Sophorae Flavescentis Radix (SFR, Sophora flavescens Ait.) and Sophorae Tonkinensis Radix et Rhizoma (STR, S. tonkinensis Gapnep.) are two commonly used traditional Chinese medicines (TCMs) from Sophora (Leguminosae) plants . Both the herbs were found to contain similar chemical constituents such as quinolizidine alkaloids, prenylated flavonoids, and oleanane triterpenoids, while their clinical applications are entirely different [2-3]. SFR is mostly used externally for the treatments of skin diseases and gynaecological diseases, such as eczema, dermatitis, and colpitis, while STR is frequently used internally to treat acute pharyngolaryngeal infections and sore throat [1-2]. Therefore, it is of great interest to find out the characteristic chemical constituents potentially leading to their unique clinical applications claimed for the two closely related herbs.
Although the alkaloids in these two herbs have received much attention, the holistic and component-based relationships between their constituents and different clinical applications are still unknown. In the Chinese Pharmacopoeia , three alkaloids (matrine, oxymatrine, and sophoridine) widely present in Sophora plants , are chosen as qualitative markers for identification of these two herbs in the thin layer chromatography tests, which is known to lack specificity in distinguishing SFR and STR. A similar situation exists in the quantitative tests for the two herbs recorded in the Chinese Pharmacopoeia . These problems may arise from the various pharmacological findings of alkaloids that are tentatively assigned as the unique active constituents of the two herbs. It is well known that the therapeutic efficacy of TCMs is often attributed to the synergic effect of their multiple bioactive components and multi-targets, and thus analysis of merely one or a few markers is not adequately representative for quality control of herbs. In the past two decades, more and more flavonoids were isolated and identified from the two herbs , some of which have been gradually revealed to possess significant bio-activities such as anticancer [5-6], anti-inflammatory , anti-diabetes  and anti-virus activities . These findings of the flavonoids prompted us to re-examine comprehensively the active components of the two herbs in terms of alkaloids and flavonoids.
HPLC fingerprinting has been used over the last decade for authentication and quality control of herbs and their preparations. This method emphasizes the whole profile of components in a complex system, and it is a strategy recommended to assess the quality of botanical products by the US Food and Drug Administration (FDA), the European Medicines Evaluation Agency (EMEA), and the State Food and Drug Administration of China (SFDA) . To the best of our knowledge, few reports have been published on systemic comparison of the alkaloid or flavonoid constituents between these two herbs using fingerprinting techniques, though many quantitative studies on the major alkaloids in SFR and STR by various chromatographic techniques have been reported [11-14]. In 2013, Ma and co-workers analyzed flavonoids in 12 batches of SFR and 4 batches of STR samples by HPLC fingerprinting and found preliminarily that these two herbs had different flavonoid profiles . To investigate systematically the characteristic chemical constituents that mayt contribute to their unique clinical applications claimed for SFR and STR, comprehensive HPLC fingerprint methods of flavonoids and alkaloids in SFR and STR are needed with more samples collected from different regions of China. The objectives of the present study were as follows: 1) to establish HPLC fingerprints of the alkaloid and flavonoid constituents for the two herbs, and 2) to compare and analyze the fingerprints by similarity calculation and hierarchical clustering analysis (HCA).Materials and Methods Plant materials
The reference SFR and STR herbs were obtained from the National Institute for the Control of Pharmaceutical and Biological Products of China (Beijing, China). All the other botanical samples studied were obtained as commercial drugs from 26 different districts in 19 provinces of China. All the samples were authenticated by one of the authors (Dr. CHEN Dao-Feng). Among them, 24 samples were identified as the roots of S. flavescens, and 19 samples were the roots and rhizomes of S. tonkinensis. The specimens were deposited in the Department of Pharmacognosy, School of Pharmacy, Fudan University, Shanghai, China. The identity, sampling part, collection source and collection time of the 43 tested samples are summarized in Table 1. All the HPLC fingerprint data were obtained in 2004, and the phytochemical isolation and statistical analysis were accomplished recently.
Cytisine, sophocarpine, matrine, sophoranol, oxymatrine, and oxysophocarpine were isolated from the roots of S. flavescens or S. tonkinensis in our laboratory, and their structures were confirmed by various spectral analyses [16-18]. Their purities were greater than 98% by HPLC-PDA analysis detected at 220 nm. Methanol and acetonitrile were of HPLC-reagent grade (Tedia, Fairfield, OH, USA). Triethylamine, phosphoric acid, chloroform, 25% aqueous ammonia, and glacial acetic acid were of analytical-reagent grade and obtained from Sinopharm Chemical Reagent Co., Ltd. (Shanghai, China). Distilled water was obtained from a Milli-Q system (Millipore, Bedford, MA, USA).Preparation of sample solutions
Sample preparation for the alkaloid analysis was conducted as follows: herbal samples (0.5 g) were pulverized, sieved through a No. 50 mesh, transferred into a 50-mL flask with stopper, and weighed accurately. Twenty milliliter of chloroform and 0.2 mL of 25% aqueous ammonia were accurately added and the flask was re-weighed. The sample was kept in dark overnight, and then extracted under ultrasonication at room temperature for 45 min. The extraction was cooled to ambient temperature and then made up to its original weight with chloroform, followed by filtration with paper filters. An accurate aliquot (10.0 mL) of the successive filtrate was collected and evaporated to dryness under reduced pressure, and the residue was then re-dissolved in 5 mL of 50% aqueous methanol. The sample solution was filtered with a 0.45 μm membrane filter prior to HPLC analysis. Five microliters of each solution were injected for analysis.
The procedure for the sample preparation of flavonoids was the same as that for alkaloids, except that the extraction solvent of methanol (40 mL) was used, and the volumes of the successive filtrate and injection were 20 mL and 20 μL, respectively.Apparatus and HPLC conditions
All chromatographic separations were conducted on an Agilent 1100 series HPLC instrument (Palo Alto, CA, USA) equipped with a photodiode array detector (PDA), and the data were recorded and analyzed by Hewlett Packard ChemStation software (Palo Alto, CA, USA).
The separation of alkaloids was performed on a YMC- pack ODS-AQ column (4.6 mm × 250 mm, 5 μm; YMC, Kyoto, Japan), and the mobile phase consisted of water, 0.1% triethylamine in methanol (V/V) and 25 mmol·L-1 phosphoric acid aqueous solution (44 : 46 : 10) at a flow rate of 0.8 mL·min-1. The column temperature was kept at 40 ℃, and the detection wavelength was set at 220 nm.
The separation of flavonoids was carried out on a Sphereclone Luna C18 column (4.6 mm × 250 mm, 5 μm; Phenomenex, Torrance, CA, USA). The mobile phase was 0.2% acetic acid aqueous solution (V/V, A) and 0.2% acetic acid in acetonitrile (V/V, B) with a gradient program as follows: 0–30 min, linear gradient 0%–45% B; 30–60 min, linear gradient 45%–85% B; 60–70 min, linear gradient 85%–96% B; 70–90 min, isocratic 96% B at a flow rate of 1.2 ml·min-1. The column temperature was kept at 40 ℃, and the PDA detector was set at 280 nm. The UV spectra were recorded on-line in the range of 200– 400 nm.Data processing and analysis
Similarity analysis was performed with a professional software named 'Similarity Evaluation System for Chromatographic Fingerprint (Version 2004A)' which was developed by Chinese Pharmacopoeia Commission. In the present study, the software was employed to synchronize and generate a simulative median fingerprint chromatogram, against which the similarity value of each chromatogram was calculated based on the approach of cosine value for vectorial angle.
The hierarchical cluster analysis (HCA) was carried out with the Statistical Program for Social Sciences (SPSS) 11.5 software (SPSS Inc., Cary, NC, USA), applying the method of average linkage between groups in alkaloid analysis and that of centroid linkage in flavonoid analysis, respectively. Squared euclidean distance was selected as similarity measurement in HCA.Results and Discussion HPLC fingerprints of alkaloids
Quinolizidine alkaloids of SFR and STR are well studied and long regarded as their main bioactive components, which have reported to exhibit significant pharmacological activities [2-4]. It is therefore of great interest to investigate the potential difference in nature of alkaloids present in these two herbs by fingerprint analysis.
In order to extract the alkaloids from the herbs effectively, a mixture of chloroform and 25% aqueous ammonia was used as the extraction solvent in the present study. This extraction solvent has been found to extract the alkaloids from the herbs well in our recent reports [11, 13]. After optimization of the HPLC conditions, the fingerprint methods of the two herbs were validated for precision, repeatability, and stability. The RSD values of all the validation tests were less than 1.65%. Then the chromatographic fingerprints of the alkaloids were generated for 18 batches of SFR and 15 of STR samples, respectively. Two simulative median fingerprint chromatograms (SMFC) were then produced for SFR and STR, respectively, by the fingerprint software. The similarity value of each chromatograph against the corresponding SMFC was thus calculated.
As shown in Fig. 1A, the alkaloid profiles of 18 batches of SFR samples were generally consistent, and 15 peaks were assigned as the common peaks in SMFC (Fig. 1B). The similarity values of these samples were found in the range of 0.80–0.98 (Table 1), indicating that similar alkaloid components were present in these samples regardless of the collection sources. In addition, Peaks 5, 6, 8, 13, and 14 were identified as oxysophocarpine, oxymatrine, sophoranol, sophocarpine, and matrine, respectively, by comparison of their retention times and UV profiles with those of the authentic standards. Peaks 5–7 and 14 were the principal alkaloids in SFR based on their peak areas, among which Peak 6 (oxymatrine) was of the highest content in all samples, followed by Peaks 5 and 7 in most of the samples.
Similar to SFR, the alkaloid profiles of 15 batches of STR samples were also fingerprinted (Fig. 2A) and analyzed. As a result, 15 common peaks were also assigned with Peaks 3, 4, 5, 8, 12, and 13 being identified as cytisine, oxysophocarpine, oxymatrine, sophoranol, sophocarpine and matrine, respectively (Fig. 2B). An intra-species similarity comparison between STR samples was performed by the fingerprint software. Most of the similarity values of STR samples were in the range of 0.83–0.97 (Table 1), indicating that the samples from the same species origin were closely similar to each other. The only exception was the sample ST-3 with a similarity value of 0.47. Together with the fingerprint results of SFR, these data indicated that the nature of the alkaloid components were very similar between these two species, both of which possessed oxysophocarpine, oxymatrine, sophoranol, sophocarpine, matrine and some unidentified peaks with the same retention times and UV spectra. However, the peak intensities of some peaks were slightly different. For example, matrine was found to be the second highest alkaloid in most of STR samples instead of oxysophocarpine in SFR. It is worthy noting that the reason for the low similarity value of the reference drug (ST-3) remained unknown, but this exceptional sample did not significantly affect the overall conclusion of the fingerprints.
Furthermore, the hierarchical clustering analysis (HCA) was conducted to compare the alkaloid profiles of these two herbs. Based on 21 alkaloid constituents recognized automatically in the HPLC chromatograms of 18 batches of SFR and 15 of STR samples, a 21 × 33 matrix was formed from HCA using SPSS 11.5 software and the results are shown in Fig. 3. It was found that the 33 tested samples could be classified into two groups. Group Ⅰ, with a relatively low content of oxymatrine (peak intensity: 629–4005 mAU), covered 9 SFR and 12 STR samples. Group Ⅱ, with a much higher content of oxymatrine (peak intensity: 4859–8108 mAU), consisted of 9 SFR and 3 STR samples. Further HCA analysis showed that group Ⅰ could be divided into three subgroups Ⅰ-A, Ⅰ-B, and Ⅰ-C, and each subgroup was made up of a single species origin. Subgroup Ⅰ-A covered 11 STR samples with a relatively higher content of oxymatrine; subgroup Ⅰ-B consisted of 9 SFR samples with a particularly high content of oxysophocarpine; and subgroup Ⅰ-C contained only one STR sample (ST-3), with the characteristic of an extremely high content of matrine but low content of oxymatrine. Group Ⅱ was slightly more complicated. One SFR sample (SF-16) and one STR sample (ST-16) were dispersed from the others, and were assigned as subgroups Ⅱ-C and Ⅱ-D, respectively, because of their corresponding high contents of matrine and oxymatrine. The remaining 10 samples were divided into subgroups Ⅱ-A and Ⅱ-B mainly based on their differences in oxysophocarpine contents. These HCA results of the alkaloids profiles suggested that the species of SFR and STR mixed together in the hierarchical clustering dendrogram (Fig. 3). Therefore, it was concluded that the alkaloid constituents of these two herbs were closely similar, with most of the alkaloids co-existed, and cannot be used as diagnostic markers to distinguish these two drugs. More importantly, the apparent lack of specificity of alkaloids may indicate other potential relations between the overall therapeutic effect and components in these two herbs. However, due to their various and notable pharmacological effects [2-4], it is still reasonable to assess their quality with the alkaloid fingerprint. In addition, our results from the alkaloid fingerprints also indicated that oxymatrine together with matrine, and oxymatrine together with oxysophocarpine, could be served as the dominant quantitative markers for quality assessments of SFR and STR, respectively. This was in good agreement with those reported in our previous studies [11, 13].
The above HPLC fingerprint studies of alkaloids showed no obvious difference between SFR and STR, and it is reasonable to speculate that alkaloids might not be the major components that lead to their different clinical applications. In the past two decades, plenty of flavonoids were reported in SFR and STR with various biological activities [3, 5-9]. This promoted us to investigate the potential difference in flavonoids between these two herbs by HPLC fingerprinting.
As shown in Figs. 4 and 5, the HPLC chromatograms of the two species demonstrated individual specific flavonoid profiles. The chromatograms of SFR samples possessed 25 common peaks dispersed in 20–52 min (Fig. 4B), while the STR samples showed a totally different profile with 23 common peaks congested in 25–65 min (Fig. 5B). An intra-species comparison of the flavonoid profiles was conducted by similarity value calculation. The results showed that the samples within the same species matched well (Table 1). For example, most of the similarity values of the SFR samples were in the range of 0.80–0.96 except for SF-16 (0.38) (Table 1). Similar observation was also achieved in the STR samples, which all the similarity values were in the rage of 0.83–0.97 except ST-4 (0.68) and ST-18 (0.27).
To better understand the nature of the flavonoids, 23 common peaks of STR samples were tentatively assigned by comparison of their online UV profiles with literature data (Data not shown). Peaks 1, 6, and 10 might be pterocarpans exhibiting maximum absorptions at 310 and 285 nm with the former stronger ; Peaks 2–5, 9, and 13 might be isoflavonoids showing band Ⅱ absorption at 260 nm due to benzoyl moieties, along with a shoulder on the right side ; Peaks 8, 11, 12, 18–20, 22, and 23 might be flavanones with a band Ⅱ maximum absorption at 280–285 nm (Peak 22 at 295 nm) and a shoulder adjacent to the long-wave side ; Peak 7 was a chalcone exhibiting a characteristic strong band Ⅰ absorption at 330–350 nm and a relative lower band Ⅱ absorption at 280 nm ; and Peaks 14–17 and 21 might be aromatic compounds with maximum absorption at 260–280 nm. This deduction was quite consistent with our recent HPLC-ESI- MS/ MS study on STR flavonoids . Thus, it could be concluded that the major sub-types of flavonoids in STR were flavanones and isoflavones, along with a few pterocarpins and chalcones. Furthermore, Peak 20 (flavanones, assigned as sophoranone ) was the dominant component, followed by Peaks 23 (flavanones), 1 (pterocarpins), 6 (pterocarpins) and 7 (chalcones) in all STR samples, except two with low similarity values (ST-4 and ST-18).
Similarly, the common peaks present in SFR were also assigned (Data not shown). Peaks 3, 5 and 11 were assigned as pterocarpins; Peak 1 was identified as an isoflavonoid; Peak 22 was a typical chalcone exhibiting a strong band Ⅰ absorption at 385 nm together with a relatively lower band Ⅱ absorption at 255 nm ; and the remaining peaks except Peaks 23–26, were all flavanones/flavanols with a band Ⅱ maximum absorption at 290–305 nm (Peak 2 at 280 nm) and a shoulder adjacent to the long-wave side . Obviously, compared to STR, SFR was a much richer source of flavanones, while relative lack of isoflavonoids and chalcones [21-22]. This is confirmed by that the most abundant peaks such as Peaks 12, 13 and 18 were all assigned as flavanones in all SFR samples.
Taken together, the fingerprint analysis of flavonoids led us to conclude that both herbs contained flavanones as their principle flavonoid components, and STR was also a richer source of isoflavones. It still needs to point out that the fine structures of the flavanones in these two herbs were quite different from each other. The flavanones in STR exhibited band Ⅱ absorption at 280–285 nm, while those of SFR showed a red shift with the maximum absorption moving to 290–305 nm, indicating that the flavanones in SFR might possess more auxochromes such as OH or OR groups in ring A than in STR . This hypothesis was supported by our phytochemical study of the herbs and the previous reports [9, 23-29]. The flavanones obtained from SFR always have an OH or OMe group attached to C-5 in ring A, while those from STR were always absent of such groups. One further piece of evidence presented to support this deduction is their retention times of the HPLC chromatograms, which most of the flavanones were eluted before 43 min in SFR instead of 45–63 min in STR, due probably to the increased polarity of the flavanones with more OH substitutes. Therefore, based on the above comparison of the retention times and the online UV spectra of the flavonoid peaks in the two herbs, it was indicated that, in term of flavonoids, they had almost no individual components in common except for two minor pterocarpins at 24.46 min (Peak 3 of SFR and Peak 1 of STR) and 33.66 min (Peak 11 of SFR and Peak 6 of STR), as well as an unknown component at 50.90 min (Peak 24 of SFR and Peak 16 of STR). This result is consistent with the systematic identification of STR flavonoids  and SFR flavonoids  by HPLC-ESI-MS/MS methods.
Again, an inter-species comparison of their flavonoid profiles was carried out by HCA. Based on 54 flavonoid components recognized automatically in the HPLC chromatograms of 24 batches of SFR and 19 of STR samples, a 54 × 43 matrix was formed for analysis. As shown in Fig. 6, except one SFR sample (SF-16) dispersed from the others because of its particularly low flavonoid content, the remaining 22 SFR samples and all 19 STR samples were clearly divided into two groups. Group Ⅰ covered all 19 STR samples and merely one SFR sample (SF-16). And group Ⅱ accounted for all the remaining 22 SFR samples with the characteristic of a much higher flavonoid content. The HCA results showed that these two herbs could be easily distinguished based on their flavonoid constituents. In summary, the results of this study indicated that flavonoids rather than alkaloids could be the diagnostic constituents between SFR and STR, and might be a key factor attributed to their different clinical applications. Quantification and biological evaluation of their flavonoids are currently under investigation.
In the present study, the chemical constituents of SFR and STR were comparatively studied by HPLC fingerprint analysis in terms of alkaloids and flavonoids. Their fingerprints were evaluated by similarity calculation and hierarchical clustering analysis. The results showed that the alkaloid constituents were not a good marker to distinguish the two herbs due to similar alkaloids co-existed, though this type of constituents was long regarded as their major active components. However, their flavonoid constituents were entirely different with few components in common, which could serve as the qualitative markers for authentication of SFR and STR, even when in absence of expected reference substances. Our study also indicated that the different flavonoids in the two herbs might be another type of potential active components that lead to their corresponding unique clinical properties claimed for SFR and STR. This was the first report to systematically compare the chemical constituents of SFR and STR by HPLC fingerprints.
Editorial Committee of Chinese Pharmacopoeia. The Pharmacopoeia of the People's Republic of China[M]. Beijing: China Medical Science and Technology Press, 2015: 27, 202.
Ding PL, Chen DF. Pharmacological actions, clinical applications and adverse effects of roots and rhizomes of Sophora tonkinensis and its preparations[J]. Chin J Clin Pharm, 2003, 12(5): 315-318.
Krishna PM, Rao KNV, Sandhya S, et al. A review on phytochemical, ethnomedical and pharmacological studies on genus Sophora, Fabaceae[J]. Rev Bras Farmacogn, 2012, 22(5): 1145-1154. DOI:10.1590/S0102-695X2012005000043
Wang XK, Li JS, Yan YN. Research progress on the lupine alkaloids and their pharmacological effects of Sophora[J]. World Phytomed, 1996, 11(1): 9-13.
Ding PL, Chen DF, Bastow KF, et al. Cytotoxic isoprenylated flavonoids from the roots of Sophora flavescens[J]. Helv Chim Acta, 2004, 87(10): 2574-2580. DOI:10.1002/(ISSN)1522-2675
Zhu H, Yang YN, Xu K, et al. Sophopterocarpan A, a novel pterocarpine derivative with a benzotetrahydrofuran-fused bicyclo[3.3.1] nonane from Sophora flavescens[J]. Org Biomol Chem, 2017, 15(26): 5480-5483. DOI:10.1039/C7OB01261F
Quang TH, Ngan NTT, Minh CV, et al. Anti-inflammatory and PPAR transactivational properties of flavonoids from the roots of Sophora flavescens[J]. Phytother Res, 2013, 27(9): 1300-1307. DOI:10.1002/ptr.v27.9
Kim JH, Cho CW, Kim HY, et al. α-Glucosidase inhibition by prenylated and lavandulyl compounds from Sophora flavescens roots and in silico analysis[J]. Int J Biol Macromol, 2017, 102: 960-969. DOI:10.1016/j.ijbiomac.2017.04.092
Woo ER, Kwak JH, Kim HJ, et al. A new prenylated flavonol from the roots of Sophora flavescens[J]. J Nat Prod, 1998, 61(12): 1552-1554. DOI:10.1021/np980103j
Li Y, Wu T, Zhu JH, et al. Combinative method using HPLC fingerprint and quantitative analyses for quality consistency evaluation of an herbal medicinal preparation produced by different manufacturers[J]. J Pharm Biomed Anal, 2010, 52(4): 597-602. DOI:10.1016/j.jpba.2010.01.018
Ding PL, Yu YQ, Chen DF. Determination of quinolizidine alkaloids in Sophora tonkinensis by HPCE[J]. Phytochem Anal, 2005, 16(4): 257-263. DOI:10.1002/(ISSN)1099-1565
Jin LX, Cui YY, Zhang GD. Determination of alkaloids of Sophora flavescens by high performance liquid chromatography[J]. Acta Pharm Sin, 1993, 28(2): 136-139.
Yu YQ, Ding PL, Chen DF. Determination of quinolizidine alkaloids in Sophora medicinal plants by capillary electrophoresis[J]. Anal Chim Acta, 2004, 523(1): 15-20.
Zhang X, Cui Z, Wang D, et al. High performance liquid chromatographic fingerprint evaluation of the quinolizidine alkaloids from commercial radix Sphorae flavescents[J]. J Asian Nat Prod Res, 2003, 5(3): 171-182. DOI:10.1080/1028602031000082043
Ma HY, Zhou WS, Chu FJ, et al. HPLC fingerprint of flavonoids in Sophora flavescens and determination of five components[J]. China J Chin Mat Med, 2013, 38(16): 2690-2695.
Dou JH, Li JS, Yan WM. Studies on alkaloids of Sophora tonkinensis[J]. China J Chin Mat Med, 1989, 14(5): 40-42.
Xiao P, Li JS, Kubo H, et al. (-)-14β-Hydroxymatrine, a new lupine alkaloid from the roots of Sophora tonkinensis[J]. Chem Pharm Bull, 1996, 44(10): 1951-1953. DOI:10.1248/cpb.44.1951
Zhang LZ, Li JS, Houghton PJ, et al. Alkaloids in Sophora alopecuroides seed and relevant tests for activity[J]. Chin J Chin Mat Med, 1997, 22(12): 740-743.
Park JA, Kim HJ, Jin C, et al. A new pterocarpan, (-)-maackiain sulfate, from the roots of Sophora subprostrata[J]. Arch Pharm Res, 2003, 26(12): 1009-1013. DOI:10.1007/BF02994750
The Department of Phytochemistry, Shanghai Institute of Materia Medica, Chinese Academy of Science. Handbook of the Systematic Identification of Flavonoids[M]. Beijing: Science Press, 1981: 399, 519, 697, 703.
He CM, Cheng ZH, Chen DF. Qualitative and quantitative analysis of flavonoids in Sophora tonkinensis by LC/MS and HPLC[J]. Chin J Nat Med, 2013, 11(6): 690-698.
Zhang L, Xu L, Xiao SS, et al. Characterization of flavonoids in the extract of Sophora flavescens Ait. by high-performance liquid chromatography coupled with diode-array detector and electrospray ionization mass spectrometry[J]. J Pharm Biomed Anal, 2007, 44(5): 1019-1028. DOI:10.1016/j.jpba.2007.04.019
Shirataki Y, Yokoe I, Komatsu M. Two new flavone glycosides from the roots of Sophora subprostrata[J]. J Nat Prod, 1986, 49(4): 645-649. DOI:10.1021/np50046a014
Yagi A, Fukunaga M, Okuzako N, et al. Antifungal substances from Sophora flavescens[J]. Shoyakugaku Zasshi, 1989, 43(4): 343-347.
Kuroyanagi M, Arakawa T, Hirayama Y, et al. Antibacterial and antiandrogen flavonoids from Sophora flavescens[J]. J Nat Prod, 1999, 62(12): 1595-1599. DOI:10.1021/np990051d
Kang TH, Jeong SJ, Ko WG, et al. Cytotoxic lavandulyl flavanones from Sophora flavescens[J]. J Nat Prod, 2000, 63(5): 680-681. DOI:10.1021/np990567x
Ryu SY, Kim SK, No Z, et al. A novel flavonoid from Sophora flavescens[J]. Planta Med, 1996, 62(4): 361-363. DOI:10.1055/s-2006-957906
Kyogoku K, Hatayama K, Suzuki K, et al. Studies on the constituents of guang-dou-gen (the root of Sophora subprostrata Chun et T. Chen) (5)-Isolation of two new flavanones and daidzein[J]. Chem Pharm Bull, 1973, 21(7): 1436-1439. DOI:10.1248/cpb.21.1436
Komatsu M, Tomimori T, Hatayama K, et al. Studies on the constituents of Sophora species. Ⅲ. Constituents of the root of Sophora subprostrata Chun. et Chen (3)[J]. Yakugaku Zasshi, 1970, 90(4): 459-462. DOI:10.1248/yakushi1947.90.4_459