Chemical basis underline the bioactivity fortication of licorice by honey-frying due to the natural deep eutectic solvent characteristics of honey

Honey has been widely used in medicine plants processing since ancient times. Honey-frying of herbal medicine is a well-known treatment in Chinese medicine. It has been reported that honey-frying can improve the immunomodulatory ecacy of licorice. However, it is still unknown why honey-frying improves the immunomodulatory activity of licorice. And our previous research demonstrated that honey has natural deep eutectic solvent (NADES) characteristics. NADES is a kind of solvent with broad polarity range and has attracted extensive attention of scholars in different elds. In the present study, we investigated chemical basis underline the possible potentiation of honey-frying on licorice to elucidate its chemical mechanism. at 1,028 cm 1 observed. The deformation vibration absorption bands of the O-H (δ O−H ) in and fructose 915 to to 919 1 in with broaden peak shape. These show that the hydroxyl groups in the glucose and fructose of NADES have inter-molecular interactions including hydrogen bonds. The sharp peaks of the deformation vibration absorption bands of the C-H (δ C−H ) in glucose shifted from 1,380 cm − to a broad peak at 1,344 cm in NADES. In several sharp peaks of the stretching vibration absorption bands of C–H (ν C−H ) in and between 2,900 cm and 3,000 DMSO, dimethyl sulfoxide; DP, declustering potential; DSC, differential scanning calorimetry; ESI, electrospray ionization; FBS, fetal bovine serum; FT-IR, Fourier transform infrared; heat electrospray ionization; UHPLC, ultrahigh-performance liquid chromatography; of lipopolysaccharide; multiple reaction monitoring; mass spectrometry; NADES, natural deep eutectic nuclear OD, optical density; partial PBS, phosphate-buffered saline; PCA, component QC, RH, relative SD, standard deviation; UHPLC, ultrahigh-performance liquid chromatography; Traditional VIP, in

water. The decoctions were stored at -80 °C before use. The decoctions were resuspended by shaking before use in the cell-based bioactivity tests.

Immunological experiment
Animals and experimental design Male ICR mice (18-22 g) were purchased from SPF Biotechnology Co. Ltd. (Beijing, China). They were housed in a speci c pathogen-free grade laboratory at room temperature (20-25 °C) and constant relative humidity (RH) under a 12-h light/12-h dark cycle. They had ad libitum access to food and distilled water. They were allowed to acclimate to their new environment for 3 d prior to the experiments. They were weighed and randomly divided into 10 groups (n = 6 per group), namely, normal, model, high-dose raw licorice (7.5 g/kg), low-dose raw licorice (5.0 g/kg), high-dose fried licorice (7.5 g/kg), low-dose fried licorice (5.0 g/kg), high-dose honey-processed licorice (7.5 g/kg), low-dose honey-processed licorice (5.0 g/kg) groups, high-dose NADES-processed licorice (7.5 g/kg), and low-dose NADES-processed licorice (5.0 g/kg) groups. Mice in group 2-10 were intragastrically (i.g.) administrated with rhubarb extract (15 g/kg, once a day) for 7 d to induce Pi-de ciency model. Then, the mice in group 1 (normal) and group 2 (model) were administered (i.g.) sterile physiological saline once daily for 7 d. Mice in the raw, fried, honey-processed, and NADES-processed licorice groups were orally administered decoctions at 7.5 mL/kg body weight (BW) once daily for 7 d. All procedures were conducted in strict accordance with the Guide for the Care and Use of Laboratory Animals of the Ministry of Science and Technology of China

Sample collection and measurement
Twenty-four hours after the nal administration, the mice were weighted and then injected in the cauda with India ink (0.01 mL/g BW). Twenty microliters blood were taken from the eyepit vein plexus at 2 min and 20 min after the India ink injection. The blood was placed in a centrifuge tube containing 2 mL of 1% (w/v) sodium bicarbonate for 1 h. The absorbances (optical density [OD]) of the mice sera were measured at 650 nm by Thermo Scienti c Multiskan GO Multiskan Spectrum (Thermo Fisher Scienti c, Waltham, MA, USA). Thirty minutes after the India ink injections, Mice were euthanized, then the spleen and thymus were excised and weighed, Data analysis The charcoal particle expurgation index (K) of the macrophage count was calculated as follows: where, t 1 and t 2 are the time points (2 min and 20 min, respectively) at which blood was drawn from the mouse eyes after the India ink injection.
The spleen index was calculated as follows [2]: Data were expressed as means ± SD. Differences between groups were evaluated with SPSS v. 25.0 (SPSS Inc., Chicago, IL, USA). P < 0.05 was considered statistically signi cant.

Biological activity testing
The estrogen receptor (ER)-α, nuclear factor erythroid 2-related factor (Nrf2), and toll-like receptor 4 (TLR4) promoter assays were conducted on the HEK293 cells. An appropriate licorice extract concentration was added to the pER-α-Luc HEK-293 cells, and the suspension in a humidi ed CO 2 (5%) incubator at 37 °C was incubated for 24 h. Then, 17-β-estradiol (Sigma-Aldrich Corp.) was dissolved in dimethyl sulfoxide (DMSO) and used as a positive control. The medium was aspirated before adding 0.5 mL lysis buffer (pH 7.8) to each well for 10 min. Luciferase assay buffer (0.1 mL/well) was added, and the luciferase activity was immediately determined. Estrogenic activity was calculated as follows [20]: For the TLR4 promoter assay, an appropriate licorice extract concentration was added to the pTLR4-Luc HEK-293 cells, and they were incubated for 4 h at 37 °C. Lipopolysaccharide (LPS) (Sigma-Aldrich Corp.) dissolved in DMSO was the positive control. For the Nrf2 promoter assay, an appropriate licorice extract concentration was added to the pNrf2-Luc HEK 293 cells, and they were incubated for 16 h at 37 °C. Andrographolide (20 μM) dissolved in DMSO was the positive control. The other steps followed were the same as those used for the ER-α assay.

Total polysaccharide and avonoid content measurement
The levels of total polysaccharides and flavonoids in the decoctions were determined by UV-Vis spectrophotometry [21]. The total polysaccharides were detected at 590 nm by the anthrone-sulfuric acid method using anhydrous glucose as a reference [22]. Liquiritin was the reference, as it is the main dihydro avone in licorice. All major avonoids in licorice have maximum absorptions at ~336 nm.

Metabolomics analysis
Non-target metabolomics analysis with UHPLC-Q-Orbitrap MS The licorice decoctions were thawed, diluted 20,000-fold with methanol, and ltered through a 0.22 μm membrane before analysis on an Ultimate 3000 ultrahigh-performance liquid chromatography (UHPLC) instrument coupled with a Thermo Q-ExActive TM Plus Orbitrap TM high-resolution mass spectrometer (Thermo Fisher Scienti c) tted with a heat electrospray ionization (HESI) interface. Samples were separated on an Acquity UPLCHSS T3 C18 column (2.1 mm × 100 mm; 1.8 μm; Waters Corp., Milford, CT, USA). The column temperature was maintained at 40 °C. The ow rate was 0.30 mL/min. Mobile phase A was 0.1% (v/v) formic acid-acetonitrile, and mobile phase B was 0.1% (v/v) formic acid. The gradient program was as follows: 0-1 min, 98% B; 1-2. The raw mass data were pre-processed with Progenesis QI v. 1.0 (Waters Corp., Milford, CT, USA) for peak alignment and selection as well as deconvolution. Individual ion fragment intensities were normalized for all compounds. The data MS matrices were imported into SIMCA-P v. 13 (Umetrics, Umeå, Sweden) for principal component analysis (PCA) by Pareto scaling to identify group distributions, perform orthogonal partial least square discriminant analysis (OPLS-DA), and identify the potential quality difference markers among groups by variable importance in projection (VIP). One-way ANOVA was conducted on the ion response strength data of the potential mass difference markers using GraphPad Prism v. 7.0 (GraphPad Software, La Jolla, CA, USA) to identify any signi cant differences among potential mass difference markers. The screened quality difference markers were identi ed by chromatographic comparison against reference substances, laboratory databases, and literature reports [6, 23,24].
Quantitative analysis with UHPLC-QqQ-MS Licorice decoctions were thawed, diluted 20,000-fold with methanol, and ltered through a 0.22 μm membrane before quantitative analysis. The assay was performed on a 6500 Plus Triple Quad LC-MS/MS system tted with an ExionLC UHPLC unit (AB SCIEX Corp., Framingham, MA, USA). Sample components were separated on an Acquity UPLCHSS T3 C18 column (2.1 mm × 100 mm; 1.8 μm; Waters Corp.). The column temperature was maintained at 40 °C, and the ow rate was 0.30 mL/min. Mobile MS data were recorded using electrospray ionization (ESI), negative ion detection, and multiple reaction monitoring (MRM) scanning. The ion spray voltage and temperature were set at 4,500 V and 500 °C, respectively. The curtain, gas 1 (nebulizer), and gas 2 (heater) gas pressures were set to 35 psi, 50 psi, and 50 psi, respectively. The collision gas pressure was set to 9 psi. The compound dwell time was set to 10 ms. The entrance potential and collision cell exit potential were set to 10 V and 18 V, respectively. Data were acquired with Analyst v. 1.6.3 (AB SCIEX Corp., Framingham, MA, USA). The MS/MS parameters for the 18 compounds are listed in Table S1. The method was validated (Supplementary Material).

Pharmacokinetic experiments Drug administration and sample preparation
Male Sprague-Dawley (SD) rats (220-250 g) were obtained from HFK Bioscience Co. Ltd. (Beijing, China). The rats were bred at 25 ℃, 60 ± 5% RH, and 12-h dark-light cycle for 3 d. They had ad libitum access to water and normal chow. All animals were fasted overnight before the experiments. The rats were randomly divided into four groups (n = 6 per group). The raw licorice decoction (R; 1.0 g/mL raw licorice equivalent), the fried licorice decoction (F; 1.0 g/mL raw licorice equivalent), the honey-fried licorice decoction (H; 1.0 g/mL raw licorice equivalent), and the NADES-processed licorice decoction (N; 1.0 g/mL raw licorice equivalent) were orally administered to the rats at 8 g/kg. Blood samples 0.3 mL in volume were collected from the angular vein into heparinized tubes at 0.08 h, 0.25 h, 0.5 h, 1 h, 2 h, 4 h, 6 h, 8 h, 12 h, and 24 h after oral decoction administration. They were immediately centrifuged at 12,000 rpm at 4 ℃ for 3 min, and the plasma was stored at -80 °C until use.
Calibration standard (CS) and quality control (QC) samples were prepared (Supplementary Materials). One hundred microliters of each plasma, CS, or QC sample was spiked with 300 μL acetonitrile (including rutin and oleanolic acid; 100 ng/mL). The mixture was vortexed for 3 min and centrifuged at 12,000 rpm and 4 ℃ for 10 min. The supernatant was collected and evaporated under a nitrogen stream in a 37 ℃ water bath. The residue was redissolved in 100 μL acetonitrile, vortexed for 3 min, and centrifuged at 12,000 rpm at 4 ℃ for 10 min.

UHPLC-QqQ-MS analysis
The assay was performed on a 6500 Plus Triple Quad LC-MS/MS system tted with an ExionLC UHPLC unit (AB SCIEX Corp., Framingham, MA, USA). Sample components were separated on an Acquity UPLCHSS T3 C18 column (2.1 mm × 100 mm; 1.8 μm; Waters Corp.). The column temperature was maintained at 40 °C, and the ow rate was 0. The declustering potential (DP) were set at -80 V. The ion spray voltage and temperature were set at 4,500 V and 500 °C, respectively. The curtain, gas 1 (nebulizer), and gas 2 (heater) gas pressures were set to 35 psi, 50 psi, and 50 psi, respectively. The collision gas pressure was set to 9 psi. The compound dwell time was set to 10 ms. The entrance potential and collision cell exit potential were set to 10 V and 18 V, respectively. Data were acquired with Analyst v. 1.6.3 (AB SCIEX Corp., Framingham, MA, USA). The method was validated (Supplementary Material).

Statistical analysis
The pharmacokinetic parameters were calculated with Drug Analysis System v. 3.2.6 (Mathematical Pharmacology Professional Committee of China). Noncompartmental analysis was used to determine the maximum time (T max ), maximum plasma concentration (C max ), and area under the curve (AUC 0-12/24h ). Data were means ± SD. Statistical analyses of all data were performed with one-way ANOVA (SPSS v. 25.0; SPSS Corp., Chicago, IL, USA).

Molecular interaction determination of honey and its mixture with licorice compounds Sample preparation
Honey and corresponding NADES were obtained as described in Subsection of sample preparation. A fructose solution was obtained by dissolving 2 g fructose in 0.5 mL water to ensure that the nal water content was the same as that in NADES. Glucose hydrate was prepared in the same manner as the fructose solution. The mixture of liquiritin and NADES was obtained by dissolving 0.12 g liquiritin in 1 mL NADES and stirring thoroughly.

Effects of licorice on mouse immune function
The effects of aqueous decoctions of raw licorice (R), fried licorice (F), honey-fried licorice (H), and NADES-fried licorice (N) on murine immunity were evaluated using a carbon clearance test according to the reference [2] with some modi cation. Compared with the normal group, mice in other groups showed obvious behavioral abnormalities after 5 days of intragastric administration of rhubarb decoction, including weight loss, dry and dull hair, listless expression and thin stool, which indicates that the Pide ciency has been successfully built. Decoctions of honey-fried licorice and NADES-fried licorice both increased the phagocytic, spleen and thymus indices of Pi-de ciency mice, while in the F group, only the high dosage group was effective in improving the expurgation indices of Pi-de ciency mice.
There were no differences between honey-fried licorice and NADES-fried licorice. Compared with the highdose raw licorice group, the expurgation and spleen indices of H and N groups at the same dosage were signi cantly higher (P < 0.05) ( Table 1), indicating that the immune enhancement effect of honey-fried and NADES-fried licorice are better than fried licorice. Honey alone had no immunostimulant effect in mice [2,3]. These ndings indicate that honey-processing and NADES-processing could enhance the licorice immunomodulation action. Table 1 The expurgation index, spleen index and thymus index of mice after oral administration of physiological saline, licorice decoction, fried licorice decoction, honey-fried licorice decoction and NADES-fried licorice decoction for 7 days. (mean ± SD, n = 6) Honey-frying improved licorice bioactivity in certain signal transduction pathways To investigate the effects of honey on licorice immunomodulation, the biological effects of R, F, H, and N on the endocrine-related ER-α, TLR4, and Nrf2 signaling pathways were compared. The bioactivities increased with H and N compared with those of the R. There were no differences between H and N. However, both showed signi cantly higher activity than F in ER-α and Nrf2 signaling pathways (Fig. 1).
Hence, NADES-frying and honey-frying substantially improved licorice bioactivity by a mechanism common to both treatments.
Effect of honey-frying on the dissolution of total avonoids and polysaccharides Figure 2a shows that there was higher total avonoid content in H and N than in R. In contrast, F had relatively lower total avonoid content. There was higher total polysaccharides content in F, H, and N than in R. There were higher levels of total polysaccharides in H and N than in F. Therefore, frying without honey or NADES decreased total avonoid content and increased total polysaccharides content in licorice. Honey and NADES avoided the total avonoid loss caused by frying, and even increased the level of total avonoid compared with R. H and N did not signi cantly differ in terms of total polysaccharide or avonoid content.
Effects of honey-frying on the chemical composition of licorice decoctions Different chemical pro les observed by non-targeted metabolomics The licorice decoctions were analyzed by UHPLC-QE-Orbitrap-MS. PCA of the MS data showed three distinct groups, representing R, F and N plus H as one group (Fig. 2b). Processing altered the chemical composition of the licorice. Moreover, the chemical composition of H was signi cantly different from that of F but resembled that of N. Thus, heating and honey signi cantly affected the chemical pro le of the licorice decoction, while honey-frying and NADES-frying had similar effects.

Quantitative analysis by LC-QqQ-MS
Eighteen potential quality markers were identi ed by analyzing the UHPLC-Q-Orbitrap MS data by OPLS-DA (Table S1). Then, LC-QqQ-MS was used to perform targeted quantitative analyses on twelve markers with reference substances. It was also applied towards pseudo-target quantitative analysis of six markers without reference substances.
The levels of the saponins glycyrrhizic acid (7), uralsaponin B (17), and 22β-acetoxylglycyrrhizic acid (18) decreased in F, H, N ( Fig. 3a and S1). However, the glycyrrhetinic acid (8) content was signi cantly increased after processing (F, H, N). Compared to control R, there was a greater decrease in the glycyrrhizic acid (7) content in F than in H or N (Fig. 3a). Obviously, the reducing sugars D-fructose (10), and D-glucose (11) were the major compounds in the honey and NADES processed samples, 5hydroxymethylfurfural (12) was formed in connection with the heating.
Effects of honey-frying on the absorption of representative chemical components in licorice A detection method for plasma liquiritin apioside (1), liquiritin (2), liquiritigenin (3), and glycyrrhetinic acid (8) was established according to the reference [25] with some modi cation and validated.  (3), could not be detected after 24 hours of oral administration.
Considering the measurement results of the levels of the major compounds in the plasma of mice ( Fig. 4 and Table S7), it seems that for the major compounds measured (1, 2, 3, and 8), the avonoids 1 and 3 showed higher plasma levels in the presence of the honey or the NADES (Fig. 4), the same applies for glycyrrhetinic acid (8). For the F samples no increase of the plasma level was observed, in fact the curve is identical with the control (Fig. 4). replaced by a broad band around 2,928 cm − 1 . These spectral characteristics of NADES reveal its supramolecular structure with multiple interactions between the glucose and fructose molecules.
FT-IR spectra were recorded for NADES, liquiritin, and a mixture of liquiritin and NADES to clarify the mechanism by which NADES affects the solutes. The spectra (Fig. 5b)

Discussion
Honey-frying enhances the immunostimulatory e cacy of licorice The pharmacological experiments con rmed that honey enhances the immunomodulatory e cacy of licorice by increasing macrophage phagocytosis and improving spleen and thymus function. Cell-based bioactivity experiments showed that the health-improving function of licorice was associated with its in uences on the signal transduction pathways ER-α, TLR4, and Nrf2. The forti ed immunomodulatory e cacy of licorice by honey-frying might be contributed to the improved content of total avonoids and polysaccharides (Fig. 2). Polysaccharides enhance phagocytic activity in macrophages, promote immune organ development [26]. Metabolomics analysis (Fig. 3a) showed that honey-processing increased chalcone (isoliquiritin apioside (4), isoliquiritin (5), and O-acetylated derivatives (14, 16)) levels which, in turn, increase the lymphocyte and macrophage abundance and CD4+/CD8 + cell ratios in the blood and regulate human immunity [27,28]. These results provide clear biological and chemical evidence for the relatively increased immunostimulatory e cacy of honey-processed licorice. These ndings provide evidence for the traditional Chinese medicine theory about the processing of licorice, which says that honey-fried licorice "invigorates the spleen and qi (energy)".
NADES characteristics of honey fortify the health-improving functions of honey-processed licorice.
Our former studies proved that the major components of honey form a natural deep eutectic solvent composed of glucose, fructose, and sucrose and certain amount of water [14,18]. Hence honey does show the characteristic of NADES, including improving the dissolution and extraction of constituents, improving their thermal stability and bioavailability. The increased level of total avonoids and polysaccharides in Fig. 2a prove that both honey and NADES enhance the dissolution and extraction of the active ingredients in the nal licorice decoction, indicating that this effect of honey is due to its NADES characteristics. The total avonoid content decreased in fried licorice and increased in honey/NADES-fried licorice (Fig. 2a) indicating honey/NADES may inhibit heat-induced decomposition of the avonoids and at the same time increase the extraction of the avonoids.
The protective effect of honey/NADES on the thermal stability of chemical constituents in licorice was veri ed in some single compounds in licorice ( Fig. 3a and S1). Dihydro avones (1 and 2) may have been isomerized to chalcone (compounds 4 and 5) due to the high temperature during processing (Fig. 3) [2, 5, 6], however, honey and NADES might suppress this transformation. Similar effect of honey /NADES on the stability of compounds was also observed. The observed decrease in glycosylated dihydro avone level (1 and 2) after frying indicates that glucoside bonds may break during processing (Fig. 3) [2, 5, 6]. The relative levels of the isoliquiritigenin (6) differed from those of glycosylated chalcone (4,5,14,16) among treatments. In contrast, the level of 2,4-dihydroxyacetophenone (9) was increased considerably by the frying, similar to the glycosylated chalcone (4) and to a less extend, the related compounds 5, 14, 16 ( Fig. 3a and S1). Isoliquiritigenin (6) might degrade to 2,4-dihydroxyacetophenone (9) (Fig. 3b), whereas, honey/NADES may inhibit this heat-induced transformation. The major saponins in licorice is glycyrrhizic acid (7). The relative lower level of glycyrrhizic acid (7) and higher level of glycyrrhetinic acid (8) in F than R (Fig. 3a) indicates that glycyrrhizic acid (7) may be degraded into glycyrrhetinic acid (8) during frying [2], whereas honey or NADES could inhibit this decomposition since higher level of glycyrrhizic acid (7) was observed in H and N.
The effect of honey/NADES on the bioavailability of constituents in honey-fried licorice were demonstrated in Fig. 4 and Table S7. A pharmacokinetic analysis (Table S7)  Mechanism for the effect of honey on licorice The FT-IR measurements show that NADES and honey have similar chemical structures (Fig. 5). This nding con rms those of previous 1 H NMR and DSC analyses that NADES and honey had similar supramolecular structures [15,18]. The FT-IR measurements demonstrated that NADES interacts with liquiritin (2). This discovery is in agreement with reported molecular interactions including hydrogen bonds between quercetin and NADES [17]. Interactions between the solutes and NADES account for the ability of NADES and honey to stabilize and solubilize the bioactive constituents of licorice as well as improve their bioavailability, which result in enhanced biological activity of honey/NDES-fried licorice.
The improved bioactivities of honey-processed and NADES-processed licorice decoctions were con rmed, which is attributed to their NADES characteristics. Honey affects the solubility of the compounds in the plant material. As a rule, solubility of the aforementioned compounds is much higher in supramolecular liquids such as NADES than in water [14][15][16][17][18]. During decocting, NADES may liberate physically bound compounds from the plant matrix, interact with them, and increase their concentrations in the nal decoction.
NADES may also stabilize licorice phenolics. The experiments herein indicated that chemical transformations occurred during processing. Figure 3b showed that isomerization, glycolysis, and decomposition of chalcones to 2,4-dihydroxyacetophenone (9) might have occurred for fried licorice. The addition of honey inhibited chalcone isomerization and breakdown to 2,4-dihydroxyacetophenone caused by high temperature. However, it had no signi cant effect on glycosidic bond breakage. Previous studies [2,5,6] reported that licorice avonoids undergo isomerization and glycolysis during processing. However, in general, phenolic compounds are relatively more stable in NADES than in water [15,16]. This study also provides bases for the effect of honey on quality improvement of natural products.

Conclusion
In conclusion, this study veri es that the ancient honey-frying processing of licorice enhances the immunomodulatory e cacy of licorice. The improvement is due to the NADES characteristics of honey, which inhibit heat-induced transformations of the bioactive compounds in licorice, improved their dissolution, and enhances their bioavailability after oral licorice decoction administration. These effects of honey are due to the mechanism that honey forms molecular interactions with compounds in licorice. These

Supplementary Files
This is a list of supplementary les associated with this preprint. Click to download.