View PDF VersionPrevious ArticleNext Article
 
DOI: 10.1039/D2CC00561A (Communication) Chem. Commun., 2022, 58, 5455-5458

Fluorene analogues of xanthenes – low molecular weight near-infrared dyes†

Marek Grzybowski *a, Olaf Morawski b, Krzysztof Nowak a and Paula Garbacz a
aInstitute of Organic Chemistry, Polish Academy of Sciences, Kasprzaka 44/52, 01-224 Warsaw, Poland. E-mail: mgrzybowski@icho.edu.pl
bInstitute of Physics, Polish Academy of Sciences, Al. Lotników 32/46, 02-668 Warsaw, Poland

Received 27th January 2022 , Accepted 15th March 2022

First published on 15th March 2022

Abstract

Fluorene-based analogues of fluorescein, rhodol, and rhodamine exhibit absorption and fluorescence beyond 800–900 nm in water, 300–400 nm red-shifted compared to the original oxygen-bridged xanthene dyes, giving potential access to low molecular weight fluorescent markers for the second biological window (NIR-II, ca. 1000–1350 nm).

Living biological tissues are almost transparent to near-infrared (NIR) light, particularly within the wavelength ranges of ca. 700–950 nm and ca. 1000–1350 nm, the biological windows NIR-I and NIR-II, respectively.1,2 Furthermore, compared to UV and visible light, NIR light has lower energy and causes less photodamage, as well as offers higher contrast due to minimized autofluorescence of living matter.1,2 These factors make NIR imaging techniques highly advantageous over more widely used methods operating in the visible range,3,4 creating an increasing demand for efficient and stable NIR fluorophores. The classic method for designing NIR dyes consists of an extension of the π-conjugated systems of known chromophores (π-expansion),5 which often leads to a substantial increase in molecular weights, thus decreasing water solubility and biocompatibility. In recent years, a much more optimal approach to NIR fluorophores has been developed, consisting of the replacement of the endocyclic oxygen atom in classic xanthene dyes, fluorescein, rhodol, and rhodamine, with moieties containing different elements,6 such as silicon,6,7 phosphorus,8–12 or sulfur.13,14 The resulting heteroxanthene dyes show strongly red-shifted absorption and emission maxima reaching the biological window NIR-I, and simultaneously retain the key properties of classic xanthenes: water solubility, intense fluorescence, and low toxicity. However, the substitution of oxygen with various element-based groups still leads to an increase in molecular weight. In this context, we wondered how a simplification of the structure via a complete removal of the bridging atom from the xanthene scaffold would affect the photophysical properties of the resulting fluorene-based analogues of fluorescein, rhodol, and rhodamine – FluFscn, FluRdl, and FluRdn, respectively (Fig. 1). Therefore, density-functional theory (DFT) calculations were performed, which showed that the elimination of the central oxygen atom leads to a significant narrowing of the HOMO–LUMO energy gap, resulting mostly from the strong decrease in the LUMO energy level and the less pronounced increase in the HOMO (Fig. 1). This should lead to significantly red-shifted absorption and emission wavelengths, despite simpler structures and lower molecular weights of the fluorene analogues compared to the original xanthene dyes.
image file: d2cc00561a-f1.tif
Fig. 1 General structures of: classic xanthene dyes (left) and fluorene-based analogues of xanthene dyes (right). Ar = aromatic substituent. In the middle are shown the HOMO and LUMO energies calculated for Ar = 2,6-dimethylphenyl in water as solvent (DFT, M06-2X/6-31+G* level of theory with SMD solvation model).

Among these fluorene analogues, until now only some FluRdn derivatives were known (Ar = phenyl and 1-naphthyl) and, indeed, they exhibited absorption maxima at ca. 950–1000 nm, red-shifted by ca. 450 nm with respect to tetramethyl-rhodamine.15–17 However, their emissive properties have not yet been studied, presumably due to the lack of suitable equipment at the time.

Herein, all three oxygen-deprived versions of xanthene dyes and some of their derivatives were synthesized and characterized. It was demonstrated that the rhodol- and fluorescein-type analogues can be obtained simply by hydrolysis of the rhodamine-type derivative with a bulky substituent at the 9 position, similar to the phosphorus-bridged xanthene dyes.10 All dyes exhibit absorption and fluorescence beyond 800 nm and 980 nm, respectively, although the extinction coefficients and fluorescence quantum yields are much lower than for the original xanthene dyes.

The synthesis of the fluorene-based analogues of the xanthene derivatives is depicted in Scheme 1 and employs derivatives of 3,6-diaminofluorenones 3 as the key intermediates. Although 3,6-bis(dimethylamino)fluorenone (3-H) is known, existing syntheses are difficult and multistep (6 or more steps from commercial substrates).18,19 Instead, 3,6-diaminofluorenes 2 were synthesized from derivatives of 4,4′-methylenebis(3-bromoaniline) 1. Reported attempts to obtain 3,6-diaminofluorene 2-Hvia copper-mediated Ullmann cyclization of 1-H and its diiodo analogue failed.18 To our delight, the cyclizations of dibromide 1-H and its derivatives 1-F and 1-In proceeded smoothly and with high yields through Yamamoto-type coupling20–23 involving in situ generation of Ni(0) species by reducing a catalytic amount of Ni(PPh3)2Cl2 with an excess of zinc metal (Scheme 1). The resulting fluorenes 2 were then oxidized to the corresponding fluorenones 3-H, 3-F and 3-In in good yields by air in the presence of tetrabutylammonium bromide (TBAB) as a phase-transfer catalyst and solid NaOH as a base, a modified procedure previously used to synthesize phosphaxanthones.11,12 The reaction of fluorenone 3-H with Grignard or organolithium reagents followed by acidification yielded the FluRdn derivatives 4-H and 4-H-CF3 bearing bulky 2,6-dimethylphenyl and 2-trifluoromethylphenyl groups at the 9-position, respectively. Yamaguchi and co-workers demonstrated that phospha-rhodamines with bulky substituents at the 9-position can be easily hydrolysed to the corresponding phospha-rhodols and -fluoresceins using aqueous NaOH.10 We found that FluRdn 4-H could also be efficiently hydrolysed to the rhodol analogue 5-H with 38% yield in 0.02 M NaOH; and to fluorescein analogue 6-H with 50% yield under harsher basic conditions (Scheme 1). Attempts to synthesize the fluorinated rhodamine analogue 4-F from fluorenone 3-F failed – the product decomposed during purification. However, 4-F could be generated in solution and directly hydrolysed to the fluorescein analogue 6-F. We have also synthesized an indoline analogue 4-In, since in the case of sila- and phospha-rhodamines such fused amino moieties led to significant bathochromic shifts of absorption and emission.12,24


image file: d2cc00561a-s1.tif
Scheme 1 Synthesis of the fluorene analogues of the xanthene dyes. Reaction conditions: (a) Ni(PPh3)2Cl2 (10 mol%), 2,2′-bipyridine (10 mol%), Zn, KBr, DMF, 60 °C; (b) Air, TBAB (20 mol%), NaOH, THF, rt; (c) ArMgBr⋯LiCl or ArLi, THF, then HCl (aq) or trifluoroacetic acid; (d) 0.02 M NaOH, H2O/MeOH, rt; (e) 0.1–0.2 M NaOH, H2O/MeOH 1/1, rt, then 0.5–1.0 M NaOH, 60 °C.

The structure of the fluorene-based fluorescein analogue 6-H was determined by a single crystal X-ray diffraction analysis (Fig. 2). Interestingly, the differences in bond lengths between the ring A and C of 6-H are very small (0.01–0.02 Å), so the π-system of the molecule is nearly symmetrical. This indicates a strong electron delocalization that may result from relatively strong intermolecular hydrogen bonds between the molecules of 6-H (O⋯O distances: 2.67 Å, Fig. S1, ESI†).


image file: d2cc00561a-f2.tif
Fig. 2 ORTEP drawing of the crystal structure of FluFscn 6-H (CCDC 2090872). Thermal ellipsoids are drawn at 50% probability and the bond lengths are shown in Å.

In analogy to fluoresceins, the photophysical properties of 6-H and 6-F are pH-dependent. The deprotonation of the phenolic groups leads to a colour change from violet/blue to green as a result of the formation of a NIR-absorptive anion. The pKa values were determined to be 7.5 and 6.6 for 6-H and 6-F, respectively (Fig. S2 and S3, ESI†), demonstrating the strong inductive effect of the fluorine atoms in 6-F. It is noteworthy that the colour change occurs near the neutral pH values.

Absorption and emission spectra of the obtained fluorene-based analogues of xanthene dyes were studied in water containing 1–5% vol. DMSO (for 6-H and 6-F with 10 mM carbonate buffer, pH = 10.0, to ensure deprotonation). The results are summarized in Table 1 and the selected spectra are shown in Fig. 3. As expected, in water all the studied dyes exhibit lowest energy absorption bands, corresponding to the S0 → S1 transitions, beyond 800 nm. Similarly as in the xanthene dyes family, the cationic rhodamine analogues exhibit the most red-shifted absorption and emission maxima, while the anionic fluorescein analogues – the least. In the structurally-related series 4-H (cation), 5-H, 6-H (anion), the absorption maxima are located at 943, 811, and 800 nm, respectively. The absorption bands are broad and have low intensities (molar absorption coefficients 4700–14[thin space (1/6-em)]700 M−1 cm−1), with only FluRdn dyes showing vibronic structures. The oscillator strengths of the S0 → S1 transition derived from the absorption spectra are within the 0.06–0.15 range. Except for 4-In, emission of which could not be detected, all molecules are weakly fluorescent in water with maxima at 980–1060 nm (Fig. 3) and fluorescence quantum yields on the order of 10−6 to 10−4. Ring fusion in 4-In and fluorine atoms in 6-F induce substantial bathochromic shifts.

Table 1 Photophysical properties of the fluorene-based analogues of xanthene dyes measured in water containing 1–5% (v/v) DMSO as a cosolvent
Dye λ abs [nm] ε max [M−1 cm−1] λ em [nm] Δνa [cm−1] Φ fl
a Stokes shift. b Measured using IR140 dye as a reference (Φfl = 0.15). c Anionic forms measured in the presence of 10 mM carbonate buffer (pH = 10.0). d Emission not detected.
4-H 943 14 700 1037 960 1.9 × 10−4
4-H-CF3 952 14 600 998 440 1.1 × 10−4
4-In 1053 5 080 nd d – nd d
5-H 811 4 720 945 440 1.7 × 10−4
6-H anionc 800 7 710 907 830 1.1 × 10−5
6-F anionc 839 5 060 1116 2960 1.1 × 10−6


image file: d2cc00561a-f3.tif
Fig. 3 Absorption (left) and emission (right) spectra of 4-H, 5-H and 6-H in water containing DMSO as a cosolvent (1–5% v/v). For 6-H, a 10 mM carbonate buffer (pH = 10) was used to maintain the dye in the anionic form.

In contrast to FluRdn and FluFsc dyes, which do not show solvatochromism (Fig. S4, ESI†), the absorption and fluorescence spectra of FluRdl 5-H shift bathochromically in polar solvents (Fig. S6, ESI†), which results from its strongly polarized donor–acceptor structure. Unusually, the Stokes shift decreases and does not correlate with the Lippert–Mataga parameter, f(ε, n), a common measure of solvent polarity (Fig. S7, ESI†).25,26 Instead, the strong bathochromic shifts depend rather on the solvent's hydrogen-bond donor acidity, a parameter proposed by Kamlet (Fig. S8, ESI†).27 This indicates the formation of hydrogen bonds between solvent molecules and the oxygen atom of 5-H. In this solvatochromic behaviour, 5-H is more similar to phospha-rhodols than the classic oxygen-rhodols.10

To evaluate the reactivity of FluRdn dyes towards nucleophiles, the changes in absorption spectra of 4-H and 4-H-CF3 were monitored in the presence of thiol (DL-homocysteine, HCys), as well as at neutral to basic pH values (pH = 7.5, 9.1, and 10.0). 4-H proved to be resistant to nucleophilic attack, as a noticeable absorption decline was observed only after 20 h at pH = 10 (Fig. S9, ESI†). On the contrary, 4-H-CF3 underwent rapid discoloration in the presence of 0.25–10 mM HCys and slowly reacted with hydroxide anions under basic conditions (Fig. S10, ESI†). These results indicate that double-sided protection of the C9 position, as with the 2,6-dimethylphenyl substituent in 4-H, is essential to ensure good chemical stability under biological conditions.

The electronic structure and properties of the new dyes were further explored by DFT calculations at the M06-2X/6-31+G* level of theory. As mentioned already, the removal of oxygen from the xanthene dyes narrows down the HOMO–LUMO energy gap by 0.71 to 0.94 eV, which is responsible for the strong bathochromic shifts of the absorption and emission maxima (Fig. 1 and Fig. S11, ESI†). In line with that, the results of the time-dependent DFT calculations (TD-DFT) predict a decrease of the S0–S1 transition energies by ∼1 eV, as well as a significant drop in the oscillator strengths from 0.74–0.94 for the xanthene dyes to 0.16–0.22 for the fluorene analogues (Table S3, ESI†), although the calculated values are higher than the experimental ones (0.08–0.15). The oscillator strengths are further reduced by 40–50% for the S1–S0 emission (Table S4, ESI†), which is consistent with low fluorescence quantum yields.

It has been proposed that the red-shifted absorption of 2,6-diaminofluorenylium cations is caused by antiaromatic character.28 Indeed, the dyes in the fluorene series 4-H, 5-H, and 6-H are all antiaromatic according to the nuclear-independent chemical shifts (NICS(1)zz)29 calculated for the three rings constituting their π-systems (Fig. 4). They show strong paratropic currents in the central cyclopentadienyl rings (NICS(1)zz = 24.5 to 34.0 ppm), which is flanked by nonaromatic benzene rings in FluFscn and FluRdn, or weakly aromatic/antiaromatic rings in FluRdl. In contrast, classic xanthenes are characterized by diatropic ring currents in all rings with more aromatic peripheral benzene rings (NICS(1)zz = −13.6 to −19.5 ppm) and weakly aromatic central pyran (NICS(1)zz = −4.9 to −6.7 ppm). NICS calculation results are in very good agreement with magnetically induced current density maps (Fig. S14–S16, ESI†).30 The antiaromaticity of the fluorene analogues explains the observed strongly red-shifted absorption and emission, since it energetically destabilizes the ground states and, conversely, stabilizes the S1 excited states,31 which should be aromatic following the Bairds rule. The excited state aromaticity manifests in more uniform bond lengths on the outer rim of the π-systems, as reflected in the decreased values of the mean absolute deviations of the bond lengths (MAD, Table S5, ESI†) in the S1vs. the S0 states. For the oxygen-bridged xanthenes, the differences are negligible.


image file: d2cc00561a-f4.tif
Fig. 4 NICS(1)zz values (in ppm) of the rings constituting the chromophores of xanthene dyes and their fluorene-based analogues.

In light of these results, the low luminescence quantum yields of the fluorene-based dyes can be attributed not only to the small S0–S1 energy gaps, which facilitates internal conversion, but also to the antiaromatic character of the fluorene core. This can open the radiationless relaxation pathway through the conical intersection between the excited and ground-state potential energy surfaces, as in the majority of known antiaromatic molecules.31–33

In conclusion, we have successfully synthesized and characterized a series of fluorene-based analogues of xanthene dyes: fluorescein, rhodol, and rhodamine. The 2,6-diaminofluorene skeleton was prepared by Yamamoto-type cyclization, a much more facile route than previously reported.18,19 Similar to phospha-rhodamines,10 the fluorene analogue can also be hydrolyzed to the corresponding rhodol and fluorescein-type dyes under basic conditions. Despite their simpler structures, the new fluorene derivatives have absorption in the NIR-I and weak fluorescence in the NIR-II biological window, 300–400 nm red-shifted compared to the classic oxygen-bridged analogs. In addition, the rhodol-type dye FluRdl exhibits pronounced solvatochromism, mostly dependent on the solvent proticity. The molecules could potentially be used as NIR fluorescent dyes for bioimaging, and FluRdl can be considered as a hydrogen bond sensor, although the detection of weak fluorescence may be a challenge in some environments. Overall, we have shown here that it is not always necessary to increase the molecular complexity to achieve highly red-shifted photophysical properties, which will be helpful in designing new NIR dyes. We envision that the fluorescence of the fluorene-based dyes may be enhanced by mitigating their antiaromatic character and we will continue the research on this topic in search of more emissive derivatives.

Conceptualization, investigation, and the major part of the synthesis: M.G.; photophysical measurements: O. M., M. G.; Synthesis of 4-In and 5-H: K. N., P. G.

This work was funded by the National Science Centre (SONATA 2018/31/D/ST5/00432). The authors thank Dr Cina Foroutan-Nejad for assistance with the quantum chemical calculations. The calculations were performed using the PLGrid Infrastructure.

Conflicts of interest

There are no conflicts of interest to declare.

Notes and references

  1. A. M. Smith, M. C. Mancini and S. Nie, Nat. Nanotechnol., 2009, 4, 710–711 CrossRef CAS PubMed .
  2. J. Bin Li, H. W. Liu, T. Fu, R. Wang, X. B. Zhang and W. Tan, Trends Chem., 2019, 1, 224–234 CrossRef PubMed .
  3. S. Wang, B. Li and F. Zhang, ACS Cent. Sci., 2020, 6, 1302–1316 CrossRef CAS PubMed .
  4. Z. Guo, S. Park, J. Yoon and I. Shin, Chem. Soc. Rev., 2014, 43, 16–29 RSC .
  5. G. Qian and Z. Y. Wang, Chem. – Asian J., 2010, 5, 1006–1029 CrossRef CAS PubMed .
  6. T. Ikeno, T. Nagano and K. Hanaoka, Chem. – Asian J., 2017, 12, 1435–1446 CrossRef CAS PubMed .
  7. M. Fu, Y. Xiao, X. Qian, D. Zhao and Y. Xu, Chem. Commun., 2008, 1780–1782 RSC .
  8. X. Chai, X. Cui, B. Wang, F. Yang, Y. Cai, Q. Wu and T. Wang, Chem. – Eur. J., 2015, 21, 16754–16758 CrossRef CAS PubMed .
  9. A. Fukazawa, S. Suda, M. Taki, E. Yamaguchi, M. Grzybowski, Y. Sato, T. Higashiyama and S. Yamaguchi, Chem. Commun., 2016, 52, 1120–1123 RSC .
  10. M. Grzybowski, M. Taki and S. Yamaguchi, Chem. – Eur. J., 2017, 23, 13028–13032 CrossRef CAS PubMed .
  11. M. Grzybowski, M. Taki, K. Senda, Y. Sato, T. Ariyoshi, Y. Okada, R. Kawakami, T. Imamura and S. Yamaguchi, Angew. Chem., Int. Ed., 2018, 57, 10137–10141 CrossRef CAS PubMed .
  12. M. Grzybowski, M. Taki, K. Kajiwara and S. Yamaguchi, Chem. – Eur. J., 2020, 26, 7912–7917 CrossRef CAS PubMed .
  13. J. Liu, Y. Q. Sun, H. Zhang, H. Shi, Y. Shi and W. Guo, ACS Appl. Mater. Interfaces, 2016, 8, 22953–22962 CrossRef CAS PubMed .
  14. K. V. Vygranenko, Y. M. Poronik, A. Wrzosek, A. Szewczyk and D. T. Gryko, Chem. Commun., 2021, 57, 7782–7785 RSC .
  15. A. Barker and C. C. Barker, J. Chem. Soc., 1954, 1307–1309 RSC .
  16. S. Nakatsuji, H. Nakazumi, H. Fukuma, T. Yahiro, K. Nakashima, M. Iyoda and S. Akiyama, J. Chem. Soc., Perkin Trans. 1, 1991, 1881–1886 RSC .
  17. S. G. R. R. Guinot, J. D. Hepworth and M. Wainwright, Dyes Pigm., 1999, 40, 151–156 CrossRef CAS .
  18. A. Barker and C. C. Barker, J. Chem. Soc., 1954, 870–873 RSC .
  19. S. G. R. Guinot, J. D. Hepworth and M. Wainwright, J. Chem. Res. Part S, 1997, 183 RSC .
  20. I. Colon and D. R. Kelsey, J. Org. Chem., 1986, 51, 2627–2637 CrossRef CAS .
  21. V. Sharma, B. Bachand, M. Simard and J. D. Wuest, J. Org. Chem., 1994, 59, 7785–7792 CrossRef CAS .
  22. M. Zembayashi, K. Tamao, J. ichi Yoshida and M. Kumada, Tetrahedron Lett., 1977, 18, 4089–4091 CrossRef .
  23. G. C. Paul and J. J. Gajewski, Org. Prep. Proced. Int., 1998, 30, 222–225 CrossRef CAS .
  24. Y. Koide, Y. Urano, K. Hanaoka, W. Piao, M. Kusakabe, N. Saito, T. Terai, T. Okabe and T. Nagano, J. Am. Chem. Soc., 2012, 134, 5029–5031 CrossRef CAS PubMed .
  25. E. Lippert, Z. Naturforsch., A: Phys. Sci., 1955, 10, 541–545 Search PubMed .
  26. N. Mataga, Y. Kaifu and M. Koizumi, Bull. Chem. Soc. Jpn., 1955, 28, 690–691 CrossRef CAS .
  27. M. J. Kamlet, J. L. M. Abboud, M. H. Abraham and R. W. Taft, J. Org. Chem., 1983, 48, 2877–2887 CrossRef CAS .
  28. F. Dietz, N. Tyutyulkov and M. Rabinovitz, J. Chem. Soc., Perkin Trans. 2, 1993, 157–164 RSC .
  29. Z. Chen, C. S. Wannere, C. Corminboeuf, R. Puchta and P. von Ragué Schleyer, Chem. Rev., 2005, 105, 3842–3888 CrossRef CAS PubMed .
  30. T. Janda and C. Foroutan-Nejad, ChemPhysChem, 2018, 19, 2357–2363 CrossRef CAS PubMed .
  31. M. Rosenberg, C. Dahlstrand, K. Kilså and H. Ottosson, Chem. Rev., 2014, 114, 5379–5425 CrossRef CAS PubMed .
  32. B. D. Rose, L. E. Shoer, M. R. Wasielewski and M. M. Haley, Chem. Phys. Lett., 2014, 616–617, 137–141 CrossRef CAS .
  33. C. K. Frederickson, J. E. Barker, J. J. Dressler, Z. Zhou, E. R. Hanks, J. P. Bard, L. N. Zakharov, M. A. Petrukhina and M. M. Haley, Synlett, 2018, 2562–2566 CAS .

Footnote

† Electronic supplementary information (ESI) available. CCDC 2090872. For ESI and crystallographic data in CIF or other electronic format see DOI: 10.1039/d2cc00561a
This journal is © The Royal Society of Chemistry 2022
Click here to see how this site uses Cookies. View our privacy policy here.