Introduction

Natural products have played a majorrole in human health care for cen-turies. The discovery that compoundsfrom natural sources are highly effec-tive against “modern” disease, such asbreast cancer, in combination with therealization that there are thousands ofuncharacterised and even unknownspecies still to be investigated, meansthat the analytical investigation of nat-ural products is of great importance.Whether the source is animal or plant,it is frequently the case that the effec-tive components are present in a richmatrix of other compounds. This pre-sents specialized problems to theinvestigating scientist. Not only shouldthe separation technique employed becapable of high resolution and highefficiency separations, the detectiontechnique chosen should be highlyselective and hopefully capable of pro-viding as much identification datapoints as possible.

AuthorsYi Li

Peking University,

Beijing, PRC

Gordon Ross,

Carsten Buhlmann

Agilent Technologies,

Waldbronn, Germany

Pharmaceutical

CE and CE/MS for the analysis ofnatural products

Application Note

Capillary electrophoresis (CE) is a highefficiency, high resolution techniquewith the added benefits of being able tohandle very complex matrices. Since CEis performed using an open tube of verysmall diameter, it is not prone to foulingas is an LC column, and can easily bewashed by flushing to prevent carry-over. This makes the system inherentlymore robust, more reliable and reducesthe sample preparation required. In thisapplication note we show a number ofapplications demonstrating the utility ofCE with UV and MS detection for theanalysis, component identification andquantification, and characterization oftraditional Chinese medicines.

ents. However, this is complicated bythe long run time necessary (30 min-utes) and contamination of the LC col-umn. The LC method is also a gradientmethod which requires column re-equilibration and washing betweenruns. The CE method described here is simple and rapid requiring only a 4-minute capillary wash between runs.

Sample extraction 1 g herb extracted with 20 mL water andheated to 80 °C for 30 minutes then filte-red through 0.2 µm pore and injected.

Results Figure 1 shows the separation ofMahuang extract. The ephedrine peak

appears after approximately 6 minutesbut the rest of the sample shows anenormous complexity. If using LC, theremaining sample must be washedfrom the column after detection of theactive ingredients. However in CE, thecapillary is simply flushed throughwith fresh buffer. Figure 2 shows theanalysis of a standard solution ofephedrine and a Mahuang extract andcompares the spectra of the indicatedpeaks. Ephedrine could therefore beidentified by its spectra and its migra-tion time and confirmed by spikingwith a known standard. For ephedrinethe analysis was linear over the range1 µg/mL to 1 mg/mL with r2 = 0.9995.

Experimental

Capillary Electrophoresis All analyses were performed using theAgilent Capillary Electrophoresis sys-tem which is equipped with diodearray detection and controlled by theAgilent ChemStation software. ForCE/MS analyses the CE instrumentwas fitted with the CE/MS adapter kit(PN G1603A) which prepares the CEfor coupling with the MS. Thisincludes an MS cassette which allowsthe capillary to be threaded throughthe UV detector interface before leav-ing the instrument for MS connectionthus providing tandem UV and MSdetection.

Mass Spectrometry An Agilent MSD single quadrupolemass spectrometer was used for allCE/MS analyses. An Agilent CE-ESI-MS Nebulizer kit G1607A) was used tocouple the CE with MS using electro-spray ionization (ESI). The sprayerwas of a triple tube design accommo-dating the separation capillary, a tubefor delivery of sheath liquid and anouter tube for delivery of the nebuliz-ing gas and is arranged orthogonal tothe MS entrance capillary in the ionsource. The sheath liquid was deliv-ered by an Agilent isocratic LC pumpequipped with a 1:100 splitter.

The traditional medicines samplesused in the described applicationswere the kind gift of Professor Liu,Peking University, Beijing- PRC.

1. Identification of Ephedrine inMahuang

Mahuang (Ephedra sinica stapf) is aChinese herbal medicine commonlyused in the treatment of asthma andrespiratorial infections. Its activeingredients (ephedrine, norephedrine,psudoephedrine and norpseu-doephedrine) are strong central ner-vous system stimulants. Mahuang canbe used alone or to potentiate theeffects of other herbal medicines. Inorder to evaluate the herbal quality,HPLC may be used to determine theconcentrations of these active ingredi-

Figure 1Analysis of a Muhuang extract.

Time [min]5 10 15 20 25

Abso

rban

ce [m

AU]

0

10

20

30

40

50

60

Figure 2Comparing the spectra of ephedrine in Mahuang with that of a standard.

Time [min]

Time [min]0 1 2 3 4 5 60

10

20

30

40

Wavelength [nm]200 250 300 350

Norm.

0

50

100

150

200

250

300 Ephedrine standard

Mahuang extract

Abso

rban

ce [m

AU]

0

20

40

60

80

0 1 2 3 4 5 6

Ephedrine standard

Mahuang extract

Chromatographic conditionsCapillary: 80.5 cm (72 cm) x 75 µgmBuffer: 25 mM borate pH 9.3Detection: 200/10 nmInjection: 500 mbar x sVoltage: 30 kV

2

2. Identification and quantitati-on of the alkaloids berberine,palmatine and jatrorrhizine inMahonia stem

Many species of the Mahonia plantare used in China as traditional medi-cines. Some of the active componentshave been identified as alkaloids andpharmacological research has deter-mined that the plant has antibacterial,antioxidant, antifungal, anticancer andantiproliferative properties. Here wedescribe the quantitation of threealkaloids in the stem of variousMahonia species.

Sample extraction 2 g of pulverized Mahonia stem wasultrasonicated in 10 mL methanol for 30minutes. This was repeated twice using5 mL methanol and ultrasonicated for 20minutes. The combined extracts werecentrifuged for 10 minutes (4000 rpm)and filtered through a 0.45-µm membra-ne before injection. Prior to each run thecapillary was flushed for 2 minutes with0.1 M NaOH, 2 minutes with water and 5 minutes with buffer. Replenishment ofbuffer vials was performed every 6thrun to achieve best reproducibility.

Results A typical separation of a Mahoniastem is shown in figure 3. Detection at265 nm confers some selectivity on theanalysis and the three alkaloids ber-berine, palmatine and jatrorrhizine arewell separated from other samplematrix components which absorb atthat wavelength. Detection at 200 nmillustrates the complexity of the sam-ple matrix. These three alkaloids canbe automatically detected and identi-fied by spectral library search using a library constructed from standards(figure 4). Detection of berberine, palmatine and jatrorrhizine was linearover the ranges 3.4–109 mg/mL, 1.4– 44 mg/mL and 1.1–37 mg/nL respectively with r2 better than 0.999in all cases.

3

Figure 3Analysis of Mahonia stem by CE with detection at 200 nm and 265 nm.

Time [min]4.5 5 5.5 6 6.5 7 7.5

Norm.

2

4

6

8

10

12

Sig = 200 nm

Sig = 265 nm

Figure 4Library search from separation of alkaloids in Mahonia stem.

Time [min]3 4 5 6 7 8 9

Absorbance[mAU]

0

20

40

60

80

100

120

4.402

4.584

5.766

7.677

Electropherogramof a mixture of 3 alkaloids

Part of the report, generated immediately after the run

Chromatographic conditionsCapillary: 50 µm id x 48.5 cm (40 cm effect.)Buffer: 50 mM phosphate, 50 mM borate containing 50 % methanol with

apparant pH adjusted to pH 8Detection: 200/10 nm and 265/20 nmInjection: 500 mbar x sVoltage: 30 kVTemperature: 20 °C

3. Quantitative analysis oftetrandrine and fangchinoline

Fangchinoline and tetrandine are twoalkaloids which are present in RadixStephaniae tetradrae S. Moore. Thesecompounds have pain-relieving activi-ties, can reduce blood pressure, haveantineoplastic and antibiotic activityand are therefore of pharmaceuticalinterest. The plant is used in variousChinese herbal preparations. Here wedescribe application of a CE method toquantitative analysis of these alkaloidsin some traditional Chinese medicines.

Sample extraction 2 g of each pulverized herbal drug wereextracted with 7 mL 50 % ethanol bystirring for 30 minutes followed by cen-trifugation (4000 rpm, 10 minutes). Theextraction was repeated two times andthe combined extracts were filteredthrough 0.45-µm pore. For electrokineticinjection, a volume of 200 mM NaClsolution equivalent to one fifth of thesample volume was added to the sam-ple to equalize the sample conductivity.

Results The medicines were separated using aMEKC system with Tween- 20 as thesurfactant2. This resulted in a very cleanelectropherogram where two peakscould easily be seen (figure 5). Due totheir similar structure tetrandrine andfang-chinoline (figure 6) have very simi-lar spectra and therefore were identi-fied by spiking experiments with purestandards. After identification the twoalkaloids were quantified in a number oftraditional Chinese medicines. Linearitywas determined for both compoundsover the range of 5 to 250 µg /mL.Linearity was greater than 0.9999 forboth analytes. Reproducibility of migrati-on times was very good (< 0.4 %). Forquantitation, the reproducibility of peakareas was acceptable (< 4 %) butdepended on the medicine and there-fore the sample matrix (table 1).

4

Figure 5Analysis of tetrandrine and fangchinoline in Radix Stephaniae tetradrae S. Moore.

Table 1Migration time and peak area reproducibility (n=3) and quantitation of tetrandrine and fangchinolinein various traditional Chinese medicines.

Wavelength [nm]200 250 300 350

Absorbance[mAU]

0

10

20

30

40

Time [min]5 10 15 20 25

3.443

25.462

27.517

29.909

tetrandrine

fangchinoline

Figure 6Structure of terandrine and fangchinoline.

fangchinoline tetrandrine

H

N

OO

OH

OH

N

O

O

H

N

OO

O

OH

N

O

O

Tetrandrine FangchinolineMedicine mg/L % RSD % RSD mg/g mg/L % RS % RSD mg/g

area time area time

Fang ji guan jie wan 28.41 2.44 0.32 0.41 17.03 3.45 0.36 0.25Qu feng gu tong lu 6.96 1.89 0.08 0.07 3.88 2.15 0.07 0.04Ling long gan mao jiao nang 16.52 1.08 0.23 0.17 9.79 1.16 0.22 0.10Xi xian feng shi wan 7.47 5.58 0.28 0.07 3.54 2.11 0.29 0.04Feng shi ton gao 7.14 0.05 0.15 0.08 3.93 1.16 0.15 0.04Shen jin dan jiao nang 10.94 3.33 0.21 0.12 5.81 3.97 0.25 0.07Radix Stephaniae tetrandrae 233.26 1.09 0.21 5.52 185.04 1.06 0.22 4.38

Chromatographic conditionsCapillary: 64,5 cm (56 cm) x 50 µmBuffer: 60 mM phosphoric acid\TAE, 50 mM Tween-20, 20 % methanol, pH 2,5,Detection: 214/10 nmInjection: 4 kV x 23 secVoltage: 20.2 kV

extract (figure 9). Table 2 shows the calculated amounts of CA in Flos

Lonicerae, some TCMs and in a coffeeextract.

4. Identification and quantifica-tion of chlorogenic acid

Chlorogenic acid (CA) (figure 7), is anester of caffeic acid and quinic acid.All three of these substances naturally

occur in many plants. For example, CAwhich is present in the surface skin ofpeaches, inhibits the cutin-digestingenzyme of the brown rot fungus,Monilinia fructicola demonstrating itsantifungal activity. It has also beendemonstrated to have antioxidantactivity. CA is an active consituent ofFlos Lonicerae and CZE has been usedfor its analysis in the plant and insome traditional medicines which con-tain this plant as a constituent 3. Herewe describe how CA can be identifiedand quantified in some Chinese tradi-tional medicines using capillary elec-trophoresis.

Sample extraction Weighed amounts of pulverized samples were soaked with 7 mL 50 %ethanol/water overnight and extractedby stirring for 30 mintes. After centrifu-gation (4000 rpm, 10 minutes) the extrac-tion was repeated two more times. Thecombined extraction volume was madeup to 25 mL and filtered through 0.45 µm.Liquid samples were simply diluted andfiltered before measurement.

Results The analysis of extracts using the abovemethod gives a complex electrophero-gram (figure 8). Identification of CA ismade more problematic because wheninjected individually as a standard itmigrates in an area occupied in thesample injection by three peaks, all ofwhich have similar spectra. By spikingsamples with pure CA it can be unequi-vocally identified in the prepared TCM

5

Table 2Calculated amounts of CA in Flos Lonicerae, some TCMs and in a coffee extract.

Medicine Chlorogenic acid Powder weight (g) Content (mg/g))L/gm( ro)L/gm(

Flos Lonicerae 670 0.946 17.83Zhi I jin hua wan 61.18 3.34 0.49Vc yin qiao pain 104.1 2.83 0.85Yin qiao jie du pain 542.9 2.9 3.5Xiao er qing re jie du kou fu ye 31.14 liquid 0.39Shung huang lian 187.8 liquid 2.1Café (hot water extract) 197.23 liquid 0.2

Figure 8Analysis of an extract of the traditional Chinese medicine Zhi zi jin wan.

Time [min]5 10 15 20 25 30 35

Abso

rban

ce [m

AU]

0

20

40

60

80

Figure 9Identification of cholorogenic acid in a traditional Chinese medicine by spiking with a pure standard.

[min]10 20 30 [min]10 20 30

Shung huang lian could be separated in chlorogenic acid and two similar phenolic acids

Shung huang lian sample was spiked with 5 times the amount of chlorogenic acid

Figure 7Structure of chlorogenic acid.

chlorgenic acid

HO

HO

HO O

OH

OH

OH

O

O

C

Chromatographic conditionsCapillary: 64.5 cm x 50 µm (56 cm effect.)Buffer: 40 mM phosphate, 80 mM boric acid containing 5 % ethanol with

apparant pH adjusted to 7.0Detection: 254 nmInjection: 500 mbar x sVoltage: 20 kV

Sample preparation Different concentrations of tetrandrineand fangchinoline ranging from 0.2 µg/mLto 200 µg/mL were prepared by dissol-ving the standard in 50 % aqueous etha-nol solution. Other samples were prepa-red as previously described.

Results The run buffer was chosen for its com-patibility with electrospray ionization.The separation achieved was not opti-mal however using the second dimensi-on of MS detection ensured that theanalytes could be selectively detectedand identified.

Linearity and detection limits Linearity and limits of detection weredetermined for tetrandrine and fangchi-noline. The low detection limit and linearrange of CE/MS indicate the possibilityof trace-level identification and quanti-tative determination of tetrandrine andfangchinoline in Chinese herbal medici-ne. Detection was linear over the range0.2 to 200 µg/mL for tetrandrine andfangchinoline with r2 values of 0.9991and 0.9986 respectively. The limits ofdetection were 3 µg/mL in TIC and 0.05 µg/mL in SIM.

RepeatabilityThe % relative standard deviation wascaculated for peak areas for five repli-cate injections of standard solution of 50 µg/mL. Reproducibility for fangchino-line was 6.1 % while that for tetrandrinewas 5.9 %. The relatively poor repeata-bility of peak area by SIM compared tothe results of the UV detection, whichshowed % RSD of ca. 3.1, was most pro-bably due to variation in the efficiencyof the ionization process. The crudedrug of Radix Stephaniae tetrandraeand three Chinese herbal preparationswhich contain Radix Stephaniaetetrandrae were analyzed.

MS ion optics, especially where thespray is "in-line" with the MS entrancecapillary, although the orthogonalarrangement of the sprayer used inthis study can help to mitigate thiseffect5. In this application a methodwas developed for the CE/MS analysisof tetrandrine and fangchinoline inChinese herbal medicines, and theirconcentrations in some Chineseherbal medicines were determined.

Experimental The preparations analyzed were:

1) Fang ji guan jie wan: Radix Stepahniae tetrandrae (1.80 g),Poria (Poria cocos (Schw.) Wolf) (1.80 g),Rhizoma atractylodis macrocepahalae(Atractylodes macrocephala Koidz) (1.2 g), Radix Aconiti (0.60 g), RadixGlycyrrhizae (Glycyrrhiza uralensisFisch.) (0.60 g), Radix Codonopsis pilosulae (Codonopsis pilosula (Franch.)Nannf.) (0.6 g).

2) Qu feng gu tong lu: Caulis Piperis futokadsurae (Piper foto-kadsura Sinb. et Zucc.) (3.4 g), RadixClematis (2.8 g) Caulis spatholobi(Spatholobus suberectus Dunn) (2.8 g),lignum sapan (Caesalpinia sappan L.)(2.0 g), Radix Stepahniae tetrandrae(2.0 g), Cortex periplocae radicis(Periploca sepium Bge.) (2.0 g), HerbaSieges-beckiae (2.0 g), Radix Aconitikusnezoffi (aconitum kusnezoffii Rchb.)(1.4 g).

3) Shen jin dan jiao nang: Lumbricus (Pheretima aspergillumPerrier) (8.3 g), Flos Carthami(Carthamus tinctorius L.) (5.83 g),Olibanum (Boswellia carterii Bierdw.)(2.5 g), Myrrha (Commiphora myrhaEngl.) (2.5 g), Radix Stephaniae terand-rae (2.5 g), Cortex periplocae radicis (2.5 g) Rhizoma drynariae (Drynariafortunei (Kunz) J. Sm.) (2.5 g).

5. Analysis of tetrandrine andfangchinoline by CE/MS

Although CE can provide good resolu-tion of these compounds in complexmatrices (see section 3) one of itsmain limitations, when used with UVdetection, is its relatively low sensitivi-ty due to the short optical path lengthwhich is the capillary bore. These two alkaloids have extremely similarUV absorption spectra making theirunequivocal identification via UV-spectral library search alone problem-atic. The online coupling of CE withmass spectrometry (MS) is a promis-ing combination. While CE provideshigh separation efficiency, MS affordshigh sensitivity and selectivity, as wellas molecular structural information.MS detection can provide unequivocalidentification of the analytes by theirmasses.

Successful coupling of CE to MS isdependent upon the efficiency of ions’transfer from the aqueous to thegaseous phase. This, in turn, dependsupon the volatility of the running bufferand therefore places some restrictionson the types and /or concentrations ofbuffer which can be used for CE/MS.

The previous separation of tetrandrineand fangchinoline was only obtainedusing MEKC conditions, with a buffersolution of 60 mM triethylamine-phos-phate containing 50 mM Tween-20 and20 % methanol, where the analyteswere separated according to differ-ences in their hydrophobicity.However, such a running buffer is notsuitable for ESI- CE/MS because itsnon-volatile components can hinderthe electrospray process by reducingthe spray efficiency and can causeintense background signals resultingin a decreased sensitivity in the TIC4.Further, these buffer components cancontaminate the ESI interface and the

6

Figure 10 shows that the presence offangchinoline and tetrandrine in the her-bal medicine “Qu feng gu tong lu”extract is unclear using CE-UV wherethe electropherogram contains a num-ber of large and very small peaks. Anumber of other peaks are present inthe area from two to four minutes in theUV trace, where the alkaloids’ peaksare expected. However, using the SIMmode of MS the two alkaloids could bedetected unambiguously. The repeatabi-lity (% RSD, n=5) of peak area of fanchi-noline and tetrandrine in “Qu feng gutong lu” were 3.5 % and 4.3 % respec-tively. Results from quantitation tetrand-rine and fangchinoline in RadixStephaniae tetrandrae and three otherChinese herbal preparations are shownin table 3

7

Figure 10Analysis of Tetrandrine and fangchinoline in Qu feng gu tong lu by CE with UV and MS detection.

Time [min]0 2 4 6 8

Absorbance[mAU]

0

2

4

6

8

10

12

14

16

Time [min]1.5 2 2.5 3

Absorbance[mAU]

0

2

4

6

Time [min]2 4 6 8

0250005000075000

100000125000150000175000 SIM 305 m/z

Time [min]2 4 6 8

0250005000075000

100000125000150000175000 SIM 312 m/z

Abundance

Abundance

Table 3Determined values (mg/g) of fanchinoline and tetrandrine in Chinese herbal medicines by CE/MS.

Fangchinoline ( mg/g) Tetrandrine (mg/g)

91.747.5 eardnartet eainahpetS xidaR62.1637.0naw eij naug ij gnaF

315.0252.0ul gnot ug gnef uQ338.0074.0gnan oaij nad nij nehS

Chromatographic conditionsCECapillary: 80 cm (22 cm UV) x 50 µm idBuffer: 100 mM formic acid pH 2.5Detection: 210 nm/16 nmInjection: 200 mbar x sVoltage: 27 kVTemperature: 20 °C

MS conditionsSheath liquid: 5 mM ammonium formate in 50 % methanolFlow rate: 5 µL/minNebulization gas pressure: 10 psiElectrospray voltage: -4.0 kV (positive ion mode)Drying gas temperature: 250 °CMS Scan: m/z 300–m/z 650 at the rate of 0.85 s/cycleMS SIM: m/z=305.2 and m/z=312.1

6. CE/MS for alkaloid finger-printing of Mahonia species

The pharmaceutically relevant com-pounds in Mahonia species are alka-loids, including berberine, palmatine,jatrorrhizhine, columbamine, isotetran-drine, berbamine, oxyacanthine, andmagnoflorine amongst others.

Experimental Two species of Mahonia: M. fortuneiand M. japonica were collected in Chinaand identified by Beijing University ofChinese Medicine.

Sample preparationA 2.0 g sample of pulverised dried stemwas extracted with 10 mL methanol byultrasonication for 30 minutes. Extractionwas repeated twice with 5 mL methanolfor 20 minutes. All methanol extractswere combined and centrifuged at 4000 rpm for 15 min, then filtered througha 0.45 µm filter.

Results The electropherograms of the extractsare shown in figure 11. The total ionelectropherogram (TIE) is shown in thelower part and the UV signal in theupper part of the figures. The mainpeaks 1, 2, and 5 in the electrophero-gram could be identified as berberine,palmatine and jatrorrhizine. Peak assig-nment was performed by comparison ofmigration time, UV spectrum and m/zwith those of standards. Compound 4showed a dominant signal at m/z 338and was tentatively assigned ascolumbamine, which has been reportedto exist in the genus of Mahonia.Compound 9 showed a major signal atm/z 312 and a minor one at m/z 623 indi-cating that it is an alkaloid with twonitrogens which exists both as a singleprotonated ion (m/z 623) and a doubleprotonated ion (m/z 312). Therefore, itwas tentatively assigned as isotetrand-rine, which has also been reported toexist in the genus of Mahonia. Peaks 3,6, 7 and 8 gave m/z values of 314, 305,

8

Figure 11CE/MS of two Mahonia species.

0 1 2 3 4

0

10

20

30

40

50

60M. fortunei 1

234

5

6

8,9UV

0 2 4 6 8 10 12 14

050000

100000150000200000250000300000350000400000

TIE

1

2 3

4

5

6,78,9

Abundance

Time [min]Time [min]2.5 5 7.5 10 12.5 15 17.5

050000

100000

150000200000250000300000

350000

TIE

0 1 2 3 4

Absorbance[mAU]

0

10

20

30

40

50

60 M. japonica

94

2,3

1 5

6

UV

94

2,3

15

6

Abundance

Absorbance[mAU]

7

Chromatographic conditionsCECapillary: 70 cm (22 cm UV) x 50 µm idBuffer: 50 mM amonium acetate

pH 9.0, 40 % methanolDetection: 200 nm/16 nmInjection: 250 mbar x sVoltage: 25 kVTemperature: 20 °C

MS conditionsSheath liquid: 0.5 % formic acid in 50 % methanolFlow rate: 5 µL/minNebulizationgas pressure: 10 psiElectrosprayvoltage: -4.0 kV (Positive ion mode)Drying gasTemperature: 200 °CMS scan: m/z 300–m/z 650 at the rate of 0.85 s/cycle

305 and 312 respectively. Since bothberbamine and oxyacanthine have m/zof 305 and have been reported as pre-sent in Mahonia, peaks 6 and 7 could beassigned either as berbamine or oxya-canthine. Peak 3 and 8 could not beidentified but due to their almost ubiqui-tous presence in a number of the spe-cies they were also monitored. Peaks 3,6, 7, and 8 could not be unequivocallyidentified but they could be specificallycharacterized by their m/z and migrationtimes. SIM mode was employed todetect all the putative alkaloid peakswhich appeared in the TIC of the eightspecies of Mahonia.

Figure 12 shows the SIM signals obtai-ned from analysis of these two Mahoniaspecies. Since CE/MS is a two-dimen-sional separation technique, compo-unds with different µe could be resolvedby electrophoresis, while compoundswith similar µe but different m/z couldbe selectively detected by using SIMmode. Therefore, although compounds 4and 5, 6 and 7, and 8 and 9 have thesame m/z, they could be separatedbased on differences in their variouselectrophoretic mobilities prior to SIMdetection.

Fingerprints of the two species ofMahonia were constructed by calcula-ting the area of each alkaloid peak as apercentage of the total area of all ninepeaks detected in SIM mode (figure 13).The profile of these nine peaks is quitedifferent in these two species. Clearlythe alkaloids berberine, palmatine andjatrorrhizine (peaks 1, 2 and 5) are themost predominant, however their pro-portions are quite different and levels ofthe other analytes help to further diffe-rentiate these two species.

9

Figure 12SIM traces from CE/MS of two species of Mahonia.

Time [min]5 10 15

0400000

800000

1200000

1600000

Time [min]5 10 15

04000080000

120000160000

Time [min]5 10 15

050000

100000150000200000250000

M. fortunei

336 m/z338 m/z

352 m/z

305 m/z

314 m/z312 m/z

ecnadnubAecnadnubA

Time [min]5 10 15

0200000400000600000800000

1000000

Time [min]5 10 15

0200000

600000

1000000

1400000

Time [min]5 10 15

020000400006000080000

100000

M. japonica

336 m/z338 m/z

352 m/z

305 m/z

314 m/z

312 m/z

Figure 13Fingerprint of components of Mahonia species detected by CE/MS. Blocks represent the peak.

0.00

20.00

40.00

60.00

80.00

1 2 3 4 5 6 7 8 9

0.00

20.00

40.00

60.00

80.00

1 2 3 4 5 6 7 8 9

M. fortunei

M. japonica

References

1. Li, Y., Ji, X., Liu, H., Yan, Y., Li,J.,Chromatographia. 51, 357–361, 2000.

2. Yang, J, Long, H, Liu H and Sun, Y. J.,Chromatogr A. 811, 274–279, 1998.

3. Long H, Yang, J-J, Liu H-W, Wang, T-S.Huang, A-J and Sun Y-L. J., ChineseMedicine, 2000.

4. Serwe, M and Ross G.G., LC-GC 18 (1),46–55, 2000.

5. “Development of a new orthogonalgeometry atmospheric pressureionization interface for LC/MS.”Agilent application note, 5968- 4465E,2007.

Conclusion

CE has been shown to be useful forthe separation of traditional Chinesemedicines. In combination with DADUV or MS detection, components ofthese sample types can be clearlyidentified using a variety of techni-ques, spectral analysis, spiking withknown standards, m/z determination.The open tubular arrangement of CEmakes it very suitable to the complexmatrices which are encountered inthe analysis of natural products.Coupling of CE with MS adds a seconddimension of separation where ana-lytes which have similar mobilitiesmay be separated by the MS whileisobaric analytes require separationprior to their detection in the MS.

10

11

www.agilent.com/chem/ce

© Agilent Technologies, Inc., 2000-2009Published March 1, 2009Publication Number 5990-3406EN