A Comparison Study on Fluorescence Properties of Formononetin and Ononin

In this work, reasons for the spectral difference between two isoavones, Formononetin (F) and ononin (FG), are explained in the viewpoint of molecular structure through a comparison study of the uorescence features of the two. The uorescence enhancement of FG in hot alkaline condition is reported for the rst time. For F, there was almost no uorescence under acidic conditions, but when pH>5, its uorescence began to increase with increasing pH due to the proton ionization of 7-OH. In the range of pH 9.3-12.0, the anion form of F produced a fairly strong and stable uorescence with maximum excitation wavelength (λ ex ) of 334 nm and emission wavelength (λ em ) of 464 nm, its uorescence quantum yield (Yf) was measured to be 0.042. And for FG, its aqueous solution uoresced weakly in a wide pH range until it was placed under hot alkaline conditions, which was presumed to the cleavage reaction of the γ-pyrone ring in FG by observing a signicant uorescence at λ ex / λ em =288 / 388nm, and Yf was determined to be 0.020. The uorescence sensitization methods of F and FG both exhibit low limits of detection (2.60 ng·mL -1 , 9.30 ng·mL -1 ) and wide linear ranges (0.0117-1.86 μg·mL -1 , 0.0146-2.92μg·mL -1 ). Although the structural relationship between F and FG is glycoside and aglycone, FG cannot be translated to F by glucoside hydrolysis under hot alkaline condition, the uorescence enhancement mechanisms of the two are essentially different. The uorescence difference between the two under different experimental conditions lays the foundation for future uorescence quantitative analysis. reaction in and the spectrum of their reaction products. The possible existing forms (molecular form, ionic form, cleavage product or hydrolyzate) of the two were explained from the intrinsic characteristics of molecular structure and spectral information by analyzing the spectral information of uorescence wavelength and uorescence intensity under different experimental conditions. The results showed that the molecular form of F had no uorescence in the aqueous solution, and the anion form could exhibit strong uorescence due to the proton ionization of 7-OH under weak alkaline conditions. FG aqueous solution was weakly uorescent, but could produce strong uorescence by C ring cleavage reaction in strong alkaline solution. These results provide some new experimental basis for the establishment of uorescence analysis methods of F and FG, and also open up a new way to expand the uorescence analysis of isoavones.


Introduction
Iso avones are a class of compounds derived from iso avone (3-phenylchromone), which widely exist in foods and drugs, and are also the active ingredient of many traditional Chinese medicine (TCM).
Eva de Rijke et al. had tested the native uorescence of 19 kinds of avonoids, observed that only some avonols and iso avones, such as, daidzein, formononetin and ononin had uorescence [14]. They studied the effects of pH and different solvents on the uorescence of F and FG, but did not discuss the relationship between uorescence properties and their molecular structure. The authors had studied the uorescence properties of some avonoids and found that some avonoids and avonoids will undergo uorescence enhancement reactions under hot alkaline conditions, which is attributed to the cleavage reaction in the C ring [15,16]. These methods and ideas have expanded the scope of uorescence analysis of avonoids to some extent.
In this paper, based on the existing research results, uorescence properties of F and FG, and the relationship with the molecular structure were studied. Focused on the effect of pH on the uorescence spectrum, the authors studied the reaction of C ring cleavage reaction or oxygen glycosides hydrolysis reaction in hot alkaline conditions and the uorescence spectrum of their reaction products. The possible existing forms (molecular form, ionic form, cleavage product or hydrolyzate) of the two compounds were explained from the intrinsic characteristics of molecular structure and spectral information by analyzing the spectral information of uorescence wavelength and uorescence intensity under different experimental conditions. The results showed that the molecular form of F had no uorescence in the aqueous solution, and the anion form could exhibit strong uorescence due to the proton ionization of 7-OH under weak alkaline conditions. FG aqueous solution was weakly uorescent, but could produce strong uorescence by C ring cleavage reaction in strong alkaline solution.

Apparatus
Fluorescence measurements were performed on a Hitachi (Tokyo, Japan) F-7000 uorescence spectrophotometer equipped with a xenon lamp and 1 cm quartz cell. The excitation and emission slits (band pass) 5 nm/5 nm were used throughout the work. Absorption spectra were recorded using a Shimadzu (Kyoto, Japan) UV-2501PC recording spectrophotometer with 1 cm quartz cell. An Orion (Beverly, USA) 868 pH/ISE meter was used for pH measurement. Five-digit analytical balance, up to 0.01mg.

Methods for spectral measurement
Effects of pH on uorescence: Added F or FG solution into a series of 10 mL volumetric asks, added B-R buffer solution and NaOH solution to adjust different pHs ,then diluted to volume with water and mixed well. After setting aside for 20 min, the two-dimensional uorescence spectrum was measured at room temperature.
Effects of solvent on F: Determined pH = 9.3, F solution and different volumes (1.0, 1.5, 2.0, 2.5, 3.0,3.5, 4.0, 4.5 mL) of MeOH were added into 10 mL volumetric asks, separately. The mixtures were diluted to the mark with water and mixed well. After setting aside for 20 min, the two-dimensional uorescence spectrum was measured at room temperature.
Effects of temperature on uorescence: Prepared series of 10 mL volumetric asks containing 1.0mL FG and 2.0mL NaOH solution. The mixtures were diluted to the mark with water and mixed well, then set for a 30min at room temperature (about 20 ℃), 40 ℃, 50 ℃, 60 ℃, 70 ℃, 80 ℃, 90 ℃ and 100 ℃, respectively. Then scanned the spectrum after cooling to room temperature.
Effects of heating time on uorescence: Added 1.0 mL FG solution and 2.0 mL 1.0 M NaOH solution into 10 mL volumetric asks. The mixtures were diluted to the mark with water and mixed well, then heated in a 100℃-water bath. Took out one every 10 minutes, then cooled to the room temperature and added water to the scale, scanned the two-dimensional uorescence spectra, and investigated the in uence of heating time on uorescence.

Measurement of uorescence quantum yield
The uorescence quantum yield (Yf) was mesured according to the reference method [17], selected quinine sulfate (Yf = 0.58 in 0.1 M H 2 SO 4 ) and L-Tryptophan (Yf = 0.13 in water) as the standards separately [18]. Appropriate quantities of quinine sulphate solution were prepared into a 25 mL ask and diluted to the scale with 0.1 M H 2 SO 4 , mixed well. Add a moderate amount of L-Tryptophan into a 25 mL ask and diluted with water. Prepared suitable amounts of F or FG into a 25 mL ask, diluted to the mark with water and mixed well. Recording absorption and uorescence spectra, calculated quantum yields.
The relative Yf of the sample is represented by Y u , and the corresponding Y s represents the known Yf of the standards. The calculation formula is: Wherein, A is the absorbance of the solution, F is the integrated area of the emission peak, s is the standard, and u is the sample. The prerequisite for the application of this formula is that the solubility of the solution should not be too large, and it is generally controlled that the absorbance is close to and not more than 0.05.

Fluorescence spectrum of F aqueous solution
The uorescence spectra of F aqueous solution at different pHs were measured, as shown in Fig. 1. F hardly emitted uorescence in the range of pH 2.0-5.0. The uorescence peak of F was signi cantly enhanced at the excitation wavelength (λ ex ) of 256 nm and 334 nm and the emission wavelength (λ em ) of 464 nm in weak acidic and weak basic conditions (pH > 5.0). Fluorescence intensity would reach its peak and remain stable in pH range of 9.3-12.0, and then quench rapidly as the pH continued to increase (pH > 12.0), but the wavelength of excitation and emission remained the same throughout the progress.
The reason why the uorescence spectrum changed with pH is that F molecules contain ionizable hydroxyl protons. There is a hydroxyl group at position 7 (7-OH) in the F molecule as shown in Scheme 1. The proton dissociation of 7-OH caused the change of uorescence spectrum when the pH of the solution changed from near neutral to weakly alkaline. The ionization constant of 7-OH proton was determined to be pK a = 7.31 ± 0.03 using the pH-F data from the Fig. 1(c) and based on pH-Fluorescence method [19].
UV absorption spectra of F at different pH conditions were studied to verify the 7-OH proton ionization of F, and the results are shown in Fig. 2. In the pH range of 2.2-11.2, three absorption peaks appeared in the absorption spectra, which were located at 255, 303 and 334 nm, respectively. As the pH increased, the absorption peaks of 255 and 334 nm increased, while the absorption peak at 303 nm decreased, in this way, three isochromatic points located at 245, 285 and 312 nm were formed. The spectral characteristics showed that the molecular form of F changed into ionic form under weak alkaline conditions. According to pH-A data in Fig. 2 (b), the 7-OH proton ionization constant of F (pK a ) was calculated by pHspectrophotometry [20], the result pK a =7.34 ± 0.01 was consistent with the pK a value obtained by the uorescence method. In addition, there were isosbestic points in Fig. 2, but no iso uorescent points in Fig. 1, which showed that both the molecule and ionic form of F can absorb light in the ultraviolet region, but only the ionic form can produce uorescence.
As shown in Fig. 1, the uorescence intensity of F increased in the pH regions of 6.0-10.0 and reduced rapidly when pH > 12.0. This was due to the cleavage reaction of pyrone ring in F with product of the ohydroxyl-phenyl-benzyl ketone derivative [21], as exhibited in Scheme 2. This product had no uorescence under experimental conditions.

Fluorescence spectrum of FG and its Cleavage product
Different from F, the uorescence of FG was very weak in the acidic, neutral or weak alkaline aqueous solutions at room temperature (only weak uorescence of FG was observed in the literature [14]). The absence of ionizable hydroxyl protons, and the presence of 7-position oxoside (7-OGlu) prevents FG from producing uorescence like F, therefore, the spectral properties did not change in varying pHs .
But in the experiment, we found that FG aqueous solution would produce strong uorescence when heated under alkaline conditions, as shown in Fig. 3. The uorescence intensity of the FG aqueous solution was still very weak after heating in the pH range of 1.0-8.8, but when pH > 8.8, the uorescence peak located at λ ex /λ em = 243 / 388 nm enhanced signi cantly, until the uorescence intensity reached the maximum at pH 12.5, then dropped sharply as the pH increased (pH > 12.5).
As the experimental results shown in Fig. 3, we can conclude that, chemical changes must have occurred in FG and produced new uorescent substances under hot alkaline conditions. According to the molecular structure of FG, it can infer that there are two possible reactions: one is the ring-opening reaction of γ-pyrone ring, as shown in Scheme 2, and the other was the hydrolysis reaction of 7-OGlu.
We inferred which kind of chemical reaction happened in FG combining with the uorescence characteristics of the reaction product. If the hydrolysis reaction of the 7-oxo-glycosidic bond had occurred, the reaction product should be like the anionic form of F after 7-OH proton ionization, as the uorescence spectrum was shown in Fig. 1, the λ ex is located at 256 nm and 334 nm, and the λ em is located at 464 nm. However, the λ ex (243 nm, 288 nm) and λ em (388nm) of the actual product shown in Fig. 3 were signi cantly different from the wavelength position in Fig. 1. These results indicated that no hydrolysis reaction occurred under hot alkaline conditions. Another speculation, if FG had undergone the ring-opening reaction of the γ-pyrone ring, the cleavage product should has a lower degree of conjugation in the molecular structure compared to the anionic form of F, and moreover, should has shorterwavelength uorescence peaks according to the theory of uorescence, the theoretical speculation was in good agreement with the results shown in Fig. 3. Therefore, it could be inferred that the ring-opening reaction of the γ-pyrone ring occurred under hot alkaline conditions, and the cleavage product is a uorescent. But due to the hydrolysis reaction of the cleavage product, its uorescence quenched rapidly with the increased pH when pH > 12.5, the hydrolyzate was the same as the cleavage product of the anion form of F in Fig. 2.
In summary, the uorescence properties of F and FG were quite different, although the two are the relationship between glycoside and aglycone from the molecular structure. In the experiment, FG cannot be converted to F by the cleavage of the glycosidic bond under hot alkaline conditions, instead, the γpyrone ring cleavage reaction occurred rst to produce uorescence, and then the glycosidic bond hydrolysed under strong alkaline conditions, resulting in uorescence quenching.

Effect of Solvent on Fluorescence of F
The effect of solvent (the volume fraction of methanol in aqueous solution) on the uorescence of F was investigated under weak alkaline conditions (pH 9.3). The results in Fig. 4 show that when the volume fraction of methanol in the aqueous solution was changed, the uorescence wavelength was almost unchanged, but the uorescence intensity was signi cantly enhanced. Therefore, in this study, the volume fraction of methanol in the aqueous solution was controlled to 10% or less when studying the effect of other experimental conditions on the uorescence.

Effect of Heating Temperature on Fluorescence of FG
As shown in Fig. 5, uorescence of FG is extremely weak under the strong alkaline solution (pH 12.5) at room temperature, then increases along with the heating temperature increases, indicating that the temperature accelerated the cleavage reaction. When the temperature is greater than 70°C, the uorescence reaches the maximum, showing that the cleavage reaction is basically completed. The uorescence spectra were similar to those in Fig. 3 (a, b), and had no changes during the entire temperature-changing progress.

The stability of uorescence
The experiments revealed that, the uorescence of F was basically stable when placed in weakly alkaline solution at room temperature, and also had no changes when the solution was continuously irradiated by the xenon lamp.
The uorescence intensity of FG increased with the extension of heating time and became stable after 1.5 h, and did not change signi cantly when continuously exposed to the xenon lamp for 200s, indicating that the uorescence properties were basically stable.

Fluorescence Quantum Yield
Yf of F in a weak alkaline solution (pH 9.3) was measured to be 0.042, and that of FG heating pyrolysis product in alkaline solution (pH 12.5) was 0.020. Though the values of Yf were not very high, we can analysis them using modern uorescent instruments with high sensitivity.

Relationship between concentration and uorescence intensity
Based on the above experimental results, a series of solutions containing different amounts of F under pH = 9.27, from 0.0117 µg·mL − 1 to 1.86 µg·mL − 1 , were prepared and their uorescence spectra were scanned, as shown in inset in Fig. 6(a). A working curve of uorescence intensity, I F (λ ex /λ em = 334 nm / 464 nm), versus concentration of F, c F , was drawn, as shown in Fig. 6(a). In the range of 0.0117-1.86 µg·mL − 1 , I F has a linear relationship with c F . The regression equation is I F = 21.9 + 2188.8c, with the correlation coe cient R 2 = 0.999 (n = 14). The blank signal was scanned and the lowest limit of detection for F was found to be 2.60 ng·mL − 1 (9.69×10 − 9 mol·L − 1 ).
Similarly, Fig. 6(b), a working curve of I F (λ ex /λ em = 288 nm / 388 nm) versus c FG was drawn by controlling the pH of the solution to 12.5, heating in a boiling water bath for 1 hour, and scanning the uorescence spectra (as shown in inset of Fig. 6b) after cooling to room temperature. The results show that the linear relationship between I F and c FG is good in the range of 0.0146-2.92µg·mL − 1 , the regression equation is I F = 18.9 + 611.6c, R 2 = 0.998 (n = 15), and the lowest limit of detection for FG was found to be 9.30 ng·mL − 1 (2.16×10 − 8 mol·L − 1 ).
It can be seen from the above results that when a sample contains both F and FG, although the structures of the two compounds are similar, the difference in uorescence between the two can be used for quantitative analysis, such as controlling the sample solution to weakly alkaline conditions for the analysis and determination of F, and the sample solution can be placed under strong alkali heating conditions when determining the content of FG.

Conclusions
In summary, F is weakly uorescent under acidic and neutral conditions, and its weakly basic solution can produce uorescence at λ ex /λ em =339/465 nm, which due to the proton ionization of 7-OH. Under strong alkaline conditions, the cleavage of the γ-pyrrole ring leads to uorescence quenching. Since there is no ionizable protons in the molecular structure of FG, its uorescence in aqueous solutions is quite weak under normal temperature conditions and regardless of the pH of the solution. But under hightemperature alkaline conditions, FG undergoes γ-pyrone ring cleavage reaction, resulting in increased uorescencethe with λ ex /λ em =288/388 nm. When heated under strong alkaline conditions, the pyrolysis products of FG will further undergo glycosidic bond hydrolysis, causing uorescence quenching. Although the structural relationship of F and FG is glycoside and aglycon, in the experiment, the glycoside will not be converted into the aglycone, so the uorescence enhancement mechanisms of the two are essentially different.

Con icts of interest
There are no con icts to declare.

Ethics approval
Not applicable.

Consent to participate
Not applicable.

Consent for publication
Not applicable.

Data availability
The data that support the ndings of this study are available from the corresponding author on request. Signed by all authors as follows: Figure 1 Fluorescence spectra (λex/λem: 334/464 nm) of F aqueous solutions (0.972 μg·mL-1) at different pHs (a, b, c) and relationship between uorescence intensity and pH(d).  In uence of methanol on uorescence intensity (λex/λem: 334/464 nm) of formononetin(0.972 μg·mL-1, pH 9.3).