A near-infrared large Stokes shift probe based enhanced ICT strategy for F- detection in real samples and cell imaging

Article abstract of DOI:10.1016/j.tet.2019.130762

In view of the few reports of the near-infrared emissive probe for fluorine ions, we herein designed and synthesized a new easy-to-get colorimetric and near-infrared emissive fluorescent probe (IS-NR-F) with a large Stokes shift (>127 nm). Based on specif

Full text of DOI:10.1016/j.tet.2019.130762

Journal Pre-proof
A near-infrared large Stokes shift probe based enhanced ICT strategy for F- detection
in real samples and cell imaging
Qinghua Hu, Qiuxiang Huang, Yu Mao, Xinlin Liu, Fangrong Tan, Yuyuan Wang,
Qiang Yin, Xumeng Wu, Hongqing Wang
PII:
S0040-4020(19)31155-X
https://doi.org/10.1016/j.tet.2019.130762
TET 130762
DOI:
Reference:
To appear in: Tetrahedron
Received Date: 15 September 2019
Revised Date: 29 October 2019
Accepted Date: 1 November 2019
Please cite this article as: Hu Q, Huang Q, Mao Y, Liu X, Tan F, Wang Y, Yin Q, Wu X, Wang H, A near-
infrared large Stokes shift probe based enhanced ICT strategy for F- detection in real samples and cell
imaging, Tetrahedron (2019), doi: https://doi.org/10.1016/j.tet.2019.130762.
This is a PDF file of an article that has undergone enhancements after acceptance, such as the addition
of a cover page and metadata, and formatting for readability, but it is not yet the definitive version of
record. This version will undergo additional copyediting, typesetting and review before it is published
in its final form, but we are providing this version to give early visibility of the article. Please note that,
during the production process, errors may be discovered which could affect the content, and all legal
disclaimers that apply to the journal pertain.
© 2019 Published by Elsevier Ltd.

A near-infrared large Stokes shift probe based enhanced ICT
strategy for F- detection in real samples and cell imaging
Qinghua Hu, Qiuxiang Huang, Yu Mao, Xinlin Liu, Fangrong Tan, Yuyuan Wang,
Qiang Yin, Xumeng Wu, Hongqing Wang*
School of Chemistry and Chemical Engineering, Hunan key laboratory for the design and
application of actinide complexes, University of South China, 28 Changsheng West Road,
Hengyang, Hunan, 421001, PR China;
* Corresponding author: E-mail address: hqwang2015@126.com
1

Abstract:
In view of the few reports of the near-infrared emissive probe for fluorine ions,
we herein designed and synthesized a new easy-to-get colorimetric and near-infrared
emissive fluorescent probe (IS-NR-F) with a large Stokes shift (>127 nm). Based on
specific F- triggered desilylation reaction induced enhanced ICT strategy involving
the donor phenolate anion and the acceptor malononitrile, the probe exhibited dual
colorimetric and fluorescent turn-on responses, and provided excellent selectivity for
fluoride ions. The fluorescent response at 665 nm displayed very good linear
relationship in the wide concentration range and deduced a low detection limit of 0.09
1
ppm. The detection mechanism was confirmed by H NMR, ESI-MS, and TLC
calculation. Moreover, probe IS-NR-F has been successfully employed to detect F- in
tap water, toothpaste samples, and fluorescent imaging of F- in HeLa cells.
Keywords: Fluoride; Near-infrared fluorescence; Large Stokes shift; Fluorescent
image
1

1. Introduction
The smallest and most negative anionic, fluorine ion (F-), has been received great
attention as an important chemical component in organisms and medical systems [1,
2]. Fluoride can be usually contacted through food, air, toothpaste, tea and
medicines[3]. The intake of trace amount fluoride can not only prevent dental caries
but help bone growth [4-6]. However, the World Health Organization (WHO)
recommends fluoride levels should be less than 1.5 ppm in drinking water [7], and
excessive ingestion of fluoride is likely to cause chronic poisoning, including dental
fluorosis, bone fluorosis, metabolic dysfunction and urolithiasis [8, 9]. Therefore,
given these widespread applications of F- in daily life and its special function in a
biological system, it has drawn greater attention to developing highly efficient and
reliable methods for the detection and monitoring of F- in drinking water as well as in
the biological system.
The traditional methods including ion-selective electrodes [10], ion
chromatography [11] and colorimetry [12] have been used for detecting fluorine ions.
However, most of them couldn’t be applied to detect F- in a biological system.
Compared to these traditional methods, the fluorescent probe technique could achieve
easy, rapid and imaging in living cells and in vivo [13-34]. Especially, the
near-infrared (NIR, 650-900nm) emissive fluorescent probe shows an excellent
imaging performance in this aspect because of that its long-wavelength photons often
perform a lower environmentally-induced light scattering, less interference from
endogenous chromophores, and less photodamage [35-40]. In recent years, several
2

NIR fluorescent probes have been developed for F- detection. However, most of them
have small Stokes shifts(<35 nm) [41, 42], which may limit their further application
in biosamples and decrease the detection sensitivity of probe to some degree because
of the interference from excitation source with the “self-absorption” effect [43]. Thus,
it remains attractive to develop NIR fluorophores from common raw materials with
large Stokes shifts.
On the other hand, at present, most of the reported fluorescence probes have
widely applied the two main strategies to detect F- (Table S1 and Fig. S1): a)
hydrogen bonding interaction between F- and OH group/NH group; b) Lewis
acid\base interaction [44-55]. However, since F- is the most electronegative anion and
has strong hydration, the hydrogen bonding strategy would make the detecting
process be largely hampered in protic solvent and could only be detected in a pure
solvent or high content organic solvent that is harmful to cells [56]. For probes based
on Lewis acid/base interaction, they could monitor fluoride in drinking water, and this
strategy mainly includes fluorine-boron complexation [57-60] and desilylation
reaction [61, 62]. Nevertheless, the fluorine-boron complexation is unstable and high
cell toxicity [63] while the fluoride-triggered desilylation reaction relative to
fluorine-boron complexation has had wide application in cell imaging and provides a
specific irreversible recognition site due to the strong affinity between fluorine ion
and silicon, thereby performing an excellent selectivity [64-67]. However, some of the
probes based on desilylation reaction are still subject to the restriction for their short
wavelength (UV-Vis light range) emission peak because they would suffer the
3

interference from endogenous chromophores and relative large photodamage to the
biological system compared to NIR fluorescent probes [68-70].
Therefore, in order to solve the problem produced by the short-wavelength
emission peak and achieve NIR fluorescent imaging in the cell, we proposed a novel
NIR emissive fluorescence probe (IS-NR-F, Scheme 1B) by incorporating with a
hydroxyl-functionalized dicyanoisophorone derivative and a tert-butyldimethylsilyl
(TBDMS) group as the trigger moiety. To test the influence of the malononitrile
group in IS-NR-F on the recognition of F-, a control probe IS-F (Scheme 1A) lack of
malononitrile group was also synthesized and tested in parallel with IS-NR-F for
their sensing ability to F-. Fortunately, as shown in Scheme 1, this strategy not only
achieved the satisfactory NIR fluorescence around 665 nm with a remarkable Stokes
shift (≥127 nm) in the existence of F-, and IS-NR-F was used to monitor F- in living
cells successfully. Moreover, the probe displayed a low detection limit of 0.09 ppm
and could be used for colorimetric detection of F- in actual water samples and kinds
of toothpaste with an excellent selectivity even in the presence of various cations and
anions.
4

Scheme 1 The proposed sensing mechanism of probe IS-F(A) and IS-NR-F (B) for F-
recognition
2. Experimental
2.1. Material and methods
All chemicals (solvents, reagents and metal salts) were analytical grade,
purchased from Aladdin (Shanghai, China) and used without further purification. The
probe IS-F and IS-NR-F were dissolved in analytical grade DMSO for a stock
3-
2-
2-
2-
solution (1 mM). The solution of anions (F-, Cl-, PO4 , SO3 , SO4 , Ac-, CO3 ,
HCO3 , NO2 , NO3 , S2-, Br-, I-, ClO-) and cations (Al3+, Hg2+, Mn2+, K+, Zn2+, Ba2+,
Mg2+, Cr3+, Ca2+, Pb2+, Ag+, Fe3+, Ni2+, Cu2+) were prepared by dissolving their
sodium salt and nitrates in deionized water, respectively.
-
-
-
The UV-vis spectra were recorded through a Hitachi U-3900 spectrophotometer
(made in Tokyo, Japan). Fluorescent spectra were measured in a Hitachi F-7000
spectrophotometer (made in Tokyo, Japan) in a 1 cm quartz cell. The 1H NMR spectra
were acquired on a Bruker 400 MHz spectrophotometer (made in Bruker, Germany)
in deuterated dimethyl sulfoxide (DMSO-d6) with tetramethylsilane as an internal
5

standard. The HR-MS was obtained on Alliance e2695-ACQUITY QDa (made in
Waters, America) and LCMS-IT-TOF instrument (made in Shimadzu, Japan). The pH
was performed on a model pHs-3C digital pH/mV meter (made in Shanghai, China).
The cell fluorescent images were performed by a TE2000 inverted fluorescent
microscope (made in Nikon, Japan).
2.2. Synthesis and characterization
The detailed synthetic procedure of probe IS-F and IS-NR-F were outlined in
Scheme 2.
Synthesis of IS-OH: To a mixture solution of a (Isophorone) (1.38 g, 10 mmol)
and p-hydroxybenzaldehyde (1.22 g, 10 mmol) in absolute EtOH (20 mL) was added
EtONa (2.57 g, 15 mmol). The mixture was refluxed for 3 h, then 5 mL water was
added to quench the reaction. The solvent was removed by vacuum rotary evaporation.
Finally, the crude product was purified by silica gel (petroleum ether/ethyl acetate =
6:1, v/v) to give IS-OH as an orange solid (1.08 g, 44.6 % yield). 1H NMR (400 MHz,
DMSO-d6, Fig. S2) δ (ppm) 9.81 (s, 1H), 7.50-7.48 (d, J = 8.0 Hz, 2H), 7.12-7.08 (d,
J = 16.0 Hz, 1H), 6.97-6.93 (d, J = 16.0 Hz, 1H), 6.82-6.80 (d, J = 8.0 Hz, 2H), 6.01 (s,
1H), 2.53 (s, 2H), 2.24 (s, 2H), 1.06 (s, 6H). 13C NMR (400 MHz, DMSO-d6, Fig. S3):
δ 199.28, 159.02, 156.11, 135.71, 129.51, 127.61, 126.62, 125.51, 116.19, 51.40,
38.61, 33.44, 28.48. HR-MS (Fig. S4): m/z (%): Calculated for C16H18O2, 242.1307,
found 265.1190 [M+Na] +.
Synthesis of IS-F: To a solution of IS-OH (121 mg, 0.5 mmol) in
dichloromethane (12 mL) and Et3N (138 ꢀL, 1 mmol), tert-butyldimethylsilyl chloride
6

(150.7 mg, 1 mmol) dissolved with 4 mL dichloromethane was added dropwise at 0
ºC. The mixture was stirred at room temperature for 12 hours. Then, the solvent was
removed under reduced pressure. The resulting residue was purified by silica gel flash
chromatography (eluted with petroleum ether/dichloromethane = 3:1, v/v) to afford
1
the desired product as a yellow solid (106 mg, 59.6 % yield). H NMR (400 MHz,
DMSO-d6, Fig. S5) δ (ppm): 7.57-7.55 (d, J = 8.0 Hz, 2H), 7.16-7.12 (d, J = 16.0 Hz,
1H), 7.04-7.00 (d, J = 16.0 Hz, 1H), 6.90-6.88 (d, J = 8.0 Hz, 2H), 6.04 (s, 1H), 2.52
13
(s, 2H), 2.25 (s, 2H), 1.06 (s, 6H), 0.98 (s, 9H), 0.23 (s, 6H). C NMR (400 MHz,
CDCl3, Fig. S6): δ 200.23, 156.84, 155.20, 134.77, 129.35, 128.65, 127.57, 126.32,
120.25, 51.42, 39.10, 33.34, 28.51, 25.64, -4.39. HR-MS (Fig. S7): m/z (%):
Calculated for C22H32O2Si, 356.2172, found 379.2056 [M+Na] +.
Synthesis of compound b: an (Isophorone) (3.455 g, 25 mmol) and malononitrile
(1.980 g, 30 mmol) were dissolved in absolute EtOH (50 mL), followed by addition
of piperidine (0.25 mL, 2.5 mmol) and glacial acetic acid (0.15 g, 2.5 mmol). Then,
after the mixture was refluxed for 6 h under N2 atmosphere, it was quenched with
water, extracted with dichloromethane, dried over anhydrous sodium sulfate, filtered
and concentrated in vacuo. Finally, the crude residue was purified by silica column
chromatography (eluted with petroleum ether/dichloromethane = 1:3, v/v) to give b as
1
a light yellow solid (3.63 g, 78 % yield). Melting point: 70.2-72.6 ºC. H NMR (400
MHz, DMSO-d6, Fig. S8) δ (ppm) 6.56 (s, 1H), 2.54 (s, 2H), 2.23 (s, 2H), 2.05 (s,
3H), 0.96 (s, 6H). 13C NMR (400 MHz, DMSO-d6, Fig. S9): δ 171.91, 162.98, 119.88,
113.93, 113.14, 76.50, 45.27, 42.42, 32.44, 27.75, 25.45. HR-MS (Fig. S10): m/z (%):
7

Calculated for C12H14N2, 186.1157, found 185.1086 [M-H] -.
Synthesis of IS-NR-OH: 20 mL acetonitrile was added to dissolve compound b
(259 mg, 1.4 mmol) and p-hydroxybenzaldehyde (170 mg, 1.4 mmol), followed by
five drops of piperidine. The mixture was refluxed for 5h under N2 atmosphere and
subsequently, the solvent was evaporated and removed through reduced pressure. The
resulting residue was dissolved in 20 mL dichloromethane, washed with water (3×20
mL), and dried by anhydrous Na2SO4. After the removal of the solvent, the crude
product was purified by column chromatography on silica gel (petroleum
ether/dichloromethane=1:1, v/v) to afford the desired product as an orange solid (335
1
mg, 85.2% yield). Melting point: 200.4-201.5 ºC. H NMR (400 MHz, DMSO-d6,
Fig. S11) δ (ppm) 9.99 (s, 1H), 7.57-7.55 (d, J = 8.0 Hz, 2H), 7.22-7.23 (d, J = 4.0 Hz,
2H), 6.81-6.79 (m, 3H), 2.60 (s, 2H), 2.53 (s, 2H), 1.02 (s, 6H). 13C NMR (400 MHz,
DMSO-d6, Fig. S12): δ 170.72, 159.81, 157.15, 138.74, 130.33, 127.59, 126.71,
121.84, 116.35, 114.58, 113.77, 75.28, 42.79, 38.67, 32.12, 27.91. ESI-HRMS (Fig.
S13): m/z (%): Calculated for C19H18N2O, 290.1419, found 313.1307 [M+Na] +.
Synthesis of IS-NR-F: To a solution of IS-NR-OH (145 mg, 0.5 mmol) in
dichloromethane (10 mL) and Et3N (138 ꢀL, 1 mmol), tert-butyldimethylsilyl chloride
(150.7 mg, 1 mmol) dissolved with 5 mL dichloromethane was added dropwise at 0
ºC. The mixture was stirred at room temperature for 12 hours. Then, the solvent was
removed under reduced pressure. The resulting residue was purified by silica gel flash
chromatography (eluted with petroleum ether/dichloromethane = 2:1, v/v) to afford
1
the desired product as a yellow solid (158 mg, 78.2 % yield). H NMR (400 MHz,
8

DMSO-d6, Fig. S14) δ (ppm): 7.63-7.61 (d, J = 8.0 Hz, 2H), 7.26 (s, 2H), 6.89-6.87 (d,
J = 8.0 Hz, 2H), 6.84 (s, 1H), 2.60 (s, 2H), 2.53 (s, 3H), 1.01 (s, 6H), 0.95 (s, 9H),
0.20 (s, 6H). 13C NMR (400 MHz, DMSO-d6, Fig. S15): δ 170.74, 157.13, 156.73,
137.94, 130.10, 128.13, 122.45, 120.81, 114.43, 113.62, 76.02, 42.78, 38.66, 32.01,
27.89, 25.96, 18.43, -4.07. ESI-HRMS (Fig. S16): m/z (%): Calculated for
C25H32N2OSi, 404.2284, found 427. 2175 [M+Na] +.
Scheme 2 The synthetic route of IS-F and IS-NR-F
2.3. UV-vis and fluorescent measurement
The IS-F and IS-NR-F were dissolved in DMSO as a stock solution (1.0 mM)
while fluorine ion and other analytes were prepared by dissolving their sodium salt
and nitrates in deionized water, respectively. All fluorescent and UV-vis spectral
measurements were investigated in DMSO/PBS=1:1 (v/v, 20 mM, pH=7.4) solution at
37 ºC. The excitation and emission wavelength of fluorescent measurement for each
fluorophore (for IS-F: λex = 440 nm, λem = 568 nm and for IS-NR-F: λex = 550 nm, λem
= 665 nm) were measured respectively while their slit widths were 10.0 nm and 10.0
nm separately. Under the voltage of 450V or 550 V.
9

2.4. F- detection in aqueous sample and kinds of toothpaste
One of the aqueous samples was collected directly from the tap water in our
laboratory, and the other one that was filtered by membrane before used was collected
from Xiangjiang River in Hengyang city. Then the pH value of the obtained solution
was adjusted to 7.4 with NaH2PO4-Na2HPO4. Besides the F- detection in aqueous
water, the amount of F- in commercially available toothpaste was also detected. To
achieve the analysis of F- in toothpaste, 1.0 g toothpaste was accurately weighed and
dissolved in 10 mL PBS buffer solution (20 mM, pH=7.4) and stirred at 70 ºC. After 3
h, the fluorine-containing solution was obtained by filtering and 40 mL PBS buffer
solution was added to the solution for further use.
2.5. Cell imaging and MTT assays
The toxicology of probe toward HeLa cells was determined by common
methylthiazolyldiphenyl-tetrazolium bromide (MTT) assays. A certain density of
HeLa cells was inoculated into 96-well plates and cultured at 37 ºC for 24 hours.
Different concentrations of the probe were then added to the cells for staining and
incubation continued for a further 24 hours. Finally, The HeLa cells were treated with
1 mg/mL of MTT (20 ꢀL/well) in an incubator for 2 hours. After completion, the
liquid in the well was carefully aspirated with a pipette, and the obtained crystal was
dissolved in 150 ꢀL DMSO, and the absorbance was recorded. Subsequently, the
fluorescent imaging of HeLa cells was obtained. The HeLa cells were incubated with
IS-NR-F for 1 hour at 37 ºC after washed with PBS buffer for two times. And then F-
was added into the cells for another 1 hour where the excessive IS-NR-F has been
10

removed with sterile PBS buffer first. The fluorescence images were collected with a
TE2000 inverted fluorescent microscope (made in Nikon, Japan).
3. Results and discussion
3.1. Synthesis and characterization
The synthetic procedure of the probe IS-F and IS-NR-F were outlined in
Scheme 2. Firstly, the isophorone based fluorophore IS-OH was readily synthesized
from isophorone via one-step reaction, the dicyanoisophorone based fluorophore
(IS-NR-OH) was synthesized according to the previous procedure [40]. Then, the
control probe IS-F and probe IS-NR-F was easily prepared by the reaction of IS-OH
or IS-NR-OH with tert-butyldimethylsilyl chloride in the presence of Et3N at room
temperature. The structures of the target IS-F and IS-NR-F and intermediates were
1
characterized and confirmed by H NMR, 13C NMR, and HR-MS, as shown in
Figures S2-S16.
3.2. Optical properties measurement
The absorption and emission spectra of fluorophores and probes (IS-F and
IS-NR-F) were measured by the UV-vis and fluorescence spectra in PBS buffer
solution (20 mM, pH=7.4, 50 % DMSO) at 37 °C. As shown in Fig.1, the fluorophore
IS-OH and IS-NR-OH displayed a maximum absorption peak at 360 nm and 430 nm
respectively, which were caused by the ICT process of the D-π-A structure. When
excited, IS-OH (λex = 440 nm) solution showed an optimal emission at 568 nm (Ф
=0.092, using Fluorescein in 0.1 M NaOH as standard) and IS-NR-OH (λex = 535 nm)
showed an optimal emission at 665 nm (Ф = 0.083). This was because the
11

modification of isophorone with the strong electron-withdrawing malononitrile group
could facilitate the ICT process, and also elongate the π-conjugation of the isophorone
derivatives, which finally caused the emission colors turned from yellow-green to red.
Meanwhile, the UV-Vis and fluorescence spectra response of control probe IS-F and
IS-NR-F toward F- were measured under the same condition. In absorbance spectra,
the probe IS-F and IS-NR-F displayed a maximum absorption peak at 345 nm and
416 nm respectively. After the addition of F- (8 ppm), the maximum absorption bands
of two probe solutions red-shifted to 360 nm and 430 nm respectively. Accordingly,
the addition of F- could produce a significant color change, which can be used for
naked-eye detection. In the fluorescent spectra, both probes IS-F (Ф = 0.0086) and
IS-NR-F (Ф = 0.0057) showed weak emission due to the introduced
tert-butyldimethylsilyl moiety occludes the ICT process. when F- (8 ppm) was added,
the emission intensity of probe IS-F and IS-NR-F solutions at 568 nm and 665 nm
were dramatically increased respectively, and we also measured the excitation
spectrum of the reaction solutions, the maximum excitation wavelengths of the IS-F +
F- and IS-NR-F+ F- were revealed at 446 nm and 538 nm respectively. The new
absorption, excitation and emission peaks were quite similar to that of IS-OH and
IS-NR-OH, which revealed that F--induced desilylation reaction products ought to be
their corresponding fluorophores, the Stokes shift of IS-OH and IS-NR-OH were
found to be 122 nm and 127 nm respectively, which were calculated as the difference
between the maximum excitation wavelengths and emission wavelengths. These
initial results indicate that IS-F and IS-NR-F can be served as colorimetric and
12

fluorescent“turn-on” probe for rapid detection of F-. Moreover, we found IS-NR-F
showed obvious NIR emission enhancement around 665 nm for F- and possesses a
super large Stokes shift (127 nm), which is highly beneficial for biological application,
as its low background interference, less photodamage, and deep penetration. In
comparison, the bioimaging applications of probe IS-F may be limited due to its
shorter emission wave. Therefore, in the following paper, we focused on the sensing
properties of probe IS-NR-F for F-.
Fig. 1 (A) Absorption spectra and fluorescence spectra changes of probe IS-F (20 µM)
upon addition of F- (8 ppm) in PBS buffer solution (20 mM, pH=7.4, 50 % DMSO) at
37 °C. (B) Absorption spectra and fluorescence spectra changes of probe IS-NR-F (20
µM) upon addition of F- (8 ppm) in PBS buffer solution (20 mM, pH=7.4, 50 %
DMSO) at 37 °C. Insert: natural light color and emission color (under a 365 nm light).
In order to further study the optical properties of two probes, the photophysical
properties of IS-F, IS-O, IS-NR-F, and IS-NR-O were performed density functional
theory (DFT) calculations. The four molecular structures as well as the orbital
electron cloud distribution, HOMO and LUMO energy were offered at the level of
RB3LYP/6-311G (d) by the Gaussian 09 program. As exhibited in Fig. 2, all electron
13

clouds on HOMO and LUMO mainly focus on the whole π-conjugated chains,
including C=O moiety and malononitrile group, and there is almost no electron cloud
distribution on the side chain. The electrons on the aromatic ring attached to the
negative ions of oxygen (O-) in the LUMO orbital of IS-O and IS-NR-O were more
dispersed than those of IS-F and IS-NR-F respectively. This could be interpreted as
the significant ICT effect from the electron-donating phenoxy oxygen anion moiety to
the carbonyl or dicyanoisophorone moiety. Nonetheless, we found the width of the
electron cloud distribution of IS-NR-F and IS-NR-O were wider than that of IS-F
and IS-O, which proved the malononitrile group elongates the π-conjugation of the
isophorone. Moreover, the probe IS-NR-F has a smaller LUMO-HOMO energy gap
(2.986 eV) than that of IS-F (3.691 eV), the result was in accord with the redshift of
maximum absorption peak from 345 nm to 416 nm. Meanwhile, the negative ions
IS-NR-O also has a smaller LUMO-HOMO energy gap (2.313 eV) than that of IS-F
(2.595 eV), which in accord with the redshift of maximum absorption peak from 360
nm to 430 nm and emission peak from 568 nm to 665 nm.
Fig. 2 The structural optimization of IS-F, IS-O, IS-NR-F and IS-NR-O
14

3.3. Sensing condition optimization
To explored the potential applications of IS-NR-F, the sensing conditions should
be optimized, such as pH, temperature, solution system and time. Firstly, the intensity
changes at 665 nm in different pH conditions of single IS-NR-F and toward F- were
investigated. As shown in Fig. S17, the results indicated that the fluorescent intensity
of single IS-NR-F and toward F- showed pH-dependence. The emission intensity of
IS-NR-F solutions at 665 nm increased slowly in the range of 6.0-8.0 and increased
drastically in the range of 8.0-11.0. When F- added, the emission intensity at 665 nm
increased drastically in the range of 6.0-8.0 and reached the peak value at pH=8.0 and
then remained a plateau at 9-11. These results indicated that IS-NR-F would be
suitable for detecting F- in biological environments. Secondly, in view of the
reaction-based IS-NR-F, we explored the influence of response time during the
identification of F-. As can be seen in Fig. 3, the fluorescence barely changed in the
absence of F-, while the fluorescent enhancement with varying degrees after a series
of F-(0.38, 1.9, 3.8, 5.7, 11.4 ppm) was added in DMSO/PBS buffer (1:1, v/v) (20 mM,
pH=7.4) solution. When 11.4 ppm of F- was added, the fluorescent intensity reached
to a plateau within 25 min, which could be achieved with a pseudo-first-order rate
constant (k´) for F- calculated to be 0.136 min-1 (Fig. S18). To achieve the F- detection
in cells and take the pH range and temperature of biologically relevant into
consideration, accordingly, the next experiment conditions were optimized to be
carried out in a mixed solvent of 37 ºC and pH value of 7.4 for 25 minutes.
15

Fig. 3 The fluorescent intensity at 665 nm of IS-NR-F (20 ꢀM) toward different F-
concentration (0.38, 1.9, 3.8, 5.7, 11.4 ppm) was continuously monitored at time
intervals in DMSO/PBS buffer (1:1, v/v) (20 mM, pH=7.4) solution.
3.4. The sensing properties of probe IS-NR-F for F-
The absorption spectra of IS-NR-F in the absence and presence of different F-
concentration was shown in Fig. 4. The IS-NR-F displayed a strong absorption peak
at 416 nm in the absence of fluoride firstly. Interestingly, two new absorption peaks
arose around 430 nm and 550 nm respectively with the addition of F-, which were
ascribed to the generation of IS-NR-OH and its deprotonation form (IS-NR-O)
separately [71]. And the absorbance at 550 nm was enhanced gradually with the
increasing concentrations of fluoride. What’s more, a color change from yellow to
brown was observed obviously in the presence of F-, which meant that IS-NR-F could
implement the “naked-eye” colorimetric recognition of F-.
16

Fig. 4 UV-vis spectra changes of IS-NR-F (20 µM) with the addition of different
concentrations of fluoride (0, 0.38, 0.76, 1.14, 1.52, 1.90, 4.56, 6.08, 6.84, 7.60 ppm)
in PBS buffer solution (20 mM, pH=7.4, 50 % DMSO), inset: left: IS-NR-F, right:
IS-NR-F+F-
The fluorescent responses of probe IS-NR-F to F- were recorded in Fig. 5 (A).
Initially, probe IS-NR-F (20 µM) itself displayed a slightly NIR fluorescence at 665
nm. It can be explained by the protection of the -OH with the tert-butyldimethylsilyl
moiety, which hindered the intramolecular charge transfer (ICT) process. However,
the NIR emissive fluorescence was augmented gradually with the titration of
increasing F- concentration. The NIR fluorescence at 665 nm was attributed to the
resulting IS-NR-O after the reaction between IS-NR-F and F-. As shown in Fig. 5 (B),
a wide detection range from 0.38-6.84 ppm with a good linearity relationship was
achieved between the fluorescence of IS-NR-F and F-, and the limit of detection
17

(LOD) was calculated to be 0.09 ppm based on the equation LOD=3 δ/k (δ was the
standard deviation of the blank solution and k was the absolute value of the slope
between fluorescence intensity and corresponding fluorine ion concentration) [72]
.
The LOD was much lower than the recommended fluoride level (1.5 ppm) in drinking
water specified by WHO, which could give the credit to that the specific reaction
induced by F- improved the sensitivity. Moreover, the color changes of probe
IS-NR-F toward F- under natural light and handheld UV lamp (365 nm) was also
performed, respectively. In natural light, the solution color gradually turned to brown
in the presence of 0.38-7.6 ppm F- and be more deepened with the increasing F-
concentration (Fig. 5 (C)). Similarly, the emission fluorescent color of IS-NR-F
gradually turned to red with the increasing F- concentration under hand-held UV lamp
(365 nm) (Fig. 5 (D)), which meant that the color depth and fluorescent intensity
would be enhanced gradually along with the increased concentration of F-. Therefore,
the probe IS-NR-F could achieve the naked eye and fluorescent colorimetric
observation in the range of 0.38-7.6 ppm F- in solution.
18

Fig. 5 (A) The changes of fluorescent emission intensity (at 665 nm) versus fluoride
concentration (0-11.4 ppm), inset (Hand-held UV lamp at 365 nm): left: probe
IS-NR-F, right: IS-NR-F+F-; (B) The intensity linear relationship diagram of
IS-NR-F (20 µM) toward different fluoride concentration (0-11.4 ppm) (at 665 nm);
(C) and (D) The color changes and emission color changes of IS-NR-F toward
various concentration of F- (0-7.6 ppm) were separately observed under naked eyes
and hand held UV lamp (365 nm).
3.5. Selective experiment and colorimetric recognition
The selectivity of IS-NR-F was inquired by the fluorescent response of probe
toward various relevant ions in organisms. When the system was in the presence of
3-
2-
2-
2-
-
-
different 25 equiv. ions, including Cl-, PO4 , SO3 , SO4 , Ac-, CO3 , HCO3 , NO2 ,
-
NO3 , S2-, Br-, I-, ClO-, H2O2, Al3+, Hg2+, Mn2+, K+, Zn2+, Ba2+, Mg2+, Cr3+, Ca2+, Pb2+,
Ag+, Fe3+, Ni2+, Cu2+, the responsive intensity of IS-NR-F toward them was quite
weakly at 665 nm while a great fluorescence enhancement was acquired upon the
addition of 25 equiv. F- (Fig. 6 (A)). The great contrast declared that IS-NR-F
19

exhibited an excellent selectivity to F-, which could be ascribed to the strong affinity
of F- toward silicon. Apart from the excellent selectivity, IS-NR-F could be employed
to the colorimetric recognition of F- in the presence of different ions. As shown in Fig.
6 (C) and (D), the specific cleavage reaction made the solution turn brown or produce
a strong red fluorescence in the existence of F- while there was no color change in the
addition of other analytes. Hence, F- could be easily recognized from other analytes
whatever they were cations or anions or observed by naked eyes or hand-held UV
lamp at 365 nm. In order to further estimate the resistant ability of IS-NR-F to other
analytes during F- detection, competitive experiments were carried out through adding
F- into the solution including different analytes directly (Fig. 6 (B)). It could be seen
that the other analytes nearly performed slight interference on F- detection. Therefore,
probe IS-NR-F could be employed to recognize F- specifically and colorimetry even
in the existence of lots of common ions.
20

Fig. 6 (A) Fluorescence responses of probe IS-NR-F (20 ꢀM) to various analytes
(25 equiv.) in a mixture of DMSO and PBS buffer solution (pH=7.4, 20 mM),
(1:1,v/v); (B) The competition experiment of IS-NR-F toward F- in the presence of
3-
2-
2-
2-
-
various analytes, including: 1:Cl-, 2:PO4 , 3:SO3 , 4:SO4 , 5:Ac-, 6:CO3 , 7:HCO3 ,
8:NO2 , 9:NO3 , 10:S2-, 11:Br-, 12:I-, 13:ClO-, 14:H2O2, 15:Al3+, 16:Hg2+, 17:Mn2+,
18:K+, 19:Zn2+, 20:Ba2+, 21:Mg2+, 22:Cr3+, 23:Ca2+, 24:Pb2+, 25:Ag+, 26:Fe3+, 27:Ni2+,
28:Cu2+, 29:F-; (C) and (D) were the colorimetric recognition of F- in different cations
and anions, respectively. All data were obtained after the reaction was conducted for
25 min at 37 ºC, λex=535 nm, λem=665nm.
-
-
3.6. Detection mechanism studies
In order to verify our thoughts, the detection mechanism was investigated.
According to the previous introduction, F- could selectively cleave the Si-O bond. To
21

ascertain whether the turn-on NIR fluorescence response was attributed to the F-
induced Si-O bonds cleavage, therefore, 1H NMR, ESI-MS and thin-layer
1
chromatography (TLC) were carried out. The H NMR spectrum was taken after
1
purifying product from the reacting system. H NMR spectrum of IS-NR-F, the
purified reaction product, and the synthesized IS-NR-OH were shown in Fig. 7.
1
When F- was added, comparing the H NMR spectra of IS-NR-F (Fig. 7a), fifteen
protons of IS-NR-F corresponding to the protons of tert-butyldimethylsilyl group
around 0.20 and 0.95 ppm disappeared (Fig. 7c), and a new proton peak emerged at
9.98 ppm, corresponding to the -OH. Especially, the purified product (Fig. 7c)
1
displayed almost the same H NMR spectrum with the synthesized IS-NR-OH (Fig.
7b). It indicated that the -OH emerged from the F- induced Si-O bonds cleavage. And
the detailed HR-MS analysis (Fig. S19) and TLC analysis (Fig. S20) further
illustrated that the product of this reaction was IS-NR-OH.
Fig. 7 Partial 1H NMR spectra in DMSO-d6: (a) only IS-NR-F, (b) the synthesized
22

IS-NR-OH, (c) the purified product after IS-NR-F reacted with F-
3.7. F- detection in aqueous sample and kinds of toothpaste
Given the widespread application of fluoride in our daily life, samples from tap
water and the filtered Xiangjiang River in Hengyang were collected. The detective
effects of probe IS-NR-F to F- were exhibited in Table 1. The collected and treated
water samples did not detect the presence of F- in the case of that there was no
additional F- added. When samples were spiked with 0.38, 0.76, 1.52 ppm F-
respectively, all of the recoveries were in the range of 100-105 % after measured five
times with the standard addition method. Besides, other ions in tap and river water
didn’t cause any obvious interference in the determination of F- due to the strong
affinity of F- toward silicon. In the view of these results, probe IS-NR-F could be
used to detect F- in the actual aqueous sample successfully and selectively. Next,
probe IS-NR-F was used to detect the content of F- in some toothpaste samples. The
tests of fluoride contents in toothpaste samples were listed in Table 2 and the
corresponding fluorescent spectra of these tests were shown in Fig. S21.
Table 1 F- detection in actual water samples
Sample
Added F- (ppm)
Detected (ppm)
Recovery (%)
RSD (%)
Tap water
0
0
-
-
0.38
0.76
1.52
0
0.39
0.78
1.55
0
102.63
102.63
101.97
-
2.28
2.17
2.30
-
Xiangjiang river
23

0.38
0.76
1.52
0.39
0.79
1.56
102.63
103.94
102.63
2.33
2.07
2.14
Table 2 The detection of F- in commercially available toothpaste
Sample
Brand
Jiajieshi
F- content (tested) mg/g
F- content (marked) mg/g
1
2
3
4
5
0.11
0.13
0.10
0.10
0.12
0.11
0.14
0.10
0.10
0.14
Colgate
Oral-B
Bamboo salt
Zhonghua
3.8. Cells imaging and MTT assays
Taking the advantages of NIR emission and large Stokes shift into consideration,
a standard cell viability protocols for investigating the biocompatibility of probe
IS-NR-F toward HeLa cells were carried out first by MTT assay. As shown in Fig.
S22, when the probe concentrations up to 50 ꢀM, at least 90% of cells remained in
good condition over 12 h. These results indicated that IS-NR-F had low cytotoxicity
when the concentration was less than 50 ꢀM, therefore, IS-NR-F was safe for the
following bioimaging applications. For the next imaging experiment, 50 percent of
HeLa cells were obtained after incubating an incubator for 12 hours, followed by the
addition of IS-NR-F (20 ꢀM contains 5 % DMSO) and incubated for 1 hour at 37 ºC.
24

And then the HeLa cells were washed with sterile PBS buffer solution for eliminating
the interference of excess IS-NR-F. Finally, the fluorescent imaging was recorded
after HeLa cells were treated with F- for another 1 hour at 37 ºC. As shown in Fig. 8,
there was a feeble fluorescence in the dark field when cells were merely treated with
probe IS-NR-F (Fig. 8 E, F, G), illustrating that IS-NR-F was accessible to combine
the F- in cells and not responsive to the inside cell species. However, an apparent
fluorescent enhancement was observed in the presence of F-, and IS-NR-F performed
an obvious red fluorescence that could decrease the background interference caused
by the cell itself (Fig. 8 I, J, K). In addition, a reverse process of incubate the cell
with fluoride first then add the probe were carried out to further confirmed the probe
is able to detect the fluoride in cells. As shown in Fig. S23, only in the presence of F-,
no fluorescence signal can be observed in the dark field, when the cells were
preincubated with F- and then incubated with IS-NR-F for 30 min, an obvious
enhanced red fluorescence can be clearly observed. Combining the characteristic of
low cytotoxicity and successfully imaging assays together, it demonstrated that the
probe IS-NR-F could be used for discerning F- in cells.
25

Fig. 8 Fluorescence images of untreated HeLa cells (A-D) and HeLa cells incubated
with 20 ꢀM of the IS-NR-F in the absence (E-H) and presence (I-L) of 3.8 ppm of
fluoride ion. Corresponding bright field images (A, E, I), fluorescence field (B, F, J),
merge images (C, G, K) and pseudo color (D, H, L) of HeLa cells were shown,
respectively.
4. Conclusion
In summary, we have successfully designed and synthesized
dicyanoisophorone derivative-based colorimetric and near-infrared fluorescent probe
IS-NR-F and an isophorone based control probe IS-F for the detection of fluorine ion
by employing F- selective cleavage of silicon-oxygen bonds. Fortunately, in the
presence of F-, the probe IS-NR-F exhibited a NIR fluorescence turn-on signaling at
665 nm with a remarkable Stokes shift (127 nm), which is more suitable for
a
bioimaging application. The probe exhibited a wide detection range and very low
26

detection limit for fluorine ion, as well as selectively recognized fluorine ion in actual
water samples including various ions present. Importantly, IS-NR-F has been
successfully applied to detect fluorine ion in actual water samples and toothpaste
extracted samples. Moreover, the probe toxicological experiments and near-infrared
fluorescence imaging indicated that IS-NR-F was cell-permeable with low
cytotoxicity and capable of real-time monitoring F- in living cells. Taken together, our
work provided a promising strategy for colorimetric and large Stokes shift
near-infrared fluorescent probe for the anions monitoring in both real samples and
living systems.
Conflicts of interest
There are no conflicts of interest to declare.
Acknowledgments
This work was supported by National Natural Science Foundation of China
(No.11804146 and 11875161); Natural Science Foundation of Hunan Province
(No.2016JJ5003); the Research Foundation of Education Bureau of Hunan Province
(No.17B223); College Students' Innovation & Entrepreneurship Training Project of
University of South China (No. 2018XJXZ332).
Reference:
[1] Dhillon A, Nair M, Kumar D. Anal. Methods. 2016; 8: 5338-5352.
[2] Beer P-D, Gale P-A. Angew. Chem. Int. Ed. 2001; 40: 486-516.
[3] Baker J-L, Sudarsan N, Weinberg Z, Roth A, Stockbridge R-B, Breaker R-R.
Science 2012; 335: 233-235.
27