LC-MS Profiling of Kakkonto and Identification of Ephedrine as a Key Component for Its Anti-Glycation Activity

A total of 147 oral Kampo prescriptions, which are used clinically in Japan, were evaluated for their anti-glycation activity. Kakkonto demonstrated significant anti-glycation activity, prompting further analysis of its chemical constituents using LC-MS, which revealed the presence of two alkaloids, fourteen flavonoids, two but-2-enolides, five monoterpenoids, and four triterpenoid glycosides. To identify the components responsible for its anti-glycation activity, the Kakkonto extract was reacted with glyceraldehyde (GA) or methylglyoxal (MGO) and analyzed using LC-MS. In LC-MS analysis of Kakkonto reacted with GA, the peak intensity of ephedrine was attenuated, and three products from ephedrine-scavenging GA were detected. Similarly, LC-MS analysis of Kakkonto reacted with MGO revealed two products from ephedrine reacting with MGO. These results indicated that ephedrine was responsible for the observed anti-glycation activity of Kakkonto. Ephedrae herba extract, which contains ephedrine, also showed strong anti-glycation activity, further supporting ephedrine’s contribution to Kakkonto’s reactive carbonyl species’ scavenging ability and anti-glycation activity.


Introduction
Protein glycation, which is the non-enzymatic reaction between reducing sugars and proteins, results in the formation of advanced glycation end products (AGEs). In the first step of glycation, reducing sugars react reversibly with free amino groups on the protein through a nucleophilic attack, leading to the formation of Schiff bases. Intermediate products such as 3-deoxyglucosone undergo slow rearrangement to form stable Amadori products or ketoamines [1][2][3][4]. Furthermore, reactive dicarbonyl compounds such as glyoxal and methylglyoxal (MGO), which are primarily produced as by-products of glycolysis, can also cause the formation of AGEs by selectively reacting with the side chains of basic amino acids such as arginine and lysine on proteins [3]. The aforementioned reactive carbonyl species, including glyoxal, MGO, and 3-deoxyglucosone, are highly reactive and represent important precursors of AGEs in the human body [5]. Accumulation of AGEs in tissues leads to the formation of irreversible cross-linked structures with proteins [6], resulting in pathological reactions in tissues and organs and causing a series of diabetic complications [7]. Furthermore, AGEs and reactive carbonyl species are also implicated in aging [8], Alzheimer's disease [9], and atherosclerosis [10]. While several compounds, including aminoguanidine [11] and metformin [12,13], have been investigated for their ability to inhibit the formation of AGEs and avert diabetic complications, their clinical application has been constrained due to apprehensions regarding their safety and efficacy. Aminoguanidine, for example, failed clinical trials of ACTION II due to its high toxicity in diabetic patients [14]. Hence, it is a critical issue to develop effective and safe anti-AGE drugs to protect people with diabetes from complications.
Kampo medicine is a traditional Japanese medicine that originated in ancient China and underwent its unique development in Japan. There are currently 294 Kampo prescriptions available as over-the-counter (OTC) medications and 148 Kampo prescriptions Subsequently, the anti-glycation activity of the 18 prescriptions was evaluated using a BSA-glyceraldehyde (GA) assay according to the manufacturer's protocol. GA is a reactive carbonyl species that is produced during sugar metabolism. Out of the 18 prescriptions, 14 showed anti-glycation activity of more than 50% inhibition at a concentration of 400 µU/mL (Table 1). Kakkonto was found to exhibit more than 80% inhibitory activity in both assays and was selected for further study. Further testing of Kakkonto was carried out at different concentrations ranging from 400 to 12.5 µU/mL in the BSA-GA assay. The results showed that Kakkonto inhibited the formation of AGEs in a dose-dependent manner, with an IC 50 value of 120 µU/mL ( Figure 1A). tive carbonyl species that is produced during sugar metabolism. Out of the 18 prescriptions, 14 showed anti-glycation activity of more than 50% inhibition at a concentration of 400 μU/mL (Table 1). Kakkonto was found to exhibit more than 80% inhibitory activity in both assays and was selected for further study. Further testing of Kakkonto was carried out at different concentrations ranging from 400 to 12.5 μU/mL in the BSA-GA assay. The results showed that Kakkonto inhibited the formation of AGEs in a dose-dependent manner, with an IC50 value of 120 μU/mL ( Figure 1A).

Identification of Chemical Constituents in Kakkonto by LC-MS
To identify the chemical constituents in Kakkonto, an LC-MS analysis was co ducted. As shown in Figure 2, a total of twenty-seven peaks were detected in the total i chromatogram, and their retention time and MS fragmentation ions are listed in Table  By carefully analyzing the chromatographic behaviors and MS data and referring to t literature [23][24][25][26][27], the presence of two alkaloids (1, 2), fourteen flavonoids (3, 4, 6, 7, 9-13, 14, and 17-21), two but-2-enolides (12,16), five monoterpenoids (5,8,15,22,23), a four triterpenoid glycosides (24)(25)(26)(27), were identified ( Figure 3). The structure elucidati based on MS fragmentation is described briefly as follows.         Peaks 13 (t R 8.50 min) and 18 (t R 11.88 min) were identified as liquiritin (13) and isoliquiritin (18), and peak 20 (t R 12.71 min) was the aglycone of 13, which was liquir-itigenin (20). These compounds are all contained in Glycyrrhizae radix [28]. Peaks 13 and 18 have the same molecular formula of C 21 H 22 O 9 , which was deduced from the ions of m/z 436.1597 [M+HN 4 ] + and m/z 419.1333 [M+H] + , respectively. In MS/MS, they also exhibited the same product ions [Aglycone+H] + at m/z 257 resulting from the detachment of the glucose moiety, and ions at m/z 137 and m/z 147 derived from the A ring and B ring, respectively ( Figure 5). Although the MS data alone could not distinguish peaks 13 and 18, they were identified based on the retention times [29]. Similarly, peak 20 was most likely to be the aglycone of 13 or 18, as evidenced by its molecular formula of C 15 H 12 O 4 and the MS/MS product ions of m/z 137 and m/z 147 from the precursor [M+H] + ion. Due to its slightly longer retention time than 18, peak 20 was identified as liquiritigenin (20) instead of isoliquiritigenin (18a) [30].
Compounds 3, 4, and 6 were considered to have one sugar attached to the isoflavone based on their molecular formula. These compounds exhibited a series of product ions due to the elimination of H2O from the sugar moiety, followed by the characteristic bond cleavage of C-glycosidide via retro-Diels-Alder reaction, as well as elimination of hydride, formyl group, H2O, and CO. On the other hand, it was revealed that 7, 11, and 14 have an apiose moiety bound to 4, 3, and 4′-methoxypuerarin (14a), respectively, since in MS/MS of 7, 11, and 14, the product ions generated by the detachment of terminal sugar corresponded to 4, 3, and 14a (Figure 7). Peaks 12 (t R 7.96 min) and 16 (t R 11.19 min) were identified as pueroside A (12) and sophoraside A (16), respectively, which are characteristic but-2-enolide derivatives of Puerariae radix [38].
In MS/MS, 12 underwent fragmentation to generate fragment ion 12a, corresponding to a demethylated 16, through the elimination of terminal rhamnose. Further fragmentation  of 12a and 16 generated 12b and 16b by the loss of a glucose, then 12c and 16c by the loss of a hydroxy group. Furthermore, 12d and 12e were generated from 12c via the elimination of benzoyl and hydroxy groups, and 16f from 16c via the elimination of a phenyl group (Figure 8). Peaks 12 (tR 7.96 min) and 16 (tR 11.19 min) were identified as pueroside A (12) and sophoraside A (16), respectively, which are characteristic but-2-enolide derivatives of Puerariae radix [38].
In MS/MS, 12 underwent fragmentation to generate fragment ion 12a, corresponding to a demethylated 16, through the elimination of terminal rhamnose. Further fragmenta- tion of 12a and 16 generated 12b and 16b by the loss of a glucose, then 12c and 16c by the loss of a hydroxy group. Furthermore, 12d and 12e were generated from 12c via the elimination of benzoyl and hydroxy groups, and 16f from 16c via the elimination of a phenyl group (Figure 8).  ions 5a and 15a via the loss of a glucose, followed by the loss of H2O and the benzoyl group to generate product ions 5b and 15b (Figure 9). On the other hand, peoniflorin (8) first produced ion 8a via the loss of H2O, followed by the loss of the benzoyl group and glucose to produce ion 8b. A series of product ions from 5b, 8b, and 15b were generated by the loss of H2O and CO (Figure 9). Paeonivayin (23) and benzoylpaeoniflorin (22) possess structures with an additional benzoyl moiety attached at C-6 of the glucose of 5 and 8, respectively. In MS/MS, product ions from 23 and 22 were observed to be the same as those from 5 and 8 (Figure 9). Nota- bly, 5, 8, 15, 22, and 23 can also be distinguished by their retention times [26]. Fragmentation of 5 and 15 generated ions 5a and 15a via the loss of a glucose, followed by the loss of H 2 O and the benzoyl group to generate product ions 5b and 15b (Figure 9). On the other hand, peoniflorin (8) first produced ion 8a via the loss of H 2 O, followed by the loss of the benzoyl group and glucose to produce ion 8b. A series of product ions from 5b, 8b, and 15b were generated by the loss of H 2 O and CO (Figure 9). Paeonivayin (23) and benzoylpaeoniflorin (22) possess structures with an additional benzoyl moiety attached at C-6 of the glucose of 5 and 8, respectively. In MS/MS, product ions from 23 and 22 were observed to be the same as those from 5 and 8 ( Figure 9). Notably, 5, 8, 15, 22, and 23 can also be distinguished by their retention times [26].

Identification of the Components That Contribute to the Anti-Glycation Activity
To identify the components contributing to the anti-glycation activity, the methanol fraction prepared from Kakkonto was reacted with GA and analyzed using LC-MS (Figure 11A). Comparison of the total ion chromatograms of Kakkonto extract with and without GA addition showed a decrease in the peak intensity of ephedrine (1) with GA addition, while three new peaks (EM1-EM3) were detected. The new peaks were most similar to products generated by the trapping of GA by 1, as indicated by their molecular formulas of C12H18O3N (EM1 and EM2) and C13H20O3N (EM3), which were deduced from the molecular ions [M+H] + observed in MS. Further analysis of MS/MS revealed that EM1-EM3 are derivatives with an oxazoline ring generated through nucleophilic addition reaction and condensation between 1 with GA [41]. In MS/MS of EM1-EM3, a series of product ions were observed, which were formed through stepwise loss of H2O, ring opening, and loss of C3H3N (Table S2, Figure 12). Since EM1 and EM2 showed the same molecular formula and fragment pattern, they were considered to be a pair of diastereomers.

Identification of the Components That Contribute to the Anti-Glycation Activity
To identify the components contributing to the anti-glycation activity, the methanol fraction prepared from Kakkonto was reacted with GA and analyzed using LC-MS ( Figure 11A). Comparison of the total ion chromatograms of Kakkonto extract with and without GA addition showed a decrease in the peak intensity of ephedrine (1) with GA addition, while three new peaks (EM1-EM3) were detected. The new peaks were most similar to products generated by the trapping of GA by 1, as indicated by their molecular formulas of C 12 H 18 O 3 N (EM1 and EM2) and C 13 H 20 O 3 N (EM3), which were deduced from the molecular ions [M+H] + observed in MS. Further analysis of MS/MS revealed that EM1-EM3 are derivatives with an oxazoline ring generated through nucleophilic addition reaction and condensation between 1 with GA [41]. In MS/MS of EM1-EM3, a series of product ions were observed, which were formed through stepwise loss of H 2 O, ring opening, and loss of C 3 H 3 N (Table S2, Figure 12). Since EM1 and EM2 showed the same molecular formula and fragment pattern, they were considered to be a pair of diastereomers.
Additionally, when the Kakkonto methanol fraction was reacted with MGO and analyzed using LC-MS, a decrease in the peak intensity of ephedrine (1) was also observed, along with two new peaks EM4 and EM5 ( Figure 11B). Peaks EM4 and EM5 were identified as a pair of diastereomers formed between 1 and MGO through nucleophilic addition, and their structures were determined based on their molecular formulas, which were obtained from the molecular ions [M+H] + observed in MS, as well as the MS/MS product ions generated through the loss of H2O, CO, and C3H7N (Table S3, Figure 13).    Additionally, when the Kakkonto methanol fraction was reacted with MGO and analyzed using LC-MS, a decrease in the peak intensity of ephedrine (1) was also observed, along with two new peaks EM4 and EM5 ( Figure 11B). Peaks EM4 and EM5 were identified as a pair of diastereomers formed between 1 and MGO through nucleophilic addition, and their structures were determined based on their molecular formulas, which were obtained from the molecular ions [M+H] + observed in MS, as well as the MS/MS product ions generated through the loss of H 2 O, CO, and C 3 H 7 N (Table S3, Figure 13).

Evaluation of Ephedrae Herba Extract for Anti-Glycation Activity
In Japan, ephedrine (1) is a regulated substance by law, which made it impossible to investigate the anti-glycation activity of 1 as a single compound. However, the inhibitory activity of an Ephedrae herba extract, which contained 1 as the main component, was evaluated using the BSA-GA assay. As a result, the extract dose-dependently inhibited glycation in the concentration range of 10-0.125 mg/mL, with an IC 50 of 0.22 mg/mL ( Figure 1B). Its activity was comparable to that of the positive control, aminoguanidine (IC 50 2.6 mM, equivalent to 0.19 mg/mL) ( Figure 1C). Based on these results, it can be concluded that ephedrine (1) is the key component responsible for the anti-glycation activity of Kakkonto, and its possible mechanism of action is trapping the reactive carbonyl species.
So far, there have been no reports on the anti-glycation activity of ephedrine or Ephedra herba, but there have been several reports that they suppress hyperglycemia in vivo. For example, it has been reported that crude extracts of Ephedra sinica or E. foeminea and l-ephedrine showed suppression of hyperglycemia in mice with diabetes induced by streptozotocin [42,43]. Furthermore, extracts of E. sinica have been shown to promote the regeneration of atrophied pancreatic islets. It has also been suggested that E. sinica could improve hyperglycemia by regenerating atrophied pancreas islets and restoring insulin secretion [42]. In addition, the anti-obesity effect and anti-hyperglycemic action of Ephedra sinica were evaluated in mice fed a high-fat diet, and Ephedra sinica was found to reduce weight gain and fasting blood glucose levels, as well as improve HDL cholesterol levels [44]. Furthermore, it was suggested that E. sinica could improve obesity and hyperglycemia by increasing PPAR-α and adiponectin and decreasing TNF-α [44]. In these reports, it was concluded that the decrease in blood sugar levels was due to improved metabolism, but considering our research results as well, it is possible that the direct trapping of sugar by ephedrine may also be involved.

Evaluation of Ephedrae Herba Extract for Anti-Glycation Activity
In Japan, ephedrine (1) is a regulated substance by law, which made it impossible to investigate the anti-glycation activity of 1 as a single compound. However, the inhibitory activity of an Ephedrae herba extract, which contained 1 as the main component, was evaluated using the BSA-GA assay. As a result, the extract dose-dependently inhibited glycation in the concentration range of 10-0.125 mg/mL, with an IC50 of 0.22 mg/mL (Figure 1B). Its activity was comparable to that of the positive control, aminoguanidine (IC50 2.6 mM, equivalent to 0.19 mg/mL) ( Figure 1C). Based on these results, it can be concluded that ephedrine (1) is the key component responsible for the anti-glycation activity of Kakkonto, and its possible mechanism of action is trapping the reactive carbonyl species.
So far, there have been no reports on the anti-glycation activity of ephedrine or

General Methods
LC-MS analysis was performed on a Vanquish UHPLC system combined with a Q-Exactive Hybrid Quadrupole Orbitrap mass spectrometer. Fluorescence intensity was measured using a multi-label plate reader from EnSpire 2300 (Perkinelmer Japan Co., Ltd., Kanagawa, Japan). A Diaion HP-20 for column chromatography was purchased from Mitsubishi Chemical Co. (Tokyo, Japan).

Materials and Chemicals
A total of 147 oral Kampo prescriptions, which are covered by health insurance in Japan, were purchased from Ohsugi Pharmaceutical Co., Ltd. (

Sample Preparation of Kampo Solutions
We defined the amount of the daily dose as 1 unit (U) for Kampo prescriptions. A Kampo prescription (2 mU) was suspended in 1 mL of purified water and extracted using sonication at room temperature for 15 min. The mixture was then centrifuged at 4800 rpm for 15 min, and the supernatant was used as the 2 mU/mL sample solution.

Assay for the Anti-Glycation Activity Using D-Ribose
An assay of AGE formation inhibitory activity between BSA and D-ribose was conducted according to a method from the literature with minor modifications [22]. The solution containing BSA solution (final concentration 10 mg/mL), each sample solution (final concentration 50 µU/mL), and D-ribose solution (final concentration 0.5 M) were added to each well of a 96-well microplate, and then fluorescence intensity (excitation wavelength: 370 nm, fluorescence wavelength: 440 nm) was measured using a fluorescence plate reader after 1 h of incubation. The glycation value of the vehicle control (0 µU/mL) was assumed to be 100%. The glycation value of each well relative to the control was determined as the glycation ratio.

Assay for the Anti-Glycation Activity Using Glyceraldehyde
The anti-glycation activity using glyceraldehyde was evaluated with an Albumin Glycation Assay Kit, Glyceraldehyde (Cosmo Bio Co., Ltd. (Tokyo, Japan), Cat. No. AAS-AGE-K01). The assay kit buffer was used, with sodium azide (3 mM) added. Each concentration of sample solution was prepared from 2 mU/mL sample solution via dilution using the assay kit buffer to each concentration (final concentrations 400, 200, 100, 50, 25, and 12.5 µU/mL).
BSA solution (50 µL), each concentration sample solution (40 µL), and glyceraldehyde solution (500 mM, 10 µL) were added to each well of a 96-well microplate, and then fluorescence intensity A (excitation wavelength: 370 nm, fluorescence wavelength: 440 nm) was measured using a fluorescence plate reader. After 24 h of incubation, fluorescence intensity B was measured. The glycation value (fluorescence intensity B-fluorescence intensity A) of the vehicle control (0 µU/mL) was assumed to be 100%. The glycation value of each well relative to the control was determined as the glycation ratio.
A Q-Exactive hybrid quadrupole orbitrap high-resolution accurate mass spectrometer system (Thermo Scientific, Waltham, MA, USA) with an ESI source was operated in the positive and negative ion modes. The calibration solutions were used to calibrate the ESI-MS to increase mass accuracy. The optimized parameters of mass spectrometry were illustrated below: spray voltage, +3.5 kV (for positive ion mode) or −2.5 kV (for negative ion mode); capillary temperature, 262.5 • C; sheath gas flow rate, 50 units; AUX gas flow rate, 12.5 units; sweep gas flow rate, 2.63 units; S-lens RF level, 50 units; and probe heater temperature, 425 • C. Data were collected in the full MS modes and full MS/data-dependent (dd)-MS/MS. In-source CID was set at 0 eV. The resolution was 70,000 for full MS and 35,000 for full MS/dd-MS/MS. The AGC was set at 1E6 for full MS and 1E5 for dd-MS/MS. Maximum IT was set at 200 ms for full MS. Scan range was set at 150 to 2000 m/z for full MS. Data-dependent scan was performed using high-energy collision with normalized collision energy (NCE) at 10 eV; 30 eV; and stepped NCE at 10, 25, and 40 eV.

Preparation of Kakkonto Solution for LC-MS
Kakkonto formulation (2.5 g) was suspended in 25 mL of purified water and extracted via sonication at room temperature for 30 min. The mixture was then centrifuged at 4800 rpm for 15 min, and the supernatant was subjected to Diaion HP-20 column chromatography to yield the water eluted part (300 mL) and methanol eluted part (300 mL). The methanol eluted part was evaporated in vacuo to yield a methanol fraction (0.29 g). The methanol fraction in methanol (1 mg/mL) was filtrated with a 0.22 mm filter and then analyzed using LC-MS.

Evaluation of the GA and MGO Trapping Capacity of Kakkonto
Kakkonto methanol fraction (10 mg) was incubated with GA (50 mM or 0 mM) in PBS buffer (pH 7.4, 100 mM) at 37 • C. After 24 h of incubation, the reaction was stopped by adding 200 µL of acetic acid. Purified water (10 mL) was added to the reaction mixture, followed by Diaion HP-20 column chromatography to yield the water eluted part (100 mL) and the methanol eluted part (100 mL). The methanol eluted part was evaporated in vacuo to yield a methanol fraction. The methanol fraction in 10 mL of methanol was filtrated with 0.22 µm filter and then analyzed using LC-MS. Kakkonto (10 mg) and MGO (50 mM or 0 mM) mixtures were reacted and analyzed using LC-MS using a similar method to that described above.

Extraction and Assay for the Anti-Glycation Activity of Ephedrae Herba
Ephedrae herba (1 g) was suspended in 25 mL of methanol and extracted via sonication at room temperature for 20 min. The mixture was then filtrated, and the filtrate was evaporated in vacuo to yield methanol extract (98 mg). The methanol extract (8 mg) was dissolved with DMSO (80 mL) and 2.5, 1.25, 0.625, and 0.3125 mg/mL of solution (final concentrations 1, 0.5, 0.25, and 0.125 mg/mL) were prepared with the buffer of the Albumin Glycation Assay Kit, Glyceraldehyde. An assay for the anti-glycation activity was evaluated using the same method as that described above.

Conclusions
In this study, Kakkonto, one of the prescribed Kampo medicines, revealed significant in vitro anti-glycation activity, and ephedrine was identified as the key component responsible for this activity. Kakkonto is clinically used in Japan for the early stages of colds, inflammatory diseases, shoulder stiffness, upper body neuralgia, and hives. Our findings suggest the potential for Kakkonto to be used for the treatment of diabetic complications based on its anti-glycation activity, but further investigation is needed.

Supplementary Materials:
The following supporting information can be downloaded at: https:// www.mdpi.com/article/10.3390/molecules28114409/s1. Table S1: Mass and MS/MS spectrometric data of compounds identified in Kakkonto extract. Table S2: Mass spectrometric data of products detected in Kakkonto extract reacted with glyceraldehyde (GA). Table S3: Mass spectrometric data of products detected in Kakkonto extract reacted with methylglyoxal (MGO).

Data Availability Statement:
The data presented in this study and samples of the compounds are available upon request from the corresponding author.