American Journal of Anal ytical Chemistry, 2011, 2, 500-510
doi:10.4236/ajac.2011.24060 Published Online August 2011 (http://www.SciRP.org/journal/ajac)
Copyright © 2011 SciRes. AJAC
Extract-Template Modeling and Pattern Recognition in
the Assessment of (Cymbopogon proximus)
Mohamed I. Abou-Shoer, Hoda M. Fathy, Abdallah A. Omar
Department of Ph arm ac og n os y , Faculty of Pharmacy, Alexandria University, Alexandria, Egypt
E-mail: aboushoerm@yahoo.com
Received April 21, 2011; revised May 25, 2011; accepted June 7, 2011
Abstract
Comprehensive analytical methodologies are exceedingly needed in order to evaluate product quality and
raw material specifications especially for botanical preparations. Advances in teaming up statistical (PCA,
CA & pattern recognition techniques) and spectroscopic (UV, IR, MS & NMR) methods of analysis have
generated a substantial impact on how to use spectroscopic instruments as intelligent modules capable of
identifying and classifying the composition of a variety of natural products. Cymbopogon proximus, a tradi-
tionally used medicinal herb claimed to be an effective remedy for renal spasms, lacks appropriate evaluation
procedures. UV assisted-PCA and PLS analysis were exercised herein to maximize the usefulness and appli-
cability of some previously established analytical specifications for herbal materials (e.g. solvent extractive
values). Hierarchical cluster analysis was also attempted to categorize and associate the generated
solvent-extracts. In addition, DE-TLC and GC were used to examine the different plant fractions in a qualita-
tive and quantitative manner.
Keywords: Chemometry, Natural Products, Cymbopogum proximus, Quality Assessment, Herbal Products,
PCA, Cluster Analysis
1. Introduction
The escalating interest in herbal therapies and its ex-
pansive involvement in the health sector is not surprising
and is undoubtfully capturing positive reception. Foliage
and plant parts like roots, leaves, flowers, etc. have been
continually used to promote health or treat diseases and
are typically marketed as herbal remedies or phytophar-
maceuticals. The challenge facing herbal analysis often
revolves around what tests to be performed and which
meaningful product specifications should be determined
as a part of an operative quality control routine. Univa-
riate measurements, which, indeed, are straightforward
direct properties of a sample, can be a valuable source of
information. For example, certificates of analysis issued
by most herbal products manufacturers use solvent-ex-
tractive values as helpful univariate measurements for
determining compliance with the established or declared
values. Meanwhile, chemical markers are playing a cru-
cial role in the evaluation of herbal preparations and are
routinely used in conventional identification, authenticity,
and standardization procedures [1].
A stimulating exercise is visualized on how to analyze,
assess or identify botanical preparations in the absence of
unique or distinct chemical markers. Nevertheless, the
intricate chemical compositional profile of the plant’s
metabolic pool, per se, constitutes a specific phenotypic
identity and a quality fingerprint. Most of the current QC
practices focus on using one or few key active compo-
nents or specific markers in the analysis while keeping a
blind eye to the synergy of the whole chemical configu-
ration. Analytical approaches that can simultaneously
track down subtle changes over a wide array of variables
will be more competent of accurately describing the
sample. Such comprehensive analytical approach can
conceive an intuitive resolution on how to evaluate natu-
ral products from a holistic perspective. Chemometric
tools are superbly useful in simultaneously monitoring
elements of change in the sample’s makeup; hence, it can
equally be exploited to monitor plant-to-plant metabolic
divergence or unexpected chemical signature change by
adulteration [2]. Multivariate data analysis (MVDA) has
proven to be exceptionally efficient in extracting hidden
information through cross-examining all the data’s array
of variables. On the other hand, conventional methods
handle observations from a univariate standpoint as the
M. I. ABOU-SHOER ET AL.
Copyright © 2011 SciRes. AJAC
501
key data processor and consequently they only expose
limited information [3].
Cymbopogon proximus, Family Gramineae, locally
known as Halfa-bar, is an aromatic densely-tufted grass
growing wildly and widely in Upper Egypt. The herb is
highly reputed in folkloric medicine as an effective di-
uretic, renal or abdominal antispasmodic agent, and for
relieving bronchial asthma as well. The herb exerts its
unique pharmacological action through relaxation of the
smooth muscle fibers without abolishing the propulsive
movement of the tissue, thus, it is traditionally used in
the expulsion of renal and ureteric calculi [4]. The ses-
quiterpene cryptomeridiol (proximadiol) has been pre-
viously reported to be responsible for its antispasmodic
activity [5,6]. The plant itself or admixed with other
herbs like Ammi visnaga, as Sekem® tea, or khellagon®
capsules, is marketed for renal ailments. Proximol® tab-
lets or effervescent granules are labeled to contain stan-
dardized C. proximus extract. A colorimetric method has
been reported to quantitatively assess proximadiol in C.
proximus after separation from the extract by preparative
TLC [7]. In addition, the amount of cryptomeridiol in the
petroleum ether or ether extracts was assessed by GLC
[8].
The study in hand is designed to explore the potential
of using chemometric approaches to analyze a family of
data sets derived from the UV spectra (UV assisted-PCA
and PLS analyses) for a collection of systematically gen-
erated C. proximus solvent extracts. The extracts were
assembled in such a manner to suit correlating chemical-
ly communicative groups pertinent to the nature of the
extract. The collected spectral data can be used for pre-
dicting the membership of samples to pre-defined chem-
ical profile (solvent-extractives). GC and DE-TLC tech-
niques were also introduced as simple and selective me-
thods for routine analysis. Moreover, the plant’s antihis-
taminic activity on guinea pig ileum was investigated to
substantiate its traditional application in asthma.
2. Experimental
2.1. Biological Materials
Plant material.The aerial parts of Cymbopogon proximus,
collected from Luxor, Egypt, were air-dried in shade
before processing. It was identified by the Department of
Botany, Faculty of Science, Alexandria University. Other
plant samples were purchased from four different herbal
stores (A - D).
In vitro testing. Guinea pig ileum was used to evaluate
the antihistaminic activity of the different extracts. Con-
tractions were expressed as a percentage inhibition of the
response to histamine dose.
2.2. Chemicals & Reagents
Light petroleum, hexane, benzene, Dichloromethane,
isopropanol, toluene, acetone, ethylacetate, methanol and
ethanol are analytical grade.
TLC analysis: Silica gel GF-254, pre-coated TLC
plates, 0.25 mm thick, E. Merck, Darmstadt, Germany.
The developed TLC plates were analyzed under UV
lamp (Hanovia lamps, Germany) at λ 254, 366, or after
spraying with either of ferric chloride or (anisaldehyde/
H2SO4) reagents. Dichloroethane: isopropanol 100:3
(system I); dichloromethane: ethylacetate: glacial acetic
acid 30:30:1 (system II); were used as mobile phases.
Silica gel (10 - 40 µ, mesh size), E. Merck, Darmstadt,
were used for column chromatography
Menthol, thymol, borneol, camphor, eugenol, and
cineol, are BDH laboratory reagents. Reference solutions
were prepared in concentrations 10 mg/ml for the solid
compounds or 0.3 ml/ml for cineol and 0.1 ml/ml euge-
nol in methanol.
2.3. Apparatus
UV-VIS spectra were carried on a Helios α Thermo-
Spectronic Spectrophotometer, England, supported with
software Vision 32, using 1 cm bath length quartz cells.
GC was carried on Perkin Elmer, using RTX-5 (0.32
mm, 30 m) capillary column, and connected to FID de-
tector, carrier gas was nitrogen with a flow rate of 15
ml/min. The detector temperature was set at 225˚C while
the injector temperature was set at 200˚C. Initial column
temperature was set 50˚C and held for 3 min. A gradient
temperature a ramp of 20˚C /min was applied to a final
temperature 210˚C that was maintained during the final
10 minutes.
GC/MS was carried on Fenningan mat SSQ 7000 with
digital DEC 3000 workstation, column DB-5 (5% phenyl)
methyl polysiloxane (30 m × 0.25 mm id) carrier gas was
helium with a flow rate of 1 ml/min, initial temperature
started with 50˚C, which was held for 3 min, and then
elevated to 300˚C with a ramp 5˚/min and maintained for
a final 5 min.
A Fujifilm Finepix A-210 X 3.2 mega-pixel digital
camera was used to capture the TLC plate-images.
Software: SIMCA-P by Umetrics and hierarchical
clustering explorer and Sorbfil for digitizing the TLC
chromatogram images [9].
2.4. Preparation of the Extracts
Two types of extracts were prepared.
Selective solvent extracts which were prepared by ex-
tracting five weights (each 10.0 g) of the powdered au-
M. I. ABOU-SHOER ET AL.
Copyright © 2011 SciRes. AJAC
502
thentic plant with 75 ml of each of the following five
solvents, light petroleum, dichloromethane, ethylacetate,
alcohol and water. Aqueous extraction was carried out
using water at ambient temperature and boiling water).
The extracts were filtered and the solvents were evapo-
rated under reduced pressure. The extraction process for
each solvent is repeated ten times, using fresh material
each time, to generate ten replicate extracts for each sol-
vent.
In parallel, successive solvent extracts were prepared
by extracting 10 grams of the plant powdered material
successively with light petroleum followed by dichloro-
methane, ethylacetate, alcohol and finally water, once
more this process was repeated ten times in order to ob-
tain ten replicates per each solvent.
All extracts that have been prepared were re-dissolved
in the appropriate solvent and neatly filtered through
0.45 µm pvp filter before processing.
Different plant extracts analyzed by GC were prepared
in a concentration of 25 mg/ml.
2.5. Preparation of Extract-Solutions (UV
Spectra Calibration Set Solutions)
All selective and successive solvent extracts—for all
replicates—were dissolved in methanol and stock solu-
tions (10%) for each different extract were prepared.
These solutions were used to generate appropriate dilu-
tions that would produce suitable UV absorbances.
Selective solvents-extractives solutions. Hot water
extracts prepared in concentrations of 0.18333mg/ml,
aqueous extract (ambient temperature) as 0.18333mg/ml
solutions, alcoholic extracts (0.0625mg/ml), ethylacetate
extracts (0.076923mg/ml), dichloromethane extracts
(0.125 mg/ml), and light petroleum extracts as (0.25
mg/ml) solutions.
Fractional solvents-extractives solutions (succes-
sive). Aqueous successive extracts prepared as 0.15714
mg/ml solutions, alcoholic successive extracts (0.0625
mg/ml), ethylacetate successive extract, (0.045 mg/ml)
and dichloromethane extracts as (0.142857mg/ml) solu-
tions.
Sample dilution (PLS study). The initial calibration
set, of PLS-1, is constructed using UV-absorbances (av-
erage of three runs) were recorded for different concen-
trations from the following extracts:
1) Ethylacetate successive extract in concentration
ranges from 0.06 - 0.12 mg/ml (10 sample concentra-
tions).
2) Alcohol successive extract, 12 samples of different
concentrations, ranging from 0.01 - 0.09 mg/ml.
3) Alcoholic extract, 10 samples of different concen-
trations, ranging from 0.06 - 0.09 mg/ml.
Extracts Electronic Models
The prediction ability of the calibration model for the
determination of the content of the extract was tested
using 3 samples of different known identity and content.
A second, PLS-2, calibration set was constructed by
using 21 different alcoholic extract samples, in concen-
tration ranges from 0.03 - 1.2 mg/ml. Prediction ability
was assessed with test set T containing 6 samples. These
samples are composed of two new concentration samples
from the alcoholic extract and four artificially reconsti-
tuted extracts. These were produced initially by succes-
sively fractionating 1 gm dried alcoholic extract of C.
proximus with the following solvents; light petroleum,
dichloromethane, ethylacetate, ethanol and water. Their
compositional ratios were found to be 3%, 13%, 22%,
51% and 11%, respectively. The reconstituted extracts
were created by adding these solvent-extractives in al-
tered ratios (different from its natural existence) to for-
mulate the modified or mutilated mixtures. Mixture 1
consisted of 1% light petroleum fraction, 33% dichloro-
methane fraction, 33% ethylacetate fraction, 16% alco-
holic fraction and 17% aqueous fraction. Mixture 2 con-
sisted of 10% dichloromethane fraction, 40% ethylace-
tate fraction, 20% alcoholic fraction and 30% aqueous
fraction, Mixture 3 consisted only of 77.3% alcoholic
fraction and 16.7% aqueous fraction. Mixture 4 consisted
of 100 % alcoholic fraction of the alcoholic extract.
2.6. Isolation of Proximadiol
The dried light petroleum extract (24 gm, oily yellowish
green residue) of air-dried powdered plant material (600
g) was chromatographed on 480 gm (10 - 40 µ silica gel)
using negative pressure. Light petroleum was used for
elution and the polarity was raised by increasing concen-
trations of ethylacetate. 110 mg of proximadiol (crypto-
meridiol) was separated from the fraction eluted with
60% ethylacetate in light petroleum. Its identification
was based on its different physical and spectral proper-
ties in comparison with the reported ones.
2.7. Preparation of the Commercial Products
Samples
Both 10 g from the herbal tea Sekem® as well as 10 g of
lab-prepared imitated compositional mixture of the same
locally available herbal constituents were extracted with
dichloromethane.
60 grams of Proximol® effervescent granules, (con-
taining 18.6 mg C. proximus extract standardized to con-
tain 8 mg proximadiol/100 g granules, in addition to
hexamine and piperazine citrate) was dissolved in 300 ml
of water, gradually added, and then the aqueous extract is
extracted three times with dichloromethane. The organic
M. I. ABOU-SHOER ET AL.
Copyright © 2011 SciRes. AJAC
503
layer was evaporated to dryness, the residue was re-
served to be used for TLC, GC, and chemometric test set
C analysis.
2.8. Selection of Internal Reference
Menthol, thymol, borneol, camphor, eugenol and cineol
were experimented under the adopted GC conditions to
explore the suitability of any of them as an internal stan-
dard. The tR of cineol (8.3 min) was found to be appro-
priate and not interfering with sample peaks. Hence, 0.2
ml of its solution was added to all extracts as the appro-
priate internal reference.
3. Results and Discussion
3.1. Chemometric Assisted-UV Evaluation of
Cymbopogon Extracts
3.1.1. PCA An al ysi s
Initially, the first multivariate calibration project was
created, using SIMCA-P (Soft Independent Modeling of
Class Analogy), to construct a mathematical model that
relates the sample’s UV-absorbances to the type of the
extract (whether selective or successive) and conse-
quently, predict its affiliate fit-in class. All of the UV
absorbance data (average of three runs) of the 100
extracts (4 different succesive solvent x 10 replicates and
6 different slective solvents x 10 replicates) produced by
different solvent extraction of C. proximus were em-
ployed to construct the training set of the PCA models.
R2X was calculated as 0.992 (close to 1 indicates an
excellent model for assigning the solvent extraction mode).
In addition, the model was electronically cross-validated.
The prediction ability of the calibration model for the
determination of the class identity of the extract was
tested using external test sets A & B. Test set A con-
tained ten new samples of different C. proximus extracts
prepared with the same extraction protocol (4 different
succesive solvent, 6 different selective solvents), whilst
test set B was prepared by deliberately extracting six
different plants, namely Ammi visnaga, Ambrosia,
Achillea, Chicory, Liquorice and Mentha. Test set B
contained 21 extracts prepared using the same selective
solvents for the mentioned plants with the same extrac-
tion routine.
The described project has proven to be valuable in
showing that the UV spectra of the different extracts
were successful in predicting the solvent-class identity
for the ten positive validation samples (test set A), as
demonstrated by the small variation of DmodX values,
and conclusively decisive in indicating that samples of
prediction set B are typical outliers as shown by their
significantly different Dmodx values (negative identity).
Furthermore, the score scatter plot described in Figure
1, clearly reveals that different extracts categorically
associated in proximity to each others, which reflects a
sensible resemblance in content-constituents. Meanwhile,
the 2D plot failed to clearly express the resolution
between the clusters, but, instead the 3D representation
(Figure 2) have successfully corrected the aberration in
the 2D data representation and displayed the distant al-
location of the clusters.
3.1.2. HCA A nal ysi s
Alternatively, HCA (hierarchical Cluster Analysis) was
explored to examine the capacity of the UV spectra of
the extract to reveal authenticity of the commercial C.
proximus samples. Hence, four C. proximus samples
Alcoholic extract Ethylacetate extract Light petroleum extract
Dichloromethane extract Hot aqueous extract Aqueous extract
Ethylacetate succ. extract Dichloromethane succ. extract Aqueous succ. Extract
Figure 1. The score scatter plot of the project.
M. I. ABOU-SHOER ET AL.
Copyright © 2011 SciRes. AJAC
504
Figure 2. The 3D score scatter plot of the project colored as Figure 1.
were purshased from 4 different herb stores, designated
as A, B, C and D, and were similarly treated with the
same extraction protocol as the authentic sample but in
three triplicates. Then, separately, each solvent-extract
for the commercial samples-along with the 10 replicates
for the equivalent selective solvent-extracts for the
authentic sample of C. proximus numbered from 1 to 10-,
were scrutinized by HCA.
Figures 3-4 clearly reveals that, regardeless of the
solvent used, each of the ten replicates were well
grouped together into one cluster and distinctably
separate from the commercial samples, as an example the
dendogram of light petroleum and alcohol extracts are
shown in Figure 3. Nevertheless, some solvent- extracts
were found to be, even, equivaluable in distinguishing
the source of the sample (as belonging to a specific
supplier). Hence, it is clearly obvious that using any
solvent for the extraction can conclusively distinguish
between the available samples.
In parallel, the same analysis was repeated on the ten
sample replicates for the successive extracts (five sol-
vents) of the authentic samples (numbered from 1-10),
along with the equivalent three replicates from the com-
mercial extracts. Figure 5 reveals that the ten replicates
were again well grouped as one cluster separate from the
commercial ones. Furthermore, the alcoholic and aqueous
successive procedure can be considered good discrimi-
nating routines as they succedded, to great extent, to
identify the source of sample. This suggests that deffating
the plant prior to the analysis can be considered a wise
step to diminish the effect of the fluctuating volatile
content due to the evaporation processing steps.
3.1.3. Quantitative Anal y si s
Additionally, PLS-1 multivariate analysis, using SIMCA-
P, was utilized to explore the possibility of using the ab-
sorbance data of the extracts in a quantitative prospect.
The linear calibration of the predicted versus observed
concentration of all the extracts of the calibration set
(Figure 6) had R2 0.98.
The predicted concentration of the calibration and
prediction set samples were of 95% - 109% of their ac-
tual values.
In Consent, the compositional constitution of any ex-
tract is expected to vary once the original sample used
for extraction is subjected to any sort of tampering by
exhaustion or adulteration. However, the plant’s mono-
graph specifications, in this regard, propose adopting or
specifying the weight of the “solvent extractive” in some
organic solvent as a preliminary indicator in these cases.
Alas, these weights are non-informative values that do
not actually reflect the chemical nature of the compounds
in the sample. However, once a more informative prop-
erty, such as an IR or UV absorption characteristics, is
related to the sample, a better understanding of the
chemical compositional pattern of such extract can be
reliably achieved.
Accordingly, constructing a PLS-2 model, has pro-
duced R2x cumulative of 0.981 and R2Y 0.928. Mixtures
1 to 4 were classified as not belonging to the model ac-
cording to their DmodX values, while the two concentra-
tion samples of the same alcoholic extract were in
agreement with the model, which was also capable of
predicting their concentrations.
3.2. GC Analy s i s
When all extracts (selective and successive) were ana-
lyzed by GC, four main peaks at tR 11.2, 14.1, 15.0 and
17.7 min were detected and were found to be in common
in all the GC chromatograms. Hence, cineol was used as
an internal standard in all the experimented extracts and
M. I. ABOU-SHOER ET AL.
Copyright © 2011 SciRes. AJAC
505
(a) (b)
Figure 3. HCA analysis for the UV spectra for (light petroleum (a) and alcohol (b) solvent extracts respectively) from authen-
tic Cymbopogon proximus samples R and suppliers A, B, C and D.
(a) (b)
Figure 4. HCA analysis for the UV spectra for (ethylacetate (a), and water (b) successive-solvent extracts respectively) from
authentic Cymbopogon proximus samples R and suppliers A, B, C and D.
(a)
(b)
Figure 5. HCA analysis for the UV spectra for (solvent selective (a) light petroleum (Pe), methylene chloride (CCl), ethylace-
tate (OAc), alcohol (OH) and aqueous (Aq), and solvent successive (b)) from authentic Cymbopogan proximus samples.
M. I. ABOU-SHOER ET AL.
Copyright © 2011 SciRes. AJAC
506
(a)
(b)
Figure 6. The linear calibration of the predicted versus observed concentration of two of the extracts of the calibration set.
their relative tR were found to be 1.34, 1.69, 1.83, 2.13
respectively and the normalized total areas (As/Ast) of
these four peaks were calculated and are listed in Table
1.
Obviously, the normalized total peak areas for the re-
solved peaks are at maximum in the light petroleum and
dichloromethane extracts, in a lesser amount in ethylace-
tate extract, minimum in alcoholic extract and barely
observed in the aqueous extract (chromatogram 1).
Subsequently, GC/MS analysis of the dichloromethane
extract has helped in the identification of piperitone
(peak 1). However, peaks 2 and 3 were proposed by the
software library of MS data, as nerolidol and β-eudesmol,
respectively. Finally, peak 4 identity was additionally
secured as proximadiol by spiking the injected samples
with reference proximadiol.
Validation of the GC Method
Accuracy and Precision
Precision, expressed as the standard deviation, ob-
tained by repetitive four injections of the dichlormethane
extract, was calculated as 0.08 and relative standard
deviation as 5.7%. Accuracy is assessed by calculating
percentage recovery using minimum three concentration
levels and three replicates of each concentration. Percen-
tage recovery was calculated by injecting five samples of
different concentrations of dichloromethane extracts (30,
45, 60, 70 and 120 mg/ml). The normalized total areas
was measured and used to construct the calibration curve.
Three additional samples with concentrations (35, 75 and
115 mg/ml) were injected in triplicate injections and the
average percentage recovery was calculated and found
88% - 90%.
The above-mentioned protocol was used to assess the
amount of Cymbopogon in a commercial herbal prepara-
tion (Sekem renal herbal tea). The percentage of Cym-
bopogon in the herbal formula was measured by calcu-
lating normalized peak areas ratio, the markers were
evident but in much lesser amount in the marketed sam-
ple as the normalized peak areas ratio for the marketed
product was calculated as 5% of the herbal tea. Despite
the fact that it is stated in the inserted pamphlet to con-
tain 20%, Halfa bar (Chromatogram 2).
3.3. Evaluation of C. proximus by Digitally
Enhanced TLC (DE-TLC)
Quantitative analysis by TLC was carried out using Sor-
bifil TLC videodensitometer. A densitometric calibration
curve was constructed by spotting serial dilutions of pro-
M. I. ABOU-SHOER ET AL.
Copyright © 2011 SciRes. AJAC
507
Chromatogram 1. GC chromatogram of the light petroleum and dichloromethane solvent extracts.
Table 1. The normalized total areas of the marker peaks
relative to the internal standard for each extract.
Solvent Ratio
Light petroleum 1.78
Dichloromethane 1.81
Ethylacetate 1.36
ethanol 0.96
water 0.24
Dichloromethane successive 0.69
Ethylacetate successive 0.51
Ethanol successive 0.50
Aqueous successive 0.05
ximadiol standard and revealing the spots with anisalde-
hyde SR (Figure 7).
Proximadiol can be evaluated quantitatively by ex-
tracting the powdered herb with light petroleum or op-
tionally by dichloromethane. The amount of proximadiol
in the Proximol® effervescent granules dichloromethane
extract was determined according to the peak area from
the calibration curve and was found 4 mg/50 g powdered
drug which is in accordance to the stated label. While,
the amount of proximadiol in the root extract was found
10.5 mg % while that of the aerial parts 5.6 mg %.
The more polar plant extracts (ethylacetate) were ana-
lyzed by TLC using solvent system II. The developed
chromatograms have shown two well-defined UV ab-
sorbing spots at Rf 0.54 and Rf 0.68; on spraying with
ferric chloride; the upper spot turned to orange color
while the lower spot acquired a green color. The ethyla-
cetate successive-solvent fractions contained both spots
in a reasonable amount. A calibration curve was built by
spotting serial dilutions of EtOAc-successive fraction
and revealing with FeCl3 SR then the peak areas for the
TLC-markers of the ethylacetate successive extract of
Sekem® was calculated (Figure 8) and was found to re-
flect the presence of plant material equivalent to ca. 33
mg per 0.77 g sample extract i.e. 4.4% of the total con-
tent.
TLC Validation
Selectivity:
To secure that the selected markers spots are distinctive
and characteristic for Cymbopogon proximus, each indi-
vidual plant of the components of Sekem renal herbal tea
which are Mentha, Ammi visnaga, Liquorice, Chicory,
Achillea, Ambrosia was successively extracted with light
petroleum, dichloromethane, ethyl acetae and ethanol.
The developed TLC plates for these extracts have re-
vealed that no other spots are interfering with the above
mentioned marker spots.
Accuracy and Precision
RSD% (n = 3) Recovery (%) Concentration
3.04 108% 24 mg/ml
3.98 95.6% 32 mg/ml
1.99 95% 40 mg/ml
Linearity
Standard solutions at three conc. Levels 20 - 60 mg/ml
were used. Each standard solution was examined three
times and the acceptability of linearity was judged by
examining the correlation coefficient R2 (>0.96) (Figure
9).
3.4. Antihistaminic Activity of Cymbopogon
proximus Extracts
In folkloric use, C. proximus is included in the herbal
mixtures used in the treatment of bronchial asthma. Early
in 1960, the alcoholic solution of the oleoresin was given
orally to guinea pigs. It was found that it gradually de-
velops a protective action against bronchial spasms pro-
duced by histamine sprays. Since then no further trials
concerning the antihistaminic properties were performed.
M. I. ABOU-SHOER ET AL.
Copyright © 2011 SciRes. AJAC
508
Chromatogram 2. GC chromatogram of the commercial drugs.
Calibration curve of proximadiol
y = 27545x + 80721
R
2
= 0.9944
0
100000
200000
300000
400000
500000
0510 15
mg
area
(a) (b)
Figure 7. (a)-TLC chromatogram, (b)-Plot of the chromatogram and of the reference proximadiol.
M. I. ABOU-SHOER ET AL.
Copyright © 2011 SciRes. AJAC
509
Figure 8. Linear calibration of the serial concentrations of EtOAc-successive fraction.
Figure 9. Regression plots.
The alcoholic extract produces the maximum inhibition
(100%), followed by aqueous extract (80%), ethylacetate
(38%), dichloromethane (12%) and light petroleum
(8%).
4. Conclusions
Authenticity and purity are mandatory key drivers of
acceptance for herbal products to consistently meet the
approval of regulatory bodies or consumer-defined qual-
ity. These requirements are strategic attributes that can
be monitored by different chromatographic and spec-
troscopic techniques. Accordingly, UV spectroscopy and
chemometric analysis were utilized to assess the quality
of Cymbopogon proximus. HCA, PCA or PLS, can defi-
nitely detect the slightest change in the composition of
the plant extract, and again it can easily distinguish any
exhaustion or adulteration attempt that might have been
encountered during processing the plant material. This
routine has proven to be valuable in showing that the UV
spectra of the different extracts were successful in pre-
dicting the solvent-class identity for the ten positive va-
lidation samples. Meanwhile, GC can be used to quanti-
tatively evaluate Cymbopogon proximus by extracting
the powdered herb with dichloromethane with using
cineol as internal standard.
5. References
[1] S. Li, Q. Han, C. Qiao, J. Song, C. L. Cheng and H. Xu,
“Chemical Markers for the Quality Control of Herbal
Medicines: An Overview,” Chinese Medicine, Vol. 3, No.
7, 2008, pp. 1-16.
[2] C. Daolio, F. L. Beltrame, A. G. Ferreira, Q. B. Cass, D.
A. G. Cortez and M. M. C. Ferreira, “Classification of
Commercial Catuaba Samples by NMR, HPLC and
Chemometrics,” Phytochemical Analysis, Vol. 19, No. 3,
2008, pp. 218-228. doi:10.1002/pca.1019
[3] M. Franssona, J. Johanssona, A. Spare´na and O. Svens-
son, “Comparison of Multivariate Methods for Quantita-
tive Determination with Transmission Raman Spectros-
copy in Pharmaceutical Formulations,” Journal of Chemo-
metrics, Vol. 24, No. 11-12, 2010, pp. 674-680.
doi:10.1002/cem.1330
[4] F. M. Abdel-Moneim, Z. F. Ahmed, M. Fayez and H.
Ghaleb, “Constituents of Local Plants XIV. The Anti-
spasmodic Principle in Cymbopogon Proximus,” Planta
Medica, Vol. 17, No. 3, 1969, pp. 209-216.
[5] H. D. Locksley, M. B. E. Fayez, A. S. Radwan, V. M.
Chari, G. A. Cordell and H. Wagner, “Constituents of
Local Plants: XXV, Constitution of the Antispasmodic
Principle of Cymbopogon Proximus,” Planta Medica,
Vol. 45, 1982, pp. 20-22. doi:10.1055/s-2007-971233
[6] F. E. Evans, D. W. Miller, T. Cairns, G. Vernon Baddeley
and E. Wenker, “Structure Analysis of Proximadiol
M. I. ABOU-SHOER ET AL.
Copyright © 2011 SciRes. AJAC
510
(Cryptomeridiol by CNMR Spectroscopy,” Phytochemi-
stry, Vol. 21, No. 4, 1982, pp. 937-938.
[7] E. N. M. El-Sayed, “Investigation of Chemical Constitu-
ents of Cymbopogon Species, Family Gramineae,” M.Sc.
Thesis, Environmental Studies, Institute of Graduate Stu-
dies and Research, Alexandria University, Alexandria,
1990.
[8] A. S. Radwan, “An Analytical Method for Proximadiol,
the Active Principle of Cymbopogon Proximus,” Planta
Medica, Vol. 27, 1975, pp. 93-97.
doi:10.1055/s-0028-1097767
[9] A. V. Hess, “Digitally Enhanced Thin Layer Chromato-
graphy: An Inexpensive, New Technique for Qualitative
and Quantitative Analysis,” Journal of Chemical Educa-
tion, Vol. 84, No. 5, 2007, pp. 842-847.