5989-9912EN HERBAL ADULTERANTS

Analysis of Herbal Medicines and Healthcare Products

Application Compendium

2

This compendium is a collection of applications on the analysis of traditional

herbal medicines and healthcare products. The applications are grouped

according to the types of substances being analyzed. These include pesticide

residues, heavy metals, residual solvents, bioactive compounds and metabolites.

All applications featured in this compendium have been developed with

complete solutions using Agilent analytical products, software and services.

The analytical techniques deployed include gas chromatography (GC), liquid

chromatography (LC) and mass spectrometry (MS) as well as hybrid techniques

such as GC/MS, LC/MS and ICP-MS. Further details about each solution can be

found on the Agilent web site. Simply visit the web address given at the end of

this compendium.

Introduction

Introduction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .2

Contents . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .3-4

Development of Analytical Methods for Quality Control . . . . . . . . . . . . . . . .5-44

Extraction and HPLC Analysis of Alkaloids in Goldenseal . . . . . . . . . . . .7

The High-Resolution Reversed-Phase HPLC Separation of Licorice

Root Extracts using Long Rapid resolution HT 1.8-µm Columns . . . . .13

Fast, high-resolution analysis of notoginseng by

rapid resolution liquid chromatography . . . . . . . . . . . . . . . . . . . . . . . . . . .19

Analysis of ginseng and American ginseng using the

Agilent 1120 Compact LC . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .25

Development of reliable quality control methods for TCM

preparations using rapid resolution LC with UV and MS detection . . .29

Analysis of Traditional Chinese Medicines with the

Agilent 1200 Series evaporative light scattering detector . . . . . . . . . . .23

Analysis of TCMs with the Agilent 1200 Series

evaporative light scattering detector . . . . . . . . . . . . . . . . . . . . . . . . . . . . .39

Monitoring of Pesticide Residues . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .45-58

Screening for 430 Pesticide Residues in Traditional Chinese

Medicine Using GC/MS: From Sample Preparation to

Report Generation in One Hour . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .47

Low Part-per-Billion Level Pesticides Screening in Traditional

Chinese Medicine Using the Agilent 7000A GC/MS/MS . . . . . . . . . . .53

Determination of Heavy Metal Content . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .59-70

Determination of Toxic Elements in Traditional Chinese

Medicine Using Inductively Coupled Plasma Mass Spectrometry . . . .61

Fast determination of five toxic elements in Traditional

Chinese Medicine (TCM) by ICP-MS . . . . . . . . . . . . . . . . . . . . . . . . . . . . .67

Quality Control for Residual Solvents . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .71-80

Better precision, sensitivity, and higher sample throughput

for the analysis of residual solvents in pharmaceuticals –

Using 7890A GC with G1888 headspace sample . . . . . . . . . . . . . . . . . . .73

Contents

3

4

Purification and Profiling of Bioactive Compounds . . . . . . . . . . . . . . . . . . .81-122

Isolation of formononetin and other phytoestrogens from

red clover with the Agilent 1100 Series purification system . . . . . . . . .83

Analysis of a complex natural product extract from ginseng –

Part I: Structure elucidation of ginsenosides by rapid resolution

LC-ESI TOF with accurate mass measurement . . . . . . . . . . . . . . . . . . . .89


Part II: Structure elucidation of ginsenosides by high resolution

ion trap LC/MS . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .97


Part III: Species differentiation of ginseng plants and authentication

of ginseng products by LC/MS . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .103

Agilent MassHunter – Fast computer aided analysis of

LC/ESI-TOF data from complex natural product extracts – Part 1:

Analysis of 6210 data with the Molecular Feature Extractor in

MassHunter Workstation software . . . . . . . . . . . . . . . . . . . . . . . . . . . . .111

Agilent MassHunter – Fast computer aided analysis of

LC/ESI-TOF data from complex natural product extracts – Part 2

Comparison of 6210 TOF data from different biological origin

with the Mass Profiler in MassHunter software . . . . . . . . . . . . . . . . . .117

Metabolic Profiling . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .123-132

An interwoven multi-algorithm approach for computer-assisted

identification of drug metabolites - Rapid identification of

drug metabolites from accurate QTOF MS and MS/MS data by

MassHunter metabolite identification software . . . . . . . . . . . . . . . . . .125

5

2 3 4 5 6 7 8

Palm

atin

e

S

S

M

Development of Analytical Methods for Quality Control

6

AuthorHolly A. Weber

Midwest Research Institute

425 Volker Blvd

Kansas City, MO 64110

USA

Maureen Joseph

Agilent Technologies, Inc

2850 Centerville Road

Wilmington, DE 19808-1610

USA

Abstract

An ambient extraction of goldenseal root powder fol-lowed by HPLC analysis of the alkaloids on a ZorbaxRapid Resolution Eclipse XDB-C18 column provides anaccurate method for the determination of key alkaloids ingoldenseal, including berberine and hydrastine. Theextraction and HPLC analysis can be applied to severalother alkaloids, including canadine, hydrastinine, andpalmatine, and may be applicable to other berberine-con-taining plant roots. The Rapid Resolution Eclipse XDB-C18column is used for an isocratic separation with high res-olution of all components in under 15 minutes.

Extraction and HPLC Analysis of Alkaloidsin Goldenseal

Application

Introduction

Goldenseal, Hydrastis canadensis L., is one of the oldestherbal medicinal plants and is of current interest as a nat-ural medicine. There are two alkaloids present in gold-enseal that are the expected active components, berberineand hydrastine. Canadine, hydrastinine, berberastine, andcanadaline are minor alkaoloids. Palmatine, which is closelyrelated to berberine, is not found in H. canadensis, but isfound in Coptis, another berberine-containing plant [1].

Goldenseal has been used as an anti-inflammatory andantibiotic. It has also been used to treat nasal congestion,cold, flu, and a variety of intestinal disorders. The wholeroot of the plant is used and is currently available in bulk(dried or powdered roots), tablets, capsules, and tinctures.Goldenseal plants have been overharvested and many arenow grown on farms for use as herbal supplements.Inconsistent quantities of alkaloids are present in the prod-ucts sold as herbal supplements. A simple process forextracting and analyzing the alkaloids is highly desirable toevaluate product quality. Figure 1 shows the structures ofthe alkaloids in goldenseal.

Consumer Products and Drug Manufacturing/QA/QC

5988-7136EN7

Experimental

Extraction Procedure

Literature reports by Betz and Anderson [1] and Burney [2]indicate that ambient extraction of alkaloids from H.canadensis is possible. This is followed by an HPLC deter-mination for accurate quantitation of the alkaloid extracts.The optimized ambient extraction conditions used for theanalysis of neat goldenseal root powder are summarized inthe steps below.

1. Weigh ~ 0.5 g of root powder

2. Mix with 100 mL of acetonitrile:water:H3PO4 (70:30:0.1,v/v/v)

3. Sonicate 5 min, shake (wrist-action shaker)10 min, centrifuge 5 min

4. Dilute extracts 1/5

5. Direct HPLC analysis of diluted extracts

HPLC Analysis

The HPLC analysis of alkaloids present in goldenseal needsto resolve the major alkaloids. In addition, it is desirable toresolve palmatine, because it is present in other berberine-containing plants. An HPLC method was developed toresolve all of these components. Optimum resolution andpeak shape were obtained using the Zorbax EclipseXDB-C18 column with an ammonium acetate buffer andacetonitrile. The separation is shown in Figure 2 with acomplete list of the optimized conditions used. For consis-tent retention times, temperature control was required at30°C [3].

N+

N+

H3CO

H3CO

H3CO

H3C

H3C

OCH3

OCH3

OCH3

OCH3

H3CO

H3CO

OCH3

O

O

N

O O

O

O

N O

O

OH

N

O

O

Hydrastine

Canadine

Berberine

Hydrastinine

Palmatine

Figure 1. Structures of key alkaloids in goldenseal and related plants.

85988-7136EN

Method Validation

This HPLC method was applied in the validation of theambient extraction method. Linearity, precision, and alka-loid recovery were investigated. Table 1 shows theseresults. The precision of the method is excellent, andrecoveries of the different alkaloids ranged from92%–102%, excellent for a quantitative method. The linear-ity was very good (Figure 3) and sensitivity was good. Fromthe standard data, the calculated LOD (limit of detection)for berberine was 0.50 mg/mL and the LOQ (limit of quan-titation) was 1.65 mg/mL, so that accurate quantitationdown to low levels is possible, making it easy to test thequality of different goldenseal products.

0 1 2 3 4 5 6 7 8

Time (min)

Resp

onse

Palm

atin

e

Can

adin

e

Ber

beri

ne

Hyd

rast

ine

9 10 11 12 13 14 15 16 17

Spiked sample

Sample

Mixed standard

Column: Zorbax Eclipse XDB-C18 (3.5 µm, 4.6 × 150 mm)Guard column: Zorbax Eclipse XDB-C18 (5 µm, 4.6 × 12.5 mm)Solvent: 68% 30 mM ammonium acetate, 14 mM TEA, pH ~ 4.85 32% acetonitrileWavelength: 230 nmFlow rate: 1 mL/minColumn temperature: 30 ˚CInjection volumn: 10 µLRun time: 17 min

Figure 2. HPLC separation of goldenseal extract on Eclipse XDB-C18 column.

5988-7136EN9

Table 1Validation Results of the Ambient Extraction Method

Figure 3Linearity of berberine and hydrastine as standards and from goldenseal extracts.

Palmatine Berberine Hydrastine Canadine

Precision (n = 10) 0.18 ±0.002(s)% 3.06 ±0.05(s)% 2.04 ±0.01(s)% 0.08 ±0.001(s)%

Alkaloid recovery(~ 0.3–2 g of GS)(~ 0.6–1 mg/mL) (n = 12)

0.18 ±0.003(s) 3.10 ±0.06(s)% 2.05 ±0.02(s)% 0.08 ±0.001(s)%

Spike and recovery(n = 3)

Spike level = ~ 0.15%92.2 ±5.5(s)%

Spike level = ~ 2.0%101.5 ±0.2(s)%

Spike level = ~ 2.0%101.9 ±0.2(s)%

Spike level = ~ 0.10%101.9 ±7.9(s)%

Linearity of Berberine and Hydrastine Standards

0

1000000

2000000

3000000

4000000

5000000

6000000

0.000 20.000 40.000 60.000 80.000 100.000 120.000

µg/mL

Are

aA

rea

BerberineHydrastine

Berberine

Hydrastine

Goldenseal LinearityBerberine and Hydrastine

0

200000

400000

600000

800000

1000000

1200000

1400000

1600000

1800000

0.5000 0.6000 0.7000 0.8000 0.9000 1.0000 1.1000

mg goldenseal/mL diluent

(s) = standard deviation

105988-7136EN

Results

Goldenseal Testing

The ambient extraction and HPLC analysis method wereapplied to six lots of goldenseal root powder from three dif-ferent vendors to determine the alkaloid content. Theresults are summarized in Table 2. These results demon-strate variability among vendors and the reason a goodquantitative HPLC method is desirable.

Table 2Results of Testing Goldenseal Root Powder from Multiple Lots and Multiple Vendors with theAmbient Extraction Method and HPLC

Vendor Lot number % Palmatine % Berberine % Hydrastine % CanadineTotal weight % ofknown alkaloids

1 AB

ndnd

3.273.29

2.362.40

0.090.07

5.945.98

2 CD

0.190.18

3.013.06

1.992.04

0.090.08

5.535.36

3 EF

ndnd

4.603.93

4.062.67

0.120.20

8.996.93

Conclusions

This ambient extraction method of goldenseal is simpleand reliable. This is followed by an isocratic HPLC analysiswith a Zorbax Rapid Resolution Eclipse XDB-C18 column,which provides high resolution and excellent peak shapeof six alkaloids in 15 minutes. This analysis provides reli-able quantitative results of the alkaloids in goldenseal,including berberine and hydrastine. The method wasapplied to goldenseal from three different vendors and maybe applicable to other berberine-containing plants.

References1. J. M. Betz, S. M. Musser, and G. M. Larkine,

“Differentiation between Goldenseal (Hydrastiscanadensis L.) and Possible Adulterants by LC/MS,”presented at the 39th Annual Meeting of the AmericanSociety of Pharmacognosy, Orlando, FL, July 19–23,1998.

2. M. L. Anderson and D. P. Burney, J. AOAC Intern.,81:1005–1110 (1998).

3. H. A. Weber, et. al., J. Liq. Chrom. and Rel. Tech. 24(1)87–95 (2001).

Acknowledgements

This work was performed by Holly A. Weber, Matthew K.Zart, Andrew E. Hodges, Kellie D. White, Roger K. Harris,and Alice P. Clark of Midwest Research Institute, 425Volker Blvd, Kansas City, MO 64110.

Diane Overstreet and Cynthia Smith of the NationalToxicology Program, 111 Alexander Drive, ResearchTriangle Park, NC 27709.

This work was funded by the National Institute ofEnvironmental Health Sciences, Contract Nos. N01-ES-55385 and N01-ES-05457.

For More Information

For more information on our products and services, visit ourWeb site at www.agilent.com/chem.

5988-7136EN11

125988-7136EN

AuthorBernard Permar and Ronald E. Majors

Agilent Technologies, Inc.

2850 Centerville Road

Wilmington, DE 19808-1610

USA

Abstract

High-resolution reversed-phase HPLC analytical studiesusing licorice, a licorice hydrolysis product, and commer-cial licorice samples, showed that resolution andthroughput using a ZORBAX 1.8-µm column greatlyexceeded that obtained using the conventional 5.0-µm column.

Introduction

Licorice is derived from the root of the Glycyrrhiza glabraplant, a 4- to 5-foot woody shrub that grows in Europe, theMiddle East and Western Asia. The root of the plant isknown to contain about 4% glycyrrhizin, the potassium orcalcium salt of glycyrrhizinic acid. The latter is a glycosideof a pentacyclic triterpine carboxylic acid (18-ß-glycyrrheticacid) with two molecules of glucuronic acid (Figure 1).

Glycyrrhizin is about 50 times sweeter than sucrose (cane sugar) but at high dosage is known to have toxicity.Upon hydrolysis, the glycoside loses its sweet taste and is converted to the aglycone glycyrrhetinic acid plus twomolecules of glucuronic acid.

The High-Resolution Reversed-Phase HPLCSeparation of Licorice Root Extracts UsingLong Rapid Resolution HT 1.8-µm Columns

Application

Extractions from any of the many species of this plant willyield a complex mixture containing more than 100 com-pounds. Several of these compounds are used as additivesin candy as sweeteners, in cough syrup as flavoringagents, and in drugs to mask a bitter taste, or for theirtherapeutic qualities, mainly in Traditional ChineseMedicines (TCMs). The medicinal properties of licoricehave been known for several centuries in China, as well asIndia, Egypt, Greece, and Rome. Uses included cough sup-pressant, laxative, and treatments for gastric ulcer, earlyAddison disease, and liver disease. Most recently, gly-cyrrhizin has been shown to have antiviral activity againstDNA and RNA viruses (influenza A and B, HIV, VZV,Hepatitis B and C) [1]. Licorice has also been used in topi-cal cosmetic applications.

Food Additive, Natural Products, and Pharmaceuticals

OCOOH

OO

COOH

COOH

O

Figure 1Structure of glycyrrhizinic acid.

5989-4907EN13

The abundance of certain compounds of interest will varygreatly according to the species of the plant, the time ofharvest, and the method of extraction. Thus, analyticalmethods to follow the active ingredients are required.Gradient elution reversed-phase HPLC has been found tobe an effective method for separating some of the impor-tant compounds in licorice [2]. This application note com-pares the traditional HPLC methodology and the newerRapid Resolution high-throughput (RRHT) columns. Wewill apply these HPLC techniques to investigate the differ-ences between two commercially available licorice rootextracts.

Experimental

Two reversed-phase (RP) columns were used in this study:

• Conventional ZORBAX StableBond (SB)-C18, 4.6 mm x 250 mm, 5 µm

• ZORBAX SB-C18 RRHT, 4.6 mm x 150 mm, 1.8 µm

The smaller particle size of the RRHT column allows use ofa shorter column to achieve the same resolution as thelonger conventional column, and also allows more rapidseparations.

Results

HPLC conditions

Instrument: Agilent 1200 Series Rapid Resolution System

Detector: Multiple wavelength detector (MWD),

254 nm/100 BW, 450 nm reference

Mobile phase: A = 1% Acetic acid in water

B = 1% Acetic acid in acetonitrile

Gradient conditions for ZORBAX SB-C18 columns:

Conventional: 4.6 mm x 250 mm, 5 µm

5% to 100% B in 50 minutes

RRHT: 4.6 mm x 150 mm, 1.8 µm

5% to 100% B in 30 minutes

Flow: 1.0 mL/min

Temperature: Ambient

The most important compound found in a typical licoriceextract is G and to a lesser extent, its hydrolysis product,GA. These substances can be purchased commercially.Although some of the other components of licorice have been identified and are available commercially, theyare quite expensive. Since our main objective was todemonstrate the advantage of using shorter, high-resolu-tion HPLC columns, we used only two standards (G andGA) to develop the initial method. Figure 2a shows the gradient separation of G and GA on the conventional col-umn (ZORBAX SB-C18, 4.6 mm x 250 mm, 5 µm) using gra-dient elution. Since the licorice extract to be examined wasquite complex, isocratic conditions were not usable to sep-arate all of the components. The G being more polar byvirtue of the additional sugar moieties eluted first while theGA came off the column much later. Using this gradient,the GA eluted in just under 42 minutes. By switching to theshorter ZORBAX SB-C18 RRHT column (4.6 x 150 mm, 1.8 µm), the separation was virtually the same, as can beseen in Figure 2b. However, the separation time was nowjust over 25 min, a savings of about 40% in time.

Standards: Purchased from Sigma Aldrich

• (G) 0.1-mg glycyrrhizic acid ammonium salt, ~75 %, dissolved in 0.5-mL mobile phase B, then brought to 1.0 mL by adding 0.5-mL mobile phase A

• (GA) 0.1-mg 18-beta-glycyrrhetinic acid, 97%, dissolved in 0.5-mL mobile phase B, then brought to 1.0 mL by adding 0.5-mL mobile phase A

Samples:

• Licorice root extract A (HERB FARM brand)

• Licorice root extract B (Newark Natural Foods)

Both extracts should be vortexed, then filtered (0.2 micron) priorto injection.

Injection volume: 5 µL of extract

145989-4907EN

min

min

0 5 10 15 20 25 30

mAU

0

40

80

120

160

0 10 20 30 40 50

mAU

0

20

40

60

80

G 14.6 min

G 22.9 min

GA 41.4 min

GA 25.5 min

ZORBAX SB-C18, 4.6- × 250-mm, 5 µm

ZORBAX SB-C18 RRHT, 4.6- × 150-mm, 1.8 µm

2a

2b

Figure 2Gradient reversed-phase separation of G and GA on the test 5.0- and 1.8-µm columns.

To investigate the use of these columns for the separationof actual licorice root extracts, Figures 3 and 4 depict thecomplex chromatograms observed by injection of filteredextracts, identified in the Experimental section. Figure 3ashows the complex chromatogram obtained using the 5-µm250-mm column. The cut-away shows the small amount ofGA that was present in the extract. Since GA is a hydroly-sis product of G, it should be at a much lower concentra-tion in a licorice extract, unless the extract was treated toenhance the concentration of the hydrolyzed product.Based on the area count, the GA concentration was lessthan 0.5% of the G concentration in extract A.

Although the actual peaks were not counted, the calculat-ed peak capacity (3) for the 5-µm column was determinedto be 290 (resolution: 1.0). Running the same extract A onthe 1.8-µm 150-mm column, one can see the finer struc-tured (that is, higher resolution) chromatogram that results(Figure 3b). The calculated peak capacity for this higher

efficiency column was determined to be 442, over 50%higher than by using the longer 5-µm column. Thus, itwould be easier to determine minor components on thisshorter rapid resolution column. The peaks per unit time(Resolution = 1.0) was calculated to be 17.7 peaks/min forthe 1.8-µm column versus 7.1 for the 5-µm column.

Figures 4a and 4b show similar runs using extract B. This extract was even more complex thanextract A which is borne out by comparing the high resolu-tion chromatograms of Figure 3b versus Figure 4b. Again,the calculated peaks per minute for the 1.8-µm columngreatly exceeded that of the 5-µm column (17 versus 7.5respectively). Based on the peak area counts for GA, it wasroughly 1% of the concentration of G in extract B.

5989-4907EN15

min0 10 20 30 40

Peak capacity (@ Rs=1.0) = 290

GA

GA 41.4 min

G 22.8 min

mAU

0

100

200

300

400

min25 30 35 40 45 50

mAU

-20

-10

0

10

20

30

40

50

Analysis of licorice root extract A using a ZORBAX

SB-C18, 4.6 × 250 mm, 5-µm column

50

min16 18 20 22 24 26 28 30 32 34

mAU

0

10

20

30

40

50

60

70

15.

896

min0 5 10 15 20 25 30

mAU

0

50

100

150

200

250

300

350

400

450


G 14.6 min

GA 25.5 min

GA

Analysis of licorice root extract A using a ZORBAX SB-C18 RRHT 4.6 × 150 mm, 1.8-µm column For conditions: see Experimental

Figure 3a and 3bThe gradient reversed-phase separation of licorice root extract A on the 5.0-µm column (A) and on the 1.8-µm column (B).

(A)

(B)

165989-4907EN

Peak capacity (@ Rs 1.0) = 306

GA 41.40 min

G 22.8 min

GA

min25 30 35 40 45 50

mAU

0

10

20

30

40

50

60

70 2

2.01

6

min05040302010

mAU

0

100

200

300

400

Analysis of licorice rootextract B using a ZORBAXSB-C18 4.6 × 250-mm column,5 µm

min0 5 10 15 20 25 30

G 14.6 min

GA 25.5 min

GA

mAU

0

50

100

150

200

250

300

350

400

450


min16 18 20 22 24 26 28 30 32 34

mAU

20

30

40

50

60

70

80

90

Analysis of licorice root extract B using a ZORBAX SB-C18 RRHT 4.6 × 150-mm column

Figures 4a and 4bGradient reversed-phase separation of licorice root extract B on the 5.0-µm column (A) and on the 1.8-µm column (B).

(A)

(B)

5989-4907EN17

Conclusions

No attempt was made to perform quantitative analysis onthe components of the licorice extracts. From our studies,it was obvious that resolution and throughput using the1.8-µm column greatly exceeded that obtained using the5.0-µm conventional column. As more complex samples ofnatural products are encountered and researchers requiremore detailed component analyses, the use of high resolu-tion, small particle columns should grow. In the investiga-tion of licorice root, other natural products, and TCMs, it isnecessary to have gradient capability and sensitive detec-tion.

References1. S. Fanali, Z. Aturki, G. D’Orazio, M. A. Raggi,

M. G. Quaglia, C. Sabbioni, and A. Rocco, (2005) J. Sep.Sci., 28, 982–986.

2. I. Kitagawa, (2002) Pure Appl. Chem. 74 (7), 1189–1198.


For more information on our products and services, visit ourWeb site at www.agilent.com/chem.

185989-4907EN

Fast, high-resolution analysis of notoginseng by rapid resolution liquid chromatography

Abstract

This Application Note describes:

• The development of rapid resolution liquid chromatography (RRLC) method for

the analysis of notoginseng.

• The results of method transfer from conventional HPLC to RRLC.

• The use of the RRLC method to shorten run times while maintaining good

resolution of complex components, thereby increasing sample throughput and

lowering costs.

• The chromatograms obtained with optimized methods for the different parts of

the notoginseng plant, which show different peak profiles and different concen-

trations of certain saponins.

Time [min]0.5 1.0 1.5 2.0 2.5 3.0

Abs

orba

nce

[mA

U]

0

200

400

600

800

1000

Zhi-xiu XuApplication Note

Agilent Equipment:

1200 Series Rapid Resolution LC system ZORBAX XDB C18 RRHT column

Application Area:

Traditional Chinese Medicine

5989-6757EN19

Introduction

Traditional Chinese Medicines (TCMs)have a long history of use in China andtheir therapeutic effects are wellknown in China and other Asia coun-tries such as Korea and Japan. InWestern countries the use of TCMs asfood supplements or nutriments isbecoming more and more popular.More than 11,000 kinds of TCMs havebeen used over time. Research andquality control of TCMs rely heavily oninstrumental separations and the per-formance of these separations.

In this study a rapid resolution liquidchromatography (RRLC) method forthe analysis of notoginseng wasdeveloped. The conventional HPLCanalysis method was transferred easily to RRLC using Agilent’s methodtranslator. Different extraction meth-ods were used to produce differentsamples from different parts of thenotoginseng plant. These sampleswere separated with optimized meth-ods and the resulting chromatogramsshowed different peak profiles. ThisApplication Note also shows thequantitative results. Using the fasterand better methods developed in thisstudy, quality control departments willbe able to reduce analysis times andincrease sample throughput. Usingthese methods also reduced the costof solvent as well as improved theoverall analysis process.

Notoginseng, which is also known asin Chinese as Sanqi or Tianqi, is animportant and highly valued traditionalmedicine in China. It has been cultivat-ed for about 400 years and more than85 % of the notoginseng productionoriginates from the Yunnan Province,China. Notoginseng is known for itsefficiency in promoting blood circula-

tion, removing blood stasis, inducingblood clotting, relieving swelling andalleviating pain. Current pharmacologi-cal studies revealed that Panax noto-ginseng possesses anticarcinogenicand hepatoprotective properties, aswell as protective effects on cardio-vascular and cerebrovascularsystems1. Compared with the wellknown Panax ginseng and Panax quin-quefolium (American Ginseng), the pro-file of the saponins in Panax notogin-seng is similar because they belong tothe same genus. Notoginseng is usedmostly in south China in different tradi-tional Chinese formulations. Thefamous traditional Chinese formula-tions Yun Nan Bai Yao and Pien TzeHuang also contain Panax notogin-seng.

The complex matrixes of TCMs alwayspresent major challenges for qualitycontrol and research. The 2005 editionof the China Pharma-copeia listsMateria Medica Sanqi and Chinesepatent medicines such as SanqiShangyao Pian as containing notogin-seng saponin R1, which must be ana-lyzed in these formulations. The typicalrun time for the analysis of notogin-seng saponin is longer than 60 min-utes. With more and more research onTCMs, scientists realized that controlof only one or two components is notenough to determine the quality of aTCM and that it is necessary to devel-op other quality control methods thatseparate more components.Separating complex samples such asTCMs requires longer analysis timesthan synthetic drugs. When samplethroughput is also an issue, havingsimilar or better performance withshorter analysis times would be anideal solution for this situation. WithRRLC it is possible to develop methodswith shorter run times and with betterperformance. RRLC methods also give

users the benefits of time and solventsavings. For manufacturing qualitycontrol, using RRLC means batches ofproducts can be released faster com-pared with using conventional HPLCmethods

Experimental

EquipmentFor development of the RRLC methodan Agilent 1200 Series RapidResolution LC system with the following modules was used:

Agilent 1200 Series RRLC system consisting of:• Agilent 1200 Series binary pump SL

with vacuum degasser• Agilent 1200 Series high-perfor-

mance autosampler SL• Agilent 1200 Series thermostatted

column compartment SL• Agilent 1200 Series diode array

detector SL with micro flow cell (2 µL volume, 3 mm path length)

• Agilent ChemStation B.02.01 SR1 fordata acquisition and evaluation

• Agilent ZORBAX XDB C18 RRHT column, 4.6 x 50 mm, 1.8 µm particle size

For comparison with conventionalHPLC a standard configuration of an Agilent 1100 Series LC with the following modules was used:• Agilent 1100 Series binary pump with

degasser• Agilent 1100 Series autosampler • Agilent 1100 Series thermostatted

column compartment• Agilent 1100 Series diode array

detector• Agilent ZORBAX SB C18

column, 4.6 x 50 mm, 5 µm particlesize

205989-6757EN

Samples and sample preparation• Notoginseng saponin R1, ginsenoside

Rg1 and ginsenoside Rb1 standardswere purchased from the NationalInstitute for Control ofPharmaceutical and BiologicalProducts (NICPBP), China.Gensenoside Re and ginsenoside Rcstandards were purchased fromSigma-Aldrich, USA.

• Notoginseng caudexes and leafextracts (NCLE) were kindly providedby a customer’s laboratory. Thecaudexes and leaves were extractedwith 60% ethanol, the solvent wasevaporated to dryness and theresidue dissolved in n-butanol, thesolvent was then evaporated to dry-ness again and the residual yellowpowder was dissolved in methanoland used for injection.

• Notoginseng root extracts (NRE)were kindly provided by a customer’slaboratory. The raw componentswere extracted with water, filteredthrough a marcoporous membraneand the filtrate evaporated to dry-ness. The residual white powder wasdissolved in methanol and used forinjection.

• Notoginseng was purchased from alocal drug store. The brand was TongRen Tang. The raw material wasextracted with water/methanol(30/70, v/v), treated ultrasonically for30 minutes, filtered through a 0.22 µmmembrane and the clear filtrate wasused for injection.

• Water, acetonitrile and methanolwere purchased from Fisher, USA.

Method translationThe conventional HPLC method usingthe column packed with 5 µm particlesneeded to be transferred to an RRLCmethod using a column packed with1.8 µm particles. Agilent’s methodtranslator (available on CD for Agilent1200 Series Rapid Resolution LC sys-tem, publication number 5989-5130EN)was used for this purpose and made

the process easy and quick (figure 1).Depending on individual requirementsit might be necessary to make furthermodifications to the transferredmethod based on the different factorslisted by the method translator. Fordetails of how to translate manually aconventional HPLC method to an RRLCmethod, further Application Notes areavailable from Agilent2,3,4.

Results and discussion

ChromatogramsAccording to results of pharmacologi-cal studies, different parts of the noto-ginseng plant have different therapeu-tic effects as a result of different com-ponents being present. Specific extrac-tion methods were used with theobjective of obtaining more effectivecomponents. The notoginseng sam-

Figure 1Agilent’s method translator for transfer of standard HPLC methods to RRLC technology.

Figure 2Chromatogram of NCLE analyzed using a conventional HPLC method on an Agilent 1100 Series LCsystem. See table 1 for chromatographic conditions.

Time [min]0 10 20 30 40 50 60

Abs

orba

nce

[mA

U]

0

20

40

60

80

100

120

5989-6757EN21

Figure 4Chromatogram of NCLE, analyzed using RRLC method 2 on an Agilent 1200 Series RRLC system.See table 1 for chromatographic conditions.

Time [min]2 4 6 8 10 12

Abs

orba

nce

[mA

U]

0

20

40

60

80

100

ples were analyzed with conventionalHPLC methods and RRLC methods toreveal different peak profiles. Figure 2shows the chromatogram of notogin-seng caudexes and leaf extracts(NCLE) analyzed using the convention-al HPLC method. Typically, this type ofseparation requires more than 60 min-utes5. As described in the experimen-tal section, the RRLC method used foranalyzing the notoginseng sampleswas developed based on the conven-tional HPLC method. The same NCLEsample was analyzed with the conven-tional HPLC method and the RRLCmethod 1. The chromatograms areshown in figures 2 and 3. The peaks infigure 3 were narrower than those infigure 2. This means that the peakcapacity increased for the complexsamples. The resolution or separationperformance of the RRLC analysis wasbetter than conventional HPLC. At thesame time the run time was reduceddramatically from 60 to 20 minutes.Another factor that should be consid-ered is the selectivity of the columnsfor conventional HPLC and RRLC,which should be kept the same. If thenew RRLC column has totally differentselectivity properties than the conven-tional HPLC column, transferring themethod makes no sense. In this studypeak profiles and elution sequence

Figure 3Chromatogram of NCLE, analyzed using RRLC method 1 on an Agilent 1200 Series RRLC system.See table 1 for chromatographic conditions.

Time [min]2.5 5 7.5 10 12.5 15 17.5

Abs

orba

nce

[mA

U]

-20

0

20

40

60

80

100

120

20

Chromatographic conditions fortraditional method with Agilent1100 Series LC:

Column: Agilent ZORBAX SB-C18, 4.6 x 250 mm,5 µm particle size

Mobile phase: A: Water; B: ACN

Gradient: 0 min, 20 %B; 10 min,30 %B; 20 min, 35 %B;30 min 40 %B; 40 min,60 %B; 50 min, 100 %B

Flow rate: 1.0 mL/min

Injection volume: 5 µLDiode array detection: 203 nm ±8 nm,

Ref. 360 nm ±100 nm

Chromatographic conditions forRRLC method 1:

Column: Agilent ZORBAX XDB C18, 4.6 x 50 mm,1.8 µm particle size


Gradient: 0 min, 20 %B; 3 min,30 %B; 7 min, 35 %B;10 min 40 %B; 13 min,60 %B; 17 min, 100 %B


Injection volume: 5 µL

Diode array detection: 203 nm ±8 nm,

Ref. 360 nm ±100 nm




Gradient: 0 min, 20 %B; 1.5 min,30 %B; 5 min, 35 %B;7 min 40 %B; 10 min,60 %B; 12 min, 100 %B




Ref. 360 nm ±100 nm




Gradient: 0 min, 30 %B; 2.5 min,60 %B; 5 min, 100 %B




Ref. 360 nm ±100 nm

Table 1Chromatographic conditions.

225989-6757EN

were the same so that it was not nec-essary to identify the peaks again inthe new chromatogram. Further, theUV spectra from the diode array detec-tor helped to do further confirmation.

Figure 4 shows the chromatogram ofthe NCLE sample analyzed by RRLCmethod 2. Increasing the flow rate to1.5 mL/min reduced the run time. Atthe same time it was possible to main-tain the same or comparable perfor-mance as the results shown in figure3. For more complex samples, shorterrun times mean less peak capacity.Hence users should choose a balancebetween peak capacity and run time.

As mentioned in the experimental sec-tion, three kinds of samples wereobtained from different origins. Figure5 shows the chromatogram of thenotoginseng root extraction (NRE)sample analyzed with the RRLCmethod 1. This sample was extractedfrom the notoginseng root using a dif-ferent procedure. The peak profilesshow that NRE and NCLE contain different kinds of saponins and othercomponents.To test if the developedRRLC method could be used with othernotoginseng samples, we purchasednotoginseng from a local TCM store.The brand was Tong Ren Tang and thenotoginseng had already been groundto powder at TCM store. The chro-matograms of the notoginseng analy-ses are shown in figures 6 and 7. RRLC method 1 was used to obtain the chromatogram in figure 6 andRRLC method 3 was used in figure 7.Compared with the conventional HPLCmethod, which needs more than 60minutes, the chromatogram in figure 7demonstrates that the analysis can bedone in 4 minutes. Notoginsengcaudexes and leaf extracts are morecomplex than the other samples ana-lyzed and RRLC method 1 or 2 should

be used to obtain analysis times whilemaintaining resolution. For less com-plex samples such as Tong Ren Tangnotoginseng extractions, the fastermethod can be used to complete theanalysis in a shorter time as shown in figure 7.

Figure 5Chromatogram of NRE, analyzed using RRLC method 1 on an Agilent 1200 Series RRLC system.See table 1 for chromatographic conditions.

Time [min]2.5 5 7.5 10 12.5 15 17.5

Abs

orba

nce

[mA

U]

0

100

200

300

400

500

20

Figure 6Chromatogram of notoginseng from Tong Ren Tang, analyzed using RRLC method 1 on an Agilent1200 Series RRLC system. See table 1 for chromatographic conditions.

Time [min]

2.5 5 7.5 10 12.5 15 17.5

Abs

orba

nce

[mA

U]

0

200

400

600

800

1000

1200

1400

1500

20

5989-6757EN23

References

1.Wei J.X., Du Y.C., “Modern scienceresearch and application of panaxnotoginseng”, Yunnan Science andTechnology Press, Kunming, China, 1996.

2.Angelika Gratzfeld-Huesgen, “Fast andUltra-fast Analysis with the Agilent1200 Series Rapid Resolution LCSystem Compared to a ConventionalAgilent 1100 Series LC System UsingSub 2-µm Particle Columns”, AgilentApplication Note, Publication Number5989-5672EN, 2006.

3.Michael Woodman, “Improving theEffectiveness of Method Translationfor Fast and High ResolutionSeparations”, Agilent ApplicationNote, Publication Number 5989-5177EN, 2006.

4.Michael Frank, “Saving analysis timeand gaining resolution by simplemeans”, Agilent Application Note,Publication Number 5989-6819EN,2007.

5.Lau A.J., Woo S.O., Koh H.L., “Analysisof saponins in raw and steamed Panaxnotoginseng using high-performanceliquid chromatography with diode arraydetection”, J. Chromatogr. A, 1011, 77-87, 2003.

Table 2Quantitative results of the saponin standards.

Standard Formula Correlation Concentration Concentration ConcentrationY: peak area in NCLE in NRE in Tong Ren Tangx: amount (µg/µL) (µg/µL) notoginseng (0.1 µg/µL) powder (µg/µL)

R1 Y = 54.03850x + 0.395722 0.99995 0.029 0.48 0.89Rb1 Y = 71.64751x – 0.159297 1.0000 0.18 4.64 2.136Re Y = 93.85331x + 3.83488 0.99982 N/A 1.28 N/ARg1 Y = 151.66579x + 0.501417 1.0000 0.07 1.52 1.54Rc Y = 11.12002x + 0.128832 0.99980 N/A 0.025 N/A

Quantitative resultsThe standards listed in the experimen-tal section were analyzed using RRLCmethod 2. Using different levels ofconcentrations of the standards, a cor-relation curve was created and used todetermine the concentrations of thethree samples. Table 2 shows thequantitative results. NCLE containedmany kinds of saponins but did notcontain Rc or Re. NRE contained fewerkinds of saponins but did contain thefive standards analyzed in this study.The Tong Ren Tang notoginsengextraction did not contain many kindsof saponins and did not contain Rc orRe. The concentrations of R1, Rb1 andRg1 were higher than in NCLE andNRE. The concentration of notogin-seng R1 ranged from 0.029 to 0.89µg/µL across all samples. The quanti-tative method developed in this studycan be deployed easily in the qualitycontrol of notoginseng with differentextraction methods.

Conclusion

Traditional Chinese Medicines arecomplex natural products and theiranalysis by conventional HPLCrequires a high performance systemand long run times. The RRLC methodsdeveloped in this study demonstratedhow to shorten run times while main-taining good resolution of complexcomponents. The chromatogramsshow the different peak profiles corre-sponding to the different notoginsengextraction samples. Taking the require-ments of the application into consider-ation, RRLC with a faster flow rate canbe deployed to achieve a faster analy-sis method. The quantitative methodsdeveloped in this study can be imple-mented easily as a quality controlmethod for notoginseng products.Using RRLC methods maintain or giveeven better performance while at thesame time reducing the run time dra-matically, thereby increasing samplethroughput and saving costs.

Figure 7Chromatogram of notoginseng from Tong Ten Tang, analyzed using RRLC method 3 on an Agilent1200 Series RRLC system. See table 1 for chromatographic conditions.

Time [min]0.5 1.0 1.5 2.0 2.5 3.0

Abs

orba

nce

[mA

U]

0

200

400

600

800

1000

Zhi-xiu Xu is an Application Chemist atAgilent Technologies, Shanghai, China.

245989-6757EN

Abstract

This Application Note describes the development of reliable quality control

methods for analysis of individual traditional Chinese medicines (TCMs) and

TCM formulations using rapid resolution liquid chromatography with UV-visible

and quadrupole mass spectrometry detection. The chromatograms obtained

with UV and MS detection for different individual TCMs are compared and the

combination of UV and MS spectra is used to qualify target compounds.

Zhixiu XuApplication Note

Agilent Equipment:

1200 Series Rapid Resolution LC system6140 Quadrupole MS

Application Area:


Development of reliable quality controlmethods for TCM preparations using rapidresolution LC with UV and MS detection

5989-7682EN25

Introduction

China has a long history of usingTCMs and TCM preparations and hasvast experience of the therapeuticeffects. However, quality control isbased on appearance only. With thedevelopment of new technologies, itwas recognized that TCMs containhundreds of compounds and that theconcentration levels are wide ranging.Further, differences in compositionresult from herbal TCMs being grownin different regional areas and harvest-ed at different seasonal times as wellas from preparation and manufactur-ing processes. As a result, controllingthe quality of the compounds in TCMshas become a tremendous challenge.Today, tracking only one or two com-ponents in TCM preparations is notenough to control the quality becauseTCM preparations with specific formu-las are often made from more thanone TCM. For example, Qishenyiqidripping pills, a traditional patentChinese medicine used to treat coro-nary diseases, include four kinds ofTCMs: • Astragali (huangqi)• Salviae miltiorrhizae (danshen)• Notoginseng (sanqi)• Lignum dalbergine ordoriferae

(jiangxiang)

Qishenyiqi dripping pills were used asthe sample in this study, in which anew method for the quality control ofTCMs was developed using usingrapid resolution liquid chromatography(RRLC) with UV-visible and MS detec-tion. The main objective was to devel-op a reliable method to determine thepotential target components that donot undergo change during heating ormixing.

Experimental

Equipment• Agilent 1200 Series Rapid Resolution

LC system comprising binary pumpSL with vacuum degasser, high per-formance autosampler SL, ther-mostatted column compartment SL,and diode array detector SL withmicro flow cell (2 µL volume, 3 mm path length)

• Agilent 6140 quadrupole MS withESI source

• Agilent ChemStation B.02.01 SR1 fordata acquisition and evaluation

• Agilent ZORBAX SB C18 RRHT col-umn, 3.0 x 50 mm, 1.8 µm particlesize

StandardsSalvianolic acid B, tanshinone I, tan-shinone IIA, cryptotansshinone, dan-shensu, notoginsenoside R1 and astra-galoside were purchased fromNational Institute of ChemicalPharmaceutical and BiologicalProducts (NICPBP), China.Ginsenoside Rg1, ginsenoside Rb1,3,4-dihydroxybenaldehyde, 3,4-dihy-droxybenzoic acid, ginsenoside Re andginsenoside Rc were purchased fromSigma-Aldrich, USA.

SolventsAcetonitrile was purchased fromFisher, USA. Water was prepared witha Milli-Q pure water system.

Samples and sample preparationQishenyiqi dripping pills, intermediateextractions of huangqi, danshen, sanqiand essential oil of jiangxiang werekindly provided by the TASLY Pharma-ceutical Company, China. The rawTCMs were purchased from a localTCM store.

The dripping pills, TCM extractions andraw TCMs were prepared by dissolvingin a 70 % methanol/water solution,treating ultrasonically for 30 minutes,and filtering through a 0.22 µm mem-brane.

RRLC method• Mobile phase:

A: Water with 0.1 % formic acid B:ACN with 0.1 % formic acid

• Gradient: 0 min, 10 %B;8 min, 38 %B; 12 min, 100 %B;hold for 3 min, then 10 %B

• Flow rate: 1.0 mL/min (passive splitter reduces flow to MSto about 0.4 mL/min)

• Column temperature: 45 °C• Detection wavelength: 203 nm• Peak width: 0.5 s• Slit width: 4 nm• Spectra: 190-400 nm, step 2 nm

MS method• Scan: 80-1400 (pos/neg)• Fragmentor: 70 (pos/neg)• Drying gas: 12 L/min• Nebulizer pressure: 50 psi• Drying gas temperature: 50 °C• Cap. Voltage: 3200 V (pos/neg)


The Agilent method translation tool(available on CD pub. no. 5989-5130EN) was used to convert the tradi-tional LC method to a rapid resolutionmethod.

Comparison of danshen sanqi,huangqi and jiangxaing extractsFigure 1 shows that the chro-matograms of danshen and sanqi werevery different from the combinedextracts when detected at 203 nm.After mixing and heating, some peaksdisappeared and new peaks emerged.

265989-7682EN

This means that target compoundscreening during quality control is animportant step to make sure the com-pounds do not undergo changes duringmanufacturing. Most researchers usedetection at 203 nm because the mainactive compounds are saponins, whichhave weak absorbance at 203 nm. Butlarge peaks appearing at 203 nm mightnot be relevant for quality control andsmall peaks might be important for theresearch of active components. As aconsequence, a complementary detec-tor with higher sensitivity is needed toprovide more information about thecomponents.

Comparison between different detectors and conditionsFigure 2 shows the differences in theresults from UV and MS detection.More peaks appeared in the MS totalion chromatogram (TIC) because somecomponents had weak or no UVabsorbance. Using MS detection addedan extra dimension and gave moreinformation about the structures of thedifferent compounds. The negativepolarity mode gave more completeinformation about the peaks andenabled the target compounds to beidentified for quality control.

Comparison of qishenyiqi drippingpills and different extractsFigure 3 shows chromatograms of danshen sanqi, huangqi and jiangxiangextracts, as well as the qishenyiqi drip-ping pills, which are made from thesethree extracts and other additives. Thenumbered peaks are the target com-pounds that were screened for qualitycontrol of this TCM preparation. Basedon the UV and MS data, differentpeaks were studied to determinewhether the compounds had under-gone changes. Table 1 shows theresults of the qualifying process.

Figure 1Chromatograms of individual TCMs danshen and sanqi, and of the combined extracts.

Figure 2Chromatograms of danshen sanqi combinations, showing the different results obtained by UV and MS detection.

Danshensample

Danshenand sanqi

extracts

Sanqi sample

Figure 3Comparison of the negative TICs for the three TCMs combinations and the dripping pills sample.

m in2 4 6 8 1 0 12

N o rm .

0

5 00 0 0

1 0 0 0 0 0

1 5 0 0 0 0

2 0 0 0 0 0

2 5 0 0 0 0

3 0 0 0 0 0

m in2 4 6 8 1 0 12

N o rm .

0

5 00 0 0

1 0 0 0 0 0

1 5 0 0 0 0

2 0 0 0 0 0

2 5 0 0 0 0

3 0 0 0 0 0

m in2 4 6 8 1 0 12

N o rm .

0

5 00 0 0

1 0 0 0 0 0

1 5 0 0 0 0

2 0 0 0 0 0

2 5 0 0 0 0

3 0 0 0 0 0

m in2 4 6 8 1 0 12

N o rm .

0

5 00 0 0

1 0 0 0 0 0

1 5 0 0 0 0

2 0 0 0 0 0

2 5 0 0 0 0

3 0 0 0 0 01

2 3

4

5

7

6

5989-7682EN27

Zhixiu Xu is an Application Chemist atAgilent Technologies, Shanghai, China.

Conclusion

The RRLC method with UV and MSdetection described in this ApplicationNote is more reliable than the currentpharmacopeia quality control methodbecause the quality of several compo-nents in the TCM can be trackedbased on the information provided bythe UV and MS detectors. Trackingthese components is importantbecause the components could under-go changes during manufacture andpreparation. The Agilent 6140 quadru-pole MS used for the mass spectrome-try measurements is easy to use andintegrates seamlessly with the Agilent1200 Series RRLC system and AgilentChemStation sofware.

References

1.Ai-hua Liu, et.al., Journal ofChromatography B, 846, 32-41, 2007.

2.Jin-huai Liu, et.al., Journal of ChinesePharmaceutical Science, 13 (4), 2004.

3.Chinese Journal of AnalyticalChemistry, 1676-1680, 2005.

4.Agilent Application Notes, publicationnumbers 5989-4506EN, 5989-5493EN,5989-6757EN

Peak Compound MS UV lq m/(nm)

1 Danshensu 197 [M–H]–, 395 [2M–H]– 280

2 Procatechualdehyde 137 [M–H]–, 275 [2M–H]– 225, 280, 310

3 Salvianolic acid B 717 [M–H]– 260, 280

4 Ginsenoside Rg1 845 [M+HCOO]– 210

5 Ginsenoside Rb1 11071[M–H]– 210

6 Astragaloside IV 829 [M+HCOO]– 210

7 Tanshinone I 295137 [M+H2O]– 230

Table 1Target compounds with structural formulas and detection details (numbers refer to the peaks infigure 3).

285989-7682EN

Abstract

The Agilent 1120 Compact LC is the system of choice for conventional, analytical

scale liquid chromatography. It is an intergrated LC designed for ease of use, per-

formance and reliability. It is ideally suited for the routine analysis of Traditional

Chinese Medicines (TCMs) on account of its capability to achieve highly precise

retention times and peak areas, and low detection limits for the analyzed com-

pounds. This Application Note shows the chromatograms obtained with optimizd

methods for the most well-known TCMs ginseng and American ginseng, which

show different peak profiles and different concentrations of certain saponins.


Agilent Equipment:

1120 Compact LCEZChrom Elite Compact softwareHC-C18(2) column

Application Area:


Analysis of ginseng and American ginsengusing the Agilent 1120 Compact LC

Rg1

Re

Rb1

5989-7458EN29

Introduction

Traditional Chinese Medicines (TCMs)have a long history of use and theirtherapeutic effects are well known inChina and other countries. Ginseng,perhaps the most well-known TCM,has long been used as a tonic, anti-fatigue, sedative and anti-gastric ulcerdrug. It is also widely used in differentTCM preparations. Another well-known TCM, American ginseng, hassimilar therapeutic effects as ginsengbut there are also some differencesbecause of the different saponin contents.

According to the method in the phar-macopeia of the People's Republic ofChina1, ginseng must be analyzed byHPLC to determine the ginsenosidesRg1, Re and Rb1. Similar requirementsexist for the determination of theseginsenosides in American ginseng.These requirements make the determi-nation of the ginsenosides Rg1, Reand Rb1 in ginseng and American ginseng important for quality controlof TCM raw materials and final prepa-rations.

In this study an HPLC analysis methodwas developed using the Agilent 1120Compact LC for the determiniation ofginsenosides in ginseng and Americanginseng.

Experimental

Equipment• Agilent 1120 Compact LC comprising

gradient pump with integrateddegasser, autosampler with vial tray,column oven and variable wave-length detector, see figure 1

• Agilent HC-C18(2), high carbon load,150 x 4.6 mm, 5 µm particle size col-umn

• Agilent EZChrom Elite Compact soft-ware

Samples and sample preparationGinseng and American ginseng werepurchased from a local TCM store andsamples prepared as follows. 1 g ofpowder was weighed and dissolved in50 mL of water saturated n-butanol.The solution was treated ultrasonicallyfor 30 minutes and centrifuged for 5minutes at 300 rpm. The solvent wasevaporated and the residue dissolvedin 5 mL methanol. The final solutionwas filtered through a 0.20 µm mem-brane before injection.

Chromatographic conditions• Mobile phase:

A: Water, B: ACN • Gradient: 0 min, 19 %B;

35 min, 19 %B; 55 min, 29 %B;70 min, 29 %B; 100 min, 40 %B

• Flow rate: 1.0 mL/min• Injection volume: 10 µL• Column temperature: 40 °C• Detection wavelength: 203 nm


The chromatogram of the ginseng separation is shown in figure 2. The method used to obtain this chromatogram was the same as themethod specified in the pharmacopeiaof People’s Republic of China. Thechromatogram shows excellent sepa-ration of all the target compounds. The ginsenosides Rg1 and Re are well separated, demonstrating that the Agilent 1120 Compact LC is wellsuited for this analysis.

The chromatogram of the Americanginseng separation is shown in figure3. The chromatogram shows excellentseparation of the target ginsenosides.

From the chromatograms of ginsengand American ginseng, it can be seenthat the profiles of the two samples aswell as the concentrations of the gin-senosides are different. For complexTCM samples such as ginseng andAmerican ginseng, the Agilent 1120Compact LC is a reliable tool to obtaingood results in routine analysis work.

Figure 1Agilent 1120 Compact LC

305989-7458EN

[ "?9+;s)"?

Although Traditional ChineseMedicines are complex natural prod-ucts, this study demonstrated that theAgilent 1120 Compact LC was capableof analyzing the active componentsand achieving excellent separation performance. The results proved thatthe Agilent 1120 Compact LC is idealfor routine quality control testing ofcomplex TCM samples.

!o1o®o?9o

1.Pharmacopeia of the People's Republicof China, Volume I, 2005. Figure 2

Chromatogram of ginseng analyzed with the CHP method on an Agilent 1120 Compact LC system.

10 20 30 40 50 60 70 80 90 100

0

10

20

30

40

50

60

70

Rg1 Re

Rb1

Time [min]

Abs

orba

nce

[mA

U]

Figure 3Chromatogram of American ginseng analyzed on an Agilent 1120 Compact LC system.

Rg1

Re

Rb1

Time [min]

0 20 40 60 80 100 120

Abs

orba

nce

[mA

U]

0

50

100

150

200

250

300


5989-7458EN31

325989-7458EN

AbstractTraditional Chinese medicines often contain components that must be detected

by HPLC, but that lack a chromophore and so do not produce signals with an ultra-

violet (UV) detector. The Agilent 1200 Series evaporative light scattering detector

(ELSD) is an excellent alternative because it detects all solutes that are less

volatile than the mobile phase. In the Chinese Pharmacopoeia 2005, several TCMs

require the ELSD as the detection method, and this Application Note illustrates

two examples – the flavonoids in Ginkgo biloba L. and astragaloside in Astragali.


Analysis of traditional Chinese medicines with the Agilent 1200 Seriesevaporative light scattering detector

Agilent Equipment

• Agilent 1200 Series Rapid Resolution LCsystem

• Agilent 1200 Series evaporative light scattering detector

Application area

• Traditional Chinese medicine (TCM)

min0 5 10 15 20 25 30 35

mV

020406080

100120

1

2

34

5989-8922EN33

Nebulization

Evaporation

Detection

Mobilephase

Gas

Photomultiplier

IntroductionThe evaporative light scattering detec-tor is increasingly being used as aquasi-universal detector for non-UV-absorbing analytes in HPLC systems.Especially in herbal medicines, anincreasing number of non-UV-absorb-ing compounds need to be detectedwithout derivatization. The Agilent1200 Series HPLC family now has anew member – the evaporative lightscattering detector – to help with suchapplications.

The Agilent 1200 Series ELSD candetect all solutes that are less volatilethan the mobile phase. It produces asignal only for the nonvolatile particlesthat are generated from the sample. If the compounds have no chro-mophores and are less volatile thanthe LC solvents, the ELSD is a gooddetector that can provide both ease of use and good sensitivity. Gradientmobile phases do not interfere withdetection. When a mobile phase pro-duces strong absorption under a spe-cific UV wavelength, the ELSD is anexcellent alternative that maintains astable baseline and produces strongsignals.

The ELSD principle of operation con-sists of three main successiveprocesses: a) Nebulization of the chromatograph-

ic eluent using nitrogen or air,b) Evaporation of mobile phase at

relatively low temperature, and c) Light scattering by the residual

particles, which ideally consist ofanalyte molecules.

Figure 1 shows a schematic of the various stages of detection.

Some components of TCMs have nochromophores, so they need a univer-sal detector for detection. Comparedwith other general detectors, theAgilent 1200 Series ELSD provides sig-nificant benefits. Mass spectrometry(MS) is more expensive and requireswell-trained, knowledgeable operators.MS is not normally used for the rou-tine analytical work that is done inquality control departments. Therefractive index detector provides uni-versal detection, but can be used onlyfor isocratic analyses, so is not fit forseparation of complex mixtures. A sys-tem with a UV detector that is con-nected in series with an ELSD can beused for simultaneous determinationof multiple components with variousstructures, with or without chro-mophores.

In the Chinese Pharmacopoeia 2005,there are several TCMs that require

ELSD as the detection method for cer-tain components. Examples includethe flavonoids in Ginkgo and astragalo-side in Astragali.

Ginkgo biloba L. and its extracts arenot only very famous in China, but alsoare used worldwide to treat cardiovas-cular and cerebrovascular diseases.The therapeutic effects are due to theginkgolides and bilobalide that are present together with the flavonoids.Quality control methods must deter-mine the amount of the flavonoids inGinkgo and its products.

Radix Astragali, in Chinese Huangqi, isone of the most widely used TCMs inChina. Pharmaceutical studies andclinical practice have demonstratedthat Radix Astragali possesses manybiological functions; therefore, it isused for the treatment of nephritis,diabetes, hypertension, and other dis-

Figure 1Cross-sectional view of the Agilent 1200 Series evaporative light scattering detector.

345989-8922EN

eases. One of the effective compo-nents, astragaloside, has no chro-mophores and needs a universaldetector to determine the amount ofthe compound.

In this Application Note, the separa-tion and the quantitative analysis ofthe Ginkgo flavonoids and astragalo-side are studied. The note also pro-vides some general advice on use ofthe Agilent 1200 Series ELSD.

Experimental

EquipmentFor development of the RapidResolution LC (RRLC) method, anAgilent 1200 Series RRLC system withthe following modules was used:• Agilent 1200 Series binary pump SL

with vacuum degasser• Agilent 1200 Series high-performance

autosampler SL• Agilent 1200 Series thermostatted

column compartment SL• Agilent 1200 Series diode array

detector SL with micro flow cell (2 µLvolume, 3 mm path length)

• Agilent 1200 Series evaporative lightscattering detector with standardnebulizer

• Agilent ChemStation B.03.02 for dataacquisition and evaluation

• Agilent ZORBAX XDB-C18 RapidResolution High Throughput (RRHT)column, 3.0 x 50 mm, 1.8 µm particlesize

System setup The HPLC modules are plumbed in theusual ways, as shown in figure 2. TheAgilent 1200 Series ELSD is connectedwith the computer through an RS-232cable. The remote cable connects theAgilent 1200 Series ELSD with anymodule of the HPLC.

System checkoutBefore starting an experiment, a testrun should be done to make sure thatthe Agilent 1200 Series ELSD is ingood condition and has sufficient per-formance.

Test run conditions:• Sample: caffeine at 250 µg/mL

(or similar concentra-tion)

• Solvent: isocratic, 80 % water, 20 % acetonitrile

• Flow rate: 1 mL/min• Injection volume: 20 µL• Column: Agilent ZORBAX

Eclipse XDB-C18, 4.6 x 150 mm, 5 µm (or other column withcomparable size)

• Thermostatted column compartment(TCC) temperature: 40 °C

• ELSD temperature: 40 °C• ELSD pressure: 3.5 bar (51 psi)• ELSD gain: 7• ELSD filter: 1 second• Typical system: Agilent 1200 Series

standard LC system (Please note that

the Agilent 1200 Series RRLC systemcan be used as the standard systemwhen conventional HPLC is needed.)

After the method has been run, theresponse of the detector should berecorded when the instrument is ingood condition after the installation.Then one should use the same methodevery time the performance check isneeded.

Nonvolatile buffers cannot be usedwith the ELSD. If nonvolatile salt goesinto the ELSD, the detector cannot dis-tinguish whether the nonvolatile parti-cles came from the sample or from thebuffer. Therefore, the baseline increas-es dramatically and the signal-to-noiseis reduced for the sample.

Because ELSD can detect only com-pounds that are less volatile than themobile phase, when new methods aredeveloped, another detector should beused to make sure all the compoundscan be eluted from the column.

Figure 2Agilent 1200 Series RRLC system with Agilent 1200 Series evaporative light scattering detector.

LC ChemStationRS 232

remote cable

degasser

pump

auto-sampler

column com-partment

DAD

inlet

waste

5989-8922EN35

Standards and sample preparationThe following standards, samples, andsolvents were used for these experi-ments:• Standards were purchased from

NICPBP (National Institute for theControl of Pharmaceutical andBiological Products).

• Solvents (methanol, acetonitrile,tetrahydrofuran) were purchasedfrom Thermo Fisher Scientific.

• Water was obtained from a Milliporepure water system.

• The Astragali sample was purchasedfrom a TCM drug store.

• Ginkgo extracts were kindly providedby a customer.

Samples were prepared by weighing 1gram sample, dissolving it in 5 mLmethanol, mixing in an ultrasonic bathfor 30 minutes, then filtering and saving the clear liquid forinjection.

Method for Ginkgo analysis• Mobile phase: A = water,

B = tetrahydro-furan/methanol with a volume ratio of 10/25

• Flow rate: 0.7 mL/min• Gradient: 0 min, 12 %B;

10 min, 16 %B; 15 min, 22 %B; 20 min, 30 %B

• ELSD: temperature = 40 °C, pressure = 50 psi, gain = 7, filter = 3 seconds

Method for Astragali analysis• Mobile phase: A = water,

B = acetonitrile• Flow rate: 0.7 mL/min• Gradient: 0 min, 20 %B;

1 min, 30 %B; 5 min, 35 %B, 30 min, 100 %B

• ELSD: temperature = 40 °C, pressure = 50 psi, gain = 5, filter = 3 seconds

Results and Discussion

Analysis of GinkgoGinkgo and its products have a verycomplex composition and require agood separation by HPLC. Isocraticmethods were used frequently in thepast because of limitations in theinstrumentation. The disadvantages ofisocratic methods include long analy-sis time and lower resolution. A gradi-ent method was used in this applica-tion to get better separation in a short-er time.

Figure 3 shows chromatograms ofstandards of Ginkgo flavonoids, andcompares the signals on ELSD and

two channels of UV. The flavonoidshave very low response at a wave-length of 210 nm and no response at254 nm. Comparison of the sensitivitybetween UV and ELSD shows why it isnecessary to use ELSD to determinethe flavonoids in Ginkgo. By checkingcomplementary UV signals, thismethod can also be used to tell if thepeaks are pure flavonoids.

The chromatogram in figure 4 showsthat there are many peaks in theGinkgo sample that can be detected byELSD. The sample composition is verycomplex and the ELSD information canhelp to find more components, includ-ing those that lack chromophores. Thefour flavonoid peaks are well-separat-ed from each other and from the othercomponents.

When running ELSD methods, evenwhen there are no more peaks in the

min0 2.5 5 7.5 10 12.5 15 17.5 20

mAU

0

10

20

min0 2.5 5 7.5 10 12.5 15 17.5 20

mAU

0

10

20

min0 2.5 5 7.5 10 12.5 15 17.5 20

mV

0

10

20

1 2 3 4

UV 210

UV 254

ELSD

Figure 3Comparison of chromatograms of Gingko flavonoids between ELSD and different channels of UV.

365989-8922EN

ELSD signal, one still needs to monitorthe UV channels to make sure all thecomponents are eluted out of the sys-tem. This avoids contamination of thecolumn and system.

Analysis of AstragaliWhen developing the HPLC conditionsfor analysis of astragaloside inAstragali, one must separate all thestandards and the other componentsin the real TCM samples. To reduce theseparation time, a gradient analysiscan be used. Isocratic analysis can beused when the separation time is nottoo long and peak capacity is suffi-cient. The chromatogram at the top offigure 5 shows a gradient analysis ofan Astragali sample, while the one atthe bottom shows the astragalosidestandard. The top chromatogramshows good separation of the astraga-loside from the other sample compo-nents.

Quantitative analysisThe ELSD is not always a linear detec-tor. For some compounds, the relation-ship between the concentration andpeak area is linear only in a certainrange. In some situations, the nonlin-ear function must be calculated.

In this application, the results shownin table 1 indicate a linear functionbecause all the samples tested exhibita good response in a linear range.

The detection limit is highly related tothe conditions developed on the ELSD.For example, the ELSD parameters forgain and filter influence the detectionlimit.

The limits listed in table 2 wereobtained using methods that weresimilar to those used for the sampleanalyses. The numbers shown in table2 give the basic idea of the amount ofeach standard that can be detectedwith the current methods.

min0 5 10 15 20 25 30 35

mV

020406080

100120

1

2

34

Figure 4ELSD chromatogram of Ginkgo sample, with the flavonoid peaks numbered.

min1 1.5 2 2.5 3 3.5 4

mV

020406080

100120

min1 1.5 2 2.5 3 3.5 4

mV

020406080

100120

Astragaloside

Figure 5ELSD chromatogram of Astragali sample (top) and astragaloside standard (bottom).

5989-8922EN37

Conclusion

Two methods that used the Agilent1200 Series RRLC and the Agilent1200 Series ELSD were developed in this application. The methodsachieved good separation of the stan-dards and TCM components. Thedetection limits of the TCM standardsalso showed that the system is capa-ble of detecting low levels of compo-nents. The quantitative analyses donehere give the holistic idea of how tomeet the requirements of the ChinesePharmacopoeia. The methods can beused for the quality control of TCMsthat contain related components.

References

1. Pharmacopoeia of the People’sRepublic of China, 2005.

2. L.-W. Qi et al., “Quality evaluation ofRadix Astragali through a simultane-ous determination of six major activeisoflavonoids and four main saponinsby high-performance liquid chromato-graphy coupled with diode array andevaporative light scattering detectors”,J. Chromatogr. A, 1134, 162-169, 2006.

3.Wang Qiao'e et al. “Progresses onEvaporative Light-scatteringDetection” J. of Instrumental analysis(Chinese), Vol. 25, No.6 ,126-132,2006.

4.Roger Trones et. al ”Modified laserlight-scattering detector for use inhigh temperature micro liquidchromatography” J. of Chrom A,814, 55-61, 1998.

5.B.T. Mathews et al. “ImprovingQuantitative measurements for theEvaporative Light Scattering Detector “Chromatographia, 60 ,625-633, 2004.

6.E.Oppenheimer et. al “Examination ofthe concentration responseEvaporative Light Scattering MassDetector” J. Chrom 323, 297-304, 1985.

7.M, Dreux et al. “The EvaporativeLight Scattering Detector- a UniversalInstrument for Non- Volatile Solutes inLC and SFC” LC-GC Int, March 1996.

8.Su Hong et al. “Determination ofJujubaside A in Ziziphi SpinosaeMixture by HPLC-ELSD” J. of ModernFood and Pharmaceuticals (Chinese),Vol. 17 No 6, 35-37, 2007.

Table 1Quantitative results.

Compound name Function Linear range (µg/mL) Regression factor

Ginkgolide A y = 31.45689x – 4.18609 20 - 2000 0.99791Ginkgolide B y = 49.66089x – 2.95274 260 - 2600 0.99932Ginkgolide C y = 75.64218 + 17.60311 20 - 2000 0.99468Bilobalide y = 89.66172x – 13.33119 15 - 1500 0.99955Astragaloside y = 15.8618x – 8.0644 11 - 1100 0.9933

Table 2Detection limits of standards.

Compound name Detection limit (ng)

Ginkgolide A 12.5Ginkgolide B 19.5Ginkgolide C 30Bilobalide 15Astragaloside 22


385989-8922EN

AuthorZhixiu Xu

Agilent Technologies

Shanghai, China

Manufacturing QA/QC

Analysis of TCMs with the Agilent 1200 Series evaporative light scattering detector

Application Note

min0 2 4 6 8 10 12

mAU

20

40

60

80

*ELS1 A, Voltage (OCT27/NOV11_BEIMU1.D)

0

*DAD1 C, Sig=210,4 Ref=off (OCT27/NOV11_BEIMU1.D)

Peimine

Peiminine

UV 210 nm

AbstractThis Application Note describes the HPLC separation methods that were developed

for five traditional Chinese medicines (TCMs) – Bulbus Fritillariae Thunbergii (zhe bei

mu), Rhizoma Anemarrhenae (zhi mu), Sanjin Tablets, Fructus Kochiae (difuzi), and

Fructus Liquidambaris (lulutong). The HPLC methods were optimized to give better

separation for the maximum number of components. Sample preparation methods are

also described, to help readers who wish to follow the entire

application.

5990-3412EN39

maximum number of peaks is also help-ful for use of fingerprint chro-matograms for analysis of otherunknown components.

An Agilent 1200 Series RapidResolution LC (RRLC) system was usedfor the separation. With this system,one can easily choose to use sub-two-micron RRLC columns or conventionalHPLC columns for different analysisrequirements for TCM samples. If a nar-row internal diameter (id) column isused, one needs an RRLC nebulizer(Agilent part number G4218-20004) fornarrow peak width. The examplesshown in this Application Note takeseparations and sample throughputinto account to balance both. One caneasily use another column with thesame packing material but different col-umn inner diameter and length tospeed up the separation or increase thepeak capacity to meet different require-ments.

Experimental

Standards and materialsLiquidambaric acid, kochioside Ic,madecassoside, sarsasapogenin,peimine, and peiminine were purchasedfrom the National Institute for theControl of Pharmaceutical andBiological Products (Beijing, China).The samples of TCMs were procuredfrom the TCM drug store in Shanghai.

InstrumentationAn Agilent 1200 Series Rapid ResolutionLC system was used for conventionaland RRLC method development. Itincluded the following modules:• Agilent 1200 Series binary pump SL

with vacuum degasser• Agilent 1200 Series high performance

autosampler SL

• Agilent 1200 Series thermostattedcolumn compartment SL

• Agilent 1200 Series diode array detec-tor SL with micro flow cell (2 µL volume, 3 mm path length)

• Agilent 1200 Series evaporative lightscattering detector with standard neb-ulizer

• Agilent ChemStation B.03.02 for dataacquisition and evaluation

Sample preparationFructus Liquidambaris (lulutong): The sample was prepared by weighingaccurately 0.6 g of the powder into a vial, adding 20 mL of dehydratedethanol, weighing the mixture, and ultra-sonicating for 15 minutes. The mixture was cooled and weighed accurately again. After replenishing thelost solvent with dehydrated ethanol, thesample was mixed well and filtered.Then 10 mL of the filtrate was measuredaccurately and evaporated to dryness.The residue was dissolved with dehy-drated ethanol to a final volume of 2 mL.The solution was filtered through a 0.22µm nylon membrane filter prior to directinjection.

Fructus Kochiae (difuzi): This sample was prepared by weighingaccurately 0.5 g of the powder into a vial and adding exactly 10 mL ofmethanol. The mixture was weighedaccurately, allowed to stand overnight,ultrasonicated for 30 minutes, cooled,and weighed again. The lost weight wasmade up with methanol to the initial weight, and the sample was mixedwell. The solution was filtered through a0.22 µm nylon membrane filter prior to direct injection.

Introduction

The Chinese Pharmacopoeia 2005 pub-lished several methods that use anevaporative light scattering detector(ELSD) for analysis of TCMs.1 Becausethese TCMs contain some compoundsthat have no chromphores or have veryweak UV absorbance, one needs a uni-versal detector like the ELSD to quanti-fy these compounds in TCMs. AnotherApplication Note demonstrated theELSD analysis of Ginkgo and Astragali,2

which are two of the TCMs on thepharmacopoeia requirement list.Bulbus Fritillariae Thunbergii(zhe bei mu), Rhizoma Anemarrhenae(zhi mu), Sanjin Tablets, Fructus Kochiae(difuzi), and Fructus Liquidambaris(lulutong) are the rest of TCMs listed inthe Chinese Pharmacopeia 2005 ver-sion that need ELSD as the detector toanalyze certain active components.

The startup procedure and the quantita-tive analysis were already introduced inthe former two Application Notes.2,3

The quantitative analysis is quite sim-ple with the pharmacopoeia require-ment, which needs only two datapoints and uses a log-transform of thetwo points to generate a curve forquantitation. So the quantitative analy-sis will not be shown in thisApplication Note. A brief summary ofthe results for the separations, as wellas gradients and instrument parame-ters, will be introduced.

Because TCMs are quite complex, nei-ther the diode array detector (DAD) northe ELSD alone can detect all of thecomponents. In this study, DAD andELSD were used in a serial arrange-ment to get maximum information forthe samples and to monitor whether allthe peaks eluted from the system.Achieving a good separation for the

405990-3412EN

Bulbus Fritillariae Thunbergii (zhe bei mu): The sample was prepared by weighingaccurately 2 g of the powder into a vialwith a cap, adding 4 mL of strong ammo-nia, and macerating for 1 hour. Then 20mL of a mixture of chloroform andmethanol (4:1) was added accurately.The sample was weighed and then heat-ed under reflux on a water bath at 80 °Cfor two hours. The sample was thencooled and weighed again, and the lostsolvent was replenished with the abovemixture. The sample was filtered and 5mL of the successive filtrate was mea-sured accurately and evaporated to dry-ness in an evaporating dish. The residuewas dissolved with methanol, trans-ferred to a vial, diluted with methanol toa final volume of 10 mL, and was mixedwell. The solution was filtered through a0.22 µm nylon membrane filter prior todirect injection.

Methods

Fructus Liquidambaris (lulutong):Column: Agilent ZORBAX Eclipse

C18, 4.6 x 150 mm, 1.8 µm Solvents: A = water (0.1 % acetic

acid), B = methanol Flow rate: 0.8 mL/minGradient: 0 min, 80 %B; 10 min,

85 %B; 30 min, 100 %BUV detector: 210 nm, 254 nmELSD temperature: 40 °CELSD pressure: 50 psiELSD gain: 9ELSD filter: 3 s

Fructus Kochiae (difuzi):Column: Agilent ZORBAX Eclipse

C18, 4.6 x 100 mm, 3.5 µmSolvents: A = water (0.1 % acetic

acid), B = methanol, isocratic, 90 %B

Flow rate: 0.8 mL/minUV detector: 210 nm, 254 nmELSD temperature: 40 °CELSD pressure: 50 psiELSD gain: 9ELSD filter: 3 s

Sanjin Tablets:Column: Agilent ZORBAX SB-C18,

3 x 150 mm, 5 µm Solvents: A = water, B = methanol Flow rate: 0.6 mL/minGradient: 0 min, 45 %B; 10 min,


Rhizoma Anemarrhenae (zhi mu):Column: Agilent ZORBAX Eclipse

C18, 3 x 50 mm, 1.8 µmSolvents: A = water, B = methanolFlow rate: 0.6 mL/minGradient: 0 min, 40 %B; 5 min,

45 %B; 10 min, 80 %B; 15 min, 100 %B

UV detector: 210 nm, 254 nmELSD temperature: 45 °CELSD pressure: 50 psiELSD gain: 9ELSD filter: 3 s

Bulbus Fritillariae Thunbergii (zhe bei mu):Column: Agilent ZORBAX Eclipse

C18, 3 x 50 mm, 1.8 µmSolvents: A = water (0.03 % ammonia,

pH 9), B = acetonitrileFlow rate: 0.6 mL/minGradient: 0 min, 45 %B; 5 min,


Sanjin Tablets: Two tablets were pulverized. The coatswere removed and the tablets wereground into a fine powder. Then 0.5 g ofthe powder was weighed accurately intoa vial, and 10 mL of methanol was addedaccurately. The solution was ultrasoni-cated for 30 minutes, cooled, mixed well,and filtered. Then 5 mL of the solutionwas measured accurately. The solutionwas washed with two 0.5 mL quantities ofammonia. The extracts were evaporatedto dryness, and the residue was dis-solved in methanol and mixed well tocreate the test solution. The solutionwas filtered through a 0.22 µm nylon membrane filter prior todirect injection.

Rhizoma Anemarrhenae (zhi mu): This sample was prepared by weighingaccurately 0.5 g of the powder into a vialwith a cap, then accurately adding 10 mLof ethanol, weighing, and maceratingovernight. The sample was ultrasonicat-ed for 40 minutes, cooled, and weighedagain. The lost weight was replenishedwith ethanol and the sample was mixedwell. Then 10 mL of water and 1 mL ofhydrochloric acid were added, and themixture was heated under reflux for twohours. The sample was cooled, and a 40% solution of sodium hydroxide wasadded dropwise with shaking until thecolor of the solution changed suddenlyfrom orange-yellow to orange-red. Thesample was extracted with two 5 mLquantities of chloroform. The extractswere combined and evaporated to dry-ness. The residue was dissolved inmethanol, diluted to a final volume of 5mL with methanol, and was mixed well.The solution was filtered through a 0.22µm nylon membrane filter prior to directinjection.

5990-3412EN41


Fructus Liquidambaris (lulutong)Figure 1 shows the chromatograms ofFructus Liquidambaris, in Chinese lulu-tong, from the UV-254 nm channel andthe ELSD channel. The liquidambaricacid is labeled in figure 1. By comparingthe two channels, one can easily findthe differences. Liquidambaric acid hasno response in the UV-254 nm channel,but responds well with the ELSD. Someother components that are eluted at thebeginning of the run have no responsewith the ELSD, but do have a responsewith UV detection at 254 nm. The 210nm channel is not shown here becauseit exhibits fewer peaks and lower abun-dances.

There are two peak pairs marked as 1 and 2, respectively. The LC methodwas optimized to separate the twopeak pairs because they are the compo-nents that have a better response inELSD than in UV at 210 nm and couldbe potential components that needELSD for quantitative analysis in thefuture, although we do not know theexact structure of them yet.

Fructus Kochiae (difuzi)Figure 2 shows the separation ofFructus Kochiae components in theELSD and UV-210 nm channels. Thekochioside Ic, which has very weak UVabsorbance at 210 nm, is called out.The ELSD detected more peaks thanthe UV detector, which shows thatFructus Kochiae has some componentsthat have no chromophore and needboth detectors in series to achieveinformation from both channels in onerun. Some early-eluting componentshave no chromophores and showedresponse in the ELSD channel only.

Figure 1Fructus Liquidambaris chromatograms with ELSD detection (red), overlaid with chromatogram fromUV channel (blue).

0

0 10 15 20 25 30 min5

20

40

60

80

120

140

mAUELS1 A, Voltage

*DAD1 A, Sig=254,4 Ref=off

UV 254 nm

Liquidambaric acid2

1

Figure 2Fructus Kochiae chromatograms from ELSD channel (red) and UV channel at 210 nm (blue).

Kochioside lc

UV 210 nm

425990-3412EN

Sanjin TabletsFigure 3 shows chromatograms fromboth the ELSD and the UV-210 nmchannels for the analysis of SanjinTablets. The active component, made-cassoside, is labeled. The UV signal at210 nm has fewer peaks than the ELSDchannel because most of the compo-nents injected onto the column have nochromophores. In addition, the baselinefor the UV signal at 210 nm cannotremain flat because the absorbance ofthe solvent used for the separationchanges during the gradient. Withoutthe ELSD, there would not be verymuch information regarding the peaksand the fingerprint of Sanjin Tablets forquality control.

Rhizoma Anemarrhenae (zhi mu)The chromatogram from the ELSDchannel showed that the TCM pur-chased from the TCM store was not ofvery good quality, as the peak ofsarsasapogenin was small. In the UVsignal at 210 nm, sarsasapogeninshowed no response. For other compo-nents of Rhizoma Anemarrhenae, theELSD channel detected more peaksthan the UV channel at 210 nm, and thechromatograms showed different pro-files. Rhizoma Anemarrhenae needsthese two complementary detectors toget more information about the compo-nents with and without chromophores.

Figure 3Sanjin Tablets chromatograms with UV-210 nm channel (blue) and ELSD channel (red).

min0 2 4 6 8 10 12

mAU

20

40

60

80

*ELS1 A, Voltage (OCT27/NOV11_BEIMU1.D)

0

*DAD1 C, Sig=210,4 Ref=off (OCT27/NOV11_BEIMU1.D)

Peimine

Peiminine

UV 210 nm

Figure 4Rhizoma Anemarrhenae chromatograms from ELSD channel (red) and UV-210 nm channel (blue).

Sarsasapogenin

UV 210 nm

5990-3412EN43

References1. Pharmacopoeia of the People’sRepublic of China, volume 1, ChinaPharmacopoeia Commission, People’sMedical Publishing House, 2005.

2. Z. Xu, “Analysis of traditional Chinesemedicines with the Agilent 1200 Seriesevaporative light scattering detector,”Agilent Technologies Application Note,publication number 5989-8922EN, 2008.

3. Z. Xu, H. Huang, X. Zhang, “Analysis ofRadix Bupleuri (Chaihu) using Agilent1200 Series LC Systems withEvaporative Light Scattering Detector,”Agilent Technologies Application Note,publication number 5989-9834EN, 2008.

Bulbus Fritillariae Thunbergii(zhe bei mu)Peimine and peiminine are two mostimportant components of BulbusFritillariae Thunbergii. The UV trace at210 nm showed very few peaks. TheELSD showed the good separation ofthe two active components, and theseparation finished in 12 minutes withthe sub-two-micron column and theAgilent 1200 Series RRLC system. Withthe current method, the two compo-nents can be easily separated. Whenpeak capacity is not an issue, one canspeed up the analysis to achieve highersample throughput with the Agilent1200 Series RRLC system.

Conclusion

In this Application Note and a previousone (Agilent publication number 5989-8922EN), analysis methods were devel-oped for TCMs that require detectionby ELSD, according to the ChinesePharmacopoeia 2005. The samplepreparation methods and the instru-mentation parameters were discussed for various samples.

ELSD is one of the best choices whenanalyzing compounds that contain nochromophore. ELSD can provide a better signal when the compounds lacka functional group that absorbs well inthe UV range. ELSD also can maintain abetter baseline when UV-absorbing sol-vents are used.

It is necessary to use UV and ELSD inseries to have signals from both chan-nels because neither of them candetect all the peaks if used alone. Themethods described in this ApplicationNote can be used in quality controldepartments because the detailedinformation is provided here.

Figure 5Bulbus Fritillariae Thunbergii chromatograms of ELSD channel (red) and UV channel at 210 nm(blue).

Peimine

Peiminine

UV 210 nm

ELS1 A, Voltage *DAD1 C, Sig=210,4 Ref=off

0

0 2 4 6 8 10 12 min

20

40

60

80

mAU

445990-3412EN

45

Monitoring of Pesticide Residues

46

Authors

Wei Luan and Zhixiu Xu

Agilent Technologies (Shanghai) Co., Ltd.

412 Ying Lun Road

Waigaoqiao Free Trade Zone

Shanghai 200131

China

Abstract

A total solution of screening for pesticide residues inTraditional Chinese Medicine (TCM) is described in thisapplication, including sample preparation procedure,instrument acquisition parameter, advanced mass spec-trum data processing, and automatic qualification andquantitation report generation. Several techniques areused in this screening method. QuEChERS is employed toshorten the sample pretreatment time, reduce its cost,and environmental contamination. A significant instru-mentation innovation is the capillary flow technology ofAgilent 7890A, which can perform the backflushing toremove chemical noise caused by dirty matrices and savecycle time. Agilent's Deconvolution Reporting Software(DRS) uses the AMDIS deconvolution algorithm, whichcan be used to screen for 430 GC-amenable pesticideswith the new Japanese Positive List RTL database in only2 or 3 minutes. Overall, this screening method, frompreparing the sample to generating a report, can beaccomplished in about 1 hour.

Screening for 430 Pesticide Residues inTraditional Chinese Medicine UsingGC/MS: From Sample Preparation toReport Generation in One Hour

Application

Introduction

Traditional Chinese Medicine (TCM), which has been usedin China for thousands of years, shows very good thera-peutic effects in China as well as other countries. MoreTCM crops than ever are being cultivated in large-scalefarming operations. The pesticides that are used on thefarm need to be monitored to make sure that the amountof residues is kept within the safe range. TCM have verycomplex matrices, several hundred kinds of pesticides areoften used, and the amount of residue is very small. Allthese facts make pesticide residue screening a very diffi-cult task to accomplish in a short time.

There are several steps to solve this problem, includingsample pretreatment, instrument acquisition, data process-ing, and report generation. The basic requirements forusing a gas chromatograph/mass spectrometer (GC/MS)are listed here:

1. An effective and easy sample preparation method.

2. Repeatable and reproducible retention times.

3. A database containing the information about retentiontime and mass spectra

QuEChERS, a sample preparation method for pesticideanalyses in foodstuff, is the acronym for Quick, Easy,Cheap, Effective, Rugged, and Safe developed by Dr.Steven J. Lehotay and his colleagues [1-5]. QuEChERS isselected here because it has many advantages over tradi-tional techniques, such as high recovery for pesticides’wide range of polarity and volatility, high sample through-put, and the use of a smaller amount of organic solventand no chlorinated solvents [1-5]. QuEChERS meets thefirst basic requirement mentioned above.


5989-9341EN47

Retention time locking (RTL) [6] is the ability to very close-ly match chromatographic retention time in any Agilent6890/7890 GC system to those in another 6890/7890 GCsystem with the same nominal column. It is remarkablethat the repeatability of retention time in Agilent 7890GC isbetter than 0.01 minute. RTL can precisely match theretention time between different instruments, which meetsthe second basic requirement mentioned above.

On November 29, 2005, the Japanese Ministry of Health,Labour, and Welfare (MHLW) published a “Positive List”system for the regulation of pesticides, feed additives, andveterinary drugs. This system is one of the most restrictiveregulations on pesticide residues all over the world.Referring to this regulation, Agilent developed a new data-base called the Japanese Positive List RTL database,which includes all GC-amenable pesticides in theJapanese Positive List, together with other pesticides thatare monitored by the Japanese Quarantine Stations. Theinformation in the Agilent database comprises the reten-tion time and spectra of each pesticide, which meets thelast basic requirement mentioned above [7].

In addition, the advanced requirement for screening is torapidly identify pesticides, with minimal false positives andnegatives.

Agilent’s Deconvolution Reporting Software (DRS) [7] isdesigned to automatically deconvolute the spectra frommatrices and generate a qualification and quantitationreport. DRS integrates information from three processes:MSD ChemStation, Automated Mass SpectralDeconvolution and Identification System (AMDIS), andNational Institute of Standards and Technology (NIST)search. DRS increases the confidence in results for com-plex matrices; the typical data processing time is about 2or 3 minutes, which perfectly meets the advanced require-ment mentioned above.

This application describes how to use the techniques men-tioned above to develop a whole process for screeningpesticides in TCM matrices. The investigated targetsinclude different officinal parts of TCM, such as FlosLonicerae Japonicae for flower, Radix Astragali for rootand stem, Folium Ginkgo for leaf, Radix et RhizomaGinseng for root, and Fructus Lycii for fruit. After about a20-minute pretreatment by QuEChERS, the extract is inject-ed into an Agilent 7890A/5975C. The system then runs theJapanese Positive List method with backflushing in about37 minutes. At the end, after a 2- or 3-minute DRS process-ing with the Japanese Positive List RTL database, a qualifi-cation and quantitation report is automatically generated.

Experimental

TCM raw materials were purchased from a local medicinestore, including Flos Lonicerae Japonicae, Radix Astragali,Folium Ginkgo, Radix et Rhizoma Ginseng, and FructusLycii. All materials are finely ground.

Sample Preparation

• Weigh 1 g of ground TCM and add it into a 50-mL Teflon tube; add 6 mL distilled water to thor-oughly soak the sample and store the sample at ambi-ent temperature for 2 hours.

• Add 4 mL 0.1% HAc/acetonitrile, and shake vigorously for 1 minute by hand, then add 1.2 g MgSO4,0.6 g NaAc, and 4 g NaCl, and shake vigorously for 1 minute by hand. Centrifuge for 3 min-utes at 4,000 rpm.

• Add 500 mg PSA (p/n 5188-5364), 250 mg C18 (p/n189-1302), and 250 mg MgSO4 in a 5-mL centrifugetube. Transfer an aliquot (1.5 mL) of the above extract’supper layer into the tube and then centrifuge for 1minute at 3,000 rpm. If the solution is dark after cen-trifugation, graphitized carbon black can be added toadsorb the pigment (for example, chlorophyll).

• Transfer an aliquot (1.0 mL) of the extract into the GCvials.

For dedicated quantitation analysis after screening, ana-lytical protectants and internal standard can be added tothe final solution.

GC/MS

The Japanese Positive List method was published byMHLW for pesticide residues on all Japanese agriculturalproducts and imports to Japan. Here, the method wasmodified to handle the dirty TCM matrices by adding thebackflushing. The detail is displayed in Table 1.

System Software Requirement

Agilent MSD ChemStation (p/n G1701EA, Ver. E.02.00);Agilent Deconvolution Reporting Software (p/n G1716AA,Ver. A.04.00); NIST05b Mass Spectral Library (p/nG1033A), includes NIST MS Search (Ver. 2.0d) and AMDIS(Ver. 2.65); Agilent Japanese Positive List Pesticide RTLDatabase (p/n G1675AA).

485989-9341EN


QuEChERS Method Optimization for Versatile TCM Matrices

In Traditional Chinese Medicine formulation, different edi-ble parts of TCM plants can be chosen to make each dose.Moreover, the TCM matrices are versatile in chemical con-stituents, such as saponins, polysaccharids, flavones, phe-nolic acid, and fatty acid. Therefore, it is a tough job todetermine the pesticide residues in such kinds of complexmatrices as TCM.

In our study, Flos Lonicerae Japonicae was investigated atfirst, then we mixed different kinds of TCM together to bemore representative in respect to chemical constituents.Flos Lonicerae Japonicae for flower, Radix Astragali forroot and stem, Folium Ginkgo for leaf, Radix et RhizomaGinseng for root, and Fructus Lycii for fruit are the fiveselected TCM.

Regarding the QuEChERS method, the amount ofabsorbent, such as PSA and C18, is critical to the recoveryof pesticides. It is demonstrated that PSA is more effective

in dealing with the acidic component in TCM extract. Byscreening for different doses of PSA, 500 mg is enough forthe analysis. A typical chromatogram of mixed TCM sam-ple after QuEChERS is shown in Figure 1.

Backflushing Using Capillary Flow Technology for Complex TCM Matrices

Backflushing is demonstrated to significantly removechemical noise from run to run by reversing the columnflow to push the higher boilers out of the inlet end of thecolumn. Traditionally, the higher boilers stay near the frontof the column until the oven temperature is high enough tomove them through the column. It is well known that bak-ing the column at high temperature is inclined to causeheavy column bleed and result in shorter column lifetime.For a dirty matrix, even baking for a long time cannot thor-oughly remove the higher boilers, which may result in aretention time shift in the following injection. Backflushingis a perfect solution to avoid high-temperature baking andretention time shift from run to run. Backflushing mini-mizes the contamination that goes into the detector, whichis especially valuable for the MSD because it reduces ionsource cleaning. Another positive attribute of backflushing

Table 1. Gas Chromatography and Mass Spectrometers Conditions

GC Agilent Technologies 7890A or 6890N

Front Inlet Split/splitless Injection type SplitlessInlet temp 280 °CInlet liner Helix double taper deactivated

(p/n 5188-5398)Pressure 17.446 psiPurge flow 50 mL/minPurge time 0.75 minTotal flow 54.51 mL/minGas saver OnSaver flow 20.0 mL/minSaver time 1.00 minGas type Helium

OvenOven ramp °C/min Next °C Hold minInitial time 50 1Ramp rate 25 125 0Ramp rate 10 300 10Post run 300 5Total run time 36.5 minEquilibration time 0.5 min

Column J&W DB-5ms Ultra Inert (p/n 122-5532UI)Length 30 mDiameter 0.25 mmFilm thickness 0.25 µmMode Constanr flowPressure 17.446 psi (2 psi for post run)Nominal initial flow 1.51 mL/min (–4.05 mL/min for post run)Inlet Front InletOutlet PCM (p/n G1530-63309)Outlet pressure 2 psi (60 psi for Post run)

Front InjectorSample washes 0Sample pumps 6Injection volume 2 µLSyringe size 10 µLPreinj solv A 0Preinj solv B 0Postinj solv A 6Postinj solv B 6Viscosity delay 1 secondPlunger speed FastPreinjection dwell 0 minutesHelium 0 minutes

MSD Agilent Technologies 5975C with Triple-Axis Detector

Acquistion mode ScanSolvent delay 3.30 minLow mass 35High mass 450Threshold 100 (or setting according to noise level) Sampling 2Quad temp 150 °CSource temp 230 °CTransfer line temp 280 °CTune file Atune.uGain factor 10Trace ion detection On

RTL System retention time locked toChlorpyrifos methyl at 13.443 min

BackflushDevice 3-way splitter (p/n G3183B)Restrictor Size 0.706 m ˘ 0.18 mm id

5989-9341EN49

is that it shortens cycle time and improves lab productivity.Details about backflushing, including a method develop-ment procedure, can be found in the Agilent application inreference 8.

In our experiment, a three-way splitter with makeup gaswas used to perform the backflushing. The device hasextra makeup gas supply tubing and four plug-in spots: onefor the analytical column and three for the restrictor tubeconnecting to three available detectors. Since only MSD isused as a detector in this application, the first two spotsare plugged in the direction of makeup flow. The third spotis used for column in and the last spot is used for restrictorin. The main purpose of this configuration is to avoid peakbroadening due to extra dead volume. The length and inter-nal diameter of the restrictor tubing is 0.706 m and 0.18mm, respectively. A photo of the device and the GC/MSsystem configuration schematics are displayed in Figure 2.

Target Pesticides Identification in TCM Matrices by DRS

Another challenge for pesticides analysis is the identifica-tion of trace-level target compounds in complex matrices,which are often disturbed by high-level chemical noise andresult in poor library match factors. Background subtrac-tion is the traditional way to improve the match factor;however, it is both matrix- and operator-dependent and canyield inconsistent results.

DRS, which Agilent introduced in 2004, is a software pack-age that combines the information from three separateprocesses into one easy-to-read report. The primary benefitof DRS is significant time saving when interpreting resultsfrom complex matrices analyses. Moreover, a suitabledatabase is critical to perform DRS successfully. TheJapanese Positive List RTL database is a new DRS-usablelibrary for pesticide analysis, and includes the retentiontime and mass spectra for 430 GC-amenable pesticides.Detailed information about the database can be found inthe Agilent application in reference 7.

Using DRS with the Japanese Positive List RTL database ishighly recommended for data processing and report gener-ation for pesticide residues analysis in TCM matrices.Figure 3 is part of a typical DRS report for a spiked mixedTCM sample.

Figure 4 shows the benefit of using DRS in the analysis ofpesticides in TCM matrices. Pirimiphos-methyl was spikedinto mixed TCM sample at a level of 10 ppb. The compoundwas missed by MSD ChemStation but successfully pickedup by DRS. The upper window is the Total Ion Chromato-gram (TIC), the middle window is the raw or dirty spectrumin the scan No. 2199 (13.981 min), and the lower windowis the comparison of the deconvoluted spectrum (the whiteplot) with the spectrum of pirimiphos-methyl in theJapanese Positive List RTL database (the black plot). Afterdeconvolution, the spectrum of the scan No. 2199 is"clean," and the Pirimiphos-methyl can be easily identifiedin the mixed TCM sample.

Figure 1 Typical Total Ion Chromatogram of mixed TCM sample after QuEChERS: Flos Lonicerae Japonicae, RadixAstragali, Folium Ginkgo, Radix et Rhizoma Ginseng, and Fructus Lycii.

505989-9341EN

3-way splitterwith make gas2 ports plugged

Restrictor out to MSD

Aux EPC in

Column in

Plugged

Plugged

7683Auto-sampler

7890AGC

Column

30 m × 250 µm × 0.25 µm DB-5MS UI

0.706 m × 180 µm restrictor

AUX EPC

5975CMSD

Figure 2 Schematics of the GC/MS system configuration and tubing connection for three-way splitter.

Conclusions

The primary interference for trace-level pesticides analysisis chemical noise in complex TCM matrices. This applica-tion introduces several techniques to eliminate the inter-ference. Well-developed sample preparation is the mainapproach to remove as much of the chemical noise as pos-sible, QuEChERS was employed to simplify the samplepreparation procedure and improve lab productivity, com-pleting the process in only about 20 minutes. However,even with a pretreatment method such as QuEChERS, thechemical noise cannot be completely removed.Backflushing is an alternative approach to eliminate thechemical noise from run to run. Although backflushing isnot a new concept, capillary flow technology can perform abackflushing much more effectively than traditional tech-niques can. In this application, a three-way splitter withmakeup gas is used for backflushing, and the data-acquisi-tion step with backflushing is almost 37 minutes. The laststep related to chemical noise in data processing is the uti-lization of a mathematic algorithm called deconvolution.Agilent's DRS can shorten the data processing time to only2 or 3 minutes, which is very helpful for pesticide analysisin TCM matrices with the new Japanese Positive List RTLdatabase. Altogether, only 1 hour is needed to screen 430pesticide residues in TCM matrices from sample pretreat-ment to screening report generation.

References1. M. Anastassiades, S. J. Lehotay, D. Stajhbaher, and F.

J. Schenck, “Fast and Easy Multiresidue MethodEmploying Acetonitrile Extraction/Partitioning and'Dispersive Solid-Phase Extraction' for the

Determination of Pesticide Residues in Produce,” J.AOAC Int., 2003; 86(2): 412

2. S. J. Lehotay, K. Mastovska, and S. J. Yun, “Evaluation ofTwo Fast and Easy Methods for Pesticide Residue Analysisin Fatty Food Matrixes,” J. AOAC Int., 2005; 88(2): 630

3. S. J. Lehotay, K. Mastovska, and A. R. Lightfield, “Useof Buffering and Other Means to Improve Results ofProblematic Pesticides in a Fast and Easy Method forResidue Analysis of Fruits and Vegetables,” J. AOACInt., 2005; 88(2): 615

4. S. J. Lehotay, A. de Kok, M. Hiemstra, and P. vanBodegraven, “Validation of a Fast and Easy Method forthe Determination of 229 Pesticide Residues in Fruitsand Vegetables Using Gas and Liquid Chromatographyand Mass Spectrometric Detection,” J. AOAC Int.,2005; 88(2): 595

5. M. Anastassiades, K. Mastovska, S. J. Lehotay,“Evaluation of Analyte Protectants to Improve GasChromatographic Analysis of Pesticides,”J. Chromatogr. A, 2003; 1015(1-2): 163

6. V. Giarocco, B. Quimby, and M. Klee, “Retention TimeLocking: Concepts and Applications,” AgilentTechnologies publication 5966-2496E

7. Philip L. Wylie, “Screening for Pesticides in Food Usingthe Japanese Positive List Pesticide Method: Benefitsof Using GC/MS with Deconvolution ReportingSoftware and a Retention Time Locked Mass SpectralDatabase,” Agilent Technologies publication 5989-7436EN

8. Chin K. Meng, “Improving Productivity and ExtendingColumn Life with Backflush,” Agilent Technologiespublication 5989-6018EN.

5989-9341EN51

Figure 3 Part of typical DRS report.

Figure 4 AMDIS display showing the total ion chromatogram of a mixed TCM sample (top window), the spectrum where pir-imiphos-methyl elutes (middle window), and the deconvoluted spectrum (in white) juxtaposed to the library spec-trum for pirimiphos-methyl (in black) (bottom window).

525989-9341EN

Low Part-per-Billion Level PesticidesScreening in Traditional ChineseMedicine Using the Agilent 7000AGC/MS/MS

The control of pesticide residue levels is an increasing global concern. Severalinternational organizations and governments have established maximum residuelimits (MRLs) for a growing list of pesticides within an ever-widening scope ofcommodities. The primary task for pesticide residue monitoring laboratories is todevelop analytical methodologies to screen for a large number of analytes attrace levels within a limited time frame. Gas chromatography coupled with massspectrometry (GC/MS) has been adopted as the standard instrumentation for thispurpose. As regulations in Japan and the European Union require lower MRLs forpesticide residues, the latest challenge has been to reach part-per-billion levelconcentrations for hundreds of pesticides in complex matrices, which in turn hasrequired greater sensitivity and efficiency in pesticide screening. The TripleQuadrupole mass spectrometer, when used in Multiple Reaction Monitoring(MRM) mode, can dramatically remove matrix interferences and significantlyincrease the effective signal-to-noise ratio (S/N). This paper describes the use ofgas chromatography coupled with triple quadrupole mass spectrometry(GC/MS/MS) to screen pesticides in Traditional Chinese Medicine (TCM) at lev-els as low as 1 ppb. The target pesticides and associated MRM conditions arelisted in Table 1.

Highlights

• Consistent ion ratios over a widelinearity range and peak arearepeatability at ultra-low concentra-tions in matrix make pesticidescreening more reliable.

• The Agilent collision cell allows forexcellent repeatability at 2 ppb levelin matrix using just 5 ms dwelltimes. A single time segment canaccommodate more MRM transi-tions for greater productivity.

• Reduction of matrix interferenceusing MRM mode substantiallydecreases the detection limits thatare required in pesticide screening.

Application Brief

Wei Luan, Melissa Churley, and Mike Szelewski

5990-3568EN53

Table 1. Target Pesticides List and MRM Conditions

Experimental

Quantitation of trace level compounds is complicated by matrix, resulting in qualifierion ratios out of range, or target ions buried in the complex background. With singlequadrupole mass spectrometry, selected ion monitoring (SIM) is often used toimprove the detection limit and quantitative reproducibility. In SIM mode, the MSmonitors only a few ions for each target compound within the retention time (RT)range that the target elutes from the column. By monitoring only a few specific ions,the signal-to-noise ratio (S/N) improves dramatically. SIM may not work well fortrace levels in matrix as the interferences in SIM are the same as scan. Triplequadrupole mass spectrometry, through further fragmentation in a hexapole collisioncell of a selected precursor ion, allows for drastic reduction or elimination of matrixinterference. This process, referred to as Multiple Reaction Monitoring or MRM, isbased on acquisition of highly selective precursor to product ion transitions that are

Collision Collision Collision Segment Compound name R.T. Quant energy Qual1 energy Qual2 energy

1 Dichlorvos 6.00 185 & 93 15 185 & 109 15 109 & 79 5

1 Methamidophos 6.10 141 & 95 5 95 & 80 5

1 Acephate 7.82 136 & 94 10 142 & 96 5

1 Dimethoate 8.53 125 & 79 5 125 & 93 15

1 Omethoate 9.53 156 & 110 5 156 & 79 20

2 Dimethipin 11.12 118 & 90 5 118 & 73 5

2 Cyanophos 11.41 243 & 109 10 243 & 116 5

3 Phosfamidon-E 12.41 264 & 127 15 127 & 109 10

3 Parathion-methyl 12.58 263 & 109 10

3 Phosfamidon-Z 13.13 264 & 127 15 127 & 109 10

3 Quinoclamine 13.23 207 & 172 15 207 & 179 12 172 & 128 10

3 Chlorpyriphos 13.52 314 & 258 15 314 & 286 5 197 & 169 15

3 Cyanazin 13.52 225 & 189 15 240 & 225 5 198 & 91 10

3 Parathion 13.54 291 & 109 10 291 & 81 10

3 Fosthiazate 1-2 13.85 195 & 103 5 195 & 139 5

4 Fipronil 14.33 367 & 225 25 367 & 224 20

4 Quinalphos 14.36 146 & 118 10 157 & 129 15

4 Endosulfan-alpha 14.83 241 & 206 15 229 & 194 10 195 & 159 10

4 Chlorfenapyr 15.77 247 & 227 15 247 & 197 20 408 & 59 10

4 Endosulfan-beta 15.90 241 & 206 15 229 & 194 10 195 & 159 10

5 Triazophos 16.38 161 & 134 5 161 & 106 10 257 & 162 5

5 Bromopropylate 17.64 341 & 183 20 341 & 185 20

5 Azinphos-methyl 18.31 160 & 77 20 160 & 132 5 132 & 77 15

6 Cafenstrole 19.92 100 & 72 5 188 & 119 25 188 & 82 20

6 Cyfluthrin 1-4 20.03 163 & 127 5 163 & 91 15 206 & 151 25

7 Flucythrinate 1 20.46 199 & 157 5 199 & 107 30 157 & 107 15

7 Flucythrinate 2 20.64 199 & 157 5 199 & 107 30 157 & 107 15

8 Fenvalerate 1 21.14 167 & 125 10 225 & 119 15

8 Fenvalerate 2 21.33 167 & 125 10 225 & 119 15

8 Difenoconazole 1 21.53 323 & 265 15 265 & 202 20

8 Difenoconazole 2 21.60 323 & 265 15 265 & 202 20

8 Delthamethrin 21.85 181 & 152 30 253 & 172 5 253 & 93 20

545990-3568EN

not likely to result from fragmentation of matrix components. Precursor selectivity isthe same as in SIM but there is a high probability that at least one of the resultantproduct ions will be a unique dissociation product of the precursor and not the inter-ference.

The primary benefit is the improvement of S/N at ultra-low analyte levels with con-sistent qualifier ion ratios over a wide concentration range, even in the most com-plex matrices. Figure 1 illustrates the Agilent GC/MS/MS workstation-Quantitativeview. The numbers inside the red frame are the qualifier ion ratios of cyanophos overthe range from 0.1 ppb to approximately 1,000 ppb in TCM matrix, showing very goodaccuracy in matrix. The numbers inside the green frame indicate the accuracy of thecalibration curve. The enlarged view of the low end of the calibration curve isshown. Table 2 is the repeatability of the peak area of 1 ppb pesticides in TCMmatrix with six parallel injections.

Figure 1. Calibration curve and qualifier ion ratio of cyanophos in TCM matrix.

The Agilent hexapole collision cell, with its linear acceleration design, is optimizedfor high-speed performance without ion ghosting or cross-talk. High-speed MRMcapability at 500 MRM/sec maximizes the number of allowable transitions as well asminimizes the dwell time for each transition in one time segment so that more com-pounds can be screened per run. The dwell time experiment described in Table 3 wasperformed using TCM matrix spiked with 2 ppb bromopropylate. As dwell time isdecreased, the RSD for replicate run peak areas remains at < 5.0% until 5 ms isreached. The results at 2 ms are less precise; however, the results are acceptable for2 ppb analysis in complex matrix.

Table 2. Peak Area Repeatability of Spiked 1 ppb Pesticides in TCM Matrix

RSD of peak area (%)(n = 6)

Cyanophos 4.83

Bromopropylate 5.12

5990-3568EN55

Figure 2 is the total ion chromatogram of the TCM extract in MRM mode byGC/MS/MS with injection volume of 1 µL. Eleven pesticides are identified in ppblevel as shown in Table 4.

Table 3. Dwell Time Experiment Results: 2 ppb Bromopropylate in TCM Matrix

100 ms 50 ms 10 ms 5 ms 2 ms

Area 1 5849 5910 6265 6704 6747Area 2 5712 6167 6189 6728 6279Area 3 5895 5941 5966 6131 6523Area 4 5921 6471 6551 6243 6397Area 5 5999 6299 6119 6504 4831Area 6 5999 6524 6415 6796 4737

Sum 35375 37312 37505 39106 35514

Avg 5895.83333 6218.667 6250.833 6517.667 5919

Std 107.618617 260.1528 209.5638 276.2496 893.2359

RSD 1.83 4.18 3.35 4.24 15.1

Figure 2. Total ion chromatogram of the TCM extract in MRM mode by GC/MS/MS.

565990-3568EN

Summary

Pesticide residue monitoring is required to develop a methodology to accomplish thescreening for hundreds of target compounds in a very limited time and ultra low con-centration levels. Gas chromatography coupled with triple quadrupole mass spec-trometry can perform very selective MRM, which can dramatically remove chemicalnoise from matrix interferences and improve the S/N, as well as detection limit. Thisapplication develops a method to screen for ppb level pesticide residue in TCM.GC/MS/MS can really help to improve the capability of pesticide residues screeningin ultra low level.

References

1. Wei Luan, and Zhixiu Xu, "Screening for 430 Pesticide Residues in TraditionalChinese Medicine Using GC/MS: From Sample Preparation to Report Generationin One Hour," Agilent Technologies publication 5989-9341EN

Table 4. Screening Result of TCM Matrix Blank by GC/MS/MS

RT Name Quant transition Matrix areas Cal result(not spiked)

5.995 Dichlorvos 185.0 & 93.0 68 —

6.100 Methamidophos 141.0 & 95.0 0 —

7.760 Acephate 136.0 & 94.0 0 —

8.526 Dimethoate 125.0 & 79.0 0 —

9.478 Omethoate 156.0 & 110.0 0 —

11.100 Dimethipin 118.0 & 90.0 0 —

11.414 Cyanophos 243.0 & 109.0 0 —

12.390 Phosfamidon-E 264.0 & 127.0 0 —

12.560 Parathion-methyl 263.0 & 109.0 1227 6.21*

13.200 Quinoclamine 207.0 & 172.0 38 —

13.500 Cyanazin 225.0 & 189.0 16 —

13.510 Chlorpyrifos 314.0 & 258.0 1224 1.99*

13.520 Parathion 291.0 & 109.0 3950 5.62*

13.840 Fosthiazate 195.0 & 103.0 0 —

14.320 Fipronil 367.0 & 255.0 0 —

14.340 Quinalphos 146.0 & 118.0 3970 1.11*

14.820 Endosulfan-a 241.0 & 206.0 156 1.89*

15.750 Chlorfenapyr 247.0 & 227.0 0 —

15.880 Endosulfan-b 241.0 & 206.0 208 1.61*

16.360 Triazophos 161.0 & 134.0 531 1.63*

17.620 Bromopropylate 341.0 & 183.0 63 —

18.290 Azinphos-methyl 160.0 & 77.0 0 —

19.900 Cafenstrole 100.0 & 72.0 1260 3.37*

20.026 Cyfluthrin 163.0 & 127.0 0 —

20.439 Flucytrinate-1 199.0 & 157.0 0 —

20.624 Flucytrinate-2 199.0 & 157.0 0 —

21.122 Fenvarelate-1 167.0 & 125.0 4188 1.79*

21.326 Fenvarelate-2 167.0 & 125.0 13070 1.79*

21.520 Difenoconazole-1 323.0 & 265.0 0 —

21.574 Difenoconazole-2 323.0 & 265.0 0 —

21.840 Delthamethrin 181.0 & 152.0 2078 3.53*

*Potential positive sample

5990-3568EN57

Wei Luan is an application chemistbased at Agilent Technologies,Shanghai, China; Melissa Churley is anapplication chemist based at AgilentTechnologies, Santa Clara, California,USA; and Mike Szelewski is an applica-tion chemist based at AgilentTechnologies, Wilmington, Delaware,USA.


For more information on our productsand services, visit our Web site atwww.agilent.com/chem.

585990-3568EN

59

Determination of Heavy MetalContent

60

Determination of Toxic Elements inTraditional Chinese Medicine UsingInductively Coupled Plasma MassSpectrometry

Abstract

This Application Note describes a method for the analysis of Be, Cr, Mn, Ni, Cu, Zn,

As, Ag, Cd, Ba, Hg, Tl and Pb in Traditional Chinese Medicines using Inductively

Coupled Plasma Mass Spectrometry (ICP-MS). The samples were dissolved by

microwave digestion. To validate the method, two certified reference materials

were digested and measured. The method detection limits for all target elements

were between 0.1-7.2 ng·g-1.

Hua ZhangYan-zhi Shi

Yu-hong ChenYing-feng Wang

Application Note

5989-5591EN61

Introduction

Traditional Chinese Medicines (TCMs)and their related products have beenwidely used in China for centuries. TheChinese government as well as acade-mic research scientists are payinggreater attention to the safety issuesof TCM in clinical use. This is a keyissue which needs to be resolved toimprove the development of TCM tomeet international standards andadvance its worldwide acceptance.With the recent developments in sci-ence and technology, people arebecoming more aware of the risksassociated with heavy metals such asmercury (Hg), arsenic (As), lead (Pb)and cadmium (Cd), which are presentin some TCMs. After these elementsenter the human body, they willseverely damage the hemopoietic,immune, nervous and reproductivesystems. Because these substancescannot be completely excreted fromthe body, they will accumulate andfinally have an impact on health.Therefore the analysis of toxic heavymetals is crucial in quality control ofTCMs and regulations have beenimplemented to restrict their levels1.The export of TCMs is under intensepressure due to strict regulations, andthe high content of heavy metals intraditional Chinese medicine hasresulted in international response.

Limiting other toxic or harmful ele-ments, such as beryllium (Be), chromi-um (Cr), nickel (Ni), silver (Ag), barium(Ba) and thallium (Tl), have not beenclearly specified in current hygienerules and regulations yet. However,with the improvement of living stan-dards and a desire for good health,people now care more about whetherthese elements offer advantages ordisadvantages for their wellbeing.Therefore, strict control of the contentof these elements in TCM is required.

In addition, manganese (Mn) is goodfor the human endocrine, nervous, andenzyme systems and is the key ele-ment of most enzymes as well as sup-porting normal myocardial metabolism.Zinc is good for the immune system.Lack of zinc (Zn) will harm normal cellmetabolism and lack of copper (Cu) isone of the reasons for coronary heartdisease2. However, if the level of theseelements is too high, they will placehuman health at risk.

By nature TCM is very complex andcontains many elements which havedifferent concentration levels – highand low. In the past, they have beenanalyzed by a combination of ASS, AFSand AES technology. These proceduresare time-consuming and costly3. ICP-MS is fast becoming the technique ofchoice for the determination of ele-ments in a wide range of samples.Additionally, it has the widest linearrange (nine orders of magnitude, from1 ng·mL-1 to 1000 µg·mL-1), the high-est sensitivity and lowest detectionlimit for metals, as well as the abilityto rapidly and accurately measure mul-tiple elements simultaneously. The lat-est revision of the Pharmacopoeia ofthe People's Republic of China(Chinese Pharmacopoeia 2005) nowincludes ICP-MS as one of the stan-dard methods for the determination ofheavy metals in herbal medicines4.

In this study, microwave digestion isused for sample preparation. High tem-perature and a closed system assurefast and complete digestion and avoid

the loss of those elements which caneasily evaporate, such as arsenic andmercury. Thirteen toxic elementsincluding Be, Cr, Mn, Ni, Cu, Zn, As,Ag, Cd, Ba, Hg, Tl and Pb were ana-lyzed using the microwave digestion-ICP-MS method. The method is validat-ed by using two certified referencematerials: shrub leaves and tea leaves.The results obtained agree with thecertified values. The recovery experi-ment of standard addition was appliedin the study to evaluate the proposedmethod. The recovery percentages arebetween 90.3 -109.7 %.

Experimental

InstrumentationAn Agilent 7500 Series c ICP-MS wasused in this study (see table 1 for theinstrument details). The samples wereprepared by microwave digestion witha CEM MARS5 microwave digestionsystem (CEM Co. Ltd, USA), includingmicrowave oven, PTFE-TFE high pres-sure vessel and fixed tray. Deionizedwater (Milli-Q ultra-pure water system)was used throughout.

ReagentsNitric acid (HNO3), trace metal grade(Merck); Hydrogen peroxide (H2O2),MOS grade;standard stock solution: 10µg ng·mL-1 mixed standard solutionincluding Be, Cr, Mn, Ni, Cu, Zn, As,Ag, Cd, Ba, Hg, Tl and Pb. Calibrationstandards were prepared by usingapplicable dilutions of standard stock

Nebulizer Babington nebulizer

Spray chamber Quartz scott-type, peltier-thermostatted to 2 ± 0.1 °CTorch Quartz, 2.5 mm IDInterface Ni cone

Table 1Agilent 7500 c ICP-MS Instrument DetailsZ.

625989-5591EN

solution with 5 % HNO3; Hg stock solution: 1000 µg·mL-1 (NationalAnalytical Center for Iron & Steel,China). Calibration standards were pre-pared by using applicable dilutions ofstandard stock solution with 5 %HNO3; Internal standard solution: 1.0µg/mL. Diluted from 10 µg·mL-1 Li, Sc,Ge, Y, Tb and Bi mixed standard solu-tion. (Agilent part number 5183-4680) with 5 % HNO3; Tune solution: 10 ng·mL-1 7Li, 59Co, 89Y, 140Ce and205Tl mixed standard solution (2 %HNO3) (Agilent part no.5184-3566);Deionized water (18.2 M_) producedwith the Milli-Q® ultrapure waterpurification system (Milllipore Corp.)was used in all standard solution andsample preparations. Certified refer-ence materials: GBW07602 (shrubleaves) and GBW08513 (tea leaves).

Sample pretreatmentExactly 0.5000 g of sample wereweighed and placed into the PTFE ves-sel; 6 µL HNO3 and 1 µL H2O2 wereadded. Because of the large organiccontent of TCM and the large amount ofsample, pre-digestion is recommended.To prevent sample loss due to an explo-sion caused by gas and pressure build-up produced in the digestion process,

the sample was dissolved in the solutionand placed in the micro-wave oven. First,the temperature was raised and set at120 °C for approximately 5 minutes, thenthe sample was cooled completely andthe pressure was allowed to drop to nor-mal. Subsequently, the digestion pro-gram specified in table 2 was followed.After digestion the vessel was cooled toroom temperature and the contentspoured into a 50-mL PET bottle. The ves-sel and cap were washed with a smallamount of distilled water several timesand this mixture was added to the PETbottle. Water was added to bring theexact weight to 50 g. The procedure forreagent blanks is identical to that fortest samples and is carried out concur-rently without a sample.

ICP-MS parameters and target ele-ment isotopesThe sample solution was analyzedunder the optimized condition. The tar-get element isotopes are summarizedin table 3; 72Ge was selected as theinternal standard element for Be, Cr,Mn, Ni, Cu, Zn and As, 115In wasselected as the internal standard ele-ment for Ag, Cd and Ba; and 209Bi wasselected as the internal standard ele-ment for Hg, Tl and Pb. ICP-MS para-

meters were automatically optimizedby the instrument. All of the specifica-tions meet the installation require-ments including sensitivity, back-ground, oxide, doubly charge, stability,etc. The parameters are listed in table 4.

Calibration curvesThe mixed stock solution and Hg stocksolution were diluted to 0.1, 0.5, 2,10 ng·mL-1 using 5 % HNO3respectively. The blank solution is 5 %HNO3. The blank and calibration solu-tions were measured under optimizedconditions. The calibration curve wasautomatically plotted by the instrument.Linear correlation coefficients (r) in allcalibration curves were better than 0.9999.

Table 2Microwave digestion program.

Stage Power Ramp /min T /°C Hold/minMax / W %

1 1200 100 5:00 120 5:002 1200 100 6:00 180 20:002

Table 4ICP-MS operating parameters.

Parameter Set value

Power 1350 WFlow rate of plasma gas 15.0 L·min-1

Flow rate of auxiliary gas 1.0 L·min-1

Flow rate of carrier gas 1.12 L·min-1

Sampling rate 0.4 mL·min-1

Sampling depth 7 mmOrifice of sampling cone 1.0 mmOrifice of skimmer cone 0.4 mmData acquisition mode Quantitative

analysisIntegration time 0.3 s /isotopeCerium oxide/Cerium <0.5%Doubly charge <2%

Table 3Target element isotopes.

Element Be Cr Mn Ni Cu Zn As Ag Cd Ba Hg Tl Pb

Isotope 9 52 55 60 63 66 75 107 114 137 202 205 208

5989-5591EN63

Results and discussionMethod detection limitThe blank sample was analyzed 11times under optimized conditions. Themethod detection limits (MDL) for eachelement were calculated (table 5).

Determination of standard materialsTo evaluate reliability and accuracy, the microwave-ICP-MS method wasapplied to the determination of twocertified reference materials:GBW07602 (shrub leaves) andGBW08513 (tea leaves). The resultsare in strong agreement with the certi-fied values, a comparison is shown intable 6.

Recovery The percent recovery for each element was determined using the standard addition method to evalu-ate the reliability and accuracy of themethod. Percent recoveries of all ele-ments were between 90.3 %-109.7 %.The results were considered satisfac-tory (table 7).

Sample analysisThis study analyzed 13 toxic elementsin seven TCMs purchased on the mar-ket. Each sample was analyzed eighttimes and the accuracy (RSD %) variedbetween 0.3 % - 6.8 % (table 8).

According to the results in Table 8, notall TCMs meet the minimum legalrequirements. The concentrations ofsome elements highly exceed theallowable limit. However, in mostTCMs, the amount of all toxic ele-ments is low.

ConclusionThe evidence suggests that the cur-rent TCM market is complex and strictmanagement and quality controls areneeded to ensure that the maximumallowable limits of toxic elements setby regulations are not exceeded.National and local drug administra-

tions also require regulations to man-age and control the production andsale of TCMs. The study usesmicrowave digestion for sample prepa-ration, With optimized working para-meters all elements in different medi-cines are analyzed simultaneouslywith an Agilent 7500 Series c ICP-MS.This method is validated with certifiedreference materials by conductingrecovery experiments using knownstandard additions. It offers manyadvantages over other alternativetechniques, such as precision andaccuracy, simplicity, rapidity, low limitsof detection and multiple elementdetermination.

Table 5Method detection limits (MDL) (Unit: ng·gPT).

Element MDL

Be 0.1Cr 7.2Mn 1.1Ni 1.9Cu 2.5Zn 3.4As 3.5Ag 0.6Cd 0.8Ba 2.4Hg 1.1Tl 0.1Pb 0.9

Table 6Comparison of found value and certified value.

Element GBW07602 (shrub leaves) GBW08513 (tea leaves) Certified value Found value Certified value Found value

ug·gPT ug·gPT ug·gPT ug·gPT

Be 0.056 ± 0.014 0.046 / 0.088Cr 2.3 ± 0.3 2.0 / 2.2Mn 58 ± 6 54 2170 ± 110 2074Ni 1.7 ± 0.4 1.7 5.09 ± 0.76 4.94Cu 5.2 ± 0.5 4.6 8.96 ± 0.59 8.23Zn 20.6 ± 2.2 19.8 22.6 ± 1.5 22.4As 0.95 ± 0.12 0.79 0.180 ± 0.049 0.134Ag 0.027 ± 0.006 0.024 / 0.022Cd 0.14 ± 0.06 0.18 0.023 ± 0.004 0.022Ba 19 ± 3 17 120 ± 10 112Hg / 42.5 0.017 0.018Tl / 0.015 / 0.016Pb 7.1 + 1.1 6.1 1.00 ± 0.05 0.91

Table 7Results of the recovery experiments.

Element Spiked value Found value Recovery percentage(ng·mLPT) (ng·mLPT) (%)

Be 10 9.37 93.7Cr 10 10.63 106.3Mn 10 10.97 109.7Ni 10 10.24 102.4Cu 10 10.68 106.8Zn 10 10.36 103.6As 10 10.89 108.9Ag 10 9.20 92.0Cd 10 9.25 92.5Ba 10 9.16 91.6Hg 5 4.87 97.3Tl 10 9.58 95.8Pb 10 9.03 90.3

645989-5591EN

Table 8Analytical results of TCMs (n = 8). Unit: µg gPT.

Name of sample Gegen Zhike Guifudi Huangli Jinsang Naodes Shugan Soup San huang Pill anshangqing Pill sanjie Pill heng Pill Pill

Element

Be Foundvalue 0.020 0.009 0.032 0.034 0.060 0.037 0.013RSD% 5.2 2.6 1.0 5.1 2.0 1.4 5.2

Cr Foundvalue 0.34 0.46 1.40 1.58 4.89 1.74 4.76RSD% 2.0 2.0 1.9 2.6 2.3 2.5 0.8

Mn Foundvalue 34.77 17.80 54.78 30.37 87.23 23.96 124.4RSD% 0.7 2.4 1.1 2.20 1.21 2.62 3.10

Ni Foundvalue 1.09 1.17 1.24 1.64 3.09 1.50 3.15RSD% 0.6 0.9 1.4 0.8 2.2 1.8 1.8

Cu Foundvalue 1.49 1.68 4.08 8.85 15.23 3.56 5.38RSD% 0.6 0.6 0.6 2.5 2.3 2.38 0.46

Zn Foundvalue 5.66 4.82 21.27 18.57 37.07 15.33 22.59RSD% 0.6 1.1 2.4 1.3 1.5 1.7 1.7

As Foundvalue N.D. 0.26 0.46 0.79 0.89 0.56 29.14RSD% 1.6 1.4 3.9 3.5 5.7 3.3 1.5

Ag Foundvalue 0.0009 0.0008 0.005 0.007 0.079 0.006 0.007RSD% 5.3 6.3 2.1 5.8 3.5 5.6 4.3

Cd Foundvalue 0.022 0.042 0.10 0.11 0.15 0.077 0.082RSD% 5.5 1.5 1.2 2.7 1.9 3.6 2.0

Ba Foundvalue 5.86 8.94 15.32 44.76 167.8 136.7 12.98RSD% 1.1 1.7 1.4 2.1 2.8 0.3 0.9

Hg Foundvalue 0.003 0.002 0.85 0.088 0.027 0.076 5.05RSD% 6.8 5.8 3.3 5.5 5.0 3.5 0.6

Tl Foundvalue 0.008 0.027 0.025 0.020 0.033 0.036 0.014RSD% 1.3 2.4 2.7 3.6 3.2 1.1 4.1

Pb Foundvalue 0.15 0.16 0.93 1.54 2.14 0.68 1.69RSD% 0.7 1.1 2.6 2.1 1.7 1.5 1.3

5989-5591EN65

References

1.Wang Xiao-ru, “Application ofInductively Coupled Plasma MassSpectrometry.” Chemical IndustryPress. Bei Jing., 102, 2005.

2.Dong Hong-bo, Han Li-qing, DongShun-fu, “Instrumentation & AnalysisMonitoring” (3): 29~30, 2002.

3.Yu Zi-li, Guang Lei, “AnalyticalTechnique of Metal Ion”. ChemicalIndustry Press. Bei Jing, 225~226,2004.

4.Determination of Lead (Pb), Cadmium(Cd), Arsenic (As), Mercury (Hg) andCopper (Cu), Ch. P 2005 (English ver-sion), Appendix IX B, p. 54-56

Hua Zhang is postgraduate student inthe chemistry department of theCapital Normal University, Yan-zhi Shiand Ying-feng Wang are the professorsat the Analytical Center of the CapitalNormal University, Beijing, China. Yu-hong Chen is Application Engineer atAgilent Technologies Co. Ltd, China.

665989-5591EN

Abstract

Five toxic elements in 10 different types of Traditional Chinese Medicines (TCM)

were analyzed by Inductively Coupled Plasma Mass Spectrometry (ICP-MS).

To do this, a sample digestion method and standard operation procedures were

developed. Reproducibility and recovery were tested for method validation.

The Agilent ICP-MS method is proven to be highly sensitive, and its fast acquisition

time makes it highly suitable to analzye trace toxic elements in TCMs.

Binfeng XiaZhuqing Lu

XinMei WangKe Wang

Shen Ji

Application Note

Fast determination of five toxic elements in Traditional Chinese Medicine (TCM) by ICP-MS

Agilent Equipment

Agilent 7500 ICP-MS

Application Areas

Pharmaceutical

5989-7590EN67

IntroductionTraditional Chinese Medicine (TCM)herbs and their manufactured prod-ucts have been used for thousands ofyears for prevention and treatment ofdisease in China. TCM materials aremade up of plants, animals and miner-als with the majority being variousparts of medical plants. The contentsof heavy metals in TCMs have beencommonly studied from the viewpointsof toxicity and bio-availability. Theroots or leaves of the plant can absorbheavy metals such as Pb, Cd, Hg andAs from the atmosphere, water andsoil, and these can finally accumulateat a particular position in the plantthrough the cytosol. Different regula-tions propose to limit the content ofheavy metals in herbs (table 1). The most stringent of these regula-tions is that from the Food and DrugAdministration (FDA). Therefore, it is important to determine these ele-ments as accurately and as unambiguously as possible.

Inductively coupled plasma massspectrometry (ICP-MS) has quicklybecome the technique of choice forthe determination of elements in awide range of samples. It has beenused widely for measurement of traceand ultra-trace elements in the envi-ronmental and biological materials.The advantages of choosing ICP-MSfor measurement include rapid andsimultaneous multi-elementaldeterminations, low spectral interfer-ences, excellent detection limits andthe ability to analyze isotopic ratios.

In this study, the ICP-MS techniquewas applied to measure five toxic ele-ments in TCM samples from differentareas in China.

ExperimentalInstrumentation and reagentsIn this study a 7500c ICP-MS instru-ment was used,which uses anOctopole Reaction System. Sampledigestion was done by microwavedigestion. Concentrated HNO3 (68%w/v, GR, Merck) was used for sampledigestion and cleaning of the digestiontubes. Deionized water was usedthroughout.

Operation parametersThe ICP-MS operation parameters arelisted in table 2.

Sample digestion and calibration

curvesThe TCMs were dried at 60 °C for fourhours, and ground into powder. 0.5 gpowder were weighed precisely into amicrowave digestion tube (50 mL, PFA),and then 10 mL 65 % HNO3 was added.The powdered TCM and HNO3 wereplaced in a microwave digestion sys-tem and subjected to a standard pro-gram of heating. Following digestionthe solubilized TCM samples weretransferred into a PTFE volumetric flaskand the digestion tube was washedthree times using pure water.

Regulations (µg/kg) Pb Cd Hg As

Proposed regulation on “TCM Quality Standard” Chinese Pharmacopeia 5000 500 200 200Chinese regulations on imported herb supplements 5000 300 200 2000French regulations on imported herb supplements 5000 200 100 5000US FDA regulations on the medicines and functional food 1000 300 26 20South-East Asia regulations on imported herb supplements 20000 N/A 500 5000

Table 1Regulations on heavy metals in TCM or herb supplements.

Working parameters

RF power 1350 WNebulizer PFA 200 µL/min nebulizerSpray Chamber Scott double pass 2±0.1 ºCData Acquisition Mode Spectrum Analysis Mode and Full Quant ModeSampling depth 7.0 mmCarrier gas flow rate 1.18 L/minTotal Acquisition Time 90 s

Table 2Instrumental parameters of Agilent ICP-MS.

STD As Pb Cd Cu Hg

Blank 0 0 0 0 01 1.0 1.0 0.5 50 0.22 5.0 5.0 2.5 100 0.53 10.0 10.0 5 200 14 20.0 20.0 10 500 25 5

Table 3Calibration curve range (µg/L).

685989-7590EN

200 µL of a 1 µg/mL Au solution wasalso added into the digested sample tostabilize Hg. The solution was finallymade up to 50 mL. The system blanksolution was made by the same proce-dure except that no TCM was added.The calibration curve standards werealso prepared from 10% HNO3.

The internal standard (1:10 dilutedAgilent Internal Standard Mix, partnumber 5183-4680) was automaticallyadded on-line to samples during analy-sis. Ge (72) was used as internal stan-dard reference element of Cu and As.In (115) was used as internal standardreference element of Cd (114). Bi(209) was used as internal standardreference element of Pb and Hg.

The samples were run according to amodified method based on USEPA200.8


Calibration linear rangeAll five element’s calibration curvesshow good linearity, R2 values arebetween 0.9990 and 0.9999.

Reproducibility of the sample preparationTen TCMs were investigated:Xiyangshen, Danshen, Huangqi,Ganchao, Huangbo, Baishao, Chenpi,Lingzhi, Jinyinhua, and Huanxieyedrugs. All samples were ground intopowder and five repeated samples (0.5 g powder for every sampling) weretaken from 10 batches. Powders weredigested by microwave digestion.Reproducibility was good for all sam-ples. The Xiyangshen medicine istaken here as an example.

The results prove that the ICP-MSmethod and the sampling and diges-tion method are very good for tracetoxic elements analysis, even at lowconcentration levels.

Blank spike recoveriesCd, As, Pb, Cu, Hg were spiked into thedigestion tube, followed by microwavedigestion and then made up to a blankspike recovery solution with concen-trations similar to TCM samples. Thespike recoveries test results areshown in table 5. A total of eight blankspike recoveries solutions were madefor statistical purposes. The results again illustrate thatmicrowave digestion is a very goodmethod for sample digestion withoutloss of target elements during digestion.

Sample spike recoveries 0.25 g powder was taken from each ofthe ten TCMs, six times parallel sam-pling was done for each material.Using the measured concentrations ofthe metals in the TCM guidelines, sixundigested samples were spiked usingstandard stock solutions. All spikeswere done in pairs and the levels cho-sen for spiking were 40 %, 50 % and 60% of the levels found in the originallyanalysed TCMs. Then the sampleswere digested and measured by ICP-MS. The results for the Xiyangshendrug are summarised in Table 6. Eventhough the spike levels were low (withrespect to the elements in the sampleitself) the recoveries are excellent andhighlight that the digestion method iseffective, without loss of material.

No Cu As Cd Hg Pb

1 5962 33.8 94.9 127 68.52 6394 29.7 91.4 141 663 6577 30.7 93.6 157 68.84 6312 32.1 92.2 137 66.95 7000 28 100.9 169 76.8AVG 6449 30.9 94.6 146 69.4RSD(%) 5.9 7.2 4.0 11.4 6.2

Table 4Reproducibilities for Xiyangshen samples (µg/kg).

No Cu As Cd Hg Pb

1 112.3 111.0 97.3 88.3 130.62 120.8 90.6 110.9 101.2 122.73 89.6 94.7 111.0 78.4 120.74 110.3 88.5 108.1 93.2 134.45 91.6 107.8 111.7 80.7 108.46 117.4 89.2 101.8 97.6 111.3Avg 107.0 97.0 106.8 89.9 121.4RSD 12.4 10.2 5.5 10.2 8.5

Table 6Spike recoveries for Xiyangshen after microwave digestion and ICP-MS analysis.

Element Cu As Cd Hg Pb

Avg(%) 98.8 106.0 102.2 110.0 100.7RSD 1.8 1.0 1.1 1.3 1.9

Table 5Blank spike recoveries.

5989-7590EN69


AverageTen types of TCMs were selected astargets, and ten samples were collect-ed from different local areas for eachtype. The samples were analyzed byICP-MS, the average results are listedin table 7. While the data in table 7suggests that all measurements werebelow the regulated levels for heavymetals, please note that these areaverages; some samples were abovethe regulated limits. Interestingly, evenfor the same TCM type, the toxic ele-ment concentrations varied from verylow trace levels to highly contaminat-ed, when the samples are taken fromdifferent sources. For example, Hg con-centrations in 10 Baishao samples var-ied from 5 µg/kg to 14 µg/g, withalmost 3000 times difference.

Statistical results for 100 TCM sam-ples, summarized in tables 8 to 10These results show how many samplesout of 100 have been detected to showhigh toxic elements content.

ConclusionAll samples contained toxic elementsbelow proposed limits, but elementanalysis by ICP-MS of ten differentTraditional Chinese Medicines (seetable 1) from different geographicallocations showed that variability forthe amount of each element was veryhigh. This might correlate with the dif-fering contamination level and differentcontamination types due to soil andwater contamination in different areas.Taking good care when growing plants,such as in Good Agriculture Practice(GAP) of TCMs will be crucial in thefuture to maintain consistent quality.The Agilent ICP-MS has been proven tobe an ideal analytical method for deter-mination of toxic elements in TCMs,since this system can determine tracelevel impurities very efficiently. In com-bination with optimized microwave

digestion, this has reduced the matrixinterferences significantly. Additionally,ICP-MS enables very fast measurementof a full set of elements, saving time forother laboratories tasks.

References1. “Determination of Toxic Elements inTraditional Chinese Medicine UsingInductively Coupled Plasma MassSpectrometry”; Agilent PublicationNumber 5989-5591EN, 2006.

2. This Paper is an extract of a Chinesepublication: Modern Instrumentation,first edition, 2004.

Binfeng Xia, Zhuqing Lu, XinMei Wang,Ke Wang, Shen Ji,Shanghai Institute for Food and DrugControl (China).

TCMs Cu As Cd Hg Pb

Baishao 5403 376 72 1775 280Huangqi 7692 313 27 4805 383Lingzhi 8162 224 65 247 154Jinyinhua 11698 3105 113 127 1885Ganchao 8968 536 12 89 278Danshen 10404 594 69 22 948Xiyangshen 5281 82 104 25 131Huanxieye 5198 276 12 159 249Huangbo 2457 323 14 77 769Chenpi 2842 328 32 166 1298

Table 7Average results (µg /kg).

Elem Range (µg/kg) 2000 3000 5000 10000 20000

Cu 1289~19038 97 83 68 20 0

Table 8Cu in 100 TCMs (sample amounts).

Elem Range (µg/kg) 100 500 1000 2000 5000

Pb 69~3909 92 34 24 6 0

Table 9Pb in 100 TCMs (sample amounts).

Elem Range (µg/kg) 100 200 300 500 1000

As 12~10348 81 62 38 23 14Cd 2.4~212 21 2 0 0 0Hg 0.4~36009 34 18 17 11 5

Table 10As, Cd, Hg in 100 TCMs (sample amounts).

705989-7590EN

71

Quality Control for Residual Solvents

72

AbstractLaboratories using headspace GC for the analysis of pharmaceutical impurities

face a number of instrument-related issues:

• The area precision in headspace analysis is impacted by atmospheric pressure

changes.

• The sensitivity is poor for some low-concentration analytes.

• The presence of high-boiling impurities noticeably extends the analysis time

per sample and may damage the analytical column.

The determination of residual solvents is the most common application for head-

space GC in pharmaceutical quality control. In this Application Note an estab-

lished Agilent 6890 GC method for residual solvents analysis is transferred to the

new Agilent 7890A GC without any major changes. The results on both systems

are compared. Overall, the Agilent 7890A GC system delivers at least the same or

better performance than the Agilent 6890N GC system. Further it is demonstrated

that the new technology implemented in the Agilent 7890A GC can significantly

improve area and retention time repeatability and sensitivity. It can drastically

reduce the overall analysis time, hence increasing sample throughput and thereby

productivity.

Application Note Albert E. GudatRoger L. Firor

Ute Bober

5 10 15 20 25

Agilent Equipment7890A GC system, G1888 headspace sampler

Application AreaPharmaceutical quality control Impurity analysis

Better precision, sensitivity, and highersample throughput for the analysis ofresidual solvents in pharmaceuticals Using the Agilent 7890A GC system with G1888 headspacesampler in drug quality control

5989-6023EN73

IntroductionBecause many solvents pose a majorrisk to human health, national andinternational regulatory bodies such as the United States Food and DrugAdministration (U.S. FDA), the UnitedStates Pharmacopoeia (USP), theEuropean Pharma-copoeia (EP), andthe International Conference onHarmonization (ICH) require analysisfor residual solvents in pharmaceuticaldrug substances, excipients and finalproducts. Solvents are divided intothree classes on the basis of possiblerisk. Class 1 solvents should beavoided. Class 2 solvents should belimited. Class 3 solvents are consid-ered to have low toxicrisk. The ongoing trend toward lowercontaminant levels designated as saferequires more sensitive and accuratemethods of analysis. New USP <467>regulations for residual solvents beginin July 2007. The goal of this initiativeis the final alignment with the ICHQ3C(R3) guideline, which has alsobeen adopted by the EP.

The analysis for residual solvents inpharmaceutical products and for sol-vents considered extracta-bles/leachables in pharmaceuticalpackaging materials is typically doneusing headspace (HS) GC with aflame-ionization detector (FID) or, foridentification and confirmation, withmass-selective detection (MSD). Thishas been covered in references 1-3.Residual solvents is the most commonapplication for headspace GC in phar-maceutical quality control.Laboratories currently using HS GCface a number of instrument-relatedissues:

• The area precision in HS analysis canbe compromised primarily due toatmospheric pressure variationsinfluencing the amount of analytesinjected from the sampling loop inthe HS gas sampling valve (GSV).

• The sensitivity is poor for some low-concentration analytes e.g., benzene.

• Sample turn-around time is excessiveand caused by late-eluting impuritiesand high-boiling solvents/diluentse.g., 1,3-dimethyl-2-imidazolidinone(DMI) with boiling point of 225 oC.

Further, when the need for new analyt-ical equipment arises, the first ques-tion is whether an established validat-ed method can be easily transferred tothe next generation of instrumentswithout any additional method devel-opment and resulting in no or minimalrevalidation effort. The purpose of thisstudy was to at the least demonstrateequivalence of the HS/7890A GC/FIDand HS/6890N GC/FID systems whenboth are operated without pressureregulation of the sampling loop con-tent of the HS GSV. But more impor-tantly, to also show how new capillaryflow technology, fifth generation pneu-matics, and state-of-art electronicsimplemented in the 7890A GC haveeffectively addressed the above issueswith significant improvements in areaand retention time precision, sensitivi-ty and productivity with increasedsample throughput for residual solventanalysis.

Note: A list of acronyms and short-hand terms used in the text, figuresand descriptions of experiments andcalculation formulas are included inthe appendix.

ExperimentalBoth 6890N and 7890A GC wereequipped with an Agilent headspacesampler, volatiles interface (VI) and FID.Table 1 gives the experimental condi-tions used with the HS/VI/6890N/FIDand HS/VI/7890A/FID systems. The7890A system is operated with andwithout backpressue regulation on theHS sampling loop, whereas the 6890Ndoes not have backpressure regula-tion. For the purpose of calculating therepeatability expressed as %RSD val-ues in area and retention time anddetermining statistical MethodDetection Limit (MDL) for each ana-lyte, 20 identical samples were pre-pared. A standard solution in waterwas first prepared in a 100 mL volu-metric flask by adding Restek class 1and class 2 standards with anEppendorf pipette. 5 mL of the aque-ous standard was subsequently trans-ferred quickly to 10 mL headspacevials containing 3 g sodium sulfate andimmediately sealed with Teflon-sealcaps. Each vial was then vortex mixedfor half a minute. Ten of these sampleswere subsequently used with the6890N system and ten with the 7890Asystem, both systems being operatedwithout pressure control on the HSsampling loop. At the end of the HSequilibration, the HS vials were pres-surized to 14.000 psi by an auxiliary(AUX) Electronic Pneumatic Control(EPC) module and injected in eitherthe 6890N GC or the 7890A GC system.The same DB 624 column was used inthe two GCs for the sequence of injec-tions in order to eliminate the influ-ence of batch-to-batch variations incolumn quality.Another set of 20 samples was pre-pared in the same way as described

745989-6023EN

above for use with the 7890A systembut now implementing backpressureregulation on the HS sampling loop.Figures 1, 2 and 3 show diagrams ofthe 7890A system where new pneu-matic features of pressure regulationof the HS sampling loop, HS vial pres-surization and backflush can beapplied. At the end of the HS equilibra-tion, the HS vials were pressurized to20.000 psi by an AUX EPC channel andthe loop was regulated at 5.000 psiwith the backpressure regulator chan-nel of the Pneumatic Control Module(PCM).

7890 GC

30 m x 0.45 mm id x 2.55 μm DB624

FID

DualmodePCM

G1888Headspace sampler

AUXmodule

Vial pressure

HS

vent

Restrictor

VI

HS TR line

Electronic BP control

Forward flow

Effluentsplitter

Figure 1Block diagram of the 7890A GC configuration with backflush capability used in the backflushexperiments.

12

34

5

6

PSVentHS VentVial pressure

from Aux module MFS

PS

He source

To volatiles interface

HS vial

Flow restrictor

Carrier

Pressure difference: (HS vial – BPR) = 10 to 15 PSI

S/SL inlet

Splitter

DB624

0.1 psi

25 psi

FID

AUX EPCSplit vent

trap

Backflush of late eluting solvent ( such as DMA or DMI)

Figure 2Headspace (HS) sampling scheme with backpressure regulation (BPR).

Figure 3Schematic diagram of the reversed column flow used for backflushing oflate eluting solvents.

Table 1Instrument conditions for residual solvents analysis.

6890N or 7890A GC G1888A Headspace Sampler

Injection port Volatiles interface Loop size 1 mLTemperature 160 °C Vial pressure 14.0 psigSplit ratio 2 : 1 Headspace oven 85 °CCarrier gas Helium Loop temp 100 °CCarrier flow 9 mL/min Transfer line temp 120 °CGC oven program Equilibration time 30 min, low shakeInitial temperature 35 °C GC cycle time 50 minInitial time 20 min Pressurization 0.15 minRate 25 °C/min Vent (loop fill) 0.5 minFinal temp 250 °C Inject 0.5 minFinal time 15 minColumn: 30 m x 0.45 mm x 2.55 µm DB-624

Agilent part number 124-1334

Standards

ICH class 1 and 2 Restek #36228 (Class 1)#36229 (class 2A)#36230 (class 2B)

5989-6023EN75

Results from the above experimentsare summarized in table 2. From the 29solvents some representative class 1and 2 solvents were selected and theresults summarized in this table.Further, the average values for all 29solvents are shown. A series of experi-ments was also performed to demon-strate potential sensitivity gains real-ized from pressurizing the HS loop.This time, however, instead of prepar-ing 5 mL aqueous standards in HS vialscontaining the 3 g of sodium sulfate, aRestek class 2B standard was used asis. A 5 µL capillary tube was filled bycapillary flow action with the standard;the outside of the tube was carefullywiped with tissue paper, quickly trans-ferred to an empty 10 mL headspacevial and immediately capped. This pro-cedure ensured accurate and repro-ducible sample preparation by eliminat-ing user bias to the extent possible inpreparing identical samples. Resultsfrom this series of experiments areshown in Figure 7. The error bars inthe figure represent a 95 % confidencelevel (±2 times standard deviation orsigma).

Finally, the system was reconfigured tofacilitate a column backflush to quicklyremove late eluting impurities or highboiling solvent or diluents. This config-uration is shown in figure 3. Ratherthan using the volatiles inlet (VI), asplit/splitless (S/SL) inlet was inter-

faced to the headspace transfer line.The VI configuration is anticipated tohave a limited backflush flow rate withthe 0.45 mm diameter column, a limi-tation not observed with the S/SLinlet. The VI configuration was nottested for backflush operation.

!os;+§sCm?#C#)s9;ss)"?

A typical chromatogram of residualsolvents is shown in figure 4 andrepeatability data for area and reten-tion time of an early- and late-elutingpeak are shown in figure 5. Animprovement in area precision by a

Figure 4Gas chromatogram of class 1 and 2 residual solvents.

5 10 15 20 25 30

1

2

3

4 5 6 7,8

9

10, 11, 12

13,1415

16

17

1819

20

21

22, 23, 24, 25, 26,27, 28 29

1 Methanol2 1,1 Dichloroethylene3 Acetonitrile4 Methylene chloride5 Trans 1,2 dichloroethene6 Hexane7 Cis 1,2 dichloroethene8 Nitrobenzene (co-elute with 7)9 Trichloromethane10 Carbon tetrachloride

11 Cyclohexane (co-elute with 10 & 12)12 1,1,1 Trichloroethane (co-elute with 10 & 11)13 Benzene14 1,2 dimethoxyethane15 1,2 dichloroethane16 Trichloroethylene17 Methyl cyclohexane18 1,4 dioxane19 Pyridine20 Toluene

21 2 hexanone22 Chlorobenzene23 Ethylbenzene24 DMF25 M-xylene26 P-xylene27 O-xylene28 N,N dimethylacetamide29 Tetralin

Excipient7890A GC with BPR* 7890A GC at Atm P (no BPR) 6890N GC at Atm P ( HS valve)

Residual solvents

limit concentration

ExcipientICH Class [ppm] tR Area

MDL#

[ppm]

benzene 1 2 0.014 2.43 0.1 0.012 5.62 0.2 0.010 9.52 0.21,2-dichloroethane 1 5 0.005 4.47 0.7 0.02 8.03 0.5 0.016 8.63 0.51,1-dichloroethene 1 8 0.013 3.24 0.8 0.011 16.63 3.3 0.022 9.82 1.1methylene chloride 2 600 0.009 2.85 54.9 0.016 7.15 61.4 0.018 8.20 62.7hexane 2 290 0.014 4.18 23.1 0.027 7.15 33.2 0.020 10.61 39.1cyclohexane 2 3880 0.042 3.59 341.0 0.012 4.29 299.9 0.018 9.79 501.1trichloroethylene 2 80 0.012 2.69 5.8 0.016 5.29 7.9 0.013 7.91 7.9toluene 2 890 0.025 2.11 46.3 0.024 5.41 85.9 0.031 7.90 90.3ethylbenzene 2 369 0.002 2.27 24.4 0.002 4.90 35.3 0.003 7.40 35.5ortho xylene 2 195 0.001 1.86 9.8 0.001 5.12 19.0 0.002 7.00 18.4Average for 29 solvents 0.013 2.83 0.017 8.77 0.021 9.34

* Backpressure regulation (HS-valve outlet pressure is regulated) # Method detection limit

Repeatability[ %RSD] N=8

tR Area

MDL#

[ppm]


Excipient

tR Area

MDL#

[ppm]


Excipient

Table 2Retention time and area repeatability and calculated MDLs of representative residual solvents for the 7890A and 6890N HS/VI/GC/FID systems.

765989-6023EN

ferences. This could happen whenrunning the same method in differentgeographic locations at different altitudes or like in this case during aturbulent stormy day with large varia-tions in atmospheric pressure. Withbackpressure regulation the gas sam-pling valve can operate under a con-stant set of conditions and precisionand sensitivity improve.

Improving retention time precisionSimilar to the results obtained for peakarea precision a positive impact on therepeatability of retention time wasobserved:• In comparison to the 6890N GC the

averaged retention time repeatabilityfor all 29 residual solvents measuredon the 7890A GC was generally better,no matter whether backpressure regu-lation was applied or not.

factor of 3 was typically observed, asin this example for 1,1-dichloroethyl-ene. However, in some cases up to afactor of 4 was determined, for exam-ple, for o-xylene. Overall, the perfor-mance of the 7890A system is betterthan the 6890N system. A summary ofperformance characteristics for somerepresentative class 1 and 2 analytesis given in table 2.

Improving peak area precisionFrom the results presented in table 2and figure 6 the following conclusionscan be drawn:• Overall, the 7890A and 6890N show

the same area repeatability when nobackpressure regulation of the HSsampling loop is applied. Under thesame conditions the results don’texhibit significant differences. Theaverage area precision [%RSD] for 29residual solvents (shown on the bot-tom line of table 2) for both systemsis 9 %.

• The 7890A GC with backpressureregulation of the HS sampling loop, is at least 3 times better than the7890A GC (or 6890N GC) operatedwithout pressure regulation of the HSsampling loop. For individual analytesan improvement by a factor of 4 wasobserved in some cases as shown infigure 5 for o-xylene.

• The observed differences are evenmore apparent when considering avery turbulent day with large varia-tions in atmospheric pressure. This iswhen a series of measurement wereperformed. The data presented in fig-ure 6 was obtained on such a stormyday. Under extreme weather condi-tions the averaged area repeatabilitycan increase to 16 % when no back-pressure regulation is applied.However, a method with optimizedbackpressure regulation fully com-pensates for the atmospheric pres-sure instabilities and the measuredaveraged repeatability returns to thesame value of 3 % that was obtainedunder stable weather conditions inthe previous experiments. The reasonfor this is that variability in loadingthe gas sampling valve can occurbased on atmospheric pressure dif-

• With backpressure regulation of theHS sampling loop the averaged reten-tion time repeatability measured onthe 7890A GC was best. It improvedby a factor of two relative to the6890N GC without any pressure regu-lation.

Optimizing sensitivity with backpres-sure regulationOverall sensitivity could be doubled byapplying backpressure regulation com-pared to the 6890N or the 7890A GCwithout backpressure regulation.Figure 7 shows the area changes for1,4-dioxane with HS-vial pressurizationand HS sampling loop pressurization. Themore we pressurize the vial, the morewe dilute the HS sample. This is clearlyshown when the x-axis is zero (where

Figure 5Examples of improved area and retention time precision by applying backpressure regulation(backpressure regulation: 5.000 PSI, headspace vial pressure: 20.000 PSI).

Improvement by a factor of ~ 3 Improvement by a factor of ~ 4

3.2 %9.8 %Area

±± 0.013 %0.022 %Retention time

7890 GC6890 GCRepeatability(%RSD)

1.9 %7.0 %Area

±0.001 %±0.002 %Retention time

7890 GC6890 GCRepeatability(%RSD)

min23.2 23.22 23.24 23.26 23.28

pA

0

1000

2000

3000

4000

5000

min1.94 1.96 1.98 2 2.02

pA

14161820222426 N=8 N=81,1-dichloroethylene

1,1-dichloroethylene (8 ppm) o-xylene (195 ppm)

o-xylene

Figure 6Backpressure regulation – Effect of reducing atmospheric pressure variation at vent. Variability inloading the gas sampling valve can occur based on atmospheric pressure differences.

Improvement by a factor of 5

5.000 PSI

10.000 PSI

No

Backpressureregulation

Averaged areaRepeatability (%RSD)

Headspace vialpressurization

2.8% (N=8)20.000 PSI

4.4% (N=7)25.000 PSI

15.8% (N=8)No

5989-6023EN77

the backpressure regulator is not usedand the loop is exposed to atmospher-ic pressure). Pressurizing the HS vialto 14, 35 and 60 psi gives the highestpeak area at 14 psi. When regulatingthe pressure in the loop with the BPR,we see an increasing area count thatreaches a maximum and then decreas-es and eventually would give zero areacounts. Once the top of the curve isreached, the depressurization of theHS vial through the HS loop (the vent-ing cycle) is opposed by the excessivehigh backpressure and the sampleflow through loop will diminish andmay even reverse. After we reach thetop of the curve, we no longer trap arepresentative sample in the loop. Thepressure difference (PHS-Vial – BPR)should be 10 to 15 psi in order to col-lect and inject a proper HS sample.

Increase efficiency with backflushThe backflush capability of the 7890AGC allows to remove late eluting com-pounds by reversing the flow. The ben-efits are:• Shorter analysis time• Extended capillary column life time.

Because this system has EPC, as soonas the last analyte of interest has elut-ed from the column, the AUX modulecan be pressure-programmed to ahigher pressure (25 psi in this exam-ple) at the same time that thesplit/splitless inlet is programmed to alower pressure (0.1 psi in this exam-ple). Now the flow in the column isreversed, backflushing the remainingeluents out through the split vent ofthe inlet. A schematic overview of the functionality is given in figure 3.For the backflush experiment samplescontaining only class 1 residual sol-vents were prepared in DMSO andDMI, both high boiling diluents (figure8). A typical chromatogram of such asample lasting more than 30 minutesis shown in figure 8A. Since all theclass 1 solvents elute in 10 minutes at 35 °C isothermal, a backflush wasinitiated from 10 to 16 minutes with an elevated column temperature of 250°C. As a result the net gain of time per

Equal in linearityTo measure linearity five dilutionswere prepared ranging from 1/10th totwo times the limit concentration. Basedon the USP <467> method where 100mg of the excipient/drug product isdissolved in 5 mL of water with 3grams of Na2SO4, the solution concen-tration in 5 mL of water was convertedto the concentration of the residualsolvent in the 100 mg of excipient with

run was 14 minutes (figure 8B). Thehigh-boiling compounds were success-fully removed as can be seen from thechromatogram of the blank run thatwas executed afterwards (figure 8C).In this example backflush improvesefficiency by almost doubling samplethroughput. The fast oven cool-downof the 7890A GC further contributes tothose time savings.

Figure 7Improving sensitivity with the 7890 GC – variation in peak area with headspace vial and samplingloop pressure.

0

50

100

150

200

0 5 10 15 20 25 30BPR (PSI) on headspace gas sampling valve loop

1,4-

diox

ane

peak

are

a

P(vial) = 14 PSIP(vial) = 35 PSIP(vial) = 60 PSI

Figure 8Example of how backflushing helps to decrease analysis time and increase workload efficiency(chromatogram of a sample containing all ICH class 1 residual solvents): A. Initial situation (no backflush) B. Reduced analysis time using backflush C. Blank run after backflush.

5 10 15 20 25

1

2,3

4 5

6

7

1. 1,1 dichloroethene2. Carbon tetrachloride3. 1,1,1 Trichloroethane4. Benzene5. 1,2 dichloroethane6. DMSO7. DMI (1,3-Dimethy-2-imidazolidinone)

No backflush

Backflush from 10 to 16 min at 250 °C oven temperature

min5 10 15 20 25

Blank run after backflush

min5 10 15 20 25

A

B

C

Net gain of time: 14 min.

785989-6023EN

the formula: ce [ppm] = 50 · cv. In thefollowing text this concentration isdescribed as excipient equivalent con-centration ce while cv is the vial solu-tion concentration.

The linearity results for some repre-sentative residual solvents for the7890A GC are compared to similarexperiments for the 6890N GC1 systemin figures 9 and 10, respectively. Overall, the 7890A and 6890N GC sys-tems compare well. All calibrationcurves are linear over a range from1/10th to 2 times the limit concentra-tion. The signal-to-noise (S/N) data,limit of detection (LOD) and limit ofquantitation (LOQ) indicate that thesystems are similar in performance(only data for the 7890A is shown herein tables 3a and 3b). Results from theMDL calculations (see table 2 and theappendix for the MDL equation) and acomparison of S/N ratios calculatedfor samples at the limit concentrationindicate that the 7890A GC system isat least two times better in sensitivitythan the 6890N GC system.

ConclusionOverall, it was demonstrated, that the7890A GC delivers better results thanthe 6890 GC. In summary:• It was possible to directly transfer an

established method from the 6890 tothe 7890A GC without any methoddevelopment or altering the perfor-mance.

• Without backpressure regulation the7890A GC shows the same or betterperformance.

• The backpressure regulation from the7890A GC eliminates the influence ofatmospheric pressure variations.

• With optimized backpressure regula-tion of the headspace sampling loopof the 7890A GC area precision[%RSD] could be improved by a fac-tor of 3 to 5

• Under the same conditions retentiontime stability increased to ±0.001 min.

• Sensitivity was doubled by pressuriz-ing the headspace sampling loop ofthe 7890A GC.

• The backflush capability of the 7890AGC significantly reduces overallanalysis time (in the example by 50 %).

• Both systems are equal in perfor-mance when evaluating linearitydata.

• The experimental setup in this appli-cation is suitable for the routineanalysis of residual solvents.However, it does not provide any fur-ther information when unknowns arepresent. The solution is to couple theGC to the Agilent 5975B Series MSD,

Figure 9Linearity plots for some residual solvents determined for the 7890A Headspace GC/FID system*

Concentrations shown are excipient equivalent concentration ce [ppm]

LOD (ppm) where S/N = 3LOQ (ppm) where S/N = 10

0200040006000

4 53

1

Methylene chloride

r2= 0.99945

y = 6.09347347x - 228.37038600 Amount [ppm] Amount [ppm]

Amount [ppm]Amount [ppm]

Amount [ppm]

0

2

LOD = 0.02LOQ = 0.06

0 20 4

Benzene

50100150200

53

12

r2 = 0.99859

y = 50.2949812x - 4.6044753

LOD = 10.2LOQ = 10.4

0 5000

100200300

4 53

12

1,4-Dioxane

r2 = 0.99606

y = 0.4214021x + 11.400823

LOD = 2.5LOQ = 8.4

6000

100200300

4 53

1 2

Chloroform

r2= 0.99362 y = 2.71598018x + 5.2606117

0 1000

1000

2000

4 53

12

Trichloroethylene

r2= 0.99967

y = 14.3361657x - 106.99264

LOD = 0.07LOQ = 0.23

LOD = 0.25LOQ = 0.82

Area

Area

Area

Area

Area

*

where you can achieve superiorresults for both identification ofunknowns and quantitation of targetcompounds.

Table 3aLinearity, LOD and LOQ results for the 7890A Headspace GC/FID system.

Methylene Chloride Benzene 1,4-Dioxane Chloroform Trichloroethylene

Linearity 0.99945 0.99859 0.99606 0.99362 0.99967Slope 6.0935 50.2950 0.4214 2.7160 14.3362Intercept 228.3704 4.6045 11.4008 5.2606 106.9926

LOD 10.2 0.02 2.5 0.25 0.07LOQ 10.4 0.06 8.4 0.82 0.23

Table 3bLinearity results for the 6890N Headspace GC/FID system.

Methylene Chloride Benzene 1,4-Dioxane Chloroform Trichloroethylene

Linearity 0.9988 0.9995 0.9996 0.9991 0.9991Slope 252.82 2106.2 15.268 192.41 535.39Intercept 19.987 0.0015 0.3239 0.1851 3.7229

5989-6023EN79

References1.Roger L. Firor, “The Determination ofResidual Solvents in Pharma-ceuticalsUsing the Agilent G1888 NetworkHeadspace Sampler,” AgilentApplication Note, Publication Number5989-1263EN, 2004.

2. Roger L. Firor and Albert E. Gudat, “TheDetermination of Residual Solvents inPharmaceuticals using the AgilentG1888 HS/6890N GC/5975 inert MSDSystem,” Agilent Application Note,Publication Number 5989-3196EN, 2005.

3. Albert E. Gudat and Roger L. Firor, “TheDetermination of Extractables andLeachables in PharmaceuticalPackaging Materials usingHeadspace/GC/MS,” AgilentApplication Note, Publication Number5989-5494EN, 2006.

Figure 10Linearity plots for some residual solvents determined for the 6890N Headspace GC/FID system.

Methylene chloride

y = 252.82x + 19.987R 22

2

2

2

= 0.99880

10002000300040005000

0 5 10 15 20 μg/mL

Area

Chloroform

y = 192.41x + 0.1851R = 0.9991

0100

200

300

400

0 0.5 1.0 1.5 2.0

Area

Benzene

Ry = 2106.2x - 0.0015 = 0.9995

020406080

100120140

0 0.01 0.02 0.03 0.04 0.05 0.06 0.07

Area

Trichloroethylene

y = 535.39x + 3.7229R = 0.9991

0200400600800

100012001400

0 0.5 1.0 1.5 2.0 2.5 3.0

Area

μg/mL

μg/mL μg/mL

1,4 dioxane

y = 15.268x + 0.3239R = 0.9996

04080

120160200

0 2 4 6 8 10 12

Area

μg/mL

* x-axis = vial solution concentration cv

Appendix

Statistical Method Detection Limit (MDL)

MDL = s · tK?PTOCTPm+'(m@&&L= s · 3.143Wheret(n-1, 1-alpha) = Student’s t value for the 99%

confidence level with n-1degrees of freedom

n = number of trialss = standard deviation of the 7 trials

USEPA Method 524.2 (Revision 4, August 1992)

Roger L. Firor and Albert E. Gudat areApplication Chemists at AgilentTechnologies, Inc., USA.Ute Bober is Program Manager atAgilent Technologies, Waldbronn,Germany.

List of acronyms

%RSD – percent relative standard deviation LOD – limit of detection (S/N = 3)[limit] – limit concentration LOQ – limit of quantitation (S/N = 10)Atm P – atmospheric pressure MDL – method detection limit (statistical)AUX – auxiliary MFS – mass flow sensorBP – back pressure min – minutesBPR – backpressure regulation MSD – mass selective detectorCv – vial solution concentration P(Vial) – headspace vial pressureCe – excipient equivalent concentration PCM – pneumatic control moduleCal – calibration PS – pressure sensorDMA – dimethyl acetamide PSI – pounds per square inch DMI – 1,2-dimethyl-2-imidazolidinone RT – retention timeDMSO – dimethyl sulfoxide S/N – signal-to-noise ratioEP – European Pharmacopoeia S/SL – capillary split/splitless inletEPC – electronic pneumatic control TR – transferFID – flame ionization detector U.S. FDA – United States Food and Drug AdministrationGC – gas chromatograph USP – United States PharmacopoeiaGSV – gas sampling valve VI – volatiles inletHS – headspace X – proportional valveICH – International Conference on Harmonization

805989-6023EN

81

Purification and Profiling ofBioactive Compounds

82

Isolation of formononetin and otherphytoestrogens from red clover withthe Agilent 1100 Series purificationsystem

Application

Abstract

Isolating active natural products from plant origin is frequent in the pharmaceu-

tical industry when searching for new drugs. In this Application Note we show

how this task can be accomplished using the Agilent 1100 Series purification

systems AS (analytical scale) and PS (preparative scale)1. The isolation of for-

mononetin and other phytoestrogens from red clover serves as an example to

demonstrate the excellent performance of the Agilent purification solution at

analytical and preparative scale flow rates.

Edgar NägeleUdo Huber

O

HO O

OH

5988-5747EN83

Time [min]0 5 10 15 20

Absorbance[mAU]

0

20

40

60

80

100

120

140

*

*

*

*

*

*

* = other phytoestrogens

Formononetin

Columns ZORBAX SB-C18 3 x 150 mm, 5 µmMobile phases: 0.1 % HOAc in water, 0.1 % HOAc in acetonitrileGradient: 20 % B to 45 % B in 20 min, 45 % B to 100 % B in 1 min

100 % B for 4.5 min, 100 % B to 20 % B in 0.5 minStop time: 25 minPost time: 5 minFlow: 0.7 ml/minInjection: 5 µlColumn temp.: 35 °CUV detector: DAD 260 nm/16 (ref. 800 nm/100) standard flow cell (10 mm pathlength)

Figure 1Chromatogram of crude red clover extract.

Introduction

Estrogens are used to treatmenopause disorders and osteoporo-sis, because these disorders arecaused by a low hormone level.Unfortunately, the steroids used havea high rate of undesired side effects,for example, thrombosis. Comparativestudies between Asian and Westernpopulations showed that these disor-ders and diseases are a lot less com-mon in Asia. This is explained by theAsian soya-based diet, which containshigh levels of phytoestrogens. Phyto-estrogens2 are currently under investi-gation especially because of their impor-tance in hormone replacement therapyand cancer prevention without side effects.

The identification of phytoestrogeneswith isoflavonoid-structure in redclover (Trifolium pratense, L., Legu-minosae) extract using fluorescence3

and UV-visible detection is describedin another Application Note4. In theApplication Note here, we describe theseparation and isolation of formononetinand other phytoestrogens from redclover in analytical and preparativescale using the Agilent 1100 Seriespurification systems AS and PS.

Equipment

The experiments were performedusing the following systems:Analytical scale system:• Agilent 1100 Series vacuum

degasser• Agilent 1100 Series

quaternary pump• Agilent 1100 Series well-plate

autosampler• Agilent 1100 Series thermo-

statted column compartment• Agilent 1100 Series diode array

detector• Agilent 1100 Series fraction

collector AS

Preparative scale system:• Two Agilent 1100 Series

preparative pumps• Agilent 1100 Series

preparative autosampler• Agilent 1100 Series column

organizer• Agilent 1100 Series diode-

array detector• Agilent 1100 Series

preparative fraction collector PS

The systems were controlled using theAgilent ChemStation (rev. A.09.01) andthe Purification/HighThruput software(rev. A.01.01).

845988-5747EN


Extraction25 g dried red clover was extractedultrasonically with 250 ml methanolcontaining 25 ml 0.1 m H2SO4 for threehours. The mixture was removed from the ultrasonic bath and stirredovernight. After filtration the solutionwas evaporated at 40 °C to about 30ml and filtered again.

Analytical method developmentAn analytical scale method was devel-oped based on the method describedpreviously4 to replace the phosphoricacid by acetic acid. Figure 1 shows the resulting chromatogram. Formo-nonetine and several other phytoestro-gens4 are marked in the chromatogram.

Volume overloading experimentSince concentration overloading wasnot possible due to the fixed concen-tration of the extract, volume overload-ing had to be done to isolate the com-pounds. Injecting up to 50 µl crudeextract sample still lead to sufficientseparation for analytical scale purifica-tion (figure 2).

0 5 10 15 20 25

Absorbance[mAU]

0

1000

2000

3000

4000

50 µl

30 µl

5 µl

Time [min]

Figure 2Volume overloading experiment.

5988-5747EN85

Isolation of phytoestrogens in analytical scale. A common method to isolate com-pounds from complex natural extractsis fractionation by time slices. Becauseof the good separation achieved in theoverloading experiments, peak-basedfractionation was used for the redclover extract. The chromatogram isshown in figure 3 – vertical lines indi-cate the collected fractions. The ana-lytical method described in figure 1was used with an injection volume of50 µl. Fractions were collectedbetween 6 and 20 minutes, based onthreshold only (500 mAU).

Isolation of higher amountsTo gain more phytoestrogen material,pooling of fractions from several runswas carried out. This means repetitiveinjections were performed from onesample vial and the resulting fractionswere collected in the same fractionvials. The pooling feature is describedin detail in the User’s Guide5. 450 µl ofsample was injected in nine 50-µlinjections and the resulting fractionswere pooled automatically. Reanalysisof the fractions showed good resultswhich demonstates the excellent per-formance of the instrument and soft-ware (figure 4).

Time [min]0 5 10 15 20

Absorbance[mAU]

0

200

400

600

800

1000

Formononetin

*

*

* *

*

*

* = other phytoestrogenes

Figure 4Re-analysis of fractions from pooling.

0 5 10 15 20

Absorbance[mAU]

0

500

1000

1500

2000

2500

Time [min]

Figure 3Analytical scale fractionation of red clover extract

865988-5747EN

Another possibility to purify morematerial is to scale-up to a larger col-umn. Based on the analytical scale col-umn overloading experiment (figure 2),scale-up calculations were done toinject 450 µl in one single injection.This was achieved on a 9.4 x 150 mmcolumn at a flow rate of 7 ml/min.Since the Agilent 1100 Series well-plate autosampler AS can only be usedup to a flow rate of 5 mL/min purifica-tion was transferred to a purificationsystem PS. Figure 5 shows the chro-matogram that was obtained. Thelower peak heights and areas resultfrom the shorter pathlength of thepreparative flow cell (3 mm) comparedto the standard flow cell (10 mm). Re-analysis of fractions showed compara-ble purities as for the pooling experi-ment. The injected sample volumeswere the same (450 µl) for both thepooling experiments and when usingthe 9.4 mm id column. The solventamounts used were also comparable(approximately 200 ml). However, themain advantage of purification on the9.4-mm id column is the time that issaved in gaining the same amount ofpurified material.

Isolation of phytoestrogens in prepar-ative scaleTo purify even higher amounts of sam-ple the method was further scaled upto a 21.2 x 150 mm column. At thisscale it was possible to inject 2300 µlof sample in a single injection. The

Figure 5Preparative scale fractionation of red clover extract.

Time [min]0 5 10 15 20

Absorbance[mAU]

0

500

1000

1500

2000

2500

Columns: ZORBAX SB-C18 9.4 x 150 mm, 5 µmMobile phases: 0.1 % HOAc in water, 0.1 % HOAc in acetonitrileGradient: 20 % B to 45 % B in 20 min, 45 % B to 100 % B in 1 min

100 % B for 4.5 min, 100 % B to 20 % B in 0.5 minStop time: 25 minPost time: 5 minFlow: 7 ml/minInjection: 450 µlColumn temp.: ambientUV detector: DAD 260 nm/16 (ref. 800 nm/100) preparative flow cell (3 mm pathlength)Fraction collection: based on threshold only (100 mAU) between 5.5 and 21 minutes

5988-5747EN87

References

1. “New perspectives in purification withHPLC and HPLC/MS” AgilentTechnologies Brochure publicationnumber 5988-3673EN, 2001.

2. Julia Barrett, “Phytoestrogen: Friendsor Foes”, Environmental HealthPerspectives, (104), 478, 1996.

3. “Sensitive and Reliable FluorescenceDetection for HPLC”, AgilentTechnologies Brochure, publicationnumber 5968-9105E, 1990.

4. “Separation of phytoestrogenes in redclover by reverse phase HPLC with UV-visible and fluorescence detection”,Agilent Technologies Application Note,publication number 5988-2399EN 2001.

5. “Agilent 1100 Series PurificationSystem”, Agilent Technologies User'sGuide, part number G2262-90001, 2001.

chromatogram and the method areshown in figure 6. Re-analysis of frac-tions showed that scale-up was possi-ble without losing any performance withregard to the purity of the fractions.

Conclusion

In this Application Note we showedthe development of an analytical scalemethod to separate the compounds ina complex crude plant extract fromred clover containing formononetinand several other phytoestrogen usingthe Agilent 1100 Series preparativesystem AS. Based on this method ananalytical scale preparative separationwith peak-based fraction collectionwas carried out. To obtain more phy-toestrogen material the pooling fea-ture of the Agilent 1100 Series purifi-cation system AS was used. The purity of the compounds gainedwas determined by the reanalysis ofthe fractions. Based on the analyticalscale results the method was scaledup. The purification was repeated ontwo different columns on the Agilent1100 Series purification system PS togain higher amounts of the desiredcompounds in single runs. The resultsin preparative scale were comparableto the results achieved on the analyti-cal scale system.

Columns ZORBAX SB-C18 21.2 x 150 mm, 5 µmMobile phases: 0.1 % HOAc in water, 0.1 % HOAc in acetonitrileGradient: 20 % B to 45 % B in 20 min, 45 % B to 100 % B in 1 min

100 % B for 4.5 min, 100 % B to 20 % B in 0.5 minStop time: 25 minPost time: 5 minFlow: 35 ml/minInjection: 2300 µlColumn temp. ambientUV detector: DAD 260 nm/16 (ref. 800 nm/100) Preparative flow cell (3 mm)Fraction collection: based on threshold only (100 mAU) between 5.5 and 21 minutes

Edgar Nägele and Udo Huber are application chemists at Agilent Technologies GmbH,Waldbronn, Germany.

885988-5747EN

0 5 10 15 20

Absorbance[mAU]

0

500

1000

1500

2000

2500

Time [min]

Figure 6Fractionation on a 21.2-mm preparative column.

Analysis of a complex natural product extractfrom ginseng – Part I: Structure elucidation of ginsenosides by rapid resolution LC–ESITOF with accurate mass measurement

Abstract

Since prehistoric times extracts from herbs have been used for medical treat-

ment of disease. Their activity and effects on humans were found by trial and

error over generations. A good example of achieved efficiency is traditional

Chinese medicine (TCM). In Western medicine drugs derived from natural origin

are gaining importance due to their potential. However, Western pharmaceutical

quality standards require a deep knowledge about the ingredients in medicine

based on natural products.

This Application Note will demonstrate the use of the rapid resolution LC with

rapid resolution high throughput (RRHT) columns for the separation of the ingre-

dients found in a complex ginseng root extract and the measurement of accu-

rate masses by an ESI orthogonal acceleration time-of-flight MS (oaTOF) for the

structure elucidation. The use of a rapid resolution LC system together with an

ion trap MS for structure elucidation will be discussed in part two of this work8.

Application Note Edgar Nägele

5989-4506EN89

medicine based on natural products.Since traditional herbal medicinesoften contain hundreds of substanceswith only a few bioactive compoundsit is necessary to develop new strate-gies to screen these plant extracts forbiologically active compounds and fortheir pharmacological effectivenesson animal or cellular models as wellas receptor and enzyme based tests4.A famous Asian herb, which has beenused in herbal medicine for more than5000 years is the ginseng (Panaxspecies) root. Pharmacological effectsof ginseng which have been reportedare, for example, stimulatory andinhibitory effects on the central ner-vous system (CNS), antistress, antihy-perglycemic, antineoplastic andimmunomodulatory effects5. The mainactive compounds of the ginsenosidesare triterpene saponins from whichmore than 80 have been isolated andcharacterized during the past years.An enormous amount of work hasbeen done during the last 30 years todevelop analytical methods for theanalysis of ginseng extracts and med-ical formulations. The method ofchoice for the analysis of complexnatural product extracts, like thosederived from the ginseng root, is highperformance liquid chromatography(HPLC)6. LC/MS equipment such asLC/ESI oaTOF for accurate mass mea-surement and LC/ion trap or LC/triplequadrupole instruments for structureelucidation by MS/MS and MSn arecurrently being used to determine thecomplex and similar structures of gin-senosides7. This Application Note willdemonstrate the use of the Agilent1200 Series Rapid Resolution LC sys-tem with RRHT columns for separat-ing the ingredients found in a complex

ginseng root extract and the measure-ment of accurate masses by an ESIoaTOF MS. For the structure elucida-tion CID fragmentation was carried outand the measured accurate masseswere used to calculate empirical for-mulas of the fragments. The use of thehigh resolution LC system togetherwith an ion trap MS for structure eluci-dation by MSn will be discussed inpart two of this work8.

Experimental

Equipment• Agilent 1200 Series binary pump SL

with degasser. This pump has thecapability to deliver a pressure of up to600 bar, which is necessary to per-form high resolving HPLC analysis ona 1.8-µm particle size RRHT columnto get the best resolution perfor-mance.

• Agilent 1200 Series high performanceautosampler SL with thermostat. Thisautosampler is especially designed to worktogether with the 1200 Series binarypump SL at the lowest delay volume.

• Agilent 1200 Series thermostattedcolumn compartment (TCC). This TCCis ready for use together with thehigh pressure binary pump withoptional separate heat exchangersand post column cooling under opti-mized delay volume conditions andalternating column regeneration withan optional 2-position/10-port valve.

• Agilent 1200 Series diode-arraydetector SL (DAD). This DAD is capa-ble of acquiring data at a samplingrate up to 80 Hz. This device has abuilt-in data storage capability.

Introduction

Crude extracts of herbal and animalorigin have been used for medicaltreatment of disease in all ancient cul-tures around the globe since prehis-toric time. Their activity for treatmentof different diseases and other effectson humans were found by trial anderror over hundreds of years and theknowledge about this medicine wasinherited from generation to genera-tion. A good example of the efficiencyachieved during this process of opti-mization is the herbal based traditionalChinese medicine (TCM). Since thesedrugs are often complex mixtures con-taining hundreds of chemically variedsubstances with different effects orsynergisms a quality control and quali-ty assurance of medical potency isvery difficult. A widely used andaccepted method by the U.S. Food andDrug Administration (FDA) and WorldHealth Organization (WHO) is chro-matographic fingerprinting1,2.In addition to the importance of thechromatographic fingerprint, which iscapable of identifying a particular herband distinguishing between closelyrelated species in a qualitative manner,the quantitative analysis of medicalplant extracts is also gaining impor-tance. For traditional medicines aquantitative analysis is crucial to theirquality control and to determine theabsolute content of pharmaceuticaleffective substances as well as poten-tially toxic undesirable natural sub-stances3. In Western medicine drugsderived from natural origins are gain-ing importance due to their potential.However Western pharmaceuticalquality standards require a deepknowledge about the ingredients in

905989-4506EN

• Agilent 6200 Series MSD TOF.Orthogonal acceleration time-of-flight mass spectrometer with dualsprayer interface for mass calibrationto acquire molecular masses withhighest accuracy. This time-of-flightmass spectrometer is capable ofacquiring data at 40 Hz and pos/negswitching.

• Picard TOF software A02.00.Software used for data acquisitionwith the TOF LC/MS system.

• Analyst Software. Software for TOFand UV data analysis.

• Column. ZORBAX SB C18, 2.1 x 150 mm, 1.8 µm

SamplePowdered freeze-dried Asian ginsengroot (1g) (Panax ginseng) was treatedultrasonically for 30 minutes in 10 mLmethanol, filtered and directly used foranalysis.

The set-up of the Agilent 6200 SeriesMSD TOF system is shown in figure 1.In this set-up the Agilent 1200 Seriesbinary pump SL is connected to theAgilent 1200 Series high performanceautosampler SL with a 0.17-mm i.d.stainless steel capillary. To reducedelay volume, the seat capillary in theAgilent 1200 Series high performanceautosampler SL has an 0.12-mm i.d.,which is the same kind of capillarythat connects the low delay volume(1.4 µL) heat exchanger in the Agilent1200 Series thermostatted columncompartment to the column. A 2-µLcell is built into the Agilent 1200Series diode-array detector SL for UVdetection. The outgoing capillary isconnected directly to the sprayer ofthe electrospray source at the time-of-flight mass spectrometer, which iscapable of aquiring spectra at 40 Hz.This instrument set-up is optimized to

achieve the highest possible resolu-tion, which is demonstrated by the UVanalysis of a complex natural productextract obtained from Asian ginsengroot (Panax ginseng) (see below). Thefull performance of the Agilent 1200Series binary system in the high reso-lution configuration is demonstratedin another publication9. It is also pos-sible to use this system in a highthroughput environment by makingminor changes10.

Methods:• The Agilent 1200 Series binary pump

SL was operated under the followingconditions: Solvent A: water + 0.1 % TFASolvent B: AcN + 0.1 % TFAFlow: 0.5 mL/minGradient: 0 min 5 % B,

1 min 5 % B, 60 min 85 % B, 61 min 95 % B, 70 min 95 % B

Stop time: 70 min Post time: 15 min

• The Agilent 1200 Series high perfor-mance autosampler SL was used tomake injections of 10 µL sample andthe samples were cooled to 10 ºC.The sample loop was switched tobypass after 1 minute to reducedelay volume.

• The Agilent 1200 Series thermostat-ed column compartment, equippedwith the 1.4-µL low delay volumeheat exchanger, was set at 50º C.

• The Agilent 1200 Series diode-arraydetector SL was operated at 80 Hzfor data acquisition at a wavelengthof 220 nm/4, ref. 360/100 with a 2-µL flow cell, 30-mm path length.

• Agilent 6200 Series MSD TOF wasoperated under the following condi-tions: Source: ESI in positive mode with

dual spray for reference mass Dry gas: 12 L/min Dry Temp.: 200 ºC Nebulizer: 35 psi Scan: 200-1300 Fragmentor: 150 V or 300 V for CID Skimmer: 60 V Capillary: 3000 V

Thermostat

Column compartment

DAD SL

Degasser

650 mm x 0.17 mm capillary 320 mm x 0.12 mm capillary

2 -μL UV flow cell

Low delayvolume heatexchanger

High-performance autosampler SL

Binary pump SL

1.8 -μm SB C18column

To MS

Figure 1Agilent 1200 Series binary LC system for MS using a low delay volume configuration.

5989-4506EN91

the species dependent ginsenosidesRf and F11 are discussed in anotherpublication11.

The ginsenoside Rb1 eluting at 27.7min was identified by its protonatedmolecular ion at m/z 1109.6129 andcalculation of the correspondingempirical formula with a mass accura-cy of -1.90 ppm relative to 2.10 mDa. Aloss of one molecule of water leads tothe ion at m/z 1091.6012 with -0.91ppm mass accuracy. A CID experimentwas performed for structure elucida-tion at an elevated skimmer voltage,

After separation of the individual com-pounds, which are components of thecrude complex extract from the gin-seng root, by means of rapid resolutionHPLC on a 1.8-µm particle size RRHTcolumn, they are subjected to accuratemass measurement by means of anESI oaTOF. The main constituentselute between 20 and 33 minutes andare marked as ginsenosides Re, Rf, F11,Rb1, Rc, Rb2 and Rd in the base peakchromatogram (figure 3). The ESIoaTOF data of the ginsenosides Rb1,Rc and Re were analyzed in moredetail for structure elucidation while


To compare the resolution perfor-mance of the Agilent 1200 SeriesRapid Resolution LC system to anAgilent 1100 Series standard LC sys-tem, the analysis of a complex naturalproduct extract was performed on anAgilent 1100 Series system with itsmaximum pressure at 400 barequipped with a 5-µm particle size col-umn and in comparison on an Agilent1200 Series system with its maximumpressure at 600 bar equipped with aRRHT column with a 1.8-µm particlesize. The system backpressure wasabout 520 bar using the 2.1 x 150 mm,1.8-µm column. The resulting UV chro-matograms acquired at 220 nm clearlydemonstrate the better resolution ofthe peaks on the Agilent 1200 Seriessystem (figure 2). The peak width(FWHM) of the majority of the peaksin the UV chromatogram is below 0.1 min with baseline separation. Thisexcellent resolution, which can beachieved with the Agilent 1200 Seriespump in combination with the RRHTcolumn results in significantly moreMS information useful for compoundidentification.

min13 14 15 16 17 18

mAU

50

100

150

200

250

300

350Agilent 1100 Series

400 bar system

Agilent 1200 Series

600 bar system

min13 14 15 16 17 18

50

100

150

200

250mAU Column:

2.1 x 150 mm ZORBAX SB C-18, 1.8 µm

Column: 2.1 x 150 mm ZORBAX SB C-18, 5 µm

17.

136

15.

411

15.

967

15.

159

15.

555 1

5.06

5

16.

746 1

5.47

2

15.

790

15.

079

14.9

3714

.787

Figure 2Ginseng extract using a 400 bar system with a 5-µm particle column and using a 600-bar systemwith a 1.8-µm particle column.

Conditions for both experiments: PumpsSolvent A: H2O + 0.1 % TFASolvent B: ACN + 0.1 % TFAGradient: 10 % to 95 % ACN in 40 min,

hold for 1 minFlow rate: 0.4 mL/minAutosamplersInjection volume: 3 µLThermostatted column compartmentsTemperature: 50° CDetectorsDAD 2-µL cell and 20 Hz, 220 nm

925989-4506EN

which presented additional informa-tion in the TOF spectrum (figure 4).The loss of one of the glucose chainsresulted in the ions at m/z 785.5047and m/z 325.1136 with mass accura-cies of 0.53 ppm and -0.39 ppm,respectively. A series of ions resultingfrom a consecutive loss of water fromthe ion m/z 785.5047 was also identi-fied with high mass accuracy. Furtherloss of the second glucose chain fromthe molecule resulted in the ion at m/z425.3784 with -0.14 ppm mass accura-cy. This ion is derived from the triter-penoid core structure common to allginsenosides. A loss of one moleculeof water resulted in the ion at m/z407.3679 with 0.10 mDa and -0.30 ppmmass accuracy. This set of CID frag-ments together with the molecular ionand the set of calculated empirical for-mulas confirm the structure of the gin-senoside Rb1. The fragmentation pat-tern is shown in figure 4 and the massaccuracies and empirical formulas aresummarized in table 1.

19.0 21.0 23.0 25.0 27.0 29.0 31.0 33.0 35.0Time, min

0.0

1.0e6

2.0e6

3.0e6

4.0e6

5.0e6

6.0e6

7.0e6

8.0e6

9.0e6

Inte

nsity

cps

Rb1

RcRb2

Rd

Rf

F11

Re

Figure 3RR-LC-TOF basepeak chromatogram of the area containing the main components of the Asianginseng root extract.

O

CH3OH

CH3CH3

CH3

CH3

CH3 CH3

O

OCH3

O

O

O

OH

OH

OH

O

OH

OH

OH

OH

OHOH

OH

O

OH

OH

OH

OH 425.3784

C30H49O = 425.3784

-H2O

C30H47 = 407.3679

785.5047

325.1136

343.1248

C42H73O13 = 785.5047

-H2O

C42H71O12 = 767.4950

-H2O

C42H69O11 = 749.4850

100 200 300 400 500 600 700 800 900 1000 1100m/z, amu

0.04.0e48.0e41.2e51.6e52.0e52.4e52.8e53.2e53.6e54.0e54.4e54.8e5

425.3784

407.3679

1109.6129

325.1136

785.5047

767.4950

749.4854

343.12481091.6012

Inte

nsity

, cou

nts

Figure 4Accurate mass measurement of ginsenoside RbT and its CID fragments by ESI oaTOF.

Table 1 Empirical formulas and achieved mass accuracies for the structure elucidation of ginsenoside RbTby ESI oaTOF.

Measured mass Calculated mass Formula Mass accuracy Mass accuracy[mDa] [ppm]

1109.6129 1109.6108 C54H93O23 2.10 -1.901091.6012 1091.6002 C54H91O22 1.00 -0.91785.5047 785.5051 C42H73O13 -0.40 0.53767.4950 767.4946 C42H71O12 0.40 -0.58749.4854 749.4840 C42H69O11 -1.40 1.88425.3784 425.3783 C30H49O 0.10 -0.14407.3679 407.3678 C30H47 0.10 -0.30343.1248 343.1240 C12H23O11 0.80 2.23325.1136 325.1135 C12H21O10 0.10 -0.39

5989-4506EN93

m/z 947.5585 with a high mass accura-cy of 0.60 mDa or -0.59 ppm (figure 6).The CID fragmentation gave furtherevidence for the identity of this sub-stance. After cleavage of the saccha-ride moieties there are two remainingmain fragments. The fragmentobtained after a loss of a molecule of glucose with m/z 767.4957 wasmeasured with a mass accuracy of -1.50 ppm. A consecutive loss of wateryields the ion at m/z 749.4855. After afurther loss of the glucose disaccha-ride the fragment of the triterpenoide

min retention time are structure iso-mers with the same empirical formulaand molecular mass. In the ginseno-side Rc the sugar arabinose is in thefuranose form and in the ginsenosideRb2 the arabinose is in its pyranoseform.

The last of the main ingredients, whichwas classified in the examined gin-seng root extract is ginsenoide Re.This compound, which elutes at 20.8 min, was identified by empiricalformula calculation from the mass at

The ginsenoside Rc with 29.3 minretention time was identified afterempirical formula calculation by itsmolecular ion at m/z 1079.5991 and aproduct obtained by the loss of a mole-cule of water at m/z 1061.5888 withmass accuracies of 1.02 ppm and 0.78ppm, respectively. The fragmentationobtained at a higher skimmer voltagestarts with a loss of the sugar residuesarabinose and glucose followed by aloss of water from the remaining mole-cule fragment (figure 5). The fragmen-tation of the sugar moieties resulted inthe initial ions at m/z 785.5060 and atm/z 755.4936 with mass accuracies of-1.12 ppm and 1.26 ppm, respectively.After complete loss of all saccharidemoieties the triterpenoide core struc-ture was detected at m/z 425.3785with high mass accuracy -0.37 ppm forthe calculated empirical formula. Thecleaved glucose chain was detected at m/z 325.1124 with 0.22 pmm massaccuracy. The fragmentation pattern isshown in figure 5 and the achievedmass accuracies for all fragments aresummarized in table 2. The ginseno-sides Rc and Rb2, which elute at 28.7

Table 2Empirical formulas and achieved mass accuracies for the structure elucidation of ginsenoside Rcby ESI oaTOF.


1079.5991 1079.6002 C53H91O22 -1.10 -1.021061.5888 1061.5896 C53H89O21 -0.88 0.78785.5060 785.50.51 C42H73O13 0.90 -1.12767.4939 767.4946 C42H71O12 -0.70 0.85749.4846 749.4840 C42H69O11 0.60 -0.81755.4936 755.4946 C41H71O12 -1.00 1.26737.4830 737.4840 C41H69O11 -1.00 1.34719.4723 719,4734 C41H67O10 -1.10 1.56425.3785 425.3783 C30H49O 0.20 -0.37407.3679 407.3678 C30H47 0.10 -0.30325.1134 325.1135 C12H21O10 -0.10 0.22

150 250 350 450 550 650 750 850 950 1050m/z, amu

0.0

2.0e4

4.0e4

6.0e4

8.0e4

1.0e5

1.2e5

1.4e5

1.6e5 1079.5991425.3776

1061

.588

8

407.

3678

755.

4936

325.

1134

767.

4939

737.

4830

719.

4723

785.

506074

9.48

46

O

CH3OH

CH3CH3

CH3

CH3

CH3 CH3

O

OCH3

O

O

O

OH

OH

OH

O

OH

OH

OH

OH

OHOH

OH

O

OHOH

OH

425.3785

C30H49O = 425.3785-H2O

C30H47 = 407.3679

425.3785

767.4939

737.4830

785.5060

325.1134

755.4936

C42H73O13 = 785.5060

-H2O

C42H71O12 = 767.4939

-H2O

C42H69O11 = 749.4846

C41H71O12 = 755.4936

-H2O

C41H69O11 = 737.4830

-H2O

C41H67O10 = 719.4723

Inte

nsity

, cou

nts

Figure 5Accurate mass measurement of ginsenoside Rc and its CID fragments by ESI oaTOF.

945989-4506EN

system equipped with a 1,8-µm parti-cle size column. Complex structures ofthree ginsenosides, which are themain compounds of the extract, couldbe elucidated by the interpretation ofthe obtained TOF spectra. The pro-posed structures were confirmed byempirical formula calculation of themolecular ions as well as for the frag-ments of the molecules obtained byCID. All measured masses have accu-racies in the lower single digit rangeand therefore confirm the structureswith highest confidence. With thisdetailed knowledge about the ingredi-ents of the natural plant extract it is

possible to use the high resolutionLC/TOF technology to monitor thecontent of an extract before usage in apharmaceutical formulation. This isachieved in a fully automated mannerand confirms the compound’s identityby measuring its mass with highestaccuracy and empirical formula confir-mation.

core structure at m/z 441.3739 with amass accuracy of 0.35 ppm occurred.After a further successive loss ofwater, this fragment yields to the ionsat m/z 423.3625 and m/z 405.3519with 0.45 ppm and 0.55 pmm respec-tively. The calculated empirical formu-las and mass accuracies of the mea-sured fragments are summarized intable 3.

Conclusion

The featured application with theAgilent 1200 Series Rapid ResolutionLC system together with the Agilent6200 Series MSD TOF proves its capa-bility for use in structure elucidation ofnatural products in highly complexextracts from plant origins. In thisapplication a highly complex extractfrom ginseng root was analyzed withthe Agilent 1200 Rapid ResolutionLC/TOF system. A comparison to thepredecessor Agilent 1100 Series sys-tem clearly demonstrated the higherresolution that is achieved by the new

O

CH3OH

CH3CH3

CH3

CH3

CH3 CH3

OCH3

OHOH

OH

OH

O

OH

O

O

OH

OH

OH

O

OHOH

OH

CH3

441.3731

C30H49O2 = 441.3733

-H2O

C30H470 = 423.36275

-H2O

C30H470 = 405.35.19

441.3731

767.4946

100 200 300 400 500 600 700 800 900m/z, amu

0.0

4.0e4

8.0e4

1.2e5

1.6e5

2.0e5

2.4e5

2.8e5

3.2e5

3.6e5

4.0e5423.3625

441.3731405.3519

947.5585

767.4957749.4855

Figure 6Accurate mass measurement of ginsenoside Re and its CID fragments by ESI oaTOF.

Table 3Empirical formulas and achieved mass accuracies for the structure elucidation of ginsenoside Reby ESI oaTOF.


947.5585 947.5579 C48H83O18 0.60 -0.59767.4957 767.4946 C42H71O12 1.10 -1.50749.4855 7494840 C42H49O11 1.50 -2.01441.3731 441.3733 C30H49O2 -0.20 0.35423.3625 423.3627 C30H47O -0.20 0.45405.3519 405.3521 C30H45 -0.20 0.55

5989-4506EN95

Edgar Nägele is Application Chemist atAgilent Technologies, Waldbronn,Germany.

References

1. U.S. Food and Drug Administration,“Guidance for Industry botanical DrugProducts”, 2000.

2. World Health Organization, “GeneralGuidelines for Methodologies onResearch and Evaluation of TraditionalMedicine”, 2000.

3. Drasar P., Moravcova J. “Recentadvances in analysis of Chinese med-ical plants and traditional medicines.”,J. Chrom. B 1-2, 812, 3–21,

4. Huang X., Kong L., Li X., Chen X., GuoM., Zou H. „Strategy for analysis andscreening of bioactive compounds intraditional Chinese medicines.”, J. Chrom. B 1-2,812, 71-84, 2004.

5. Attele A.S., Wu J.A., Yuan C.S.Biochem. Pharmacol. 58, 1685-1693,1999.

6. Fuzzati N. “Analysis methods of gin-senosides.”, J. Chrom B1-2, 812, 114-133, 2004.

7. Wang X., Sakuma T., Asafu-Adjaye E.,Shiu G. K. „Determination of ginseno-sides in plant extracts from Panax gin-seng and Panax quinquefolius L. byLC/MS/MS.”, Anal. Chem. 71, 1579-1584, 1999.

8. Naegele, E., “Examination of a com-plex natural product extract from gin-seng - Part II: Structure elucidation ofginsenosides by high resolution LC-iontrap by MSn” Agilent ApplicationNote, publication number 5989-4705EN, 2006.

9. “Performance of Agilent 1200 SL LCsystem for highest resolution.” AgilentApplication Note, publication number5989-4489EN, 2006.

10. “Performance of the Agilent 1200 SLHPLC System for Ultra-Fast LCApplications with 2.1-mm i.d.columns.” Agilent Application Note,publication number 5989-4502EN,2006.

11. Naegele, E, “Examination of a complexnatural product extract from ginseng -Part III: Species differentiation of gin-seng and authentification of ginseng products by LC/MS”, AgilentApplication Note, publication number5989-4706EN, 2006.

965989-4506EN

Analysis of a complex natural productextract from ginseng – Part II: Structure elucidation of ginsenosides by high resolution ion trap LC/MS

Abstract

Since prehistoric time extracts from herbs have been used for medical treatment

of disease. Their activity and effects on humans were found by trial and error

over generations. A good example of achieved efficiency is traditional Chinese

medicine (TCM). In Western medicine drugs derived from natural origins are

gaining importance due to their potential. However, Western pharmaceutical

quality standards require a deep knowledge about the ingredients in medication

based on natural products. This Application Note will demonstrate the use of

the Agilent 1200 Series Rapid Resolution LC system with Rapid Resolution High

Throughput (RRHT) columns for the separation of the ingredients found in a

complex ginseng root extract. The information obtained with an ion trap MSn

and MRM is used to determine the structure of the ingredients and for quality

control purposes.

Application Note

Edgar Nägele

5989-4705EN97

This Application Note will demon-strate the Agilent 1200 Series RapidResolutioin LC system with 1.8-µmcolumns for the separation of theingredients found in a complex gin-seng root extract. The informationobtained with an ion trap MSn andMRM is used to determine the struc-ture of the ingredients and for qualitycontrol purposes. The use of a highresolution LC system together with anESI oaTOF for accurate mass mea-surement is described in Part I in thisseries of Application Notes4.

Experimental

Equipment• Agilent 1200 Series binary pump SL

with a degasser. This pump has thecapability to perform high resolutionHPLC analysis on a 1.8 -µm particlesize RRHT column and achieve thebest performance.

• Agilent 1200 Series high perfor-mance autosampler SL with a ther-mostat. This autosampler is espe-cially designed to work with theAgilent 1200 Series binary pump SLto ensure lowest delay volumes.

• Agilent 1200 Series thermostattedcolumn compartment (TCC). The TCCis ready for use with the Agilent1200 Series binary pump SL andoptional separate heat exchangers, aswell as post column cooling, underoptimized delay volume conditions,together with alternating columnregeneration with an optional 2-position/10-port valve.

• Agilent 1200 Series diode arraydetector SL (DAD). This DAD iscapable of acquiring data with asampling rate of up to 80 Hz. In case

of network problems, this device hasa built-in data storage capability.

• Agilent 6330 Ion Trap LC/MS.The iontrap is operated with a standard ESIsource and able to scan with 26,000m/z /sec. The MSn spectra areacquired data dependent and fullyautomated.

• The software used for instrumentcontrol was ChemStation B01.03, iontrap software 5.3, and for data analy-sis the ion trap data analysis soft-ware 3.3.

• Column: ZORBAX SB C18, 2.1 x 150 mm, 1.8 µm

SamplePowdered freeze-dried Asian ginseng root (1g) (Panax ginseng) wastreated ultrasonically for 30 minutes in10 mL methanol, filtered and directlyused for analysis.

The set-up of the Agilent 6330 Ion TrapLC/MS system is shown in figure 1.The Agilent 1200 Series binary pumpSL is connected to the Agilent 1200Series high performance autosamplerSL (ALS SL) with a 0.17-mm i.d. stain-less steel capillary. To reduce delayvolume, the seat capillary in the ALSSL has an i.d of 0.12 mm. The samekind of capillary connects to the lowdelay volume (1.6 µL) heat exchangerin the Agilent 1200 Series thermostat-ted column compartment, which isconnected to the column. A 2-µL cell isbuilt into the Agilent 1200 Seriesdiode-array detector SL for UV detec-tion. The outgoing capillary is directlyconnected to the sprayer of the elec-trospray source of the 6330 Ion Trap LC/MS, which is capable ofacquiring data with a scan speed of26,000 m/z/sec. This instrument set-up

Introduction

Crude extracts of herbal and animalorigin have been used for medicaltreatment of disease in all ancient cul-tures around the world since prehis-toric time. Their activity for treatmentof different diseases and other effectson humans were found by trial anderror over hundreds of years and theknow-ledge about this medicine waspassed on from generation to genera-tion. A good example of the efficiencyachieved during this process of opti-mization is the herbal based traditionalChinese medicine (TCM). A famousAsian herb, which has been used inherbal medicine for more than 5000years is the ginseng root (Panax gin-seng). The main active compounds ofthe ginsenosides are triterpenesaponins of which more than 80 havebeen isolated and characterized duringthe past years. A lot of work was doneduring the last 30 years to developanalytical methods for the analysis ofginseng extracts and medical formula-tions. The method of choice for theanalysis of complex natural productextracts such as those derived fromthe ginseng root is high performanceliquid chromatography (HPLC)1.LC/MS equipment, e.g LC/ESI oaTOFfor accurate mass measurement andion trap LC/MS or triple quadrupoleLC/MS instruments for structure eluci-dation by MS/MS and MSn are cur-rently used to determine the complexand similar structures ofginsenosides2. In particular, it is possi-ble to confirm the authenticity of thepharmaceutical ginseng products anddifferentiate between their activeingredients using the ion trap MSn

fragmentation patterns3.

985989-4705EN

is optimized to achieve the highestpossible resolution, which is demon-strated by the UV analysis of a com-plex natural product extract obtainedfrom Asian ginseng root (Panax gin-seng). To illustrate the performance, acomparative analysis on an Agilent1100 Series LC system (5-µm particlesize column) and on an Agilent 1200Series Rapid Resolution LC system(1.8-mm particle size column) is pre-sented4. The resulting UV chro-matograms acquired at 220 nm clearlydemonstrate the better resolution ofthe peaks on the Agilent 1200 SeriesRapid Resolution LC system. The peakwidth (FWHM) of the majority of thepeaks in the UV chromatogram isbelow 0.1 min with baseline separa-tion. The full performance of theAgilent 1200 Series Rapid ResolutionLC system in the high resolution con-figuration is documented in a separateApplication Note5. It is also possible touse this system, with minor changes,in a high throughput environment,which is described in anotherApplication Note6.

Methods• The Agilent 1200 Series binary pump

SL was operated under the followingconditions: Solvent A: water + 0.1 % TFASolvent B: AcN + 0.1 % TFA Flow: 0.5 mL/min Gradient: 0 min 5 % B,

1 min 5 % B, 60 min 85 % B, 61 min 95 % B, 70 min 95 % B


• The Agilent 1200 Series high perfor-mance autosampler SL was used forinjections of 10 µL sample and thesamples were cooled to 10 ºC. Thesample loop was switched to bypassafter 1 minute to reduce delay volume.

• The Agilent 1200 Series thermostat-ted column compartment SL wasadjusted to 50 ºC equipped with thelow delay volume heat exchanger.

• The Agilent 1200 Series diode-arraydetector SL was operated at 80 Hzfor data acquisition at a wavelengthof 220 nm/4, ref. 360/100 with the 2-µL flow cell, 30-mm path length.

• The 6330 Ion Trap LC/MS was oper-ated under the following conditions: Source: ESI in positive mode.Dry gas: 5.0 L/min Dry temp.: 300 ºC Nebulizer: 15 psi Target: 125,000 Max. accum. time: 100 msScan: 200-1300 Averages: 2 MSn: Automated MS/MS

and MS3


After separation of the individual com-pounds, which are components of thecrude extract, from the ginseng root bymeans of high resolution HPLC on a1.8-µm particle size column, they aresubjected to fragmentation for MSn.The major ingredients elute between20 and 33 minutes and are recorded asginsenosides Re, Rf, F11, Rb1, Rb2, Rc,Rd in the base peak chromatogram(figure 2). The high resolution of thecolumn used resolves a large amountof minor compounds from the ginsengextract, which may also be analyzedbecause the high scan rate of 26,000m/z/sec allow sufficient ion trapMS/MS and MSn data to be acquired.The ion trap MS/MS and MS3 data ofthe ginsenosides Re, Rb1 and Rc wereinvestigated in more detail for struc-ture elucidation while the species-dependent ginsenosides Rf and F11 isdiscussed in another part of thisstudy7.

cooler

ColCom

DAD SL

degasser

650 mm x 0.17 mm capillary 320 mm x 0.12 mm capillary

2 µl UV flow cell


h-ALS-SL

binary pump SL

1.8 µm SB C18column

To MS

Figure 1Agilent 1200 Series Rapid Resolution LC system for MS in low delay volume configuration.

5989-4705EN99

The simple MS scan delivers the massof the molecular ion in its protonatedand sodiated form (figure 3). The ratioof protonated and sodiated ionsdepends on the electrospray sourcetemperature because the sodiatedcomplexes are more stable at highertemperatures than the protonatedions, which decompose due to a lossof water and other fragmentations.The MS scan delivers the ions at m/z 928.9 [M+H-H2O]+, 946.9 [M+H]+,969.1 [M+Na]+ for ginsenoside Re; at m/z 1090.1 [M+H-H2O]+, 1108.5[M+H]+, 1131.1 [M+Na]+ for gin-senoide Rb1 and at m/z 1059.71 [M+H-H2O]+, 1077.75 [M+H]+, 1101.1 [M+Na]+ for ginsenoside Rc. The conditions for the MS/MS andMS3 fragmentation in the 6330 IonTrap LC/MS were essentially chosento produce sodiated ions. The frag-mentation of these ions gave muchclearer fragmentation and additionalstructural information compared to theprotonated ions, whose fragmentswere similar to the CID fragments dis-cussed in part 1 of this study4. Only one fragment is produced inMS/MS for the ginsenodide Re at m/z789.9 by a loss of a molecule of glu-

15 20 25 30 35 40Time [min]

1

2

3

47x10

Intens.

0

Re

Rf

F11

Rb1

Rb2Rc Rd

Figure 2Base peak chromatogram of ginseng extract by high resolution LC ion trap on a 1.8 µm particlecolumn.

+MS, 27.3 minBPC 100-1300 m/z 1108.5 [M+H]+

m/z 1131.1 [M+Na] +

m/z 1090.1 [M+H-H2O] +

2O] +

2O] +

1108.5

1131.11090.1

0.0

0.2

0.4

0.6

0.8

1.0x10

200 400 600 800 1000 1200 1400 m/z

Ginsenoside Rb1

1077.7

0.0

0.5

1.0

1.5x10

200 400 600 800 1000 1200 1400 m/z

+MS, 28.5 minBPC 100-1300m/z 1077.7 [M+H]+

m/z 1101.1 [M+Na] +

m/z 1059.7 [M+H-H

1101.1

Ginsenoside Rc

946.9

0

1

2

3

4

57

7

7

x10

Intens.

Intens.

Intens.

200 400 600 800 1000 1200 1400 m/z

969.1

+MS, 19.8 minBPC 100-1300m/z 946.9 [M+H] +

m/z 969.1 [M+Na] +

m/z 928.9 [M+H-HGinsenoside Re

928.9

1059.7

Figure 3Mass spectra of ginsenosides Re, Rc and RbT.

1005989-4705EN

cose from the sodiated molecular ion(figure 4). In MS3 this sodiated frag-ment is cleaved into two parts, where-as the detected ion at m/z 349.2comes from the cleaved disaccharidemoiety. In comparison, for the ginseno-side Rc there are two fragmentsobtained by MS/MS. There is also oneat m/z 789.5 and another one at m/z335.1 for the arabinose saccharide (fig-ure 5). The dissacharide at m/z 365.1 iscleaved off from the ion at m/z 789.5in the MS3 stage. The ginsenoside Rb1has another different fragmentationpattern where the molecular ion is cleaved to the ions at m/z 789.5 andm/z 365.1 by a loss of the glucose chain in the MS/MS stage. Inthe MS3 stage the fragment obtainedat m/z 789.5 also loses the same dis-accharide at m/z 356.1 (figure 6).

It is possible to distinguish betweenthe different ginsenosides contained inthe ginseng root extract with theseMS/MS and MS3 fragmentation pat-terns because the different saccharidemolecules connected to the triter-penoide core structure will be cleavedoff in a different but characteristicway. If the specific fragmentation pat-tern in MS/MS and MS3 mode of thevarious ginsenosides is known, ithelps to detect the presence of a spe-cial compound in the plant extract in avery specific manner with the ion trapMRM. MS/MS-MRM on the sodiated

789.5 +MS2 (969.1), 19.8 min

02468

7x10Intens.

200 600 1000 1400 1800 m/z

349.2 +MS3 (969.1->789.5), 19.9 min

0.00.20.40.60.8

6x10Intens.

200 600 1000 1400 1800 m/z

O

CH3OH

CH3CH3

CH3

CH3

CH3 CH3

OCH3

OHOH

OH

OH

O

OH

O

O

OH

OH

OH

O

OHOH

OH

CH3

789.5

349.2

Figure 4MS/MS and MSV fragmentation for structure elucidation of ginsenoside Re with ion trap LC/MS.

365.1 +MS3 (1101.1->789.7), 28.6 min

02468

6x10Intens.

200 600 1000 1400 1800 m/z

335.1

789.5 +MS2 (1101.1), 28.5 min

0.00.20.40.60.8

7x10

Intens.

200 600 1000 1400 1800 m/z

O

C H 3O H

C H 3C H 3

C H 3

C H 3

CH 3 C H 3

O

OC H 3

O

O

O

O H

OH

OH

O

O H

O H

OH

OH

O HOH

O H

O

OHO H

O H

789.5335.1

365.1

Figure 5MS/MS and MSV fragmentation for structure elucidation of ginsenoside Rc with ion trap LC/MS.

365.1

789.5+MS2 (1131.4), 27.3 min

0

2

4

67x10

Intens.

200 400 600 800 1000 m/z

365.1+MS3 (1131.4->789.5), 27.4 min

0123456x10

Intens.

200 400 600 800 1000 m/z

O

CH3OH

CH3CH3

CH3

CH3

CH3 CH3

O

OCH3

O

O

O

OH

OH

OH

O

OH

OH

OH

OH

OHOH

OH

O

OH

OH

OH

OH365.1 789.5

365.1

Figure 6MS/MS and MSV fragmentation for structure elucidation of ginsenoside RbT with ion trapLC/MS.

5989-4705EN101

5.“Performance of the Agilent 1200SL HPLC System for Ultra-Fast LCApplications with 2.1-mm i.d.columns.” Agilent Application Note,publication number 5989-4502EN,2006.

6.“Performance of Agilent 1200 SLLC system for highest resolution.”Agilent Application Note, publicationnumber 5989-4489EN, 2006.

7. “Examination of a complex naturalproduct extract from Ginseng - Part III:Species differentiation ginseng andauthentification of ginseng products by LC/MS”, Agilent Application Note,Publication Number 5989-4706EN,2006.

References


2. Wang X., Sakuma T., Asafu-Adjaye E.,Shiu G. K. „Determination of ginseno-sides in plant extracts from Panax ginseng and Panax quinquefolius L. by LC/MS/MS.”, Anal. Chem. 71,1579-1584, 1999.

3. Chan T.W.D., But P.P.H., Cheng S.W.,Kwok I.M.Y., Lau F.W., Xu H.X.“Differentiation and authentication ofPanax ginseng, Panax quinquefolius,and ginseng products by usingHPLC/MS.” Anal. Chem. 72, 1281-1287, 2000.

4. “Examination of a complex naturalproduct extract from Ginseng – Part I:Structure elucidation of Ginsenosidesby high resolution LC – ESI oaTOF withaccurate mass measurement”, AgilentApplication Note, Publication number5989-4506, 2006.

molecular ions of the ginsenosides Re,Rb1 and Rc shows exactly their pres-ence in the crude ginseng root exact (figure 7). The obtained MS/MS spec-tra are in accordance with the spectraobtained in the experiments describedabove (figures 4-6). In addition, theisomeric ginsenosides Rb2 and Rd arealso detected.

Conclusion

The Agilent 1200 Series RapidResolution LC system in combinationwith the Agilent 6330 Ion Trap LC/MSproves its capability in structure eluci-dation of natural products in highlycomplex extracts from plant origin. Inthis study a highly complex extractfrom ginseng root was analyzed usingthe Agilent 1200 Series RapidResolution LC system and the Agilent6330 Ion Trap LC/MS system. Complexstructures of three ginsenosides,which are the main compounds of theextract, could be elucidated by theinterpretation of the MS/MS and MS3

ion trap data obtained. The detailedknowledge of the different fragmenta-tion data can be used to control thequality of natural extracts or pharma-ceutical products by ion trap MRMwhen applied to special ingredients.

18 20 22 24 26 28 30 32Time [min]

25

50

75

100

125 Intens.mAU TIC +All MSnUV chromatogram, 216-224 nm

0

1

2

36x10

Intens.

0

ReRb1 Rb2

Rc

Rd

Figure 7MS/MS-MRM of the sodiated molecular ions of ginsenosides Re, RbT and Rc together with theUV chromatogram obtained from a crude ginseng root.

Edgar Nägele is Application Chemist atAgilent Technologies, Waldbronn,Germany.

1025989-4705EN

Analysis of a complex natural product extract from ginseng – Part III: Species differentiation of ginseng plants and authentication of ginseng products by LC/MS

Abstract

Since prehistoric times herbal extracts have been used for the treatment and

prevention of disease. The same plant often shows different patterns of active

ingredients depending on the region, climate, growing conditions or sub-species.

Some plants used in traditional Chinese medicine (TCM) are a good example of this

behavior. A typical plant is the ginseng root, which is widespread in various sub-

species on the entire Asian and American continents. This Application Note will

demonstrate the use of the Agilent 1200 Series Rapid Resolution LC system with

Rapid Resolution High Throughput (RRHT) columns for the separation of the ingre-

dients found in a complex ginseng root extract. Detection and identification of the

different compounds by LC/ESI orthogonal acceleration time-of-flight (oaTOF) and

LC/ion trap mass spectrometry was implemented for the differentiation of the

regional or biological origin. The possibility to provide evidence for the authenticity

of a ginseng product is also demonstrated by determining the ingredient profile

based on known structural information of the ingredients.

Edgar Nägele

Application Note

5989-5493EN103

IntroductionTraditional natural medicine from dif-ferent and regionally separated cul-tures often uses the same plantspecies for the preparation of herbal-based extracts and tinctures for thetreatment of disease. The plants used,which are often cultivated as sub-species or grown under different con-ditions, show a distinctive pattern ofactive ingredients depending on theseinfluences. A typical botanical exam-ple for this behavior is the ginsengroot (Panax spec.), which has beenused in traditional Chinese medicine(TCM) for thousands of years. Thisherb is widespread in various sub-species on entire Asian and Americancontinents1. The method of choice forthe analysis of complex natural prod-uct extracts like the ginseng root ishigh performance liquid chromatogra-phy (HPLC)2. The main active com-pounds are triterpene saponins calledginsenosides of which more than 80have been isolated and characterizedduring the past several years. LC/ESIoaTOF for accurate mass measure-ment and LC/ion trap or LC/triplequadrupole instruments for structureelucidation by MS/MS and MSn arecurrently used for the determination ofcomplex and similar structures of gin-senosides3. It is possible to differenti-ate between the origin and the gin-seng sub-species and to confirm theauthenticity of pharmaceutical ginsengproducts using fragmentation patternsobtained by means of tandem massspectrometry of the pharmaceuticalactive ingredients4. This Appli-cationNote will demonstrate the use of theAgilent 1200 Series Rapid ResolutionLC system with Rapid Resolution HighThrough-put (RRHT) columns for theseparation of the ingredients found ina complex ginseng root extract fromAsian ginseng (Panax ginseng) and

American ginseng (Panax quinque-folius). Detection and identification ofvarious compounds by LC/ESIoaTOFaccurate mass measurement andLC/ion trap MSn mass spectrometrywas implemented for the differentia-tion of the regional or biological origin.The possibility to provide evidence forthe authenticity of a ginseng productis also demonstrated by determiningthe ingredient profile based on knownstructural information of the ingredi-ents. The detailed structure elucida-tion of ginsenoides by means of ESIoaTOF for accurate mass measure-ment including CID fragmentation andion trap MSn is described in detail inparts I and II of this study5,6.

ExperimentalEquipment• Agilent 1200 Series binary pump SL

with degasser. This pump has thecapability to perform high resolvingHPLC analysis on a 1.8-µm particlesize column to get the best resolu-tion performance.

• Agilent 1200 Series high performanceautosampler SL (h-ALS SL) with ther-mostat. This autosampler is especial-ly designed to work together with thebinary pump SL with lowest delayvolumes.

• Agilent 1200 Series thermostattedcolumn compartment (TCC). The TCCis ready for use with the binary pumpSL. Optional extras include separateheat exchangers for pre-column heat-ing and post-column cooling underoptimized delay volume conditionsand a 2-position/10-port valve foralternating column regeneration.

• Agilent 1200 Series diode arraydetector SL (DAD). The DAD is capa-ble of acquiring data with a samplingrate of up to 80 Hz.

• Agilent 6210 MSD TOF. Orthogonalacceleration time-of-flight mass

spectrometer with dual sprayer inter-face for mass calibration to acquiremolecular masses with highest accuracy. This time-of-flight massspectrometer is capable of acquiringdata at 40 Hz with pos/neg switching.

• Agilent 6330 Ion Trap. Ion trap massspectrometer for MSn tandem massspectrometric experiments with scanrate up to 26,000 m/z per second andautomated data-dependent MSn

capabilities. • The software used for LC/ion trap

instrument control was ChemStationB01.03, ion trap software 5.3, and fordata analysis the ion trap data analy-sis software 3.3.

• The software used for LC/TOF instru-ment control was the Mass HunterWorkstation A02.00 for data acquisi-tion and AnalystQS for data analysis.

• Column: ZORBAX SB C18, 2.1 X 150 mm, 1.8-µm

Sample1. Powdered freeze-dried Asian gin-

seng root (1 g) (Panax ginseng) andpowdered freeze-dried American gin-seng root (1 g) (Panax quinquefolius)were treated in an ultra sonic bathfor 30 minutes in 10 mL methanol,filtered and directly used for analysis.

2. Syrup-like Korean ginseng extractpharmaceutical product (ILHWA Co.,LTD, Korea) was dissolved (1 g) in100 mL water/MeOH (1/1, v/v) andused directly after filtration.

System set-upThe set-up of the LC/MS system isshown in figure 1. The Agilent 1200Series binary pump SL is connected tothe Agilent 1200 Series h-ALS SL witha 0.17-mm i.d. stainless steel capillary.To reduce delay volume, the seat capil-lary in the h-ALS SL has 0.12 mm i.d.The same kind of capillary connects tothe low delay volume (1.6 µL) heat

1045989-5493EN

exchanger in the TCC, which is con-nected to the column. For UV detectiona 2-µL flow cell is built into the DADSL. The outgoing capillary is directlyconnected to the sprayer of the elec-trospray source at the ESI oaTOF or tothe ion trap mass spectrometer. Thisinstrument set-up is optimized toachieve the highest possible resolu-tion, which is demonstrated by thecomparative UV analysis of a complexnatural product extract obtained fromAsian ginseng root (Panax ginseng)5.Using the 2.1 x 150 mm, 1.8-µm col-umn, the system back pressure wastypically about 560 bar. The peak width(FWHM) of the majority of the peaks inthe UV chromatogram was below 0.1min with baseline separation. The fullperformance of the Agilent 1200RR LCsystem in the high resolution configu-ration is outlined in a separate perfor-mance note7. It is also possible to usethis system with minor changes in ahigh throughput environment8.

Methods• The Agilent 1200 Series binary pump

SL was operated under the followingconditions: Solvent A: Water + 0.1 % TFASolvent B: AcN + 0.1 % TFAFlow: 0.5 mL/min. Gradient: 0 min 5 % B

1 min 5 % B 60 min 85 % B 61 min 95 % B 70 min 95 % B

Stop time: 70 min. Post time: 15 min.

• The Agilent 1200 Series high perfor-mance autosampler SL was used tomake sample injections of 10 µL andthe samples were cooled to 10 ºC.The sample loop was switched tobypass after one minute to reduce delay

volume.• The Agilent 1200 Series thermostat-

ted column compartment was adjust-ed to 50 ºC, equipped with the 1.6-µLlow delay volume heat exchanger.

• The Agilent 1200 Series diode arraydetector SL was operated at 80 Hz fordata acquisition at a wavelength of220 nm/4, ref. 360/100 with the 2-µL flow cell, 3-mm path length.

• The TOF mass spectrometer wasoperated under the following conditions: Source: ESI in positive mode

with dual spray for reference mass.

Dry gas: 12 L/min. Dry temp.: 200 ºC Nebulizer: 35 psi. Scan: 200-1300 Fragmentor: 150 V or 300 V

for CID Skimmer: 60 VCapillary: 3000 V

• The ion trap mass spectrometer wasoperated under the following condi-

tions: Source: ESI in positive

mode. Dry gas: 5.0 L/min. Dry temp.: 300 ºCNebulizer: 15 psiTarget: 150,000Max. accum.time: 100 msScan: 200-1300Averages: 2. Automated

MS/MS and MS3

Results and discussionThe separation of the individual com-pounds in the samples of Asian gin-seng (Panax ginseng) and Americanginseng (Panax quinquefolius) usinghigh resolution HPLC on a 1.8-µm par-ticle size column unravels the differ-ences in the individual composition ofthe crude plant extracts from the dif-ferent ginseng sub species. Obviously,the differences lie not only in the con-centrations of the ginsenoides but alsoin the presence of different com-pounds in the individual species. The

Cooler

ColumnCompartment

DAD SL

Degasser

650 mm x 0.17 mmcapillary

320 mm x 0.12 mm capillary

2 μL UV flow cell


h-ALS-SL

Binary pump SL

1.8 μm SB C18column

To MS

Figure 1Agilent 1200 binary LC system for MS in low delay volume configuration.

5989-5493EN105

main components elute between 20and 40 minutes and are evidently thecommon ginsenosides Re, Rg1, Rb1,Rb2, Rc and Rd as well as the specialginsenoides Rf and F11 in the ion trapMS base peak chromatogram (figures 2 and 3).Despite the good resolution of the 1.8-µm particle column, their resolutioncan be improved depending on thetemperature. For instance, at a columntemperature of 50 °C the geniseno-sides Re at m/z 946.5 [M+H]+ and gin-senoside Rg1 at m/z 823.5 [M+Na]+

are not resolved (figure 2).Nevertheless, they are clearly separat-ed at 80 °C and are contained in bothginseng samples (figure 3). To distin-guish the subspecies of Asian andAmerican ginseng by LC/MS the iso-meric ginsenosides Rf and the pseudo-ginsenoside F11 are very useful. Theanalysis of Asian ginseng shows theginsenoside Rf in protonated and sodi-ated form at m/z 801.5 and m/z 823.5,respectively. The protonated form ofthe isomeric pseudoginsenoside F11 atm/z 801.5 is only detectable as a tracecompound (figure 3). In contrast, theAmerican ginseng species only con-tains the pseudoginsenoside F11 in alarger amount and not the ginsenosideRf. Both isomeric compounds have thesame empirical formula (C42H72O14).The compounds are constitutional iso-mers only differentiable by their struc-tural formula. To distinguish betweenboth isomeric forms of these species-typical compounds, it is necessary toperform MS/MS experiments. The iontrap MS/MS experiments of theAmerican ginseng sample show thetypical sequence of a loss of five watermolecules at m/z 475.4, 457.4, 439.3,421.3 and 403.3 after the cleavage ofthe glucose-rhamnose disaccharidemoiety at m/z 309.1, which is charac-teristic for the molecular constitution

of pseudoginsenoside F11 obtained at a retention time of 28.5 min (figure4). Ultimately the distinctive furan ringfragment is released at m/z 143.1. Toconfirm the structural identity of theproposed fragments, the experimentwas repeated with a high resolutionLC/MS oaTOF instrument for highmass accuracy measurement andempirical formula confirmation (figure5). The measurement of the accuratemass of m/z 801.4997 confirmed the

identity of the molecular ion with 0.41ppm mass accuracy. Four of the empir-ical formulas for the fragments result-ing from the typical consecutive lossof five molecules of water resulting inthe ions at m/z 475.3782, 457.3676,439.3569 and 421.3464 were confirmedwith mass accuracies less than 2 ppm.The cleaved disaccharide fragmentwas confirmed by the accurate massat m/z 309.1176 with 3.10 ppm accura-cy and the small furan ring fragment at

Figure 2Base peak chromatogram of the separation of compounds contained in a crude extract fromAmerican ginseng (Panax quinquefolius) on a RRHT column at 50 °C.

20.0 22.5 25.0 27.5 30.0 32.5 35.0 37.5 40.0Time [min]

0

2

4

67x10

Intens.

Re/Rg1 F11Rb1

Rc

Rb2

Rd

Figure 3Base peak chromatograms from the separation of compounds in a crude extract from A) American ginseng (Panax quinquefolius) B) Asian ginseng (Panax ginseng) sub species on a RRHT column at 80 °C.

20.0 22.5 25.0 27.5 30.0 32.5 35.0 37.5 40.0 Time [min]0

1

2

3

4

20.0 22.5 25.0 27.5 30.0 32.5 35.0 37.5 40.0 Time [min]0

1

2

3

4

20.0 22.5 25.0 27.5 30.0 32.5 35.0 37.5 40.0 Time [min]0

1

2

3

4

B

Re

Rf

F11

Rb1 Rc

Rb2Rd

Rg1

20.0 22.5 25.0 27.5 30.0 32.5 35.0 37.5 40.0 Time [min]012345

x10

20.0 22.5 25.0 27.5 30.0 32.5 35.0 37.5 40.0 Time [min]0123457

Intens

x107

Intens

F11

Re Rb1 Rb2

Rc

Rd

Rg1

A

1065989-5493EN

ride was measured at m/z 325.1134with 0.22 ppm mass accuracy. The tablein figure 6 summarizes the measuredmasses and the calculated accuraciesfor all obtained fragments derived fromginsenoside Rf. Other ginsenosideswere detected by ion trap MS/MSand/or accurate mass

measurement with empirical formulacalculation of the molecular ions by

m/z 143.1065 with 4.90 ppm. The tablein figure 5 summarizes the measuredmasses and the calculated accuraciesfor all obtained fragments derived frompseudoginsenoide F11.In comparison,the compound obtained at a retentiontime of 27.7 minutes which is typicalfor Asian ginseng, is ginsenoside Rf.This compound also contains a proto-nated molecular ion of m/z 801.5 inthe LC/MS analysis. The fragments ofthis molecule are typical and can beused to distinguish it from ginsenosideF11. After the cleavage of the diglucosechain at m/z 325.0, a consecutive lossof four molecules of water is identifiedby MS/MS analysis (figure 6). Themass of the resulting ions at m/z459.4, 441.4, 423.4 and 405.4 is differ-ent from the pattern obtained for themolecule F11. The fragment at m/z143.1 was not observed. To confirm theidentity, a LC/MS oaTOF analysis wasperformed. The TOF spectrum showsthe molecular ion at m/z 801.4999with a high mass accuracy of 0.16 ppm(figure 7). The molecules obtainedafter the loss of water at m/z459.3834, 441.3726, 423.3620 and405.3515 were confirmed with accura-cies between 0.9 and 1.6 ppm. Themass of the cleaved glucose disaccha-

means of ESI oaTOF as indicated infigure 3. The detection of the typicalginsenosides Rf and F11 by LC/MS oneither ion trap or ESI oaTOF can beused to determine the species of a gin-seng plant and therefore the origin of apharmaceutical ginseng preparation.This was performed with a black,syrup-like ginseng extract purified forpharmaceutical use, which was pur-chased from a Korean vendor. The

Figure 4MS/MS of 24(R) pseudoginsenoside FTT typical for American ginseng (Panax quinquefolius).

125.1

143.1

207.1 309.1 403.3421.3

439.3

457.4

475.4603.3

+MS2(801.5), 28.5min

01236x10

Intens.

100 200 300 400 500 600 700 800 m/z

M(C42 H72 O14 )=800.49[M+H]+ = 801.50[MH-(-Glc-Rha)-H2O] + = 475.37[MH-(-Glc-Rha)-2H2O] += 457.36[MH-(-Glc-Rha)-3H2O] += 439.35[MH-(-Glc-Rha)-4H2O] += 421.34[MH-(-Glc-Rha)-5H2O] += 403.33

O H

C H3C H 3

CH3

CH3 C H3

OH

O

O

O

O HOH

OH

O

O HO H

OH

C H3

O

CH3

C H3CH3

O H

C8H15O2

143.10

C30H 51O4

475.37

C12H21O9

309.11

Figure 5HR-LC ESI oaTOF from an extract of American ginseng (Panax quinquefolius) 24(R) pseudoginsenoside FTT.

+TOF MS: 28.530 min

150 250 350 450 550 650 750 850m/z, amu

0.0

4.0e4

8.0e4

1.2e5

1.6e5

2.0e5

2.4e5

2.8e5

3.2e5457.3676

801.4997

143.1065 439.3569

783.4892421.3464

309.1176 475.3782 765.4791

0.34-0.30C42H71O13783.4895783.4892

-0.250.20C42H69O12765.4789765.4791

3.10-1.00C12H21O9309.1186309.1176

1.52-0.60C30H43O1421.3470421.3464

1.60-0.70C30H47O2439.3574439.3569

1.25-0.60C30H49O3457.3682457.3676

4.90-0.70C8H15O2143.1072143.1065

1.12-0.50C30H51O4475.3787475.3782

0.41-0.35C42H73O14801.5000801.4997

Massaccuracy

[ppm]

Massaccuracy

[mDa]FormulaCalculated

massMeasured

mass

Inte

nsity

cou

nts

5989-5493EN107

analysis shows the pattern of gin-senoides Rb1, Rb2, Rc and Rd typicalfor the Asian ginseng sub speciesPanax ginseng (figure 8). The peaks for the ginsenosides Re and Rg1 areclearly separated by RRHT columnchromatography. In the end theappearance of the characteristic gin-senoside Rf supplies the evidence forthe Asian ginseng species. It is possi-ble to use the ion trap MRM on one or more specific masses of typical ginsenoides to look only for character-istic ginsenoides in a sample6.

ConclusionThe Agilent 1200 Series RapidResolution LC System together withthe Agilent 6330 ion trap and theAgilent 6210 TOF are powerful analyti-cal tools for the detection of struc-turally complex compounds containedin crude natural product extracts alsofound in the purified pharmaceuticalproducts manufactured thereof. In thisapplication a highly complex root

Figure 6MS/MS of ginsenoside Rf typical for Asian ginseng (Panax ginseng).

CH3

O H

CH3CH3

CH3

CH3

CH3 CH3

OH

OH CH3

OO

O

O HOH

OH

O

O H

OHOH

OH

C12H21O10

325.11

C30H51O3459.38

325.1 405.4423.4

441.4

459.4 603.3 621.1 763.2

+MS2(801.5), 27.7min

01234x10

Intens.

300 350 400 450 500 550 600 650 700 750 800m/z

M(C42H72O14 )=800.49[M+H]+ = 801.50[MH-(-Glc-Glc)-H2O] + = 459.38[MH-(-Glc-Glc)-2H2O] += 441.37[MH-(-Glc-Glc)-3H2O] += 423.36[MH-(-Glc-Glc)-4H2O] += 405.35

Figure 7HR-LC ESI TOF of an extract of Asian ginseng (Panax ginseng) ginsenoside Rf .

Inte

nsity

cou

nts

0.85-0.70C42H71O13783.4895783.4888

1.18-0.9C42H69O12765.4789765.4780

0.22-0.10C12H21O10325.1135325.1134

1.54-0.60C30H45405.3521405.3515

1.63-0.70C30H47O1423.3627423.3620

1.48-0.70C30H49O2441.3733441.3726

0.91-0.40C30H51O3459.3838459.3834

0.16-0.10C42H73O14801.5000801.4999

Massaccuracy

[ppm]

Massaccuracy

[mDa]FormulaCalculated

massMeasured

mass

150 250 350 450 550 650 750 850m/z, amu

0.0

4.0e4

8.0e4

1.2e5

1.6e5

2.0e5

2.4e5

2.8e5

3.2e5

3.6e5

4.0e5423.3620

765.4780

441.3726

783.4888325.1134

405.3515

459.3834 801.4999

Figure 8Ion trap MS analysis of a purified pharmaceutical ginseng extract from Asian ginseng (Panax ginseng) from Korea.

22 24 26 28 30 32 34 36 Time [min]

0.250.500.751.001.25

7x10Intens.

....TIC +All MS

0

ReRg1 Rf

Rb1

RcRb2 Rd

38

1085989-5493EN

extract from different ginseng specieswas analyzed using an Agilent 1200Series Rapid Resolution LC/Trap sys-temn and an Agilent 1200 Series RapidResolution LC/TOF system. With theresulting highly resolved compoundsof the extract, it was possible to identi-fy compounds which are typical for theregional species of ginseng as well asthe different compositions of the com-mon ginsenosides in ginseng rootextracts and pharmaceutical products.Complex structures of two species-typ-ical ginsenosides could be elucidatedby the interpretation of the obtainedMS/MS ion trap data and could beconfirmed by accurate mass measure-ment using the ESI oaTOF. Thedetailed knowledge about the differentfragmentation was used to determinethe regional origin of a pharmaceuticalginseng extract.

References1. Chuang W.-C., Wu H.-K., Sheu S.,Chiou S.-H., Chang H.-C., Chen Y.-P.,Planta Med. 61, 459pp., 1995.


3. Wang X., Sakuma T., Asafu-Adjaye E.,Shiu G. K. „Determination of ginseno-sides in plant extracts from Panax gin-seng and Panax quinquefolius L. byLC/MS/MS.”, Anal. Chem. 71, 1579-1584, 1999.

4. Chan T.W.D., But P.P.H., Cheng S.W.,Kwok I.M.Y., Lau F.W., Xu H.X.“Differentiation and authentication ofPanax ginseng, Panax quinquefolius,and ginseng products by usingHPLC/MS.” Anal. Chem. 72, 1281-1287, 2000.

5. “Analysis of a complex natural productextract from Ginseng - Part I: Structureelucidation of Ginsenosides by rapidresolution LC – ESI TOF with accuratemass measurement”, AgilentApplication Note, Publication number 5989-4506EN, 2005.

6. “Analysis of a complex natural productextract from Ginseng - Part II:Structure elucidation of ginsenosidesby high resolution ion trap LC/MS”Agilent Application Note, PublicationNumber 5989-4705EN, 2006.

7. “Performance of Agilent 1200 SL LCsystem for highest resoution.” AgilentApplication Note Publi-cation number5989-4489EN, 2005.

8. “Performance of the Agilent 1200 SLHPLC System for ultra-fast LC applica-tions with 2.1-mm i.d. columns.”Agilent Application Note, PublicationNumber 5989-4502EN, 2005.

Edgar Nägele is Application chemist atAgilent Technologies, Waldbronn,Germany

5989-5493EN109

1105989-5493EN

Abstract

This Application Note describes:• Accurate mass measurement with LC/ESI-TOF for the determination of complex

natural compound structures• Fast computer aided identification and elucidation of chemical structures of new

natural product compounds• The use of the Agilent 1200 Series Rapid Resolution LC (RRLC) system with rapid

resolution high throughput (RRHT) columns for separation of ingredients in acomplex ginseng root extract, and the use of the Agilent 6210 ESI-TOF massspectrometer for accurate molecular mass measurement.

• Processing of TOF data with the Molecular Feature Extractor in the MassHunterWorkstation software, to generate molecular features, which are discrete molec-ular entities defined by the combination of retention time and mass.

Edgar NaegeleApplication Note

Agilent MassHunter – Fast computer aidedanalysis of LC/ESI-TOF data from complex natural product extractsPart 1: Analysis of Agilent 6210 TOF data with the MolecularFeature Extractor in MassHunter Workstation software

Agilent Equipment:1200 Series Rapid Resolution LC system6210 Time-of-Flight MSMassHunter Workstation software

Application Area:Natural product analysis in drug discovery

5989-5928EN111

IntroductionSince prehistoric times, herbalextracts have been used for the treat-ment of disease. Their effects onhumans were found by trial and errorover generations. A good example forthe efficiency achieved is traditionalChinese medicine (TCM). In Westernmedicine, drugs of natural origin aregaining importance due to their poten-tial. But Western pharmaceutical qual-ity standards require a deep knowl-edge about the ingredients in medi-cine based on natural products. Afamous Asian herb, which has beenused in herbal medicine for more than5000 years, is the ginseng root (Panaxspecies). Pharmacological effects ofginseng that have been reported are,for example, stimulatory and inhibitoryeffects on the central nervous system(CNS), antistress, antihyperglycemic,antineoplastic and immunomodulatoryeffects1. The main active compoundsof the ginsenosides are triterpenesaponins, of which more than 80 havebeen isolated and characterized duringthe past years.

A lot of work has been done duringthe last 30 years to develop analyticalmethods for the analysis of ginsengextracts and medical formulations. Themethod of choice for the analysis ofcomplex natural product extracts, likethe ginseng root, is high performanceliquid chromatography (HPLC)2. Forthe determination of the complex andsimilar structures of ginsenosides,modern LC/MS equipment such as theLC/ESI-TOF for accurate mass mea-surement and LC/ ion trap or LC/triplequadrupole instruments for structureelucidation by MS/MS and MSn arecurrently used3. However the timeconsuming bottleneck is the examina-tion of the acquired MS data to identi-fy and elucidate chemical structures ofnew natural product compounds. This Application Note demonstrates

the use of the Agilent 1200 SeriesRapid Resolution LC (RRLC) systemwith rapid resolution high throughput(RRHT) columns for the separation ofingredients in a complex ginseng rootextract, in combination with theAgilent 6210 ESI TOF mass spectrome-ter for accurate measurement of theirmolecular masses. The TOF dataobtained were processed by theMolecular Feature Extractor (MFE) ofthe MassHunter Workstation softwareto generate molecular features, whichare discrete molecular entities definedby the combination of retention timeand mass. Natural product compoundswere identified from this simplified MSdata. The composition differencesbetween ginseng sub-species is eluci-dated by software-aided data compari-son4 and an example for an automatedapproach to analyze complex naturalproduct extracts is given5.

ExperimentalEquipment• Agilent 1200 Series binary pump SL

with degasser. This pump has thecapability to perform high resolvingHPLC analysis on a 1.8 µm particlesize column for best resolution perfor-mance.

• Agilent 1200 Series high performanceautosampler SL (h-ALS) with thermo-stat. This autosampler is especiallydesigned to work together with theAgilent 1200 Series binary pump SL.

• Agilent 1200 Series thermostattedcolumn compartment (TCC). This TCCis ready for use with the binary pumpSL with optional separate heatexchangers and post column coolingunder optimized delay volume condi-tions, as well as alternating columnregeneration with an optional 2-position/10-port valve.

• Agilent 1200 Series diode arraydetector SL (DAD SL) which canacquire data at a sampling rate of upto 80 Hz.

• Agilent 6210 MSD TOF. Orthogonal

acceleration time-of-flight (TOF)mass spectrometer with dual sprayerinterface for mass calibration toacquire molecular masses with high-est accuracy. This TOF mass spec-trometer is capable of acquiring dataat 40 Hz and with positive/negativeswitching.

• Column: ZORBAX SB C18, 2.1 x 150 mm, 1.8 µm particle size.

• Software: TOF instrument controlsoftware MassHunter WorkstationA02.00 for data acquisition, Analystsoftware for data review andMolecular Feature Extractor (MFE)software for data processing.

In this LC/TOF instrument set-up, thebinary pump SL, which is in the lowdelay volume configuration, is con-nected to the h-ALS SL with a 0.12-mmID stainless steel capillary. To reducedelay volume, the seat capillary in theh-ALS SL has a 0.12 mm ID. The samekind of capillary connects to the lowdelay volume (1.6 µL) heat exchangerin the TCC, which is connected to thecolumn. For UV detection, a 2-µL cell isbuilt into the DAD SL. The outgoingcapillary is directly connected to thesprayer in the electrospray source ofthe TOF mass spectrometer. Thisinstrument set-up is optimized toachieve the highest possible resolution.

Sample preparationPowdered freeze-dried Asian andAmerican ginseng root (1g) (Panaxginseng and Panax quinquefolius) wasultrasonically treated for 30 minutes in10 mL methanol, filtrated and directlyused for analysis.

Method• The Agilent 1200 Series binary pump

SL was operated under the followingconditions:

1125989-5928EN

Solvent A: Water + 0.1 % TFA;Solvent B: AcN + 0.1 % TFA. Flow: 0.5 mL/min. Gradient: 0 min 5 % B,

1 min 5 % B, 60 min 85 % B, 61 min 95 % B,70 min 95 % B.

Stop time: 70 min. Post time: 15 min.

• The Agilent 1200 Series high perfor-mance autosampler SL was used forinjections of 10 µL sample and thesamples were cooled to 10 ºC. Thesample loop was switched to bypassafter 1 minute to reduce delay volume.

• The Agilent 1200 Series thermostatedcolumn compartment was adjustedto 50 ºC and equipped with the lowdelay volume heat exchanger.

• The Agilent 1200 Series DAD SL wasoperated at 80 Hz for data acquisitionat a wavelength of 220 nm/4, ref.360/100 with the 2-µL flow cell, 3-mm path length.

• The 6210 TOF MS was operatedunder the following conditions: Source: ESI in positive mode

with dual spray for reference mass.

Dry gas: 12 L/min Dry Temp.: 200 ºC Nebulizer: 35 psi Scan: 200-1300. Fragmentor: 150 V or 300 V for

CID Skimmer: 60 V Capillary: 3000 V

Results and discussionThe ingredients of a ginseng rootextract from an Asian ginseng root(Panax ginseng) were separated withthe Agilent 1200 Series RRLC systemon a RRHT column (1.8-µm particlesize) with subsequent ESI TOF massspectrometry (Agilent 6210 TOF). Thehigh resolution LC provided an excel-lent separation of the major and minoringredients of the natural productextract (figure 1). The acquired TOFdata were extracted by the MFE soft-

ware and the identified molecular ionsof the comprised compounds wereseparated from undefined backgroundions (figure 2). In this process, theidentified ions were clustered to mole-cular features comprising isotope com-pounds and adducts. From the

processed data, the feature group #7at the retention time of 26.3 min willbe examined in more detail to identifythese compounds in the ginsengextract. At this retention time, 20 dif-ferent features, molecular ions togeth-er with their different adducts and iso-

22.0 24.0 26.0 28.0 30.0 32.0 34.0 36.0 38.0Time, min

0.0

2.0e6

4.0e6

6.0e6

8.0e6

1.0e7

1.2e7

1.4e7

1.6e7

1.8e7

,

Inte

nsity

, cps

15 20 25 30 35

0.5

1.0

1.5

2.0

2.5

0.0

Processed TIC

min

15 20 25 30 35

0100200300400500600700800

Processed

EIC: Group #7 (RT=26.312) --- 20 Features

min

A

B

Inte

nsity

(103 )

Inte

nsity

(106 )

Figure 2A) Software processed TIC of Asian ginseng extract and grouped molecular features. B) Feature group # 7 at retention time 26.3 min groups 20 molecular features.

Figure 1High resolution LC on a 1.8 µm particle column with an extract from Asian ginseng root (Panaxginseng) for MS analysis by ESI-TOF.

5989-5928EN113

200 400 600 800 1000 1200

0100200300400500

600700

Original RT = (26.266, 26.358)

m/z

200400 600 800 1000 1200

0

100200

300400500

600700

Processed Group #7 (RT=26.312) --- 20 Features

m/z

B

1109.6104

1131.5926

785.5045

767.4935425.3775

A

Inte

nsity

(103 )

Inte

nsity

(103 )

Figure 4A) Unprocessed mass spectrum at 26.3 min. B) Processed mass spectrum at 26.3 min showingmolecular features, which belong to the same group.

fragments of the molecular ion aregenerated by collision induced decay(CID) for structure elucidation (figure4B). These fragments, which elute atthe same time, are included in the

The significant simplification of themass spectra resulting from the gener-ation of molecular features and elimi-nation of chemical noise allows aneasier interpretation, especially when

topes, are grouped together. Theprocessed EIC of feature group #7 at aretention time of 26.3 min is shown infigure 2B. The capability to extractgroups of ions as molecular featurescan be optically displayed by a detailed2D contour plot of the spectrum, whichshows retention time and the mass ofthe ions (figure 3). The unprocessedchromatograms of the protonated gin-senoside Rb1 ion at m/z 1109.6 andthe sodiated ion at m/z 1131.4, both ata retention time 26.3 min, are showinga lot of unspecific chemical noise ions(figures 3A and 3C). These unspecificions can not be grouped to a molecular feature because of an unclear isotopicdistribution and are eliminated duringdata processing. The ions remaining atthe same retention time window aregrouped together to molecular featurescomprised of adducts and isotopes ofthe basic ion (figures 3B and 3D). Themagnified window (figures 3C and 3D)clearly demonstrates the simplificationof the data. The data enhancement canalso be demonstrated by the compari-son of the original mass spectrum ofginsenoside Rb1 (figure 4A) to theprocessed mass spectrum (figure 4B).

A BC D

Figure 32D contour plot of TOF spectrum of ginseng extract, retention time vs. mass. A) Unprocessed 2D contour plot at retention time 24.8 – 28.0 min andmass 900 – 1250. B) Magnified contour plot around the unprocessed ginsenoside Rb1 ions. C and D) Contour plot of processed data of the same massand retention time area.

1145989-5928EN

Figure 5Structure of ginsenoside Rb1 obtained by interpretation of the processed mass spectrum withmolecular features included.

same group of molecular featurestogether with the molecular ion. Theinterpretation of the spectrum fromginsenoside Rb1 for structure elucida-tion is shown in figure 5. The protonat-ed molecular ion was found at m/z1109.6104 and the sodium adduct atm/z 1131.5926. The fragment at m/z785.5045 is derived from the loss ofone glucose moiety. An additional lossof a molecule of water results in thefragment at m/z 767.4935. The loss ofthe second glucose moiety yielded theion at m/z 425.3775 from the ter-penoide core structure. The identifiedmolecular features for ginsenoside Rb1and its CID fragments are summarized intable 1. The protonated, sodiated andpotasiated adducts are found with highmass accuracies of -0.14, -0.41 and -0.32 ppm, respectively. The calculatedmass accuracy of the molecule is -0.41ppm. In addition, the mass accuraciesfor the obtained CID fragments andtheir adducts are calculated (table 1).This confirms the empirical formulaand the proposed molecular structure.The content of a natural productextract often depends on the sub-species, the growth region and growthconditions. To decide if a natural plantextract contains the right active ingre-dients in proper concentration, a com-parison to a standard extract is neces-sary. For ginseng the different sub-species contain common ginsenosidesas well as specific ginsenosides. Atypical example is Asian ginseng(Panax ginseng) (figure 2) in compari-son to American ginseng (Panax quin-quefolius). The simplification achievedby the generation of molecular fea-tures unravels the differences (figure6). Especially the high content of gin-senoside F11 eluting at 24.5 minutesand the absence of ginsenoside Rf(eluting at 24 minutes in figure 2) istypical for the American ginseng sub-species5.

O

CH3OH

CH3CH3

CH3

CH3

CH3 CH3

O

OCH3

O

O

O

OH

OH

OH

O

OH

OH

OH

OH

OHOH

OH

O

OH

OH

OH

OH

425.3775

785.5045C 42 H 73 O 13 = 785.5045

-H 2OC 42 H 71 O 12 = 767.4935

C 30 H 49 O = 425.3775

20 25 30 35

0

1

2

3

4 Processed TIC

min

F11

Inte

nsity

(106 )

Figure 6Software processed TIC of American ginseng extract and grouped molecular features.

5989-5928EN115

References1. Attele A.S., Wu J.A., Yuan C.S.,Biochem. Pharmacol. 58, 1685-1693,1999.

2. Fuzzati N., “Analysis methods of gin-senosides.” J. Chrom. B1-2, 812, 114-133, 2004.

3. Wang X., Sakuma T., Asafu-Adjaye E.,Shiu G. K., „Determination of ginseno-sides in plant extracts from Panax gin-seng and Panax quinquefolius L. byLC/MS/MS.” Anal. Chem. 71, 1579-1584, 1999.

4. Edgar Naegele: “Agilent Mass- Hunter– Fast computer-aided analysis ofLC/ESI-TOF data from complex naturalproduct extractsPart 2: Comparison of Agilent 6210TOF data from different biological ori-gin with Mass Profiler in MassHunterWorkstation software“, AgilentApplication Note, Publication number5989-6076EN, 2007.

5. Edgar Naegele: “Agilent Mass- Hunter–Fast computer aided analysis ofLC/ESI-TOF data from complex naturalproduct extracts Part 3: Automatedanalysis of Agilent 6210 TOF data fromcomplex natural product extracts”,Agilent Application Note, Publicationnumber 5989-6077EN, 2007.

ConclusionThe ingredients of highly complex nat-ural product samples can be accurate-ly separated by the Agilent 1200 SeriesRRLC system with 1.8 µm particle sizecolumns. By connecting the ESI-TOFMS, a large amount of accurate massdata for empirical formula calculationand compound structure elucidationcan be created. This data is processedby the Molecular Feature Extractor ofthe MassHunter Workstation soft-ware, which generates molecular fea-tures, including the molecular ion, itsisotopes and adducts, to simplify thelarge amount of data and to help withdata interpretation. The highly accu-rate measured masses of the differentmolecular species of molecular ionsand CID fragments, included in themolecular features, are used for empir-ical formula calculation and structureelucidation.

+ 0.10 ppm784.497226.31- 0.41 ppm1108.603426.31M

--------------------- 0.32 ppm1108.60331147.566426.30M+K

--------------------M+H+4

- 0.41 ppm

- 0.14 ppm

Accuracy

-----

26.30

26.31

26.31

26.31

26.31

26.31

26.31

26.31

RT

-----

810.4938

809.4950

808.4902

807.4486

788.5134

787.5108

786.5081

785.5045

m/z

784.4976

784.4972

Mass

0.39 ppm

0.10 ppm

AccuracyMassm/zRTSpecies

1108.6034

1108.6031

1135.599826.31M+Na+4

1134.602126.31M+Na+3

1133.59926.31M+Na+2

1132.596426.31M+Na+1

1131.592626.31M+Na

1112.621926.31M+H+3

1111.617626.31M+H+2

1110.615226.31M+H+1

1109.610426.31M+H

----------------------------------------

+ 0.75 ppm424.370226.31+ 0.7 ppm766.486226.31

----------771.496526.31

0.42 ppm

0.68 ppm

Accuracy

-----

-----

-----

-----

-----

26.31

26.31

26.31

26.31

RT

-----

-----

-----

-----

-----

428.3882

427.3850

426.3813

425.3775

m/z

-----

424.3702

Mass

-----

0.75 ppm

AccuracyMassm/zRT

766.4864

766.4862

----------

----------

791.488326.30

790.480926.31

789.475626.31

770.502726.31

769.500426.31

768.497626.31

767.493526.31

Table 1Identified molecular features and calculated mass accuracies for empirical formula confirmation. (Ginsenoside Rb1, CXWH&UOUV , M 1108.6029;Fragment CWUHZUOTV , M 784.4973; Fragment CWUHZSOTU, M 766.4867; Fragment CV SHW&O, M 424,3705).

1165989-5928EN

Edgar Naegele is Application Chemistat Agilent Technologies, Waldbronn,Germany.

Abstract

This Application Note describes:• Fast computer-aided identification of differences in complex natural product

extracts• Accurate mass measurement with LC/ESI-TOF for the determination of complex

natural compound structures• The use of the Agilent 1200 Series Rapid Resolution LC (RRLC) system with

Rapid Resolution High Throughput (RRHT) columns for separation of ingredi-ents in a complex ginseng root extract, and the use of the Agilent 6210 ESI-TOFmass spectrometer for accurate molecular mass measurement

• Processing of TOF data with the Mass Profiler of the Agilent MassHunterWorkstation software to identify statistically significant differences in a set ofsamples from two different ginseng subspecies

• Identification of the different compounds by empirical formula calculationsbased on highly accurate mass measurement

Edgar NaegeleApplication Note

Agilent MassHunter – Fast, computer-aided analysis of LC/ESI-TOF data from complex natural product extractsPart 2: Comparison of Agilent 6210 TOF data from different biological origin with the Mass Profiler in MassHunter software

Agilent Equipment:1200 Series Rapid Resolution LC system6210 Time-of-Flight MSMassHunter Workstation software

Application Area:Analysis of complex natural products indrug discovery

5989-6076EN117

IntroductionHerbal extracts have been used sinceprehistoric times for the treatment ofdisease. The effects of these extractson humans were determined by simpletrial and error over generations. Agood example for the efficiencyachieved is Traditional ChineseMedicine (TCM). In Western medicine,drugs of natural origin are gainingimportance due to their therapeuticpotential. However, Western pharma-ceutical quality standards require adeeper know-ledge of the ingredientsin medicines based on natural prod-ucts. A famous Asian herb, which hasbeen used in herbal medicine for morethan 5000 years, is the Asian ginsengroot (Panax ginseng).

Apart from the Asian subspecies,other subspecies of the ginseng plantare available such as the Americansubspecies (Panax quinquefolius). Thepharmaceutically active compounds ofeach subspecies have different pat-terns of occurrence and concentration1.Therefore, it is important to know thecomposition of the plant extract,depending on its biological origin, toanticipate its medical activity beforeuse. The main active compounds ofginsenosides are triterpene saponins,of which more than 80 have been iso-lated and characterized during thepast years.

Much work has been done during thelast 30 years to develop analyticalmethods for the analysis of ginsengextracts and medical formulations. Themethod of choice for the analysis ofcomplex natural product extracts suchas the ginseng root is high-perfor-mance liquid chromatography (HPLC)2.For determination of the complex andsimilar structures of ginsenosides,modern LC/MS equipment is currently

used such as LC/ESI-TOF for accuratemass measurement and LC/ion trap orLC/triple quadrupole systems forstructural elucidation by MS/MS andMSn3. However, the time-consuming bottleneck remains the examination ofthe acquired MS data.

This Application Note describes theanalysis of TOF data with the MassProfiler of the Agilent MassHunterWorkstation software for fast and easyidentification of differences in theextracts from different ginseng sub-species. This note also describes theuse of the Agilent 1200 Series RapidResolution LC (RRLC) system withRapid Resolution High Throughput(RRHT) columns for the separation ofthe ingredients in complex ginsengroot extracts together with the Agilent6210 ESI-TOF mass spectrometer foraccurate measurement of their molec-ular masses.

Processing of the TOF data with theMolecular Feature Extractor (MFE) ofthe Agilent MassHunter Workstationsoftware to generate molecular fea-tures for the identification of naturalproduct compounds is described inpart 1 of this study4. An example for anautomated approach to analyze com-plex natural product extracts isdescribed in part 3 of this study5.

ExperimentalEquipment

• Agilent 1200 Series binary pump SLwith degasser. This pump is capableof performing high resolution HPLCanalysis on 1.8 µm particle sizecolumns for best resolution perfor-mance.

• Agilent 1200 Series high performanceautosampler SL with thermostat. Thisautosampler is designed specificallyto be used with the binary pump SL.

• Agilent 1200 Series thermostattedcolumn compartment. This columncompartment is supplied ready foruse with the binary pump SL. Alsoavailable are heat exchangers forpost-column cooling under optimizeddelay volume conditions, as well as a2-position/10-port valve for alternat-ing column regeneration.

• Agilent 1200 Series diode arraydetector SL. This detector is capableof acquiring data at a sampling rateup to 80 Hz.

• Agilent 6210 Time-of-Flight massspectrometer. This orthogonal accel-eration TOF MS has a dual sprayerinterface for mass calibration andacquisition of molecular masses withhighest accuracy, and is capable ofacquiring data at 40 Hz and with pos-itive/negative switching

• Column: ZORBAX SB C18, 2.1 x 150 mm, 1.8 µm particles.

• Software: TOF instrument controlsoftware Agilent MassHunterWorkstation revision A.02.00 for dataacquisition, Agilent Analyst Softwarefor data review, and Agilent MassProfiler software for data processing.

In this LC/TOF system setup, the bina-ry pump SL is configured for low-delayvolumes and connected to the highperformance autosampler SL with a0.17 mm ID stainless steel capillary. Toreduce delay volume, the seat capillaryin the high performance autosamplerSL has an ID of 0.12 mm. The sametype of capillary is used to connect thelow delay volume (1.6 µL) heatexchanger, which is connected to thecolumn, in the thermostatted columncompartment. For UV detection, a 2 µLcell is built into the diode array detec-tor SL. The outgoing capillary is con-nected directly to the sprayer in theelectrospray source of the mass spec-trometer. This instrument setup is opti-mized to achieve the highest possibleresolution.

1185989-6076EN

Sample preparationPowdered freeze-dried Asian andAmerican ginseng root (1 g each) wastreated ultrasonically for 30 minutes in10 mL methanol, filtered and directlyused for analysis.

Method• The Agilent 1200 Series binary pump

SL was operated under the followingconditions: Solvent A: Water + 0.1 % TFASolvent B: ACN + 0.1 % TFA Flow: 0.5 mL/min Gradient: 0 min, 5 %B

1 min, 5 %B 60 min, 85 %B 61 min, 95 %B70 min, 95 %B


• The Agilent 1200 Series high performance autosampler SL wasused to make sample injections of 10 µL sample and the samples werecooled to 10 °C. The sample loop wasswitched to bypass after one minuteto reduce delay volume.

• The Agilent 1200 Series thermostatedcolumn compartment was adjustedto 50 °C and equipped with the lowdelay volume heat exchanger.

• The Agilent 1200 Series DAD SL wasoperated at 80 Hz for data acquisitionat a wavelength of 220 nm/4 nm, ref-erence 360 nm/100 nm with the 2 µLflow cell, 3 mm path length flow cell.

• The Agilent 6210 TOF MS was oper-ated under the following conditions: Source: ESI in positive

mode with dualspray for reference mass

Dry gas: 12 L/min Dry temp.: 200 °C Nebulizer: 35 psi Scan: 200-1300.

Fragmentor: 150 V Skimmer: 60 V Capillary: 3000 V

Results and discussionThe ingredients of the Asian andAmerican ginseng root extracts wereseparated with the Agilent 1200 SeriesRRLC system using an RRHT column(1.8 µm particle size) with subsequentESI-TOF mass spectrometry (Agilent6210 TOF). The high resolution LCfacilitated excellent separation of themajor and minor ingredients of the nat-

ural product extracts (figure 1). For thedata analysis with Mass Profiler, fiverepeated injections of each extractsample were measured by the sameLC/MS TOF method as described inthe experimental section. The acquiredTOF data were extracted by theMolecular Feature Extractor softwareas described4. In this process, theidentified ions were clustered to mole-cular features comprising isotope compounds and adducts. The filesobtained were grouped according tothe biological origin of the measured

(A) TIC of TOF-MS from Asian ginseng

8.0 9.0 10.0 11.0 12.0 13.0 14.0 15.0 16.0 17.0 18.0 19.0 20.0 21.00.0

5.0e5

1.0e6

1.5e6

2.0e6

2.5e6

3.0e6

3.5e6 Re

Rg Rf

Rb

1

Rb

Rc Rd

8.0 9.0 10.0 11.0 12.0 13.0 14.0 15.0 16.0 17.0 18.0 19.0 20.0 21.0

Rf

8.0 9.0 10.0 11.0 12.0 13.0 14.0 15.0 16.0 17.0 18.0 19.0 20.0

Rf

(B) TIC of TOF-

8.0 9.0 10.0 11.0 12.0 13.0 14.0 15.0 16.0 17.0 18.0 19.0 20.0 21.00.0

5.0e5

1.0e6

1.5e6

2.0e6

2.5e6

3.0e6

3.5e6

4.0e6F11

Re

Rg1

Rb1

Rb2

Rc

Rd

(B) TIC of TOF-

8.0 9.0 10.0 11.0 12.0 13.0 14.0 15.0 16.0 17.0 18.0 19.0 20.0 21.00.0

5.0e5

1.0e6

1.5e6

2.0e6

2.5e6

3.0e6

3.5e6

4.0e6F11

Re

Rg1

Rb1

Rb2

Rc

Rd

(B) TIC of TOF-MS from American ginseng

8.0 9.0 10.0 11.0 12.0 13.0 14.0 15.0 16.0 17.0 18.0 19.0 20.0 21.0Time [min]

5.0e5

1.0e6

1.5e6

2.0e6

2.5e6

3.0e6

3.5e6

4.0e6F

Re

Rg

Rb

Rb

Rc

Rd

21

1

11

1

2

Intensity [cps]

Intensity [cps]

Figure 1TIC of TOF-MS from an Asian (A) and an American (B) ginseng root extract between 8 and 21 minutes showing the main ginseosides.

5989-6076EN119

10 12 14 16 18 20 22 248

10

12

14

16

18

20

22

24

26

-4X

-2X

+4X

+2X

1X

Pseudoginsenoside F11

Ginseoside Rc

OH

CH3CH3

CH3

CH3 CH3

OH

O

O

O

OHOH

OH

O

OHOH

OH

CH3

O

CH3

CH3CH3

OH

O

CH3OH

CH3CH3

CH3

CH3

CH3 CH3

O

OCH3

O

O

O

OHOH

OH

O

OH

OHOH

OH

OHOH

OH

O

OHOH

OH

Log2 Abundance of Asian ginseng

Log 2

Abu

ndan

ce o

f Am

eric

an g

inse

ng

Figure 2Plot of mass against retention time shows all 671 molecular features found in the Asian (red) and American (blue) ginseng extracts and molecularfeatures of ginsenoside RbT (CXWH&UOUV at M = 1108.6029 and RT = 11.90).

Mass vs. Retention Time

0 5 10 15 20 25 30

Retention Time (min)

0

200

400

600

800

1000

1200

Mas

s

O

CH3OH

CH3CH3

CH3

CH3

CH3 CH3

O

OCH3

O

O

O

OHOH

OH

O

OH

OHOH

OH

OHOH

OH

O

OH

OHOH

OH

Retention time SD = 0.006Mass SD = 0.0004Abundance RSD = 0.23

Mass accuracy = 1.38 ppm

Retention time [min]

Mas

s [M

U]

Mass vs. Retention Time

0 5 10 15 20 25 30

Retention Time (min)

0

200

400

600

800

1000

1200

Mas

s

O

CH3OH

CH3CH3

CH3

CH3

CH3 CH3

O

OCH3

O

O

O

OHOH

OH

O

OH

OHOH

OH

OHOH

OH

O

OH

OHOH

OH

Retention time SD = 0.006Mass SD = 0.0004Abundance RSD = 0.23

Mass accuracy = 1.38 ppm

Retention time [min]

Mas

s [M

U]

Figure 3Differential analysis of Asian and American ginseng, showing pseudoginseoside FTT exclusively in American ginseng and higher concentration ofginsenoside Rc in Asian ginseng.

displayed in a logarithmic (log2) plotshowing the abundance ratio of Asianand American ginseng extracts (figure 3). In the plot there are five linesfor selected levels of abundance differ-ence in the two sample groups.Molecular features lying on the line inthe middle (1x) are equal in bothgroups, molecular features within the2x margins are up to twice the abun-

1108.6029 and RT = 11.90) wereenlarged. The standard deviations formass and retention time, the relativestandard deviation of the abundance ineach group as well as the high relativemass accuracy of 1.38 ppm for thisparticular compound demonstrate thehigh quality of the data.For the differential analysis of bothgroups, the features of each group are

extract samples into two respectivegroups for Asian and American gin-seng and loaded into the Mass Profilersoftware. All 671 identified molecularfeatures were displayed in a plot ofmass against the retention time plot inthe Mass Profiler software to inspectthe quality of the data (figure 2). Themolecular features for the known gin-senoside Rb1 (C54H92O23 at M =

1205989-6076EN

dance in one group and within the 4xmargins up to four-fold. Beyond thesemargins a feature is nearly unique orexclusively present in one group. Anexample for a compound which has ahigher concentration in the Asian gin-seng sample group is the ginsenosideRc (figures 1 and 3). The comparisonof the abundances in each sample ofthe two groups clearly shows a signifi-cant eight-fold higher occurrence of

ginsenoside Rc in Asian ginseng (table1). The calculated relative mass errorsare in the low single digit ppm range.The average for the Asian ginsengsample group is 1.71 ppm and theaverage for the American ginsengsample group is 0.90 ppm. The molecu-lar feature, which is easily recogniz-able as almost exclusively present inthe American ginseng samples, is thespecial the compound pseudoginseno-

side F11 (figure 3). This compound hasmore than 100-fold higher abundancein the samples from the American gin-seng root extract (table 2). For thiscompound, the average relative massaccuracy for the Asian ginseng samplegroup is 1.71 ppm and the average forthe American ginseng sample group is1.26 ppm.

ConclusionThe ingredients of highly complex nat-

ID Name RT Mass Abundance Mass error [mDa] Rel. mass error [ppm] Av. rel. mass error [ppm]1 American Ginseng_1 12.550 1078.5909 143,961 -1.50 1.362 American Ginseng_2 12.505 1078.5912 243,837 -1.20 1.093 American Ginseng_3 12.557 1078.5921 240,121 -0.30 0.254 American Ginseng_4 12.491 1078.5919 266,422 -0.50 0.445 American Ginseng_5 12.493 1078.5904 252,872 -2.00 1.836 Asian Ginseng_1 12.538 1078.5905 2,092,954 -1.90 1.737 Asian Ginseng_2 12.534 1078.5912 2,023,553 -1.20 1.098 Asian Ginseng_3 12.545 1078.5898 2,058,617 -2.60 2.389 Asian Ginseng_4 12.545 1078.5906 2,052,396 -1.80 1.6410 Asian Ginseng_5 12.534 1078.5905 2,042,638 -1.90 1.73

0.99

1.71

Table 1Retention times, abundancies and mass accuracies of ginsenoside Rc (CXV H&SOUU at M = 1078.5924) in Asian and American ginseng.

ID Name RT Mass Abundance Mass error [mDa] Rel. mass error [ppm] Av. rel. mass error [ppm]1 American Ginseng_1 11.328 800.4900 2,835,414 -2.20 2.702 American Ginseng_2 11.347 800.4910 2,836,847 -1.20 1.503 American Ginseng_3 11.337 800.4924 2,815,218 0.20 -0.244 American Ginseng_4 11.333 800.4909 2,797,743 -1.30 1.605 American Ginseng_5 11.335 800.4900 2,864,285 -2.20 1.606 Asian Ginseng_1 11.337 800.4914 23,132 -0.80 1.007 Asian Ginseng_2 11.337 800.4930 22,875 0.80 -1.008 Asian Ginseng_3 11.344 800.4927 20,865 0.50 -0.609 Asian Ginseng_4 11.346 800.4942 23,578 2.00 -2.4810 Asian Ginseng_5 11.342 800.4912 22,556 -1.00 1.25

1.53

1.26

Table 2Retention times, abundancies and mass accuracies for Pseudoginsenoside FTT (CWUHZUOTW at M = 800.4922), present nearly exclusively in Americanginseng samples.

5989-6076EN121

ural products can be separated withvery low standard deviations of reten-tion times using the Agilent 1200Series RRLC system and 1.8 µm parti-cle size columns. Connection to theAgilent 6210 ESI-TOF MS facilitatesacquisition of highly accurate andrepeatable mass data. With these sys-tem prerequisites, the data can beprocessed by the Mass Profiler soft-ware for statistical evaluation of thedifferences in the abundance of themolecular features in the differentsample groups. This facilitates identi-fication of differences in concentra-tion and abundance of single com-pounds in very complex samples suchas natural product extracts.

References1. Chuang W.-C., Wu H.-K., Sheu S.,Chiou S.-H., Chang H.-C., Chen Y.-P.,Planta Med. 61, 459 pp, 1995.

2. Fuzzati N., “Analysis methods of gin-senosides.” J. Chrom. B1-2, 812, 114-133, 2004.

3. Wang X., Sakuma T., Asafu-Adjaye E.,Shiu G. K., “Determination of ginseno-sides in plant extracts from Panax gin-seng and Panax quinquefolius L. byLC/MS/MS.” Anal. Chem. 71, 1579-1584, 1999.

4. Edgar Naegele: “Agilent Mass- Hunter –Fast computer-aided analysis ofLC/ESI-TOF data from complex naturalproduct extracts – Part 1:Analysis ofAgilent 6210 TOF data with theMolecular Feature Extractor inMassHunter Workstation software“,Agilent Application Note, Publicationnumber 5989-5928EN, 2007.

5. Edgar Naegele: “Agilent Mass- Hunter –Fast computer aided analysis ofLC/ESI-TOF data from complex naturalproduct extracts – Part 3: Automatedanalysis of Agilent 6210 TOF data fromcomplex natural product extracts”,Agilent Application Note, Publicationnumber 5989-6077EN, 2007.

Edgar Naegele is Application Chemistat Agilent Technologies, Waldbronn,Germany.

1225989-6076EN

123

Metabolic Profiling

124124

Abstract

This Application Note is based on scientific poster MOP 253 presented at the

ASMS conference of mass spectrometry, June 4, 2007, in Indianapolis, Indiana,

USA. The note describes:

• Rapid, computer-assisted identification of drug metabolites using a multi-algo-

rithm approach.

• High resolution chromatographic separation of drug metabolites from an in

vitro experiment.

• High mass accuracy measurement of drug metabolites by QTOF MS and

MS/MS.

Application Note

Agilent Equipment:

1200 Series Rapid Resolution LC system6510 quadrupole time-of-flight LC/MSZORBAX RRLC columnMassHunter metabolite identification software

Application Area:

Metabolite identification in drug discoveryand development

An interwoven, multi-algorithm approachfor computer-assisted identification of drugmetabolitesRapid identification of drug metabolites from accurate QTOF MSand MS/MS data by Agilent MassHunter metabolite identificationsoftware

Edgar Nägele, Frank Wolf, Uwe Nassal, Rainer Jäger,

Horst Lehmann, Frank Kuhlmann,Karina Subramanian

5989-7375EN125

Introduction

In modern pharmaceutical drug dis-covery it is of crucial importance toidentify all possible metabolites of anew chemical entity because of possi-ble toxic effects on humans and toevaluate its potential as new drug sub-stance. Today, high resolution and highmass accuracy QTOF MS and MS/MSdata1, which are acquired from in vitroas well as in vivo metabolism experi-ments, are used for metabolite identifi-cation in the different stages of thedrug discovery and developmentprocess. To make use of all potentialinformation contained in such data, itis essential to use different and com-plementary computer algorithms fordata analysis. Each of these algo-rithms provides an individual resultbased on the specific functionality andthe analyzed part of the data. The realadvantage of computer-assisted dataanalysis comes into being when thesealgorithms work together in an inter-woven fashion and contribute to aoverall result of higher confidence. Todo so, all scores produced by the indi-vidual algorithms are combined andweighted with user-settable factors.This Application Note describes anexample for the analysis of QTOF datafrom a metabolite identification experi-ment using the drug compoundNefazodone2 by means of a theAgilent MassHunter metabolite identi-fication software which uses thedescribed approach.

Experimental

Equipment• Agilent 1200 Series Rapid Resolution

LC system, including degasser, binarypump SL, high performance autosam-pler SL with thermostat, thermostat-ted column compartment and diodearray detector SL

• Agilent ZORBAX SB-C18 column,2.1 x 150 mm, 1.8 µm particle size

• Agilent 6510 quadrupole time-of-flight LC/MS system

Sample preparation• Stock solutions:

20 mg/mL S9 liver homogenatepreparation 0.1 mg/mL nefazodone inwater 1.6 mg NADP in 1.6 mL 0.1 M phosphate buffer at pH 7.450 mM isocitrate/MgCl2 (203 mgMgCl2.6H2O + 258.1 mg isocitrate in20 mL water) Isocitrate dehydrogenase0.33 U/mL

• NADPH regeneration system:1.6 mL NADP solution + 1.6 mL isoci-trate solution + 100 µL isocitratedehydrogenase solution

• Incubation mixture:3.85 µL substrate + 200 µL NADPHregeneration system + 746.15 µLphosphate buffer + and 50 µL S9 liverhomogenate

Incubation was carried out at 37 °C for60 minutes, a 100 µL aliquot was takenat 0 and 60 min. The reaction wasstopped by adding 6 µL perchloric acidand 100 µL acetonitrile to the aliquotsfollowed by centrifugation for 15 minat 14.000 g. The supernatant wasevaporated to dryness using aSpeedVac concentrator and reconsti-tuted with water containing 0.1 %formic acid (FA) for LC/MS analysis asdescribed below. Incubations stoppedat 0 min were used as controls.

Figure 1System configuration for the metabolite identification experiment, comprising the Agilent 1200Series Rapid Resolution LC with 1.8 µm particle size column and Agilent 6510 QTOF LC/MS.

1265989-7375EN

High resolution LC/MS method• Agilent 1200 Series binary pump SL

Solvent A: Water + 0.1 % FA,Solvent B: ACN + 0.1 % FAFlow rate: 0.5 mL/min

Gradient: 0 min, 5 %B; 15 min, 75 %B; 15.1 min, 95 %B; 16 min, 95 %B

Stop time: 16 minPost time: 10 min

• Agilent 1200 Series autosampler SLInjection volumes:

1-10 µL with needle wash

Sample temperature: 4 °CAutomated delay volume reduction

• Agilent 1200 Series diode arraydetector SL Detection wavelength:210 nm, (± 4 nm)Reference wavelength: 360 nm (± 16 nm)Flow cell: 2 µL volume,

3 mm path length• Agilent 1200 Series thermostatted

column compartment Column temperature: 60 °C

QTOF MS and MS/MS method• Agilent 6510 QTOF LC/MS system

Source: ESI, positive mode with dual spray for reference mass solution

Dry gas: 12.0 L/min (m/z 121.005and m/z 922.00)

Dry temperature: 300 °CNebulizer pressure: 60 psiMass range: 100-1000Fragmentor voltage: 200 VSkimmer: 60 VCapillary voltage: 4000 VCollision energy: 35 VData dependent MS/MS: 2 com-pounds, 2 MS/MS spectra, exclusion for 0.03 min

Data analysisThe first step in the analysis of thedata comprises the introduction of thecompound structure formula and cal-culation of the molecular weight fol-lowed by the comparison between thedata file that contains the metabolitecompounds (sample, incubation timet>0) and the data file that containsonly the parent drug (control, incuba-tion time t=0). In this comparison, alldetectable mass signals are extractedfrom the MS level data using theMolecular Feature Extraction (MFE)algorithm. Related compound isotopemasses and adduct masses aregrouped together into discrete molecu-lar features, and chemical noise isremoved. The compound lists of themetabolized sample and the controlare then compared. All compounds

which are new or increased in themetabolized sample are consideredpotential metabolites and are subject-ed to further analysis by different algo-rithms, which can be specified by theuser (figure 2). The algorithms canidentify and qualify new metabolites orcan just qualify metabolites found byanother algorithm. The results of allmetabolite identification algorithmsare weighted and combined into a finalidentification relevance score.Metabolites are qualified when theirfinal score is above a defined rele-vance threshold. The results from allalgorithms are collected in a resultstable and can be inspected at-a-glance.

Figure 2Setup of metabolite relevance score for each individual algorithm and weighted overall identificationrelevance score threshold.

NN

NO

CH3

N

N

OCl

Nefazodone

5989-7375EN127


The metabolite identification workflow(figure 3) used interwoven multiplealgorithms to populate a results tablewith potential metabolites. A pluralityof different procedures was used toidentify the metabolites. Each time anew metabolite candidate was found,a new row was added to the table. For each individual algorithm that was used to find or qualify a potentialmetabolite, an additional column in theresults table was used to display theresults. The following points describethe individual algorithms and theirinteraction to generate the final resulttable.

1. Sample-control comparison (figure4) – In a sample comparison table,the compounds found in themetabolite sample (time > 0, parentdrug was metabolized) and the con-trol (time = 0) are compared and

aligned by mass and RT. Thisallowed detection of both expectedand unexpected metabolites in themetabolized sample.

Figure 3Metabolite identification software workflow, featuring different data analysis algorithms. Each time a new metabolite candidate is found, a new rowis added to the table. Columns are added to the table to confirm an existing metabolite candidate and to show the result of the individual algorithm.

MFEMFE

…

………

–C2H6+O

+O2

+HF

–H2O

+O6+S2

…

–C2H6+O

+O2

+HF

–H2O

+O6+S2

…

C12 H14 N9 ClC13 H23 N3 P S ClC14 H24 N2 S2 ClC14 H22 N O3 S ClC14 H16 N6 O ClC15 H25 O P S ClC16 H19 N3 P ClC17 H20 N2 S ClC18 H21 O P Cl


10

5

1

4

2

3

98

67

14

12

15

13

16

1711

CH319

H

H

H

H

CH318

98

CH37

14

12

15

13

16

1711

23

2022

2425

NH

28 26

H

HCH321

H

CH327H

CH318

10

5

1

4

2

3

98

67

14

12

15

13

16

1711

CH319

H

H

H

H

18

98

CH37

14

12

15

13

16

1711

23

2022

2425

NH

28 26

H

HCH321

H

CH327H

CH318

O

NO

FF

F

N S

O

N

O ON

O

O

NO

FF

F

N S

O

N

O ON

O

O

NO

FF

F

N S

O

N

O O

m/z 568, major

O

NO

FF

F

N S

O

N

O O

m/z 554, major

O

NO

FF

F

N S

O

N

O ON

O

O

m/z 625, major

m/z 655, 2 in rat species

Parent, [M+H]+ m/z 639

O

NO

FF

F

N

m/z 341, major

O

NO

FF

F

N S

O

N

O ON

O

O

NO

FF

F

N S

O

N

O ON

O

m/z 611

m/z 597

O

NO

FF

F

N S

O

N

O O

O

m/z 584

O

NO

FF

F

N S

O

N

O O

O

m/z 570

O

NO

FF

F

N S

O

N

O ON

O

O

m/z 641

O

NO

FF

F

N S

O

N

O O

m/z 540

- CH2

- CH2

- CH2

+ O2

- CH2

+ O2

+ O2

MFE

Compound Correlation

CpdList

CpdList

MFE

Fragment Pattern Matching

…

MS/MS

……

Isotopic Pattern Matching

…

MS

EIC of expected masses

256

129

83

401

UV Chrom.

–C2H6+O

+O2

+HF

–H2O

+O6+S2

…

Built-inBiotransf.

–C2H6+O

+O2

+HF

–H2O

+O6+S2

…


Molecular Formula assignment


Molecular structure elucidation

10

5

1

4

2

3

98

67

14

12

15

13

16

1711

19H

H

H

CH318

98

CH37

14

12

15

13

16

1711

23

2022

2425

NH

28 26

H

HCH321

H

CH327H

CH318

10

5

1

4

2

3

98

67

14

12

15

13

16

1711

CH319

H

H

H

H

18

98

CH37

14

12

15

13

16

1711

23

2022

2425

NH

28 26

H

HCH321

H

CH327H

CH318

RAD Chrom.

Metabolite Prediction

O

NO

FF

F

N S

O

N

O ON

O

O

NO

FF

F

N S

O

N

O ON

O

O

NO

FF

F

N S

O

N

O O

m/z 568, major

O

NO

FF

F

N S

O

N

O O

m/z 554, major

O

NO

FF

F

N S

O

N

O ON

O

O

m/z 625, major

m/z 655, 2 in rat species

Parent, [M+H]+ m/z 639

O

NO

FF

F

N

m/z 341, major

O

NO

FF

F

N S

O

N

O ON

O

O

NO

FF

F

N S

O

N

O ON

O

m/z 611

m/z 597

O

NO

FF

F

N S

O

N

O O

O

m/z 584

O

NO

FF

F

N S

O

N

O O

O

m/z 570

O

NO

FF

F

N S

O

N

O ON

O

O

m/z 641

O

NO

FF

F

N S

O

N

O O

m/z 540

- CH2

- CH2

- CH2

+ O2

- CH2

+ O2

+ O2

Mass defect filter

Metabolites

Figure 4Comparison table between control sample and metabolite sample with e.g. di hydroxyl anddechlorination metabolites at mass 451.2578 and 501.2140 respectively.

1285989-7375EN

Figure 6MS/MS fragment pattern matching between protonated parent drug (m/z 470.2323) and proto-nated hydroxy metabolite (m/z 486.2265) with biotransformation mass shift assignment. TheMS/MS fragments at m/z 274.1550 from the parent drug and at m/z 290.1500 are related by ashift of 15.9999 for the metabolic hydroxylation reaction.

2. Isotopic pattern matching (figure 5) – The isotopic pattern of a metabolite coming from anexpected biotransformation wascompared to the theoretical pattern of the biotransformed parentdrug, while the pattern of an unex-pected metabolite was compared tothe theoretical isotope pattern of theparent drug.

3. Fragment pattern matching orMS/MS correlation (figure 6) – Thisprocedure correlated the MS/MSspectrum of each potential metabo-lite with the MS/MS spectrum ofthe parent drug. Using this proce-dure mass shifts in the fragmentions due to biotransformations couldbe detected and visualized.

A

B

NN

NO

CH3

N

N

O

Cl

OH

NN

NO

CH3

N

N

OOH

NN

NO

CH3

NN

OCl

OH

C15H20N3O3

290.1500

C15H20N3O2

274.1550

Figure 5A) Isotopic pattern matching for dechlorinated metabolite by comparison with calculated isotopicpattern (CIP) after application of biotransformation to parent drug formula. B) Isotopic pattern matching of chlorinated hydroxymetabolite by comparison to the calculatedisotopic pattern of the chlorinated parent drug.

5989-7375EN129

4. Extraction of chromatograms (figure7) – This included generation ofextracted ion chromatograms (EIC)directly from the data, and genera-tion of extracted compound chro-matograms (ECC) from extractedmolecular features.

5. Compound search in RAD (radioac-tivity detection) chromatograms (notpresented here).

6. Compound search in UV (ultraviolet)chromatograms or other detectionmethods (not presented here).

7. Biotransformation labeling (figures 6 and 10) – Expectedmetabolites were confirmed by com-parison of parent ion mass shiftswith a table of known biotransfor-mations and these compounds werelabeled with the name of biotrans-formation reaction in the result table (figure 10).

8. Molecular formula assignment (figures 8 and 9) – Molecular formu-la assignment was based on theassumption that only one elementalcomposition fits to the measuredaccurate mass of the product andthat subsets of the same elementalcomposition must explain the prod-uct fragment masses and their neu-tral losses in the MS/MS spectrum.

9. Mass defect filter – Potentialmetabolites with a mass defect out-side a defined mass defect windowaround the parent drug were filteredout.

Figure 8Calculated formula for highest score hydroxyl metabolite and abundance for isotopic pattern andmasses. The relative mass error was calculated to be 2.42 ppm.

Figure 7The extracted ion chromatograms (EIC) were directly obtained from the measured data for thecontrol as well as metabolite sample, and the extracted compound chromatograms (ECC) wereobtained from extracted molecular features – shown here shown for two different hydroxymetabolites at m/z 486.1903 at RT 7.2 and 7.9 min. The metabolites were not present in the controlsample.

1305989-7375EN

10. Metabolite prediction – Structuresfrom manual or computer-assistedmetabolite prediction wereassigned to the identified com-pounds in the result table.

11. Molecular structure elucidation –All data were consolidated forstructure elucidation and structureformula assignment.

12. Population of the final metaboliteresult table (Figure 10): The identi-fied metabolites were collected in aresult table, which shows themajor information about the com-pounds and the qualification fromeach individual algorithm as wellas additional available information.There is the possibility to produceand to report more detailed resulttables.

[ "?9+;s)"?C

Metabolites from the drug compoundNefazodone were automatically identi-fied by means of a computer-assistedapproach, which applied several inter-leaved algorithms to QTOF MS andMS/MS data. The comparison of con-trol and metabolized sample wasbased on molecular feature extraction(MFE) to extract metabolites. High

mass accuracy MS and MS/MS datawere acquired with low single digit rel-ative mass error and used for molecu-lar formula generation (MFG). In thefinal at-a-glance result table, an overallrelevance score was created for theidentified metabolites, which was cal-culated from the weighted relevancescore from each algorithm.

Figure 9Measured masses of MS/MS fragments and calculated fragment formulae for the the protonatedhydroxy.metabolite at m/z 486.2255. The loss masses as well as the calculated fragment formulaof the loss are displayed. The relative mass error of fragment m/z 290.1500 was calculated to be -0.20 ppm.

Figure 10Final result table for overall qualified metabolites, which are related to a known biochemical metabolic reaction. Qualified results form the algorithms“sample control comparison”, “isotopic pattern matching”, “fragment pattern matching” and “mass defect filter” are marked in green. Additionalinformation such as “assigned biotransformation2”, “calculated formula” and “available MS/MS” spectra are marked in blue. Structures can beassigned manually. Metabolites not related to a known biotransformation can also be extracted from the data by the same algorithms.

5989-7375EN131

References

1.Hopfgartner G., Chernushevich I. V.,Covey T., Plomley J. B., Bonner R.,“Exact mass measurement of productions for the structural elucidation ofdrug metabolites with a tandemquadrupole orthogonal accelerationtime-of-flight mass spectrometer”, J.Am. Soc. Mass Spectrom., 10, 1305-1314, 1999.

2.Peterman S. M., Duczak J. N.,Kalgutkar A. S., Lame M. E., Soglia J.R., “Application of a linear iontrap/orbitrap mass spectrometer inmetabolite characterization studies:Examination of the human liver micro-somal metabolism of the non-tricyclicanti-depressant nefazodone using data-dependent accurate mass mea-surement”, J. Am. Soc. MassSpectrom., 17, 363-375, 2006.

Edgar Nägele is Application Chemist,Frank Wolf, Uwe Nassal, Rainer Jägerand Karina Subramanian are SoftwareR&D Scientists, Horst Lehmann andFrank Kuhlmann are Product Managers,all at Agilent Technologies inWaldbronn, Germany.

1325989-7375EN

133

www.agilent.com/chem/tcm

© Agilent Technologies, Inc., 2010Published April 1, 2010Publication Number 5989-9912EN

Learn morewww.agilent.com/chem/tcm

Buy onlinewww.agilent.com/chem/store

Find an Agilent customer center in your countrywww.agilent.com/chem/contactus

U.S. and Canada1-800-227-9770, agilent_inquiries@agilent [email protected]

Asia [email protected]

Documents

5989-9912EN HERBAL ADULTERANTS