensity are from TD-DFT results (see text); ground state is shaded rows, the + + conformer data are in italics
Our modeled ground state configuration (knightletin – –) can be compared to 2-aminophenol (2AP), which has the same vicinal substituents coupled by the resonance of the benzene ring. The recent work of Capello et al. gives both experimental and theoretical results for jet-cooled gas-phase 2AP [21]. They report IR-UV hole burning action spectra showing IR features corresponding to NH symmetric and antisymmetric stretches and an OH stretch of 3669 cm–1 clearly indicating the OH is not involved in an intramolecular HB, a so-called ‘free’ OH stretch. This OH stretch is red-shifted from the OH stretch of MeOH(g) of 3681 cm–1 [22], showing that the context of being bound to the benzene ring and vicinal to NH2 does influence the OH vibrational frequency even when ‘free.’ The Capello data supports the 2AP conformer that corresponds to knightletin – –. Their modeling, which is at a higher level of theory than our approach, indicates the two 2AP conformers are nearly isoenergetic (favoring the free OH conformer when zero-point energy is considered) agreeing well with our own modeling of 2AP predicting the free OH conformer as more stable by less than 1 kJ/mol. And yet, the 2AP experimental data clearly points to favoring the conformer with the un-bound OH, rather than a gas-phase mixture of the two conformers. Interestingly, the Capello account of the observed 2AP conformer describes the NH···O HB as nonexistent, which raises the question to what extent the study of knightletin will reveal the impact of the NH···O HB in the knightletin – – conformer. Further modeling work will help in that regard, for example to study rotational barriers. We note the IR-action spectroscopic study by Carrascosa et al. on the cryogenically cooled aesculetin-H+ cation [23] shows that it is possible to observe gas-phase IR action spectra of substituents on coumarins similar to knightletin.
In our modeling of the knightletin excited state geometry in the gas phase, the NH2 group is planar for the optimized – – conformer. Figure 2 shows the gas-phase frontier orbitals revealing a charge transfer of HOMO electron density located on the benzene ring shifting to the pyrone ring in the LUMO, which is consistent with favoring planar geometry in the excited state. The HOMO to LUMO transition accounts for 94% of the molecular orbitals contributing to the calculated first electronic excitation. The observation that the dipole moment increases upon excitation to the LUMO also fits with a charge transfer interpretation. The planar NH2 geometry is fully realized in the gas phase (dihedral angles equal zero) and this planar geometry is approached in solution as the solvent polarity increases. Also concordant, the C–N bond length becomes shorter with the transition to the electronic excited state and increasing solvent polarity.
It is interesting to note that the higher energy ground state knightletin + + conformer exhibits a potential double HB because the OH orientation puts the H atom 2.342 Å from the N acceptor, while the H atom in NHb is oriented towards the O atom and is 2.608 Å from the O acceptor. Figure S4 is a visualization of the ground state knightletin + + conformer model. It is also worth noting that the optimized excited state geometry of the + + conformer is lower in energy than the optimized excited state – – conformer by 12 kJ/mol, thus the modeling suggests the potential for excited state dynamics involving reversing the hydrogen bond upon excitation of the – – conformer. The supplemental information contains the gas phase XYZ coordinates for the optimized ground state and first electronic excited state of both knightletin conformers.
Calculated spectroscopic properties
Table 3 also shows the results from applying TD-DFT methods to predict absorbance and fluorescence characteristics of knightletin. In general, the model underestimates the absorbance and fluorescence maximum wavelengths in the solvents by an average of 18 and 34 nm, respectively, but the shift to longer wavelengths with increasing polarity matches experiment. The calculated Stokes shift is low by about 10% compared to experiment, but again the observed increase in Stokes shift with increasing solvent polarity is reproduced in the model. Zhao argues for C120 modeling that some of the spectral mismatches in protic solvents are due to lack of explicit hydrogen bonding in the model [5], a claim that is worth further investigation for knightletin.
Other modeled molecular properties
For knightletin, the anion formed by removal of the OH proton is lower energy compared to the anion with the NH proton removed, predicting that deprotonation will occur at position 6. This is consistent with our preliminary estimate of the pKa for knightletin described in the next section on future study. Calculating the pKa of di-substituted coumarins is the subject of a current study in our lab.
The presence of the intramolecular hydrogen bond may influence the ability of the NH2 group to rotate, which in turn may influence the fluorescence quantum yield. Flexibility of the amino group in aminocoumarin laser dyes is tied directly to the so-called TICT state (twisted intramolecular charge transfer) associated with compromised fluorescence quantum yield [10]. Efforts to inhibit amino group rotation by addition of bulkier ethyl groups or inclusion of cyclic molecular structures are shown to improve the quantum yield [10].
The Zhao study [5] emphasizes the need for including explicit water molecules in modeling of the absorbance spectra. Knightletin represents another important molecular example that now includes an intramolecular HB for testing the ability of models to describe spectroscopic characteristics of coumarins.
Future study
A key question is how knightletin behaves as a weak acid and perhaps as a photoacid like aesculetin, which has been shown to be one of the strongest photoacids of the natural coumarins [2]. Preliminary studies point to knightletin having a pKa ~9, which is like the pKa of grevillone (6-hydroxycoumarin) [24], where the only ionizable proton is in position 6. Studying knightletin in a broader spectrum of solvents may reveal more molecular details and physical properties, although we note that knightletin does not readily dissolve in cyclohexane. Studies will include water and water/alcohol mixtures to observe solvent and pH effects on absorbance and fluorescence properties. Further elucidation of the role of hydrogen bonding in knightletin is a key goal.
Knightletin is a novel coumarin, which exhibits intramolecular hydrogen bonding between the vicinal amino and hydroxyl groups on the benzene ring. It has a high quantum yield that exceeds C120 suggesting the presence of OH modestly improves the quantum yield. The compound also provides an interesting comparison to aesculetin, recently shown to be a very strong photoacid, pointing to further investigation of the acid and anionic character of knightletin.
Acknowledgements
The authors are grateful for the following support: We thank Dr. Gary Blanchard, Michigan State University, for his generous time, collaboration and providing access to his TRF system. The MSU Mass Spectrometry and Metabolomics Core provided LC-high resolution mass spec data.
Funding – National Science Foundation grant 1956223, Beckman Scholars Program (LK), NSF-funded Calvin Computer cluster (MRI-grant 1726260), Calvin Research Fellowship program (MM), Luke and Pauline Schaap Summer Research Fellowship (GD), the Thedford P. Dirkse Summer Science Research Fellowship (IJ), Calvin Honors Program (ED)
Conflicts of interest/Competing interests – the authors declare no conflicts of interest or competing interests
Authors' contributions – LK edited the draft for general write up and performed experiments and modeling, ED and GD performed DFT modeling, IJ performed the experiments, LH performed the experiments and assisted with synthesis, RB performed the synthesis, MM wrote the manuscript in consultation with the other co-authors.
Data availability – All data generated or analyzed during this study are included in this published article and its supplementary information
Ethics approval/declarations – not applicable
Consent to participate – not applicable
Consent to publish – not applicable
Code availability – not applicable
- Brackmann U (2000) Lambdachrome(R) Laser Dyes, Third edition. Lambda Physik AG, Goettingen · Germany
- Knoor ALH, Du Laney GR, Jonker IB et al (2022) Aesculetin Exhibits Strong Fluorescent Photoacid Character. J Fluoresc 32:307–318. https://doi.org/10.1007/s10895-021-02842-w
- Pal H, Nad S, Kumbhakar M (2003) Photophysical properties of coumarin-120: Unusual behavior in nonpolar solvents. J Chem Phys 119:443–452. https://doi.org/10.1063/1.1578057
- Moog RS, Kim DD, Oberle JJ, Ostrowski SG (2004) Solvent Effects on Electronic Transitions of Highly Dipolar Dyes: A Comparison of Three Approaches. J Phys Chem A 108:9294–9301. https://doi.org/10.1021/jp0486088
- Zhao W, Pan L, Bian W, Wang J (2008) Influence of Solvent Polarity and Hydrogen Bonding on the Electronic Transition of Coumarin 120: A TDDFT Study. ChemPhysChem 9:1593–1602. https://doi.org/10.1002/cphc.200800131
- Vanden Broeck A, McEwen AG, Chebaro Y et al (2019) Structural Basis for DNA Gyrase Interaction with Coumermycin A1. J Med Chem 62:4225–4231. https://doi.org/10.1021/acs.jmedchem.8b01928
- Lakowicz JR (2006) Principles of Fluorescence Spectroscopy. Springer US. 3rd edn.https://doi.org/10.1007/978-0-387-46312-4
- Pham HT, Yoo J, VandenBerg M, Muyskens MA (2020) Fluorescence of Scopoletin Including its Photoacidity and Large Stokes Shift. J Fluoresc 30:71–80. https://doi.org/10.1007/s10895-019-02471-4
- Quaranta M, Murkovic M, Klimant I (2013) A new method to measure oxygen solubility in organic solvents through optical oxygen sensing. Analyst 138:6243–6245. https://doi.org/10.1039/c3an36782g
- Jones G, Jackson WR, Choi CY, Bergmark WR (1985) Solvent effects on emission yield and lifetime for coumarin laser dyes. Requirements for a rotatory decay mechanism. J Phys Chem 89:294–300. https://doi.org/10.1021/j100248a024
- Frisch M, Trucks G, Schlegel H et al (2016) Gaussian 16, Revision B.01. Gaussian, Inc., Wallingford, CT
- Schmidt J, Polik W, WebMO Pro WebMO, LLC, Holland, MI, USA. Available from http://www.webmo.net
- Marenich AV, Cramer CJ, Truhlar DG (2009) Universal Solvation Model Based on Solute Electron Density and on a Continuum Model of the Solvent Defined by the Bulk Dielectric Constant and Atomic Surface Tensions. J Phys Chem B 113:6378–6396. https://doi.org/10.1021/jp810292n
- Kaufman KD, McBride DW, Eaton DC (1965) Synthetic Furocoumarins. VII. Oxazolocoumarins from 6-Hydroxy-4-methylcoumarin. J Org Chem 30:4344–4346. https://doi.org/10.1021/jo01023a524
- Mewada GS, Shah NM (1956) Die Nitrierung des 6-Hydroxy-4-methyl-cumarins und seines Methyläthers. Chem Ber 89:2209–2211. https://doi.org/10.1002/cber.19560890924
- Ganguly NC, Datta M, De P, Chakravarty R (2003) Studies on Regioselectivity of Nitration of Coumarins with Cerium(IV) Ammonium Nitrate: Solid-State Nitration of 6-Hydroxy-Coumarins on Montmorillonite K-10 Clay Support Under Microwave Irradiation. Synth Commun 33:647–659. https://doi.org/10.1081/SCC-120015821
- Rechthaler K, Köhler G (1994) Excited state properties and deactivation pathways of 7-aminocoumarins. Chem Phys 189:99–116. https://doi.org/10.1016/0301-0104(94)80010-3
- Pina J, de Castro CS, Delgado-Pinar E, Sérgio Seixas de Melo J (2019) Characterization of 4-methylesculetin and of its mono- and di-methoxylated derivatives in water and organic solvents in its ground, singlet and triplet excited states. J Mol Liq 278:616–626. https://doi.org/10.1016/j.molliq.2019.01.083
- Simkovitch R, Huppert D (2015) Photoprotolytic Processes of Umbelliferone and Proposed Function in Resistance to Fungal Infection. J Phys Chem B 119:14683–14696. https://doi.org/10.1021/acs.jpcb.5b08439
- David V, Grinberg N, Moldoveanu SC (2018) Long-Range Molecular Interactions Involved in the Retention Mechanisms of Liquid Chromatography, Chap. 3. In: Advances in Chromatography. pp 73–110
- Capello MC, Broquier M, Ishiuchi S-I et al (2014) Fast Nonradiative Decay in o-Aminophenol. J Phys Chem A 118:2056–2062. https://doi.org/10.1021/jp411457v
- Shimanouchi T (1972) Tables of Molecular Vibrational Frequencies, Consolidated Volume 1. National Bureau of Standards
- Carrascosa E, Pellegrinelli RP, Rizzo TR, Muyskens MA (2020) Cryogenic Infrared Action Spectroscopy Fingerprints the Hydrogen Bonding Network in Gas-Phase Coumarin Cations. J Phys Chem A 124:9942–9950. https://doi.org/10.1021/acs.jpca.0c06430
- Nowak PM, Woźniakiewicz M, Piwowarska M, Kościelniak P (2016) Determination of acid dissociation constant of 20 coumarin derivatives by capillary electrophoresis using the amine capillary and two different methodologies. J Chromatogr A 1446:149–157. https://doi.org/10.1016/j.chroma.2016.03.084
Schemes 1-4 are available in the Supplementary Files section.
No competing interests reported.
- SupplementalinformationforKLmanuscript3may22submitted.docx
- Scheme1.png
a) Coumarin core structure with numbering scheme, b) 6-hydroxy-7-amino-4-methylcoumarin (knightletin) in the ‘– –‘ (minus-minus) conformer indicating a hydrogen bond (dotted line) formed by the donor NH in position 7
- Scheme2.png
Structures of compounds 1a – 1f
- Scheme3.png
Defines dihedral angles formed by a sequence of atoms in the knightletin – – conformer, (a) Hb-N-C7-C6, (b) Hf-N-C7-C8, (c) H-O-C6-C5, where b designates bound in the hydrogen bond and f designates free; in the knightletin + + conformer
- Scheme4.png
Resonance structures of knightletin with the pyramidal amino group in structure I and planar amino group in structure II
