hplc-dad and hplc-esi-ms separation, determination and identification of the spin-labeled...

J. Sep. Sci. 2009, 32, 1323 –1332 L. Zhao et al. 1323

Lei Zhao1, 2

Zhen-Ling Liu1

Peng-Cheng Fan3

Zhi-Wei Zhang1

Xiong Liu2

Yun-Jing Zhan2

Xuan Tian1

1State Key Laboratory of AppliedOrganic Chemistry, LanzhouUniversitry, Lanzhou, China

2Department of Pharmacy, GansuCollege of Traditional ChineseMedicine, Lanzhou, China

3Department of Pharmacy,General Hospital of Lanzhou,Lanzhou, China

Original Paper

HPLC-DAD and HPLC-ESI-MS separation,determination and identification of the spin-labeleddiastereoisomers of podophyllotoxin

Spin-labeled nitroxide derivatives of podophyllotoxin had better antitumor activityand less toxicity than that of the parent compounds. However, the 2-H configura-tions of these spin-labeled derivatives cannot be determined by nuclear magneticresonance (NMR) methods. In the present paper, a high-performance liquid chroma-tography-diode array detection (HPLC-DAD) and a high-performance liquid chroma-tography-electrospray ionization tandem mass spectrometry (HPLC-ESI/MS/MS)method were developed and validated for the separation, identification of fourpairs of diastereoisomers of spin-labeled derivatives of podophyllotoxin at C-2 posi-tion. In the HPLC-ESI/MS spectra, each pair of diastereoisomers of the spin-labeledderivatives in the mixture was directly confirmed and identified by [M+H]+ ions andion ratios of relative abundance of [M-ROH+H]+ (ion 397) to [M+H]+. When the [M-ROH+H]+ ions (at m/z 397) were selected as the precursor ions to perform the MS/MSproduct ion scan. The product ions at m/z 313, 282, and 229 were the common diag-nostic ions. The ion ratios of relative abundance of the [M-ROH+H]+ (ion 397) to[M+H]+, [A+H]+ (ion 313) to [M-ROH+H]+, [A+H-OCH3]+ (ion 282) to [M-ROH+H]+ and [M-ROH-ArH+H]+ (ion 229) to [M-ROH+H]+ of each pair of diastereoisomers of the deriv-atives specifically exhibited a stereochemical effect. Thus, by using identical chro-matographic conditions, the combination of DAD and MS/MS data permitted theseparation and identification of the four pairs of diastereoisomers of spin-labeledderivatives of podophyllotoxin at C-2 in the mixture.

Keywords: Diastereoisomers separation / High-performance liquid chromatography – electro-spray tandem mass spectrometry / Nitroxide / Podophyllotoxin / Relative abundance /

Received: November 21, 2008; revised: January 22, 2009; accepted: January 22, 2009

DOI 10.1002/jssc.200800674

1 Introduction

The ever-increasing knowledge on the involvement ofradicals in diverse anti-tumor drugs (such as thio-TEPA, 5-fluorouracil, nitrosourea, rubomycin, and 6-mercapto-purine) [1] has expanded the search for more efficientanti-tumor drugs with tumor-specific action and lowgeneral toxicity. Podophyllotoxin, an interesting lignan,results in pronounced cytotoxic activity. In order to findcompounds with superior bioactivity and less toxicity,our group has been involved for many years in the isola-tion and the chemical transformation of podophyllo-toxin and its analogues. Especially, a larger number ofspin-labeled nitroxide derivatives of podophyllotoxin

were synthesized, which had significant anti-tumoractivity with marked decrease in toxicity compared withthat of the parent compounds [2–14]. Research on struc-ture–activity relationships (SARs) indicated that the bio-logical activity and toxicity of podophyllotoxin and itsderivatives are stereochemically specific. The base-cata-lyzed isomerization of the rigid, strained podophyllo-toxin structure with 2b-H configuration to flexible picro-podophyllin with a 2a-H configuration (Fig. 1) led to theloss of toxicity and antimitotic activity [15]. Compara-tively, the transfused c-lactone D ring in podophyllo-toxin and its analogues are susceptible to isomerizationto the biologically inactive picropodophyllin whenexposed to a mild base. From the chemotherapeuticpoint of view, these epimerization metabolisms areundesirable because limiting the physiological lifetimeof these compounds would set an upper limit to their bio-logical effectiveness. It is therefore important to developa method for the rapid detection and identification of

Correspondence: Dr. X. Tian, State Key Laboratory of AppliedOrganic Chemistry, Lanzhou Universitry, Lanzhou, 730000, Chi-naE-mail: [email protected]: +86-931-891-2582

i 2009 WILEY-VCH Verlag GmbH & Co. KGaA, Weinheim www.jss-journal.com

1324 L. Zhao et al. J. Sep. Sci. 2009, 32, 1323 – 1332

diastereoisomers of spin-labeled derivatives of podophyl-lotoxin in the chemical reactions and pharmaceuticalmetabolism, among other things. The epimerizationfrom 2b-H configuration to 2a-H configuration was inves-tigated by proton nuclear magnetic resonance (1H NMR)spectrometry [16]. However, the 2-H configurations ofthese spin-labeled derivatives cannot be determined byNMR because conventional NMR spectra of nitroxide-con-taining molecules are quite broad and are therefore oflimited use for monitoring the changes in structure tak-ing place during a chemical synthesis [17]. Hence, also anurgent subject needs a new analytical method for ascer-taining the 2-H configurations of these diastereoisomers.Except for Liu et al. [18] who reported the method ofmicellar electrokinetic chromatography (MEKC), fewstudies have been published for determining the epimersof spin-labeled derivatives of podophyllotoxin at C-2 posi-tion. High-performance liquid chromatography (HPLC)as an efficient separation and analysis system has beenused to analyze lignans from various sources [19–24].Lim [25] studied the retention behaviors of the 4-hydroxygroup substituted by the diastereoisomers of some podo-phyllotoxin through HPLC. HPLC methods for the deter-mination of VM-26 and VP-16 body fluids have also beenreported [26–33]. Recently, Puri et al. [34] and Schmidt etal. [35] have identified two epimers at C-4 position relatedto podophyllotoxin and the lignans of podophyllotoxin,respectively, from a natural source through HPLC/MS.Zhao et al. [36] studied the separation and determinationof eight pairs of diastereoisomers of podophyllotoxinand its esters at C-2 position through HPLC-ESI/MS. How-ever, to our knowledge, there have been no reports onthe separation and identification of the diastereoisomersof spin-labeled derivatives of podophyllotoxin lignans atC-2 position by HPLC-DAD and HPLC-ESI/MS. In thepresent study, we report the HPLC-DAD and the HPLC-ESI/MS/MS separation, determination, and identification offour pairs of diastereoisomers of spin-labeled derivativesof podophyllotoxin at C-2 position (1a (1)–4a (4)). The

aims of the study were to develop a simple and stableHPLC-DAD and HPLC-ESI/MS method for the rapid separa-tion and detection of the diastereoisomers of spin-labeled derivatives of podophyllotoxin lignans at C-2position, such as degradation products, metabolites, andsynthetic compounds. The investigated methods wereable to separate successfully and identify directly thefour pairs of epimers of spin-labeled derivatives of podo-phyllotoxin lignans at C-2 position in the mixture. Themigration behaviors of the compounds with differenttypes were regular, i.e., every 2a-H compound migratedbefore its corresponding 2b-H isomer under optimumconditions. In addition, each pair of epimers in the mix-ture was directly detected and identified by the [M+H]+

ions and the ion ratios of relative abundance of [M-ROH+H]+ (ion 397) to [M+H]+ in HPLC-ESI/MS spectra.When the [M-ROH+H]+ (ion 397) was selected as the pre-cursor ions to perform the MS/MS product ion scan. Theproduct ions at m/z 313, 282, and 229 were the commondiagnostic masses. The ion ratios of relative abundanceof the [M-ROH+H]+ (ion 397) to [M+H]+, [A+H]+ (ion 313) to[M-ROH+H]+, [A+H-OCH3]+ (ion 282) to [M-ROH+H]+ and [M-ROH-ArH+H]+ (ion 229) to [M-ROH+H]+ of each pair of dia-stereoisomers of the derivatives specifically exhibited astereochemical effect. The results of this work couldserve as an effective tool for the separation and detectionof other spin-labeled derivatives of podophyllotoxin andanalogues.

2 Experimental

2.1 Reagents

HPLC grade acetonitrile and methanol (E. Merck, Darm-stadt, Germany) were used for the HPLC analysis. Deion-ized water was purified by Milli-Q system (Millipore, Bed-ford, MA, USA). All other chemicals were of analyticalgrade and were purchased from Tianjin Reagent Com-pany (Tianjin, China).

2.2 Sample preparation and identification

All melting points were taken on a Kofler melting pointapparatus and uncorrected. Molecular weights weredetermined by ESI on APEXTM II 4.7 Tesla mass spectrom-eter (Bruker Daltonics, Billerica, MA, USA). Samples wererun in ESI positive mode by direct injection with asyringe mass spectrometer (ESI/MS). FT-IR spectra weremeasured on a Nicolet NEXUS 670 FT-IR spectrometer(Nicolet, Wisconsin, USA). CD was performed on theJASCO-810 (JASCO, Tokyo, Japan) (cell path length 1 cm),simultaneously monitoring at one specific wavelength(range 250–300 nm). The electro spin resonance (ESR)spectra were obtained from 10 – 4 M alcohol solution,


Figure 1. Configuration of the pair of spin-labeled diaster-eoisomers of podophyllotoxin lignans. 4a: 4-a-O-(29,29,69,69-Tetramethyl-49-carboxy-piperidine-19-oxyl)-picropodophyllicester, 4: 4-a-O-(29,29,69,69-tetramethyl-49-carboxy-piperidine-19-oxyl)-podophyllic ester.

J. Sep. Sci. 2009, 32, 1323 –1332 Liquid Chromatography 1325

using a Bruker A300 X-band EPR spectrometer (Bruker,Rheinstetten, Germany).

Podophyllotoxin were isolated from Chinese medici-nal herb with the botanical nomenclature Podophyllumemodi Wall var. chinensis. Picropodophyllin was synthe-sized from podophyllotoxin according to the literature[37] in our group. A mixture of the appropriate stablenitroxide acid R1COOH (1.0 mmol), podophyllotoxin orpicropodophyllin (1.0 mmol) and dimethylaminopyri-dine (DMAP, 40 mg) was stirred in dry dichloromethane(20 mL) for 5 min at room temperature under argon. N,N-Dicyclohexylcarbodiimide (DCC, 208 mg) was added andthe reaction mixture was stirred for 3–40 h and moni-tored by TLC. The reaction mixture was filtered and thefiltrate was evaporated. The residue was subjected toflash column chromatography (dichloromethane/ace-tone 20:1) on silica gel to purify the desired product.Their structures were confirmed by high-resolution massspectrometry (HRMS), ESR, CD, IR.

4-a-O-(29,29,59,59-Tetramethyl-39-carboxypyrroline-19-oxyl)-podophyllic ester (1): m.p. 119–1218C; ½a�25

D –1268(c 0.5 CHCl3); CD (MeOH) De (nm): (7.60 (286), –24.8 (278);IR (KBr) 2931, 2851, 1780, 1713, 1627, 1582, 1507, 1485,1461, 1337 (N–O), 1280, 1241,1189, 1039, 1127, 1001,931 cm – 1; ESR: An = 14.80 G, g0 = 2.0060; HRMS (ESI):598.2511 for [M+NH4]+ (calcd 598.2521 for C31H38O10N2).

4-a-O-(29,29,59,59-Tetramethyl-39-carboxypyrroline-19-oxyl)-picropodophyllic ester (1a): m.p. 207–2098C; ½a�25

D

+68 (c 0.5 CHCl3); CD (MeOH) De (nm): +15.0 (289), +5.23(276); IR (KBr) 2933, 2839, 1775, 1726, 1628, 1590, 1504,1478, 1460, 1331 (N–O), 1296, 1242,1184, 1128, 1040,1002, 933 cm – 1; ESR: An = 14.84 G, g0 = 2.0061; HRMS(ESI): 598.2528 for [M+NH4]+ (calcd 598.2521 forC31H38O10N2).

4-a-O-(29,29,69,69-Tetramethyl-49-carboxy-19,29,59,69-tetra-hydropyridine-19-oxyl)-podophyllic ester (2): m.p. 114–1168C; ½a�25

D –1228 (c 0.5 CHCl3); CD (MeOH) De (nm): (24.6(285), (51.2 (277); IR (KBr) 2934, 2839, 1780, 1714, 1664,1589, 1507, 1485, 1461, 1329 (N–O), 1270, 1241, 1178,1127, 1037, 997, 931 cm – 1; ESR: An = 15.46 G, g0 = 2.0062;HRMS (ESI): 612.2665 for [M+NH4]+ (calcd 612.2677 forC32H40O10N2).

4-a-O-(29,29,69,69-Tetramethyl-49-carboxy-19,29,59,69-tetra-hydropyridine-19-oxyl) –picropodophyllic ester (2a):m.p. 178–1798C; ½a�25

D +118 (c 0.5 CHCl3); CD (MeOH) De

(nm): +14.1 (286), +1.77 (271); IR (KBr) 2935, 2837, 1776,1708, 1663, 1589, 1506, 1485, 1461, 1326 (N-O), 1270,1244, 1177, 1126, 1039, 1007, 934 cm – 1; ESR: An = 15.50G, g0 = 2.0062; HRMS (ESI): 612.2668 for [M+NH4]+ (calcd612.2677 for C32H40O10N2).

4-a-O-(29,29,59,59-Tetramethyl-39-carboxypyrrolidine-19-oxyl)-podophyllic ester (3): m.p. 117–1198C; ½a�25

D –1128(c 0.5 CHCl3); CD (MeOH) De (nm): –2.80 (285), –41.3 (276);IR (KBr) 2934, 2839, 1780, 1736, 1659, 1589, 1507, 1485,1461, 1331 (N–O), 1297, 1240, 1189, 1127, 1038, 999,

930 cm – 1; ESR: An = 14.80 G, g0 = 2.0060; HRMS (ESI):600.2669 for [M+NH4]+ (calcd 600.2677 for C31H40O10N2).

4-a-O-(29,29,59,59-Tetramethyl-39-carboxypyrrolidine-19-oxyl)-picropodophyllic ester (3a): m.p. 151–1538C; ½a�25

D

+178 (c 0.5 CHCl3); CD (MeOH) De (nm): +20.2 (287), –15.1(265); IR (KBr) 2933, 2856, 1781, 1736, 1659, 1584, 1509,1460, 1330 (N–O), 1253, 1194, 1125, 1034, 987, 937 cm–1;ESR: An = 14.86 G, g0 = 2.0060; HRMS (ESI): 600.2667 for[M+NH4]+ (calcd 600.2677 for C31H40O10N2).

4-a-O-(29,29,69,69-Tetramethyl-49-carboxy-piperidine-19-oxyl)-podophyllic ester (4): m.p. 116–1188C; ½a�25

D –948 (c0.5 CHCl3); CD (MeOH) De (nm): –7.21 (287), –44.8 (276); IR(KBr) 2936, 2838, 1780, 1733, 1589, 1507, 1484, 1460,1329 (N–O), 1289, 1240, 1190, 1127, 1038, 1001,930 cm–1; ESR: An = 15.96 G, g0 = 2.0063; HRMS (ESI):614.2829 for [M+NH4]+ (calcd 614.2834 for C32H42O10N2).

4-a-O-(29,29,69,69-Tetramethyl-49-carboxy-piperidine-19-oxyl)-picropodophyllic ester (4a): m.p. 186–1888C; ½a�25

D

+118 (c 0.5 CHCl3); CD (MeOH) De (nm): +32.2 (286), –5.81(266); IR (KBr) 2932, 2851, 1776, 1727, 1591, 1508, 1481,1460, 1324 (N–O), 1245, 1165, 1127, 1038, 1003,931 cm–1; ESR: An = 16.06 G, g0 = 2.0063; HRMS (ESI):614.2830 for [M+NH4]+ (calcd 614.2834 for C32H42O10N2).

2.3 High-performance liquid chromatography

An Agilent series 1100 HPLC instrument (Agilent, Wald-bronn, Germany) equipped with a binary pump, a diode-array detector, an autosampler and a column compart-ment was used. The sample was separated on an ODS-2Hypersil C18 column (5 lm, 250 mm64.6 mm, Thermo).The column was held at 60% solvent A (methanol/aceto-nitrile (36/24, V/V%)) and 40% solvent B (water). Themobile phase flow-rate for all separation was 0.8 mL/min.The detector was at 285 nm and the column temperaturewas set at 258C.

2.4 Mass spectrometry

HPLC-ESI/MS/MS experiments were performed using theAgilent 1100 HPLC system described above combinedwith a Bruker Esquire 3000plus ion trap mass spectrometer(Bruker Franzen Analytik GmbH, Bremen, Germany)equipped with an ESI interface. Instrument control anddata acquisition were performed using Esquire 5.0 soft-ware. The ion source temperature was 3008C and the ESIneedle voltage was always set at 5.0 kV. Nitrogen wasused to the drying and nebulizer gas at a flow rate of 8 L/min and a back-pressure of 25 psi. Helium was intro-duced into the trap with an estimated pressure of6610 – 6 mbar to improve trapping efficiency to act asthe collision gas. For HPLC-ESI/MS analysis, the flow wassubject to a split of 1: 5 before being introduced into theion source. There was a short delay of about 0.2 min



between the DAD and MS systems in consequence of theconnecting tubing between HPLC and MS.

3 Results and discussion

3.1 HPLC-DAD separation of the spin-labeleddiastereoisomers of podophyllotoxin

One goal of this study is to separate the mixture of dia-stereoisomers of spin-labeled derivatives of podophyllo-toxin (see structures in Fig. 2) through reversed-phasechromatography. The retention behavior of aryltetrahy-drona lignans in the reversed-phase chromatographywas organic-modifier-selective and -specific [25]. It is rea-sonable to assume that the solvent selectivity and specif-icity effects will also be the important factors to considerwhen developing an optimized system for the separationof the spin-labeled diastereoisomers of podophyllotoxin.With the acetonitrile –water and methanol–water atdifferent ratios as the eluent, respectively, on the ODS-2Hypersil C18 column, the peaks remained merged. Theaddition of methanol to the acetonitrile–water mixturehad an immediate effect on the separation of the fourpairs of diastereoisomers. To optimize the selectivity,methanol/acetonitrile (19:41, 65:35 v/v%) was used asphase A, respectively, water was used as solvent B and thepercentage of phase A was 60%. The mobile phase wasapplied at 0.8 mL/min flow-rate. However, it is not possi-ble to separate compounds 1a and 4a (Figs. 3b and c).Therefore, solvent rate were readjusted. Figure 3 Chro-matogram (a) depicts methanol/acetonitrile (36:24 v/v%)as phase A, water as phase B and the percentage of phaseA was 60%, flow-rate at 0.8 mL/min. As chromatogram (a)showed, compounds 1a and 4a with different molecular

weight were partly separated while the four pairs of dia-stereoisomers were completely separated. The integra-tion results of the separation were given in Table 1. Allpeaks were separated and the critical peak pair has a res-olution of 0.76. Peak symmetry is between 0.65 and 0.92.When the mixture was analyzed by HPLC-DAD, clear UVspectra were obtained for the spin-labeled diastereoiso-mers of podophyllotoxin, revealing that could distin-guish them from other types of compounds. The UV spec-tra of each compound with different free radical substi-tution at C-2 were similar. The intense absorption bandsof them were about at 205, 240, and 290 nm (Fig. 4). Avalid and stable HPLC was achieved through a ternarymobile phase under the conditions described in theexperimental section. The four pairs of diastereoisomerswere eluted in the migration order of 3a, 1a, 4a, 2a, 3, 1,4, and 2 (Fig. 4).


Figure 2. Structure of the four pairs ofspin-labeled diastereoisomers of podophyl-lotoxin lignans. 1: 4-a-O-(29,29,59,59-Tetra-methyl-39-carboxypyrroline-19-oxyl)-podo-phyllic ester , 1a: 4-a-O-(29,29,59,59-tetra-methyl-39-carboxypyrroline-19-oxyl)-picro-podophyllic ester, 2: 4-a-O-(29,29,69,69-tetra-methyl-49-carboxy-19,29,59,69-tetrahydropyri-dine-19-oxyl)-podophyllic ester, 2a: 4-a-O-(29,29,69,69-tetramethyl-49-carboxy-19,29,59,69-tetrahydropyridine-19-oxyl)-picropodophyllicester, 3: 4-a-O-(29,29,59,59-tetramethyl-39-carboxypyrrolidine-19-oxyl)-podophyllicester, 3a: 4-a-O-(29,29,59,59-tetramethyl-39-carboxypyrrolidine-19-oxyl)-picropodophyl-lic ester, 4: 4-a-O-(29,29,69,69-tetramethyl-49-carboxy-piperidine-19-oxyl)-podophyllicester, 4a: 4-a-O-(29,29,69,69-tetramethyl-49-carboxy-piperidine-19-oxyl)-picropodophyl-lic ester.

Table 1. Chromatographic data for the separation of the fourpairs of spin-labeled diastereoisomers of podophyllotoxinlignans shown in Fig. 4

Compound Symmetrya) Peak widthb)

(min)Resolution Selectivity

3a 0.71 0.31 4.06 1.151a 0.65 0.36 4.93 1.284a 0.67 0.38 0.76 1.072a 0.76 0.41 1.06 1.083 0.80 0.44 2.60 1.091 0.80 0.50 4.14 1.154 0.78 0.50 2.16 1.132 0.92 0.57 2.76 1.17

a) Symmetry (USP: tailing) calculation: T ¼ w0:05

2f; w0:05 – peak

width at 5% peak height, f – distance from the peak maximumto the leading edge of the peak, the distance measured at w0:05.

b) Peak width at 50% peak height.


First, the migration order of the isomers in each of thefour pairs was always a 2a-H compound before its corre-sponding 2b-H epimer. Second, isomers with saturatednitroxide radicals were eluted before those with unsatu-rated nitroxide radicals when their molecular weightwere similar (3a (3) A 1a (1), 4a (4) A 2a (2)). Finally, isomerswith five-membered nitroxides were eluted before thosewith six-membered nitroxides. (3a (3), 1a (1) A 4a (4), 2a(2)) (Fig. 4). These phenomena are explained as follows.

On one hand, each of the four pairs of diastereoiso-mers, owing to the conformational flexibility of 2a-Hcompounds [38], is able to arrange itself in conformation,allowing for the maximum interaction of its polargroups with an organic modifier [25]. Its partition intothe mobile phase should be higher than that of its corre-sponding rigid, inflexible 2b-H. On the other hand, 2a-Hcompound exists with the E ring in an equatorial confor-mation, which takes a more planar structure as com-pared to that of its corresponding 2b-H isomer (see Fig. 1)

[38]. Its configuration is probably less hindered than itscorresponding 2b-H epimer.

The general increase in retention time and theimprovement in resolution are attributed to the abilityof methanol, an H–bonding organic modifier, to form adifferent degree of H–bond with different compoundsdepending on their structure and stereo-configurations[23]. Isomers with unsaturated nitroxide radicals seem toform a stronger H-bond with the layer of methanol“sorbed” onto the C18 stationary-phase surface than thosewith saturated nitroxide radicals. This may be explainedthe sequence of the saturated nitroxide radicals wereeluted before those with unsaturated nitroxide radicals(3a (3) A 1a (1), 4a (4) A 2a (2)).

The reason that isomers with five-membered nitrox-ides were eluted before those with six-membered nitrox-ides (3a (3), 1a (1) A 4a (4), 2a (2)) may be explained by thesteric hindrance. The five-membered ring nitroxides (3a(3), 1a (1)) exist a priority close planar conformation,


Figure 3. Chromatograms of the separation of the four pairs of spin-labeled diastereoisomers of podophyllotoxin lignans (1a(1)–4 a (4)). (a) methanol/acetonitrile (36:24 v/v%) as solvent A, water as solvent B and the percentage of solvent A was 60%,(b) Methanol/acetonitrile (19:41, v/v%) as solvent A, water as solvent B and the percentage of solvent A was 60%, (c) methanol/acetonitrile (65:35, v/v%) as solvent A, water as solvent B and the percentage of solvent A was 60%.


while six-membered ring nitroxides (4a (4), 2a (2)) exist apriority chair conformation. The conformation of theformer is probably less hindered than that of the latter.

3.2 HPLC-ESI/MS/MS analyses of the spin-labeleddiastereoisomers of podophyllotoxin

To confirm the above analyses further and investigate acoupling method that could be used to directly detect

and identify the spin-labeled diastereoisomers of podo-phyllotoxin in the mixture, HPLC-ESI/MS/MS experi-ments were performed. Positive and negative ion modeswere tried for the spin-labeled diastereoisomers of podo-phyllotoxin, and the results suggested that the positiveion mode was more sensitive. The ESI source gave [M+H]+

ions as quasimolecular ions afforded a diagnostic ion atm/z 397 (Fig. 7). The [M-ROH+H]+ ions (at m/z 397) wereselected as the precursor ions to perform the MS/MS


Figure 4. Chromatograms and UV of the four pairs of spin-labeled diastereoisomers of podophyllotoxin lignans (1a (1)–4a (4)).

Figure 5. TIC of the four pairs of spin-labeled diastereoisomers of podophyllotoxin lignans (1a (1)–4a (4)).


experiments, and multiple product ions were observed(Fig. 8). The multistage tandem mass spectral data ofeach pair of diastereoisomers of the derivatives specifi-cally exhibited stereochemical effects. The effects of con-figuration of the condensed lactone ring D on the rela-tive abundances of characteristic ions in the mass spectraof the four pairs of diastereoisomers were compared anddiscussed. The results support the research of the litera-ture [36], revealing that provided a basis for identifyingspin-labeled diastereoisomers of podophyllotoxin and itsand analogues at C-2 position in the mixture.

3.2.1 HPLC-ESI/MS identification of the spin-labeled diastereoisomers of podophyllotoxin

HPLC analyzed with both DAD detection (Fig. 4) and posi-tive ion mode ESI/MS total ion current (TIC) monitoring(Fig. 5) indicated the four pairs of lignans in the mix-tures. It was easy to observe the quasimolecular [M+H]+

ions in the HPLC-ESI/MS spectra for all four pairs of com-pounds. The assignment of HPLC peaks was confirmed bycomparing the retention times and [M+H]+ peaks with

authentic standards. Additionally, positive ion modeextracted ion current (EIC) analyses were performed toconfirm the above results. As shown in Fig. 6, the selectedion monitoring (SIM) of the diastereoisomers of thelignans at m/z 581, 595, 583, and 597 exhibited in the EICchromatograms and two different epimer peaks wereidentified, respectively, as 1a, 1; 2a, 2; 3a, 3; 4a, 4. What iseven more important and interesting, as shown in theHPLC-ESI/MS spectra (Fig. 7 and Table 2), is that the rela-tive intensity ratio of [M-ROH+H]+ to [M+H]+ of each pair ofdiastereoisomers exhibited obvious distinction, that is,the ratio for the picropodophyllin series was more thanone, but it was the opposite for the podophyllotoxin ser-ies. The reason is that the picropodophyllin series (com-pounds 1a to 4a) contains a 2b, 3b cis-fused lactone ring,wherein the ring C is at a half-boat conformation thatmakes the loss of ROH from the 2, 4-positions easy. Mean-while, the podophyllotoxin series (compounds 1 to 4)contains a 2a, 3b trans-fused lactone ring, wherein ring Cis at a half-chair conformation that makes the loss ofROH difficult from the 2, 4-positions (Fig. 1) [36].


Figure 6. (a) EIC spectrum of [M+H]+ m/z 581 (1a, 1); (b) EIC spectrum of [M+H]+ m/z 595 (2a, 2); (c) EIC spectrum of [M+H]+

m/z 583 (3a, 3); (d) EIC spectrum of [M+H]+ m/z 597 (4a, 4).

Figure 7. Positive ion ESI-MS spectra of the pair of spin-labeled diastereoisomers of podophyllotoxin lignans. 4a: 4-a-O-(29,29,69,69-Tetramethyl-49-carboxy-piperidine-19-oxyl)-picropodophyllic ester, 4: 4-a-O-(29,29,69,69-tetramethyl-49-carboxy-piperi-dine-19-oxyl)-podophyllic ester.


A study to test the repeatability of relative intensityratio of [M-ROH+H]+ to [M+H]+ for HPLC-ESI/MS was per-formed. For a given sample, the RSD-values of them forfive replicate injections were below 10%. Thus, by usingidentical sample solutions and chromatographic condi-tions (including the same columns), HPLC coupling DAD,and ESI/MS, the four pairs of the spin-labeled diaster-eoisomers could be rapidly separated, unambiguouslyidentified, and confirmed in the mixture.

3.2.2 HPLC-ESI/MS/MS analyses of the spin-labeled diastereoisomers of podophyllotoxin

In the first-order spectra (MS) of all compounds, the qua-simolecular ion [M+H]+ afforded a diagnostic ion [M-ROH+H]+ at m/z 397 (a) (Fig. 7), which corresponded to theloss of ROH from the ion [M+H]+. When ion a was selectedas the precursor ion to perform the MS/MS experiments,multiple product ions were observed (Fig. 8). Amongthem, ions [A+H]+ at m/z 313 (b), [M–ROH–ArH+H]+ at 229

(c) and [A+H–OCH3]+ at 282 (d) were characteristic ions, b,c corresponding to the loss of a crotonolactone molecule(C4H4O2, 84 Da) and 1, 2, and 3-trimethoxybenzene (ArH,168 Da), respectively, from ion a [36, 39], d correspond-ing to the loss of methoxy (CH3O) from ion b. It was easyto observe the formation of characteristic ions a, b, c,and d provide information allowing the distinctionbetween the spin-labeled diastereoisomers, picropodo-phyllin, and podophyllotoxin series at C-2 position (Fig. 8and Table 3). The former series exhibited an [A+H]+ (b)fragment with a higher intensity than [M–ROH+H]+ (a),that is, the relative intensity ratio of [A+H]+ (b) to [M–ROH+H]+ (a) was more than one. In the latter series, theresult was the opposite, while the [M–ROH+H]+ was moreintense (usually the base peak). Meanwhile, the relativeintensity ratio of [M–ROH–ArH+H]+ (c) to [M–ROH+H]+ (a)also exhibited a distinction between the two types of ser-ies, that is, the ratio in the former series was higher thanthat in the latter series. In addition, the relative intensity


Figure 8. Positive ion MS/MS of spectra. 4a: 4-a-O-(29,29,69,69-Tetramethyl-49-carboxy-piperidine-19-oxyl)-picropodophyllic ester,4: 4-a-O-(29,29,69,69-tetramethyl-49-carboxy-piperidine-19-oxyl)-podophyllic ester m/z 597 fi 397.


ratio of [A+H-OCH3]+ (d) to [M–ROH+H]+ (a) for the diaster-eoisomers were similar with that of [A+H]+ (b) to [M–ROH+H]+ (a).

The reasons maybe that the former series contains the2b, 3b cis-fused lactone ring, wherein ring C is at a half-boat conformation that makes the loss of crotonolactonemolecule (C4H4O2, 84 Da) easy. However, the latter seriescontains the 2a, 3b trans-fused lactone ring (Fig. 1)wherein ring C is at a half-chair conformation. It is diffi-cult to eliminate the crotonolactone molecule (C4H4O2,84 Da). The same applies when explaining the differenceof the relative intensity ratio of [M–ROH–ArH+H]+ (c) to[M–ROH+H]+ (a) between the two types of series. The lossof 1, 2, and 3-trimethoxybenzene (ArH, 168 Da) via syn-elimination in the former series is easier than that of thelatter via trans-elimination [36].

4 Conclusion

It has been demonstrated that the optimized HPLC-DADand HPLC-ESI/MS method can provide valuable informa-tion for the separation and identification of the diaster-eoisomers of spin-labeled derivatives of podophyllotoxinand its analogues. This enables the analysis of the diaster-

eoisomers of podophyllotoxin and its analogues with acomplex composition that needs a highly efficient chro-matographic separation technique. In this paper,through HPLC-DAD and HPLC-ESI/MS experiments whereidentical chromatographic conditions allowed for arapid and simple analysis, the diastereoisomers of spin-labeled derivatives of podophyllotoxin became accessibledirectly from the matrix. Therefore, the method could beused in rapidly identifying the purity and monitoringthe epimerization of 2-H of podophyllotoxin and its ana-logues from natural products, chemical reactions, andpharmaceutical metabolism.

This research was supported in part by the Natural Science Foun-dation of Gansu Province (3ZS061-A25-O24). Our hearty thanks goto Professor Cui-Rong Sun and Ms. Juan-Juan Chen (Zhejiang Uni-versity) for performing the HPLC-ESI/MS analysis.

The authors declared no conflict of interest.

5 References

[1] Wang, Y. G., Tian, X., Chen, Y. Z., Acta Pharmaceutica Sin. 1988, 23,792.


Table 2. The relative abundance of the characteristic ions of HPLC-DAD-ESI/MS from the four pairs of spin labeled diaster-eoisomers of podophyllotoxin lignans

Compounds Molecular mass C-2 Lactoneconfiguration

m/z (Relative abundance%) Ion abundance ratio

[M + H]+ [M – ROH + H]+ [M – ROH + H]+/[M + H]+

1a 580 b 581 (42) 397 (100) 2.31 580 a 581 (100) 397 (50) 0.502a 594 b 595 (13) 397 (100) 7.72 594 a 595 (100) 397 (58) 0.583a 582 b 583 (5.6) 397 (100) 183 582 a 583 (100) 397 (25) 0.254a 596 b 597 (8.7) 397 (100) 11.54 596 a 597 (100) 397 (55) 0.55

Table 3. The fragmentation and relative abundance of the characteristic ions of ESI/MS/MS from the four pairs of spin labeleddiastereoisomers of podophyllotoxin lignans

Compound Molecularmass

C-2 Lactoneconfiguration

Fragment ion (relative abundance %) Ion abundance ratio

[M – ROH+ H]+

[A + H]+ [A + H– OCH3]+

[M – ROH– ArH + H]+

[A + H]+/[M– ROH + H]+

[M – ROH –ArH + H]+/[M– ROH + H]+

[A + H – OCH3]+/[M – ROH + H]+

1a 580 b 397 (15) 313 (70) 282 (100) 229 (7.5) 4.7 0.50 6.71 580 a 397 (100) 313 (70) 282 (65) 229 (15) 0.70 0.15 0.652a 594 b 397 (7.5) 313 (80) 282 (100) 229 (7.5) 11 1.0 13.42 594 a 397 (100) 313 (77) 282 (89) 229 (34) 0.77 0.34 0.893a 582 b 397 (11) 313 (92) 282 (100) 229 (10) 8.3 0.99 9.13 582 a 397 (100) 313 (77) 282 (61) 229 (19) 0.77 0.19 0.614a 596 b 397 (6.8) 313 (68) 282 (100) 229 (9.0) 10 1.3 14.74 596 a 397 (100) 313 (71) 282 (64) 229 (21) 0.71 0.21 0.64


[2] Chen, Y. Z., Zhang, C. J., Tian, X., Scientia Sinica (series B). 1987, 30,1070 – 1079.

[3] Chen, Y. Z., Tian, X., Wang, Y. G., Life Sci. 1989, 45, 2569 – 2571.

[4] Wang, J. Z., Tian, X., Tsumura, H., Anti-Cancer Drug Des. 1993, 8,193 – 202.

[5] Tian, X., Yan, Z. Q., Li, J. X., Chen, Y. Z., Chem. Res. Chin. Univ. 1995,11, 79 – 83.

[6] Tian, X., Yang, M. G., Chen, Y. Z., Chem. Res. Chin. Univ. 1996, 12,304 – 308.

[7] Lu, K. K., Wang, Y. G., Chen, Y. Z., Synthetic Commun. 1997, 27,1963 – 1968.

[8] Tian, X., Wang, Y. G., Yang, M. G., Chen, Y. Z., Life Sci. 1997, 60,511 – 517.

[9] Wang, Y. G., Pan, J. L., Shi, J. F., Chen, Y. Z., Life Sci. 1997, 61, 537 –542.

[10] Pan, J. L., Wang, Y. G., Shi, J. F., Chen, Y. Z., Chin. Chem. Lett. 1997,8, 207 – 208.

[11] Tian, X., Zhang, F. M., Li, W. G., Life Sci. 2002, 70, 2433 – 2443.

[12] Liu, Y. Q., Tian, X., Chin. J. Org. Chem (S). 2005, 25, 17.

[13] Liu, Y. Q., Tian, X., Synthetic Commun. 2005, 27, 1963 – 1968.

[14] Jin, Y., Chen, S. W., Tian, X., Bioorg. Med. Chem. 2006, 14, 3062 –3068.

[15] Ayres, D. C., Lim, C. K., J. Chem. Soc. 1976, Perkin Trans I, 832 – 836.

[16] Aso, Y., Hayashi, Y., Yoshika, S., Takeda, Y., Kita, Y., Nishimura,Y., Chem. Pharm. Bull. 1989, 37, 422 – 424.

[17] Keana, J. F. W., Chem. Rev. 1978, 78, 37 – 64.

[18] Liu, S. H., Tian, X., Chen, X. G., Hu, Z. D., J. Chromatogr. A 2002,959, 263 – 268.

[19] Treppendahl, S., Jakobsen, P., J. Chromatogr. A. 1980, 189, 276 –278.

[20] Cairnes, D. A., Kingston, D. G. I., Rao, M. M., J. Nat. Prod. 1981, 44,34 – 37.

[21] Allen, L. M., Drug Metab. Rev 1978, 8, 119 – 135.

[22] Choi, Y. H., Kim, J. Y., Ryu, J. H., Yoo, K. P., Chang, Y. S., Kim, J.,Planta. Med. 1998, 64, 482 – 483.

[23] Lim, C. K., J. Chromatogr. A 1996, 722, 267 – 271.

[24] Kartal, M., Konuklugil, B., Indrayanto, G., Alfermann, A. W., J.Pharm. Biomed. Anal. 2004, 35, 441 – 447.

[25] Lim, C. K., Ayres, D. C., J. Chromatogr. 1983, 255, 247 – 254.

[26] Shirazi, F. H., Bahrami, G., Stewart, D. J., Tomiak, E., Delorme, F.,Noel, D., Goel, R., J. Pharm. Biomed. Anal. 2001, 25, 353 – 356.

[27] Allen, L. M., J. Pharm. Sci. 1980, 69, 1440 – 1441.

[28] Liliemark, E., Pettersson, B., Peterson, C., Liliemark, J., J. Chroma-togr. B 1995, 669, 311 – 317.

[29] Robieux, I., Aita, P., Sorio, R., Toffoli, G., Boiocchi, M., J. Chroma-togr. B 1996, 686, 35 – 41.

[30] Manouilov, K. K., McGuire, T. R., Gordon, B. G., Gwilt, P. R., J. Chro-matogr. B 1998, 707, 342 – 346.

[31] Cai, X., Woo, M. H., Edick, M. J., Relling, M. V., J. Chromatogr. B1999, 728, 241 – 250.

[32] Stremetzne, S., Jaehde, U., Schunack, W., J. Chromatogr. B 1997,703, 209 – 215.

[33] Strife, R. J., Jardine, L., Colvin, M., J. Chromatogr. B 1981, 224, 168 –174.

[34] Puri, S. C., Handa, G., Bhat, B. A., Dhar, K. L., Spiteller, M., Qazi,G. N., J. Chrom. Sci. 2006, 44, 239 – 243.

[35] Schmidt, T. J., Hemmati, S., Fuss, E., Alfermann, A. W., Phytochem.Anal. 2006, 17, 299 – 311.

[36] Zhao, L., Tian, X., Fan, P. C., Zhan, Y. J., Shen, D. W., Jin, Y., J. Chro-matogr. A 2008, 1210, 168 – 177.

[37] Gensler, W. J., Gatsonis, C. D., J. Org. Chem. 1966, 31, 3224 – 3227.

[38] Brewer, C. F., Loike, J. D., Horwitz, S. B., Sternlicht, H., Gensler,W. J., J. Med. Chem. 1979, 22, 215 – 221.

[39] Li, Q. R., Yan, G. Q., Ge, T. F., Rapid Commun. Mass Spectrom. 2007,21, 2843 – 2852.


hplc-dad and hplc-esi-ms separation, determination and identification of the spin-labeled...

Documents