Determination of saikosaponin derivatives in Radix bupleuri and in pharmaceuticals of the Chinese multiherb remedy Xiaochaihu-tang using liquid chromatographic tandem mass spectrometry

Article abstract of DOI:10.1021/ac0499423

Saikosaponins are bioactive oleanane saponins derived from the Chinese medicinal herb Radix bupleuri ("chaihu" in Chinese). An LC-MS/MS-hased method has been developed for characterization and quantification of 15 saikosaponin derivatives (saikosaponin a, saikosaponin b₁, saikosaponin g, saikogenin A, saikogenin H, saikosaponin C₂, saikosaponin B₂, saikosaponin i, prosaikogenin C₂, prosaikogenin B₂, saikogenin C, saikogenin B, saikosaponin d, saikosaponin b₂, and saikogenin D) in one chromatographic run. Optimization of the ionization process was performed with electrospray and atmospheric pressure chemical ionization techniques in both positive and negative ion modes. Negative ion ESI was adopted for generation of the precursor deprotonated molecules to achieve the best ionization sensitivity for the analytes. In addition, the most abundant fragment ion was chosen for each analyte to give the best CID sensitivity. Because some of the saponin derivatives are isomeric, complete resolution for the whole analytes was achieved both chromatographically and mass spectroscopically. Furthermore, optimal internal standard was successfully discovered for determination of the analytes by making use of a combinatorial chemistry approach. Good linearity over the range ～1.65 or 4.98 to 1200 ng/mL for the analytes was observed. The intraday accuracy and precision at nominal low, intermediate, and high concentration varied between 0.8 and 11.8% and between 80 and 116%, respectively, whereas those for interday assay were between 1.1 and 15.5% and between 86 and 119%, respectively. The lower limits of quantitation for the test compounds were ～16.5 to 49.4 pg on-column. The new method offered higher sensitivity and greater specificity than previously reported LC methods. After the validation, the applicability of the method for determination of these chemicals present in a variety of crude chaihu roots and in different brands of the Chinese multiherb remedy Xiaochaihu-tang (or Shosaiko-to) extract granules has been demonstrated. The sensitivity and specificity of the technique will be the basis of a method for the accurate quantification of the saikosaponin derivatives in biomatrixes.

Full text of DOI:10.1021/ac0499423

Anal. Chem. 2004, 76, 4208-4216
De t e rm in a t io n o f S a ik o s a p o n in De riva t ive s in
Ra d ix b u p le u ri a n d in P h a rm a c e u t ic a ls o f t h e
Ch in e s e Mu lt ih e rb Re m e d y Xia o c h a ih u -t a n g Us in g
Liq u id Ch ro m a t o g ra p h ic Ta n d e m Ma s s
S p e c t ro m e t ry
Yua nw u Ba o, Chua n Li,* Hongw u She n, a nd Fa jun Na n
Shanghai Institute of Materia Medica, Shanghai Institutes for Biological Sciences, Graduate School of the Chinese Academy
of Sciences, Chinese Academy of Sciences, Shanghai 201203, China
Saikosaponins are bioactive oleanane saponins derived
from the Chinese medicinalherb Ra d ix bu pleu r i (“chaihu”
in Chinese). An LC-MS/ MS-based method has been
developed for characterization and quantification of 1 5
saikosaponin derivatives (saikosaponin a, saikosaponin
Xiaochaihu-tang (or Shosaiko-to) extract granules has
been demonstrated. The sensitivity and specificity of the
technique will be the basis of a method for the accurate
quantification of the saikosaponin derivatives in bioma-
trixes.
b
1
, saikosaponin g, saikogenin A, saikogenin H, saikosa-
ponin c, saikosaponin h, saikosaponin i, prosaikogenin
, prosaikogenin B , saikogenin C, saikogenin B, saiko-
saponin d, saikosaponin b , and saikogenin D) in one
Chaihu (Radix bupleuri, Chinese thorowax root) is a favorite
medicinal herb in traditional Chinese medicine (TCM) and is used
as a key ingredient herb for many Chinese multiherb remedies.
C
2
2
2
1
chromatographic run. Optimization of the ionization pro-
cess was performed with electrospray and atmospheric
pressure chemical ionization techniques in both positive
and negative ion modes. Negative ion ESI was adopted
for generation of the precursor deprotonated molecules
to achieve the best ionization sensitivity for the analytes.
In addition, the most abundant fragment ion was chosen
for each analyte to give the best CID sensitivity. Because
some of the saponin derivatives are isomeric, complete
resolution for the whole analytes was achieved both
chromatographically and mass spectroscopically. Fur-
thermore, optimal internal standard was successfully
discovered for determination of the analytes by making
use of a combinatorial chemistry approach. Good linearity
over the range ∼1 .6 5 or 4 .9 8 to 1 2 0 0 ng/ mL for the
analytes was observed. The intraday accuracy and preci-
sion at nominal low, intermediate, and high concentration
varied between 0 .8 and 1 1 .8 % and between 8 0 and
As described in the Chinese pharmacopoeia, chaihu is the dried
roots of Bupleurum chinense DC., known in commerce as Bei-
chaihu, or Bupleurum scorzonerifolium Willd., known as Nan-
chaihu, (family Umbelliferae), which are gathered in Spring or
2
Fall. Bei-chaihu is considered to be superior to Nan-chaihu. In
Japan, the roots of Bupleurum falcatum L. (known as Mishima-
saiko) have also been used as a source of chaihu (“saiko” in
Japanese). Chaihu has been employed traditionally to relieve fever,
to smooth the liver, and to cure drooping and ptosis. The herb
contains a mixture of bioactive oleanane saponins, principally
saikosaponins a (SSa), c (SSc), and d (SSd),3 which are
composed of a pentacyclic triterpene aglycone (saikogenin)
substituted with a sugar side chain of glucose-fucose- or
glucose-(rhamnose)-glucose- as monodesmosides (Scheme 1).
SSa, SSc, and SSd contain an unstable allyl oxide linkage and are
readily converted into diene saponins by mild acid treatment4
or on heating.7,8 The transformed saponins possess a heteroan-
nular diene structure at C-11,13(18) or homoannular diene
-5
-6
1
16%, respectively, whereas those for interday assay were
(
1) The State Pharmacopoeia Commission of P. R. China. Pharmacopoeia of
the People’s Republic of China Chemical Industry Press: Beijing, 2000; pp
163-164, (English Edition 2000, Volume I).
2) Du, X.-W.; Jiang C-M.; Liu M-Y. Zhongguo Zhongyao Zazhi (China J. Chin.
Mater. Med.) 1 9 9 1 , 16, 652-655, 701.
3) Shibata, S.; Kitagawa, I.; Fujimoto H. Chem. Pharm. Bull. 1 9 6 6 , 14, 1023-
1033.
4) Shibata, S.; Kitagawa, I.; Takahashi, R.; Fujimoto H. Yakugaku Zasshi 1 9 6 6 ,
between 1 .1 and 1 5 .5 % and between 8 6 and 1 1 9 %,
respectively. The lower limits of quantitation for the test
compounds were ∼1 6 .5 to 4 9 .4 pg on-column. The new
method offered higher sensitivity and greater specificity
than previously reported LC methods. After the validation,
the applicability of the method for determination of these
chemicals present in a variety of crude chaihu roots and
in different brands of the Chinese multiherb remedy
(
(
(
(
8
6, 1132-1137.
5) Kubota, T.; Hinoh, H. Tetrahedron Lett. 1 9 6 8 , 303-306.
(6) Shimaoka, A.; Seo, S.; Minato, H. J. Chem. Soc., Perkin Trans. 1 1975, 2043-
048.
2
*
Corresponding author address: Shanghai Institute of Materia Medica,
(7) Zhang, W.-X.; Guo, Z.-M.; Wang, X.-H.; Wang, Y.-H. Zhongyao Tongbao (Bull.
SIBS, Chinese Academy of Sciences, 555 Zuchongzhi Road, Zhangjiang
Hi-Tech Park, Shanghai 201203, China. Fax: +86 21 50807088. E-mail: chli@
mail.shcnc.ac.cn.
Chin. Mater. Med.) 1 9 8 5 , 10, 511-513.
(8) Kimata, H.; Hiyama, C.; Yahara, S.; Tanaka, O.; Ishikawa, O.; Aiura, M. Chem.
Pharm. Bull. 1 9 7 9 , 27, 1836-1841.
4208 Analytical Chemistry, Vol. 76, No. 14, July 15, 2004
10.1021/ac0499423 CCC: $27.50 © 2004 American Chemical Society
Published on Web 05/29/2004

Ta b le 1 . He rb Co m p o s it io n o f Xia o c h a ih u -t a n g
amt structural
(g) hierarchy
herb
family
Radix bupleuri (chaihu, Chinese
thorowax root)
Radix scutellariae (huangqin,
baical skullcap root)
Rhizoma pinelliae (banxia,
pinellia tuber)
Umbelliferae
12
9
prime
Labiatae
Araceae
minister
adjutant
9
Radix ginseng (renshen, ginseng) Araliaceae
9
6
adjutant
adjutant
Radix glycyrrhizae (gancao,
Lequminosae
Zingiberaceae
Rhamnaceae
liquorice root)
Rhizoma zingiberis recens
9
emissary
Fig u re 1 . Comparison of different ion sources for ionization of the
5 analytes in terms of signal-to-noise (S/N) ratios. The most intense
(
shengjiang, fresh ginger)
1
Fructus jujubae (dazhao,
12a emissary
peak in each spectrum was chosen for the comparison.
Chinese date)
a
As described by Zhang Zhongjing in his treatise, the number is
2 Chinese dates.
structure at C-9(11),12. The ether saponins could also be trans-
formed into the diene derivatives using gastric juice. In addition,
1
9
both the underivatized and the diene-transformed saikosaponins
could be stripped of their sugar moieties by intestinal flora.9,10
Saikosaponins are of interest because of their broad spectrum of
biological and pharmacological activities coupled with low toxic-
or on cell lines have demonstrated that Xiaochaihu-tang or its
ingredients possess cytoprotective effects on experimental liver
injuries,25-27 an antiinflammatory effect on chronic hepatitis,28,29
ity,11 such as antihepatotoxic activities,12,13 antiinflammatory
preventive and therapeutic effects on experimental hepatic fibrosis
activities,1
4-16
immunomodulatory activities,17,18 and anticancer
via the inhibition of hepatic stellate cells,30-32 and lipid peroxidation
in hepatocytes.32 The herbal remedy also exhibits anticancer
activities.19,20
Xiaochaihu-tang is an important multiherb remedy in TCM and
was first described by the famous Chinese physician Zhang
Zhongjing (150 to 219 A. D. in the Chinese Eastern Han Dynasty)
in his Shang Han Lun, a treatise on febrile diseases. The traditional
remedy is a combination of seven herbs (Table 1), and chaihu is
a key ingredient herb. Xiaochaihu-tang (or Shosaiko-to in Japa-
properties that inhibit chemical hepatocarcinogenesis in animals33
and suppresses the proliferation of hepatoma cells by inducing
apoptosis and arrest at the G0/ G1 phase.34-36
To understand the relationship between administration of
traditional Xiaochaihu-tang and liver dysfunction, a good investiga-
tion of the bioavailability and metabolism of the active constituents
presen, such as saikosaponin derivatives, is needed. Developing
sensitive and reliable quantitative methods for analysis of a range
of components present in Xiaochaihu-tang is a prerequisite to the
drug metabolism and pharmacokinetic (DMPK) evaluation of the
herbal remedy.
nese) has attracted a great deal of attention for its possible healing
effects on chronic hepatitis2
1-23
and its beneficial effect to prevent
the development of hepatocellular carcinoma in patients with
cirrhosis of the liver.24 In addition, many studies in animal models
(
9) Shimizu, K.; Amagaya, S.; Ogihara, Y. J Pharmacobio-Dyn. 1 9 8 5 , 8, 718-
25.
10) Kida, H.; Akao, T.; Meselhy, MR.; Hattori, M. Biol. Pharm. Bull. 1 9 9 8 , 21,
88-593.
11) Tang, W.; Eisenbrand, G. Chinese Drugs of Plant Origin. Chemistry,
Pharmacology, and Use in Traditional and Modern Medicine; Springer-
Verlag: Berlin, 1992; pp 223-232.
7
Several methods for the separation and quantification of
saikosaponin derivatives in plant materials have been described.
Although the chromatographic methods TLC densitometry,3
droplet counter-current chromatography (DCCC),41 and micellar
(
(
5
7-40
(
(
(
(
(
(
(
12) Abe, H.; Sakaguchi, M.; Yamada, M.; Arichi, S.; Odashima, S. Planta Med.
(25) Yamamoto, K.; Araki, N.; Ogawa, K. Acta Histochem. Cytochem. 1 9 8 5 , 18,
403-418.
(26) Ohta, Y.; Nishida, K.; Sasaki, E.; Kongo, M.; Hayashi, T.; Nagata, M.;
Ishiguro, I. Am. J. Chin. Med. 1 9 9 7 , 25, 333-342.
(27) Ono, M.; Miyamura, M.; Kyotani, S.; Saibara, T.; Ohnishi, S.; Nishioka, Y.
J. Pharm. Pharmacol. 2 0 0 0 , 52, 111-118.
1
9 8 0 , 40, 366-372.
13) Abe, H.; Orita, M.; Konishi, H.; Arichi, S.; Odashima, S. J. Pharm. Pharmacol.
9 8 5 , 37, 555-559.
14) Yamamoto, M.; Kumagai, A.; Yamamura, Y. Arzneim.-Forsch. 1 9 7 5 , 25,
021-1032.
1
1
15) Wang, B.-Z.; Wang, G.-Z.; Liu, A.-J. Zhongguo Yaoli XueBao (Acta Pharmacol.
Sin.) 1 9 8 1 , 2, 60-68.
16) Ohuchi, K.; Watanabe, M.; Ozeki, T.; Tsurufuji, S. Planta Med. 1 9 8 5 , 50,
(28) Kakumu, S.; Yoshioka, K.; Wakita, T.; Ishikawa, T. Int. J. Immunopharmacol.
1 9 9 1 , 13, 141-146.
(29) Yamashiki, M.; Nishimura, A.; Huang, X.-X.; Nobori, T.; Sakaguchi, S.;
Suzuki, H. Dev. Immunol. 1 9 9 9 , 7, 17-22.
2
08-212.
17) Yamaguchi, N.; Kohno, H.; Tawara, M.; Odashima, S.; Abe, H. Int. J.
(30) Sakaida, I.; Matsumura, Y.; Akiyama, S.; Hayashi, K.; Ishige, A.; Okita K. J.
Immunopharmacol. 1 9 8 5 , 7, 827-832.
18) Kato, M.; Pu, M.-Y.; Isobe, K.; Iwamoto, T.; Nagase, F.; Lwin, T.; Zhang,
Y.-H.; Hattori, T.; Yanagita, N.; Nakashima, I. Cell Immunol. 1 9 9 4 , 159,
Hepatol. 1 9 9 8 , 28, 298-306.
(31) Kayano, K.; Sakaida, I.; Uchida, K.; Okita, K. J. Hepatol. 1 9 9 8 , 29, 642-
649.
1
5-25.
19) Chou, C.-C.; Pan, S.-L.; Teng, C.-M.; Guh, J.-H. Eur. J. Pharm. Sci. 2 0 0 3 ,
9, 403-412.
20) Tsai, Y.-J.; Chen, I.-L.; Horng, L.-Y.; Wu, R.-T. Phytother. Res. 2 0 0 2 , 16, 117-
21.
(32) Shimizu, I.; Ma, Y.-R.; Mizobuchi, Y.; Liu, F.; Miura, T.; Nakai, Y.; Yasuda,
M.; Shiba, M.; Horie, T.; Amagaya, S.; Kawada, N.; Hori, H.; Ito, S. Hepatology
1 9 9 9 , 29, 149-160.
(
(
1
(33) Okita, K.; Kurokawa, N.; Yamasaki, T.; Furukawa, T.; Li, Q.; Murakami, T.
1
Transgenica 1 9 9 4 , 1, 39-44.
(
(
21) Mizuta, M.; Murata, K.; Morimoto, T. Kan Tan Sui 1 9 8 6 , 12, 77-87.
22) Hirayama, C.; Okumura, M.; Tanikawa, K.; Yano, M.; Mizuta, M.; Ogawa,
N. Gastroenterol. Jpn. 1 9 8 9 , 24, 715-719.
(34) Okita, K.; Li, Q.; Murakamio, T.; Takahashi, M. Eur. J. Cancer Prev. 1 9 9 3 ,
2, 169-175.
(35) Yano, H.; Mizoguchi, A.; Fukuda, K.; Haramaki, M.; Ogasawara, S.;
Momosaki, S.; Kojiro, M. Cancer Res. 1 9 9 4 , 54, 448-454.
(36) Motoo, Y.; Sawabu, N. Cancer Lett. 1 9 9 4 , 86, 91-95.
(37) Akabori, N.; Kagawa, K.; Shimaoka, A. Shoyakugaku Zasshi 1 9 7 5 , 29, 99-
105.
(
(
23) Tajiri, H.; Kozaiwa, K.; Ozaki, Y.; Miki, K.; Shimuzu, K.; Okada, S. Am. J.
Chin. Med. 1 9 9 1 , 19, 121-129.
24) Oka, H.; Yamamoto, S.; Kuroki, T.; Harihara, S.; Marumo, T.; Kim, S. R.;
Monna, T.; Kobayashi, K.; Tango, T. Cancer 1 9 9 5 , 76, 743-749.
Analytical Chemistry, Vol. 76, No. 14, July 15, 2004 4209

electrokinetic capillary chromatography (MEKC)42 have been
reported, high-performance liquid chromatography (HPLC) on
reversed-phase C18 columns with UV detection was the most
frequently used technique for saikosaponin determination.8
The HPLC analysis, however, is limited due to the ether saiko-
saponins SSa, SSc, and SSd not possessing any chromophore
necessary for UV detection. The chromatographic separation of
these saponins had to be traced at shorter UV wavelengths
ranging from 203 to 210 nm. The specificity at shorter wavelengths
was poor, and the sensitivity could also be affected in the case of
using methanol-water gradients. An alternative to shorter wave-
length UV detection is pre-HPLC derivatization of ether saponins
into diene saponins in order to generate a characteristic chro-
mophore that facilitates UV detection at longer wavelength.8,45 The
diene-transformed saponins possess a strong absorption at ∼250
nm. Such methods, however, are not suitable for the direct
analysis of samples containing both types of the saponins. More
recently, the LC-MS technique52 was applied to analyze saiko-
saponins in a chaihu-containing Chinese multiherb remedy. The
method was, however, only focused on analysis of SSa and SSd.
The poor sensitivity and lack of specificity of the above methods
demand developing a more advanced analytical procedure that
can be the basis of a method that is amenable to measurement of
a broad range of saikosaponins and their metabolites at low-level
concentrations in plasma or other biomatrixes following oral
administration of Xiaochaihu-tang.
an LC-MS/ MS-based method that is capable of determining 15
saikosaponin derivatives in one chromatographic run and its
application to analysis of the constituents present in crude drugs
of chaihu and in phytopharmaceuticals of Xiaochaihu-tang or
Shosaiko-to. Because some of the saikosaponins are isomeric, the
compound separation was achieved both chromatographically and
mass spectroscopically. On the basis of the investigation of the
ESI- or APCI-generated ionization of the test compounds as well
as their fragmentation, the new method provides the best
sensitivity and specificity for determination of the saikosaponin
derivatives so far.
,43-51
EXPERIMENTAL SECTION
Chemicals. Reference standard saikosaponin a (SSa, Wako
code no. 194-08411, >98.0%), saikosaponin c (SSc, Wako code no.
1
94-08421, >98.0%), and saikosaponin d (SSd, Wako code no. 194-
8431, >98.0%) were obtained from Wako Pure Chemical (Osaka,
0
Japan). The methods described by Shimizu et al.58 and by Nose
et al.59 were modified for synthesis of saikosaponin b
1
1
(SSb ),
saikosaponin g (SSg), saikogenin A (SGA), and saikogenin H
(
SGH) in acidic conditions from SSa, whereas saikosaponin h
SSh), saikosaponin i (SSi), prosaikogenin B (PSB ), saikogenin
(PSC ), and saikogenin C (SGC) were
made from SSc, and saikosaponin b (SSb ) and saikogenin D
SGD) were from SSd (as depicted in Scheme 1). The reaction
(
2
2
B (SGB), prosaikogenin C
2
2
2
2
(
products were individually applied to HPLC for purification. The
1
H NMR data of the purified saikosaponin derivatives (>99%) were
Recent success with the use of liquid chromatography com-
bined with tandem mass spectrometry (LC-MS/ MS) for char-
compared with the reported data,5
8-62
which were consistent with
the structures shown in Scheme 1. The synthesized products were
also used as reference standards for analytical purposes.
5
3-57
acterizing and quantifying complex plant extracts
suggests
that the technique might also be effective in the comprehensive
determination of multiple saikosaponin derivatives in complex
mixtures. The report presented here details the establishment of
SSa, SSb
dissolved in 50% CH
stock solutions, whereas SGA, SGH, SGD, SGB, and SGC were
in CH CN. These primary stock solutions were pooled and mixed
1
, SSg, SSc, SSh, SSi, PSB
2 2 2
, PSC , SSd, and SSb were
3
CN at 600 µg/ mL to prepare their primary
(
(
(
38) Luo, Y.-Y.; Lin, M.-M. Yaowu Fenxi Zazhi (Chin. J. Pharm. Anal.) 1 9 8 7 , 7,
04-106.
39) Yen, M.-H.; Lin, C.-C.; Chuang, C.-H.; Liu, S.-Y. J. Ethnopharmacol. 1 9 9 1 ,
4, 155-165.
3
1
to obtain an intermediate stock solution at 40 µg/ mL for each of
the 15 test compounds. The stock solutions were stored at -80
3
40) Lin, C.-C.; Yen, M.-H.; Chen, J.-Y.; Chuang, C.-H.; Namba, T. Am. J. Chin.
°
C when not in use.
Med. 1 9 9 1 , 3-4, 265-274.
Corticosterone 21-acetate (CSA) was used as an internal
(
(
(
(
(
41) Otsuka, H.; Kobayashi, S.; Shibata, S. Planta Medica 1 9 7 8 , 33, 152.
42) Hsieh, Y.-Z.; Huang, H.-Y. J. Chromatogr., A 1 9 9 7 , 759, 193-201.
43) Shimizu, K.; Amagaya, S.; Ogihara, Y. J. Chromatogr. 1 9 8 3 , 268, 85-91.
44) Shimizu, K.; Amagaya, S.; Ogihara, Y. J. Chromatogr. 1 9 8 4 , 307, 488-492.
45) Yamaji, A.; Maeda, Y.; Oishi, M.; Hirotani, Y.; Kishi, H.; Kiraoka, E.; Yoneda,
K. Yakugaku Zasshi 1 9 8 4 , 104, 812-815.
standard (IS) for the quantification of 15 saikosaponin deriva-
tives. CSA was detected neither in a crude drug of chaihu nor in
phytopharmaceutical products of Xiaochaihu-tang. A primary IS
stock solution was prepared using 50% MeOH as solvent. An IS
spiking solution (100 ng/ mL) was prepared from the primary IS
solution using 50% MeOH for the dilution. The IS stock solution
and the IS spiking solution were stored at -80 °C when not in
use.
(
(
(
(
(
46) Kanazawa, H.; Nagata, Y.; Matsushima, Y.; Tomada, M. J. Chromatogr. 1990,
5
07, 327-332.
47) Luo, Y.-Y.; Zhang, S.-Q.; Lin, M.-M. Zhongguo Yaoxue Zazhi (Chin. Pharm.
J.) 1 9 9 2 , 27, 215-217.
48) Ma, L.; He, L.-Y.; Song, W.-Z.; Sagara, K.; Ohshima, T.; Maruta, T. Yaowu
Fenxi Zazhi (Chin. J. Pharm. Anal.) 1 9 9 4 , 14 (2), 6-7.
49) Li, G.-H.; Luo, Y.-Y.; Wang, Y.; Yuan, C.-Q.; Wang, N.-H. Zhiwu Ziyuan Yu
Huanjing (J. Plant Resour. Environ.) 1 9 9 5 , 5 (3), 59-60.
50) Liu, D.; Yang, L.-X.; Cui, S.-L. Zhongguo Zhongyao Zazhi (China J. Chin. Mater.
Med.) 1 9 9 8 , 23, 92-93.
3
HPLC-grade acetonitrile (CH CN, 99.9%) was obtained from
Merck (Darmstadt, Germany). Other organic solvents and chemi-
cal reagents used were of analytical grade and were purchased
from Shanghai Chemical Reagent Co. (Shanghai, China). HPLC
water was made by distilling predeionized water twice in this
laboratory.
(
(
51) Park, I.-S.; Kang, E. M.; Kim, N. J. Chromatogr. Sci. 2 0 0 0 , 38, 229-233.
52) Liu, M.-X.; Wu, Z.-P.; Yang, C.-D.; Liu, G.-W. Zhipu Xuebao (J. Chin. Mass
Spectrom. Soc.) 2 0 0 0 , 21, 77-78.
(
(
(
(
(
53) Wang, X.; Sakuma, T.; Asafu-Adjaye, E.; Shiu, G. K. Anal. Chem. 1 9 9 9 , 71,
1
579-1584.
(58) Shimizu, K.; Amagaya, S.; Ogihara, Y. Chem. Pharm. Bull. 1 9 8 5 , 33, 3349-
3355.
(59) Nose, M.; Amagaya, S.; Takeda, T.; Ogihara, Y. Chem. Pharm. Bull. 1 9 8 9
37, 1293-1296.
54) Fuzzati, N.; Gabetta, B.; Jayakar, K.; Pace, R.; Peterlongo, F. J. Chromatogr.,
A 1 9 9 9 , 854, 69-79.
55) da Costa, C. T.; Dalluge, J. J.; Welch, M. J.; Coxon, B.; Margolis, S. A.;
Horton, D. J. Mass Spectrom. 2 0 0 0 , 35, 540-549.
(60) Tori, K.; Seo, S.; Yoshimura, Y.; Nakamura, M.; Tomita, Y.; Ishii, H.
56) Zeeb, D. J.; Nelson, B. C.; Albert, K.; Dalluge, J. J. Anal. Chem. 2 0 0 0 , 72,
Tetrahedron. Lett. 1 9 7 6 , 46, 4167-4170.
5
020-5026.
(61) Ishii, H.; Nakamura, M.; Seo, S. Chem. Pharm. Bull. 1 9 8 0 , 28, 2367-2383.
(62) Liang, H.; Zhao, Y.-Y.; Qiu, H.-Y.; Huang, J.; Zhang, R.-Y. Yaoxue Xuebao
(Acta Pharm. Sin.) 1 9 9 8 , 33, 37-41.
57) Chan, T. W. D.; But, P. P. H.; Cheng, S. W.; Kwok, I. M. Y.; Lau, F. W.; Xu,
H. X. Anal. Chem. 2 0 0 0 , 72, 1281-1287.
4210 Analytical Chemistry, Vol. 76, No. 14, July 15, 2004

S c h e m e 1 . S t ru c t u re s o f t h e S a ik o s a p o n in De riva t ive s a n d t h e Ch e m ic a l Co n ve rs io n s o f t h e Ole a n a n e
S a p o n in s
Liquid Chromatography-Mass Spectrometry. The liquid
chromatography-mass spectrometry system consisted of an
Agilent 1100 series liquid chromatograph (including a vacuum
degasser, a quaternary pump, and an autosampler; Waldbronn,
Germany) coupled to an Applied Biosystems/ MDS SCIEX API-
was changed back to 31% at 24.1 and maintained from 24.1 to 32
min for the analysis of the next sample. The effluent was delivered
at 0.2 mL/ min during the whole gradient program. The injection
volume was 10 µL, and the UV absorbance of the effluent was
monitored from 195 to 400 nm.
3
000 triple-quadrupole mass spectrometer equipped with a Tur-
Tuning solutions of the individual analyte at the concentration
of 10 µg/ mL were prepared by diluting the corresponding primary
stock solutions with 50% CH CN. The tuning solution was
3
boionspray ion source (Foster City, CA). The LC-MS/ MS system
was controlled by Analyst software. In addition, another LC-MS
system consisting of an Agilent 1100 LC pump and an autosampler,
and a Finnigan LCQ Deca ion-trap mass spectrometer fitted with
an electrospray ionization (ESI) source or an atmospheric pressure
chemical ionization (APCI) source (San Jose, CA) was used in
the early phase of the method development. The LCQ Deca
spectrometer was also fitted with a motorized six-pot divert valve
for diverting the LC flow between the mass spectrometer and
waste.
The saikosaponin derivatives were separated using a 5-µm
Zorbax XDB-C18 column (50 mm × 2.1 mm i.d.; Chadds Ford,
PA), before which a 0.2-µm filter (Upchurch Scientific, Oak
Harbor, WA) was used. The LC mobile phases were CH
H
delivered at 5 µL/ min by a syringe pump and was combined
through a Peek tee union with a mixture of solvents A and B
delivered at 0.2 mL/ min by the LC pump. The ratio of solvents A
and B varied for different test compounds according to the ratios
in the postcolumn effluents containing the corresponding com-
pounds. Instrumental parameters of the MS spectrometers were
optimized to achieve the maximum ionization of the analyte
molecules and the generation of the characteristic fragment ions.
Quantitative analysis of the saikosaponin derivatives in the
herbal samples was performed using the triple-quadrupole LC-
MS/ MS system. The operating parameters of the mass spectrom-
eter included the compound-dependent and source-dependent
considerations. The optimized source-dependent parameters con-
sisted of the flow rates of the nebulizer gas, the curtain gas,
collision gas, the ionspray voltage, and the temperature of the
heater gas. The compound-dependent parameters for the test
compounds and IS, including the declustering potential, the
focusing potential, the entrance potential, the collision energy, and
3
CN/
2
O (10:490, v/ v) for solvent A and CH CN/ H O (450:50, v/ v)
3
2
for solvent B. The LC binary gradient program consisted of an
initial 7-min linear gradient segment of increasing B from 31 to
3
7
4%, followed by an isocratic segment maintaining B at 34% from
to 11 min. Then the linear gradient was changed progressively
by increasing B to 42% at 18 min and 100% at 24 min. Finally, B
Analytical Chemistry, Vol. 76, No. 14, July 15, 2004 4211

the collision cell entrance potential, were tuned for each test
compound to achieve the highest instrument response. The mass
spectrometer was operated at low mass resolution for both Q1
and Q3 in multiple reaction-monitoring (MRM) mode.
To construct standard curves for the quantification of the
saikosaponin derivatives in the herbal materials investigated,
calibrator pool solutions containing the 15 test compounds were
prepared from the intermediate standard stock solution (40 µg/
mL) by dilution with 50% MeOH to concentrations of 1200, 400,
133, 44.4, 14.8, 4.94, and 1.65 ng/ mL. Ten microliters of the IS
spiking solution (40 ng/ mL) was added to each variety of the
calibrator pool solution (100 µL). Calibration graphs were con-
structed using a linear regression of the test compound/ IS peak
area ratio (Y) to nominal concentration of the test compound (X,
ng/ mL) with weighting of the reciprocal concentration (1/ X). To
determine the intraday and interday accuracy and precision of
the analytical method described here, 50% MeOH solutions
containing the 15 test compounds at three different nominal
concentrations (14.8, 44.4, and 400 ng/ mL) were analyzed, and
the quality control values were calculated from the regression
equations (data not shown).
Quantification of the Saikosaponin Derivatives in Herbal
Materials. A reference drug sample of pulverized dry roots of
Bupleurum chinense (Chaihu-ref, Catalog No. 0992-200102) was
obtained from the National Institute for the Control of Pharma-
ceutical and Biological Products (Beijing, China). Five other crude
drugs of chaihu were also investigated. Chaihu-1 was identified
as the dry roots of B. chinense, which was obtained from the
Institute of Chinese Materia Medica, China Academy of Traditional
Chinese Medicine (Beijing, China). Chaihu-2 and Chaihu-3 were
obtained from Shanxi Pharmaceutical Company (Taiyuan, China),
which had been gathered from the wild plants (Bupleurum spp.)
growing in Lingchuan County (Shanxi Province, China) and from
stands of B. chinense cultivated at a local farm in compliance with
good agriculture practice (GAP) in the same county, respectively.
Chaihu-4 was purchased from Shanghai Lei’s Pharmaceutical Co.,
Ltd. (Shanghai, China) with unknown habitat. Chaihu-5 was
obtained from Tasly Group (Tianjin, China). The identity of these
commercial crude drugs was reestablished organoleptically as
Bupleurum spp. by the specialists from the Department of
Pharmacognosy at Shanghai University of Traditional Chinese
Medicine (Shanghai, China). Garbling the crude drugs of chaihu
consisted of the removal of extraneous matter, such as other parts
of the plant, dirt, and added adulterants and was carried out before
pulverizing the crude drugs. The herbal samples were stored at
RESULTS AND DISCUSSION
MS and MS/ MS Optimization. Using the Zorbax XDB-C18
3 2
column and the mobile phase of CH CN/ H O provided sym-
metrical and sharp chromatographic peaks for the test saikosa-
ponin derivatives. The effect of different mobile phase additives,
such as formic acid, acetic acid, ammonium formate, and am-
monium acetate, on the ionization in LC-MS was not further
studied. To find the most sensitive ionization method for the
analytes, ESI and APCI were tested with the same LC mobile
phase at flow rates of 0.2 or 0.5 mL/ min. The pseudomolecular
ions formed were identified from the mass spectra obtained, which
were achieved by applying the individual analyte to the LC-UV-
MS using the C18 column and the mobile phase, leading to short
-20 °C until use.
For each variety of chaihu, ∼200 mg of the pulverized crude
drug was steeped at ambient temperature (∼23 °C) in 10 mL of
chromatographic retention for the test compound (t
The peak appearing in the ion chromatogram extracted from the
R
, ∼2 to 3 min).
5
0% MeOH. The mixture was vortexed for 3 min, ultrasonicated
for 2 min, and vortexed for 3 min again. The supernatants were
then separated following centrifugation at 1175g for 10 min. After
a second extraction with 10 mL of 50% MeOH, the combined
methanol extract (∼19 mL) was filtered through a 0.45-µm nylon
filter (Shanghai, China). The resultant filtants were stored at -80
LC-MS in full-scan mode was compared with the peak appearing
in the on-line UV chromatogram (at 210 nm) in terms of their t
R
.
If no peak emerged at the t in the extracted ion chromatogram,
R
the suspected ion could not be derived from the test compound.
Since the MS spectrometer was connected behind the UV
detector, there was an ∼0.1-min delay of the peak retention time
in the ion chromatogram, as compared with that in the corre-
sponding UV trace.
°C pending LC-MS/ MS analysis. Further extraction of the herbal
samples under the same conditions afforded only a trace of
material, indicating that the two-step extraction was enough. The
formation of the diene saponins from the oleanane saikosaponins
was negligible under the extracting conditions, and thus, pyridine
was not added into the extracting solvent.
In positive ion ESI experiments, the saikosaponins showed a
very high tendency to form alkali metal adducts. The sodiated
+
molecule seen at m/ z [M + Na] was the base peak in the spectra
Seven brands of Xiaochaihu-tang phytopharmaceuticals (X-1,
X-2, X-3, and X-4) and Shosaiko-to phytopharmaceuticals (S-1, S-2,
and S-3) were purchased from drugstores in Shanghai (China)
and Tokyo (Japan), respectively, and used for research purposes
only. They were tea granules of the multiherb remedy prepared
by various pharmaceutical manufactures. On the product labels,
the manufacturers claimed that the herbal remedies were com-
posed of the ingredient herbs according to the recipe of Xiaochaihu-
tang described in Zhang Zhongjing’s Shang Han Lun. The
phytopharmaceuticals purchased were stored at -20 °C until use.
For LC-MS/ MS analysis, each variety of Xiaochaihu-tang or
Shosaiko-to granules (500 mg) was dissolved in 10 mL of 50%
methanol. After centrifugation at 1175g for 10 min, the supernatant
was filtered through the 0.45-µm nylon filter. The resultant herbal
solution was kept frozen at -80 °C when not in use.
of all saikosaponins, together with less abundant adduct ions [M
K]+ and [2M + Na] . In addition, ion [M + H - H O]+ was
+
+
2
also detected for most saikosaponins, but in some cases, small
]+ and [M + H - 2H O]+ were visible
peaks for ions [M + NH
4
2
in the spectra, too. The ionization profiles of the saikogenins in
the positive ion ESI mode were quite different from those of the
saikosaponins. Alkali metal adducts were not formed in the ESI
source. Instead, the main ions derived from the saikogenins were
+
+
[M + H - H O] and [M + H - 2H O] , whereas ion [2M +
2
2
+
H] was also detected. The high tendency for the adduct formation
substantially reduced the formation of protonated molecules so
that peaks for [M + H]+ were very weak or even not seen in the
spectra of the saikosaponin derivatives.
In the negative ion ESI mode experiments, the peak due to
the deprotonated molecule [M - H]- was detected in all spectra,
4212 Analytical Chemistry, Vol. 76, No. 14, July 15, 2004

Ta b le 2 . Io n s o f S a ik o s a p o n in De riva t ive s De t e c t e d in
LC-MS S p e c t ra in Ne g a t ive ES I Mo d e
iona
chemical [M - H]- [2M - H]- [M + Cl]- [M + HCOO]-
SSa
SSb1
SSg
SGA
SGH
SSc
779 (100)
779 (100)
779 (100)
471 (60)
471 (100)
925 (100)
925 (100)
925 (100)
779 (100)
779 (100)
455 (80)
1559 (20)
1559 (30)
1559 (15)
815 (7)
815 (8)
815 (25)
825 (10)
825 (10)
825 (5)
517 (100)
517 (30)
971 (15)
971 (14)
971 (22)
825 (80)
825 (65)
501 (100)
501 (50)
825 (18)
825 (15)
517 (100)
1853 (12)
1853 (8)
1853 (7)
1559 (50)
1559 (30)
961 (12)
961 (10)
961 (26)
SSh
SSi
PSB2
PSC2
SGB
SGC
SSd
455 (100)
779 (100)
779 (100)
471 (10)
1559 (50)
1559 (15)
815 (10)
815 (7)
SSb2
SGD
a
m/ z with relative abundance (%) in parentheses.
which was the most intense peak for the analytes (except the three
saikogenins SGA, SGB, and SGD). Meanwhile, small peaks for
-
-
-
[
2M - H] , [M + Cl] , and [M + HCOO] were also seen for
the test saikosaponins. As shown in Table 2, formate adduct [M
HCOO]- peak was abundant for the saikogenins studied,
+
especially for SGA, SGB, and SGD, which was the base peak.
The phenomenon of abundant adduct formation also occurred
in the positive ion APCI mode. The main adduct for the saikosa-
+
ponins was sodium adduct [M + Na] , but in a few cases, the ion
Fig u re 2 . Discovery of an optimal internal standard for the mix-
ture of the 15 test saikosaponin derivatives using a combinatorial
chemistry approach. For panel 1, corticosterone (CCR) was tested
using the LC-MS/MS method developed for the mixture of the 15
analytes. For panel 2, CCR was used as a starting material and
reacted with nine acid chlorides. For panel 3, LC-UV-ESI-MS was
used to trap and detect IS candidate(s) from the reaction mixture.
For panel 4, the (-) ESI-MS and MS/MS spectra were obtained for
IS candidate A.
+
[
H
M + H - H
2
O] was also detected. Both the ions [M + H -
O] were abundant in the spectra of
the saikogenins. In negative ion APCI experiments, the chlorine
adduct [M + Cl] was the main peak for all saikosaponins (except
SSb
), and the deprotonated molecule [M - H]- was also
detected. The saikogenins seemed to be very poorly ionized in
+
+
2
O] and [M + H - 2H
2
-
1
the APCI source.
On the basis of the above MS experimental results (Figure
1
), negative ion ESI was adopted for generation of the precursor
Discovery of An Optimal Internal Standard. Since its
chemical structure is close to those of the analytes, corticosterone
(CCR) was first examined for IS discovery. Under the MS/ MS
conditions (in the negative ion ESI mode) developed for the
quantification of the saikosaponin derivatives, CCR exhibited
abundant generation of the deprotonated molecule ([M - H]-,
m/ z 345) (Figure 2/ panel 1) and efficient production of its
characteristic fragment ion (m/ z 327) (data not shown). The
deprotonated molecule [M - H]- in MS/ MS experiments of this
work, which gave the highest signal-to-noise ratio for most of the
analytes. Formate adduct [M + HCOO]- could be a relevant
precursor ion for SGA, SGB, and SGD in the future. The CID
spectra from the deprotonated saikosaponins exhibited a frag-
mentation pattern corresponding to successive loss of glycosidic
units, which allows a straightforward interpretation of the spectra.
In two-stage tandem MS experiments, the most abundant ions
were due to loss of the terminal rhamnose from SSc, SSh, and
SSi or the glucose from the other saikosaponins. The negative
ion MS/ MS spectra showed a clear difference between the
heteroannular and the homoannular saikogenins, in which the
characteristic fragment ions of the deprotonated molecules
chromatographic retention of CCR (t
too short and out of the retention time range of the 15 analytes
(t
, ∼5.8 to 27.7 min, see Figure 3).
Instead of the traditional way of serially testing compounds
R
, 5.2 min), however, was
R
for IS discovery by trial and error, that is, giving up CCR and
testing another compound, CCR was retained as an IS lead for
further optimization. With a combinatorial chemistry approach,
about 100 CCR derivatives were rapidly synthesized as a mixture
in the same reaction vessel. In brief, CCR was used as starting
material and was dissolved in dry dichloromethane at 0 °C. A
mixture of nine acid chlorides (0.6 mmol for each, Figure 2, panel
-
-
included the ions [M - H - H
-
for the individual transition, the fragment ions seen at m/ z [M -
2
3
O] , [M - H - CH OH] , or [M
-
H - CH
2
O - H
2
O] . After the optimization of the conditions
-
H - CH
2
O - H
2
O] were most abundant for SGA and SGC. In
contract, those at m/ z [M - H - H
2
O]- were most abundant for
SGH and SGB. The main fragment ion for SGD was ion [M - H
2 2
2) and triethylamine (12 mmol) were added to the CH Cl solution
-
-
-
CH
3
OH] , together with the less abundant ion [M - H - H
2
O]
of CCR (0.15 mmol). The reaction mixture was stirred until the
disappearance of CCR (checked by TLC) and then mixed with
-
and the poorly formed ion [M - H - CH
2
O - H
2
O] .
Analytical Chemistry, Vol. 76, No. 14, July 15, 2004 4213

abundant for the corresponding analytes and gave the best
sensitivity. Because some of the saikosaponins are isomeric,
showing the same MS/ MS fragmentation pattern, compound
resolution cannot be achieved spectroscopically only. An advanced
liquid chromatographic separation system was needed to ad-
equately resolve the isomers between SSa and SSb
2
(m/ z 779 f
6
17), between SSg and SSb (m/ z 779 f 617), and between SSc
1
and SSi (m/ z 925 f 779). As depicted in Figure 3, complete
resolution was successfully achieved both chromatographically
and spectroscopically for the 15 analytes in a 32-min chromato-
graphic run. Due to the LINAC collision cell technology of API-
3000, varying the dwell times for the ion transitions and grouping
the analytes in chromatographic time windows for MRM analysis
did not exhibit evidently enhanced sensitivity of the detection
method. Matrix-induced interference resulting in suppression
of ionization of test compounds in LC-MS has been well-
recognized.6
3-65
Under the LC-MS/ MS conditions developed in
this work, however, no obvious interference was observed in the
analysis of the complex herbal extract samples.
Quantification was performed using internal standard method.
The introduction of IS for biomatrixes before sample cleanup will
improve precision and accuracy for the quantitative analysis. Using
CSA as IS, good correlations were achieved for all the 15 analytes
relative to the compound with calibration curve regression
coefficients >0.99 spanning a concentration range of ∼1.65 or 4.94
to 1200 ng/ mL. The precision of this analytical method was
determined by calculating the relative standard deviation (RSD)
of the concentrations measured on the same day (n ) 5) and on
different days (n ) 5) for the calibration standard samples of
different nominal concentrations. As shown in Table 3, the RSD
never exceeded 15.5% at the concentrations examined, indicating
good assay precision. Meanwhile, the accuracy ranged from 80
to 119% for the 15 analytes. The method provided a lower limit of
quantification (LLOQ) of 16.5 or 49.4 pg on-column for the 15
test saikosaponin derivatives, which is quite sensitive and can be
the basis of method utilized for DMPK studies of Xiaochaihu-
tang or other chaihu-containing herbal remedies.
Fig u re 3 . Extracted ion chromatograms of the mixture of the
saikosaponin derivatives SSa, SSb1, SSg, SGA, SGH, SSc, SSh, SSi,
PSB2, PSC2, SGB, SGC, SSd, SSb2, SGD, and IS (CSA) from the
LC-MS/MS experiment (MRM mode) with negative ESI. The ion
transitions used for extraction are marked for each chromatogram
-
(
[M - H] f fragment ion).
1
0 mL of H
with 10 mL of EtOAc. After centrifugation, the combined EtOAc
extract was treated in turn with 1 N HCl, H O, brine, and MgSO
2
O. The resultant mixture was extracted three times
2
4
and then concentrated by evaporation under reduced pressure to
yield an oily mixture of esters. The resultant reaction mixture was
then applied to LC-MS/ MS using the chromatographic and
ionizing conditions developed for the mixture of the 15 test
saikosaponin derivatives. To trap the IS candidates, the divert valve
on the LCQDECA MS spectrometer was set to introduce the eluent
flow during 13-20 min to the ion source with the other eluent
flows to waste. As depicted in Figure 2, panel 3, there was one
R
compound detected at t 16.1 min, the chromatographic peak that
Application to Analysis of Samples of Chaihu and
Xiaochaihu-tang. The quantified results obtained for six crude
chaihu roots are present in Table 4. The test crude drugs except
Chaihu-ref were garbled before sampling. Each sample was
analyzed in duplicate. Chaihu-ref and the other chaihu samples
analyzed in this work ranged from ∼4.18 to 6.84 mg total
saikosaponin derivatives/ g of pulverized sample. The major
saikosaponins detected in Chaihu-ref and the other chaihu samples
were the oleanane saponins SSa, SSc, and SSd, whereas the other
saikosaponins derivatives, such as the diene saponins and the
hydrolyzed products, were scarcely detected (<LLOQ of our
present analytical method) or not found in any sample. The levels
of the individual saikosaponins SSa, SSc, and SSd present in
Chaihu-ref were 2.54, 0.99, and 1.06 mg/ g, respectively. A
comparable content of the individual saikosaponin in samples was
found for garbled Chaihu-1 to be approximately -5, i.e., ∼2.13 to
was labeled “IS candidate A”. The IS candidate A (ISC-A) showed
-
an abundant [M - H] at m/ z 387 and an abundant [M +
-
HCOO] at m/ z 433 in the ESI-MS spectrum (Figure 2, panel 4),
the ionization pattern of which was similar to those of the analytes
(
see Table 2). The difference of the molecular weight between
ISC-A and CCR is 42 mass units, suggesting that ISC-A might be
an acetylated product of CCR. The CID spectrum of the depro-
tonated molecule, obtained in the full scan MS/ MS mode,
exhibited a characteristic fragment ion at m/ z 345. To confirm
the chemical structure of ISC-A, CCR was reacted with acetyl
chloride. Fortunately, ISC-A was also synthesized from the
1
reaction and purified by HPLC. On the basis of H NMR analysis
(
(
data not shown), ISC-A was identified as corticosterone 21-acetate
CSA). CSA was found by LC-MS/ MS in neither test herbal
samples of this study nor blank plasma, urine, or tissue homo-
genate samples of this laboratory, suggesting it is amenable to
being used as IS for the analysis.
3
.62 mg/ g for SSa, ∼1.10 to 2.10 mg/ g for SSc, and ∼0.88 to 1.12
Development and Validation of Quantification Method.
The optimal conditions were applied for the TurboIonSpray
ionizations (an ESI) in negative ion mode and ion transitions. The
fragment ions chosen for quantification purposes were most
(63) Pascoe, R.; Foley, J. P.; Gusev, A. I. Anal. Chem. 2 0 0 1 , 73, 6014-6023.
(
64) Matuszewski, B. K.; Constanzer, M. L.; Chavez-Eng, C. M. Anal. Chem.
2
0 0 3 , 75, 3019-3030.
(
65) Tang, Y.-B.; Gu, J.-K.; Wang, Y.-G.; Zhong, D.-F. Asian J. Drug Metab.
Pharmacokinet. 2 0 0 3 , 3, 110-115.
4214 Analytical Chemistry, Vol. 76, No. 14, July 15, 2004

Ta b le 3 . In t ra d a y a n d In t e rd a y Va ria t io n s o f S a ik o s a p o n in De riva t ive Ca lib ra t io n s a
low concn (ng/ g)
RSD
intermediate concn (ng/ g)
high concn (ng/ g)
RSD
chemical
FCb
Ac
FC
RSD
A
FC
A
Intraday (n ) 5)
SSa
SSb1
SSg
SGA
SGH
SSc
15.4 ( 1.8
16.3 ( 0.4
13.2 ( 1.2
15.2 ( 1.5
15.2 ( 0.6
13.9 ( 1.4
14.6 ( 1.2
14.1 ( 0.2
14.3 ( 0.8
15.1 ( 0.2
15.3 ( 1.2
11.9 ( 1.2
14.9 ( 0.7
14.7 ( 1.2
14.9 ( 1.4
11.8
2.7
9.3
10.2
4.2
9.9
7.9
1.7
5.5
1.3
7.9
10.1
4.6
8.3
9.2
104
110
89
103
103
94
99
95
97
102
103
80
100
99
45.1 ( 2.3
5.1
7.7
7.1
5.3
4.4
0.8
6.6
4.2
5.2
3.1
6.5
3.6
2.3
2.2
6.5
102
104
104
90
116
104
101
102
106
104
95
409.0 ( 15.8
406.6 ( 25.6
392.5 ( 24.6
433.7 ( 13.2
431.6 ( 16.7
415.8 ( 11.3
410.5 ( 13.0
409.6 ( 4.8
406.0 ( 12.9
407.7 ( 12.9
391.8 ( 28.0
374.1 ( 37.9
409.9 ( 28.8
405.4 ( 15.8
438.9 ( 20.3
3.8
6.3
6.3
3.0
3.9
4.6
3.1
5.1
3.2
3.2
7.1
10.1
7.0
4.0
4.6
103
102
98
46.1 ( 3.6
46.1 ( 3.3
40.0 ( 2.1
51.4 ( 2.2
46.3 ( 0.4
44.8 ( 3.0
45.3 ( 1.9
47.1 ( 2.5
46.1 ( 1.7
42.1 ( 2.7
36.2 ( 1.3
43.7 ( 1.0
48.5 ( 1.1
46.0 ( 3.0
108
108
104
103
102
102
102
98
SSh
SSi
PSB2
PSC2
SGB
SGC
SSd
82
98
109
104
94
103
101
109
SSb2
SGD
100
Interday (n ) 5)
SSa
SSb1
SSg
SGA
SGH
SSc
12.8 ( 0.4
13.6 ( 1.6
15.9 ( 0.9
15.3 ( 2.0
16.6 ( 0.8
13.6 ( 1.6
13.1 ( 1.1
13.3 ( 1.0
13.4 ( 1.3
15.0 ( 0.7
15.7 ( 1.5
13.5 ( 1.5
13.1 ( 0.9
13.5 ( 0.4
13.4 ( 1.1
3.2
11.6
5.5
13.5
4.9
11.0
8.3
7.5
8.8
4.9
9.3
11.1
7.0
86
92
107
103
112
92
88
90
91
101
106
91
88
91
42.8 ( 1.5
43.2 ( 4.6
43.0 ( 5.5
41.0 ( 2.3
53.0 ( 1.6
44.5 ( 2.2
45.3 ( 3.0
44.4 ( 1.1
49.2 ( 1.3
44.8 ( 1.3
39.4 ( 3.0
38.6 ( 3.9
39.2 ( 2.9
48.6 ( 1.3
42.9 ( 2.4
3.4
10.7
12.7
5.7
3.0
5.0
6.7
2.5
2.6
2.8
7.5
10.1
7.3
2.6
5.6
96
97
97
390.7 ( 20.7
397.5 ( 32.9
414.3 ( 49.6
415.3 ( 58.8
401.7 ( 21.2
414.2 ( 21.2
406.7 ( 10.8
404.7(4.5
410.4 ( 32.8
409.4 ( 15.5
392.1 ( 60.9
358.5 ( 44.0
418.3 ( 34.7
403.1 ( 15.1
415.9 ( 55.3
5.3
8.3
12.0
14.2
5.3
5.1
2.7
1.1
7.9
98
99
104
104
100
104
102
101
103
102
98
92
119
100
102
100
111
101
89
87
88
109
97
SSh
SSi
PSB2
PSC2
SGB
SGC
SSd
4.8
15.5
12.3
8.3
3.7
13.3
90
105
101
104
SSb2
SGD
2.7
8.3
91
a
Nominal low, intermediate, and high concentrations were 14.8, 44.4, and 400 ng/ mL, respectively. b FC: found concentration in ng/ mL. c A:
accuracy in %.
Ta b le 4 . Qu a n t ific a t io n o f S a ik o s a p o n in s P re s e n t in Va rio u s Cru d e Dru g s o f Ch a ih u a
chemical
Chaihu-ref
Chaihu-1
Chaihu-2
Chaihu-3
Chaihu-4
Chaihu-5
SSa
SSc
SSd
Total
2.54 (55.3%)
0.99 (21.6%)
1.06 (23.1%)
4.59
2.36 (49.2%)
1.52 (31.7%)
0.92 (19.1%)
4.80
3.62 (52.9%)
2.10 (30.7%)
1.12 (16.4%)
6.84
3.34 (55.4%)
1.63 (27.0%)
1.06 (17.6%)
6.03
2.13 (51.0%)
1.17 (28.0%)
0.88 (21.0%)
4.18
2.23 (51.8%)
1.10 (25.5%)
0.98 (22.7%)
4.31
a
Values are the measured levels expressed in mg/ g. Values in parentheses are the percentage of the chemical in the total saikosaponin measured.
mg/ g for SSd (Table 4). The ratio of SSa, SSc, and SSd was around
homoannular isomers SSg and SSi, respectively. The ratios of SSb
1
2
.6:1.4:1 (w/ w/ w).
The quantified results obtained for the seven brands of
and SSg and that of SSh and SSi were 1.61 ( 0.10 and 1.76 (
0.05 for Chinese granules (n ) 4) and 1.39 ( 0.29 and 1.94 (
0.10 for Japanese granules (n ) 3), respectively. The major
saikosaponins contained in the test Xiaochaihu-tang and Shosaiko-
Xiaochaihu-tang or Shosaiko-to phytopharmaceuticals are pre-
sented in Table 5. Besides the oleanane saponins SSa and SSc,
the diene saponins SSb
1
, SSg, SSh, SSi, and SSb
2
were also found
2 1
to samples were SSb (26.9%), SSa (25.8%), SSb (22.4%), SSg
in the samples of the herbal extract granules, suggesting the
formation of the diene saikosaponins during Xiaochaihu-tang or
Shosaiko-to granule manufacturing. There is no data supporting
any occurrence of further hydrolysis of these saponins during the
manufacturing course. SSd was not detected in most of the granule
samples, except for the Japanese Shosaiko-to samples S1 and S2,
in which only a quite low amount of SSd was found. SSd seemed
(14.3%), SSc (6.9%), SSh (5.4%), and SSi (2.9%). Although the daily
ingested amount of the total saikosaponin for the four brands of
Chinese Xiaochaihu-tang granules (20.32 ( 8.55 mg, RSD 42.1%,
n ) 4) was comparable to that of the three brands of Japanese
Shosaiko-to granules (14.01( 5.21 mg, RSD 36.5%, n ) 3), the
saikosaponin composition of these phytopharmaceuticals was
inconsistent, probably as a result of the discrepancy in raw
materials used and during the herb extraction process. The
Chinese manufacturers used a greater amount of excipients in
their Xiaochaihu-tang granules than Japanese manufacturers did,
resulting in a bigger size of the products for daily administration.
to be transformed into SSb
oleanane saponins coexisted with their transformed diene saponins
to various extents. The heteroannular diene products SSb and
SSh seemed to be more abundant than their corresponding
2
. In the cases of SSa and SSc, the
1
Analytical Chemistry, Vol. 76, No. 14, July 15, 2004 4215

for these chemicals was achieved both chromatographically and
spectroscopically. Making use of combinatorial chemistry ap-
proach successfully created an optimal internal standard for
quantification. The glycosidic linkages can be determined from
the CID multistage MS experiments using [M - H]- as the
precursor ions. The applicability of the method for determination
of these chemicals present in a variety of crude chaihu roots and
in different brands of Xiaochaihu-tang (or Shosaiko-to) extract
granules has been demonstrated. The sensitivity and specificity
of the described LC-MS/ MS technique will be the basis of a
method for the accurate quantification of the saikosaponin deriva-
tives in biological fluids and could be applied to the DMPK
evaluation and studies of the Chinese multiherb remedy Xiaochaihu-
tang.
Ta b le 5 . Qu a n t ific a t io n o f S a ik o s a p o n in s P re s e n t in a
Va rie t y o f P h yt o p h a rm a c e u t ic a ls o f Xia o c h a ih u -t a n g o r
S h o s a ik o -t o
a
X1
X2
X3
X4
S1
S2
S3
b
chemical (60 g) (60 g) (60 g) (60 g) (7.5 g) (6 g) (7.5 g)
SSa
7.2
66.1
42.5
TA
16.5
9.3
ND
63.8
205.4
90.8
75.7 204.8
50.0 123.3
4.6
5.4
3.1
ND
71.9 109.4
301.5 540.6 307.1 2370.1 2680.3 1092.1
72.9
9.6
95.4
55.3
2.1
29.1
15.9
ND
673.7 928.2 352.0
238.8 214.6 132.3
SSb1
SSg
SSc
SSh
SSi
SSd
SSb2
Total
183.7 186.1
77.5
c
4.0
16.5
9.7
389.0 341.6 176.0
124.2 119.0
64.3
32.9
ND
67.1
7.7
58.3
70.0
d
ND
99.7
685.9 762.5 257.1
a
The values are the measured levels in µg/ g granules. The amount
of the saikosaponin for administration per day can be calculated by
multiplying the measured level by the amount of the granules for daily
b
c
ingestion. The amount of the granules for ingestion per day. TA,
d
ACKNOWLEDGMENT
trace level detected (lower than the LLOQ). ND, not detected.
This study was supported, in part, by Grant 30271596 of the
National Natural Science Foundation of China, Grant 01DJ19008
of the Science and Technology Commission of Shanghai Munici-
pality, and Grant 2003AA2Z347A of the Chinese Ministry of
Science and Technology. Some experiments of this work were
performed in Prof. Y-M. Yang’s laboratory at this research institute
or in Prof. C. Yu’s laboratory at Shanghai Xuhui Central Hospital.
The authors thank Dr. J. Zheng from Northeastern University for
the helpful suggestions.
Full evaluation of Xiaochaihu-tang products depends not only on
the analysis of saikosaponins but also on the active principles of
the other herbs. The methods for the quantification of the
medicinal constituents in the other herbs of Xiaochaihu-tang have
also been developed in this laboratory for determining the quality
of the multiherb products and will be reported elsewhere.
CONCLUSION
In this work, a new LC-MS/ MS-based methodology for
Received for review January 9, 2004. Accepted April 27,
characterization and quantification of a broad range of 15 saiko-
saponin derivatives is described. Negative ion ESI-MS/ MS method
was chosen for the saikosaponin derivatives. Complete resolution
2
004.
AC0499423
4216 Analytical Chemistry, Vol. 76, No. 14, July 15, 2004