Journal of Chromatographic Science, 2019, 1–7

doi: 10.1093/chromsci/bmz058
Article

Downloaded from [Link] by guest on 05 September 2019

Article

HPLC Coupled with Chemometric Analysis and

LC–MS Studies of Three Flavonoids in Tephrosia
purpurea (L.) Pers Revealed Impact of
Chemodiversity on the Quest for the Chemical
Markers
Sagar Katakam1 , Rajeshwari Rathod2 , Poojadevi Sharma2 ,
Dharmesh Kachhadiya2 , Sheetal Anandjiwala2 , Sonal Sharma2 and
Neeta Shrivastava2,*
1 National
Institute of Pharmaceutical Education and Research (NIPER), Ahmedabad, 380054, India and 2 B.V. Patel
Pharmaceutical Education and Research Development (PERD) Centre, Thaltej, Ahmedabad, 380054, India.
∗ Authorto whom correspondence should be addressed. B.V. Patel Pharmaceutical Education and Research Development
(PERD) Centre, Thaltej, Ahmedabad, 380054, India. E-mail: neetashrivastavaperd@[Link]
Received 4 August 2018; Revised 24 April 2019; Editorial Decision 28 June 2019

Abstract
Globally, Tephrosia purpurea (L.) Pers is used as an important component in herbal drug for-
mulations for liver health. The present study is aimed to develop a suitable analytical approach
for simultaneous analysis of three flavonoids (rutin, deguelin and rotenone) to establish quality
control methods for plant. A novel High-performance liquid chromatography photodiode array
detector (HPLC-PDA) method has been developed to quantify these flavonoids in T. purpurea.
The method was validated, and data were subjected to chemometric analysis to select most
optimal marker compound. The method that was found linear with R2 values ranges from 0.996
to 0.998 with good recoveries. Intra- and inter-day precision values were <2. HPLC analysis
revealed high level of chemodiversity. Quantity of all the three chemical markers was found
significantly disparate in samples from different locations. Deguelin was detectable only in three
out of total eight samples. However, liquid chromatography–mass spectrometry analysis was found
sufficiently sensitive to detect all the compounds in all samples. Thus, results suggest to apply
combination of approaches to enhance confidence in chromatographic methods for quality control
of herbal drugs. Principal component analysis ranked the markers as Rutin>Rotenone>Deguelin.
This comprehensive approach employing multichromatography platforms can be successfully
utilized in analysis of these bioactive markers and routine standardization of herbal material and
formulations containing T. purpurea.

Introduction
the various popular liver formulations as Tephroli, Yakrifit, Livonil,
Tephrosia purpurea (L.) Pers (Family:Fabaceae), commonly known etc. (2). Along with the effect on gastrointestinal ailments, the plant
as “Sharapunkha”, is a popular herbal drug used to improve liver has antioxidant, antimicrobial, anti-inflammatory, anti-viral, wound-
health owing to its hepatoprotective activity (1). The plant is used in healing, immunomodulatory, anti-diabetic and anti-cancer properties

© The Author(s) 2019. Published by Oxford University Press. All rights reserved. For permissions, please email: [Link]@[Link]
This article is published and distributed under the terms of the Oxford University Press, Standard Journals Publication Model ([Link]
1
access/funder_policies/chorus/standard_publication_model)
2 Katakam et al.

Downloaded from [Link] by guest on 05 September 2019

Figure 1. Chemical structure of marker compounds.

Table I. Description of plant samples and their voucher specimen numbers

Serial Number Plant Original source Voucher Specimen Numbers

1 Tephrosia purpurea Ahmedabad (Gujarat) NIPER/A/NP0813/02

2 Karimnagar (Andhra Pradesh) NIPER/A/NP0913/03
3 Rajkot (Gujarat) NIPERA/NP1013/05
4 Vadodara (Gujarat) NIPERA/NP1013/06
5 Valsad (Gujarat) NIPERA/NP1013/07
6 Deesa (Gujarat) NIPERA/NP1113/08
7 Pune (Maharashtra) NIPERA/NP1113/09
8 Gandhinagar (Gujarat) NIPERA/NP1213/16

(3–5). These medicinal activities can be linked to glycosides, markers as well as on the bioefficacy of T. purpurea flavonoids,
rotenoids, isoflavones, flavanones, chalcones, flavanols and sterols quantitative HPLC data of all the compounds in plant samples from
found in different parts of the plant (6). In addition, it is also used different geographical locations have been subjected to principal
as a botanical insecticide in organic agriculture (7, 8, 9). Therapeutic component analysis (PCA) and hierarchial cluster analysis (HCA).
importance of T. purpurea in healthcare system incites a need to Statistical analysis has helped elucidate the significance and classify
develop stringent quality control methodology to certify reliability, the marker compounds as per their impact to be used as finest
bioequivalence, safety and reproducibility of its therapeutic chemical markers among these three (13).
effects (10). The present method is based on the combined employment of the
The quality of final herbal product, at the foremost, depends HPLC and LC–MS and then chemometric analysis that can provide
upon the use of scientifically authenticated and validated starting more reliability and confidence to the analysis (10). The approach is
material to ensure drug safety. Qualitative and quantitative analysis easy and reproducible to assess the quality of significant medicinal
of one or multiple chemical components is a widely recommended plant T. purpurea starting material.
approach to establish the identity of the herbal material (11, 12).
Accordingly, in present study, we have developed a comprehen-
sive analytical approach using multichromatographic platforms for
determination of rutin, rotenone and deguelin (Figure 1) in plant Experimental
using High-performance liquid chromatography (HPLC) and liquid Reagents, standards and solvents
chromatography–mass spectrometry (LC–MS) (10). Rutin (Natural Remedies, India), rotenone (Sigma Aldrich, USA)
Rutin, rotenone and deguelin are among the major flavonoids and deguelin (Santa Cruz Biotechnology, Germany) were used as
of the plant (5). Therefore, these compounds can be used as suitable standards for the analysis. Solvents like methanol, ethyl acetate,
chemical markers in developing analytical methods for quality con- ethanol and formic acid were procured from Fischer Scientific (India).
trol studies. A reversed-phase HPLC method has been optimized for Solvents of HPLC grade were used for HPLC analysis. Chemical
simultaneous quantification of these compounds. Calibration curve ammonium acetate (Himedia) was of HPLC grade.
has been prepared, and linearity of the method has been examined.
Inter- and intra-day precision, limit of detection (LOD), limit of quan-
tification (LOQ) and accuracy of the method have been measured as Plant material
the method validation parameters. LC–MS has been used to validate Tephrosia purpurea plants were collected from eight different habi-
the HPLC detection. The quality of herbal drugs also depends upon tats in India (Table I) during the months of September and October.
the consistent supply of bioactive compounds in medicinal material, Plant material was authenticated with the help of a taxonomist, and a
which is a very important aspect for higher efficacy of drug. However, voucher specimen was submitted to Department of Natural Products,
content of the bioactive compounds/markers can be affected by their NIPER, Ahmedabad (Table I). The whole-plant materials collected
own genetic makeup, environment and growth conditions. In order from various habitats were dried, powered, sieved, weighed and used
to examine the effect of chemodiversity on the selection of chemical for extraction.
HPLC Coupled with Chemometric Analysis and LC-MS Studies 3

Downloaded from [Link] by guest on 05 September 2019

Figure 2. HPLC chromatogram of standard chemical marker compounds (A) (rutin, quercetin, rotenone and deguelin) and HPLC chromatogram of sample at
361 nm and 270 nm (B and C, respectively).

HPLC conditions Phenomenex® 100RP18e (4.6 mm i.d. × 250 mm,5 μm) as stationary
The combination of solvents, buffers and columns was tried for phase. Test compounds were resolved by gradient analysis using
simultaneous detection of three marker compounds in HPLC. The mobile phase A of composition—ACN: methanol: 0.1% formic acid
final HPLC analysis was performed on a Shimadzu-LC-2011 (Kyoto, (35 : 5 : 60) and—mobile phase B of ACN: 0.1% formic acid (60:40).
Japan) with photodiode array (PDA) detector using a Kinetex® The total run time of 30 min was started with mobile phase A lasting
4 Katakam et al.

Table II. Method validation results

Precision (n = 3)
LOD LOQ
Compounds Regression equation R2 Accuracy
(ng) (ng) Concentration (ng) Intra-day Inter-day
Coefficient of variation CV (<2%)
(CV) (<2%)

Downloaded from [Link] by guest on 05 September 2019

1.5 1.13 0.13 100.4%
Rutin Y=13537x + 3989.1 0.998 0.2 0.5 25 0.58 0.14 102.4%
80 0.46 0.37 100.4%

1.5 0.30 0.39 101.9%

Rotenone Y =10946x + 11660 0.998 0.2 0.5 25 0.40 0.10 97.5%
80 0.06 0.81 103.9%

1.5 0.06 0.51 97.5%

Rotenone Y = 11721 + 8145.4 0.997 0.2 0.5 25 1.59 0.73 103.9%
80 0.09 0.14 99.8%

Table III. HPLC quantitation and LC–MS detection of targeted metabolite analysis in T. purpurea sample extracts

Location Rutin Rotenone Deguelin Rutin Rotenone Deguelin

%w/w ± SD %w/w ± SD %w/w ± SD m/z609.0 m/z393.2 m/z393.2

HPLCa LC–MSb

Ahmedabad 0.72 ± 0.009 0.021 ± 0.003 ND + + +

Andhrapradesh 0.98 ± 0.005 0.052 ± 0.003 0.063 ± 0.004 + + +
Rajkot 0.23 ± 0.008 0.022 ± 0.004 0.023 ± 0.003 + + +
Baroda 0.51 ± 0.009 0.008 ± 0.006 ND + + +
Deesa 0.27 ± 0.009 0.015 ± 0.007 0.034 ± 0.004 + + +
Pune 0.15 ± 0.005 0.007 ± 0.004 ND + + +
Gandhinagar 0.73 ± 0.008 0.024 ± 0.003 ND + + +
Valsad 0.85 ± 0.00046 0.024 ± 0.0046 ND + + +
a ND—not detected;
b In case of LC MS/MS, “+” indicates presence. Rotenone and deguelin have common parent ion. Cannot be differentiated by mass. These compounds were separated on the basis of tR in
HPLC-PDA method developed in the study.

for 6 min, then switching to mobile phase B for the next 6 min Intra-day precision and accuracy were evaluated on the same day
and reverting back to mobile phase A for the remaining run time at three different time points while inter-day precision and accuracy
with flow rate of 1 mL/min; injection volume was 5 μL, column were evaluated for 3 days by analysis of three different concentration
temperature was 25 ± 2◦ C and detection was at suitable wavelength levels (n = 3) along with linearity. For the study of accuracy, agreement
range (190–400 nm). between applied and measured value was determined, and percent
accuracy was noted for every level of standards.
Standard and sample preparation
Stock solutions of all the standard compounds were prepared in
Liquid chromatography–mass spectrometry conditions
methanol (1 mg/10 mL). Eight-point linearity curves were prepared
Sample preparation was the same as the one used in HPLC. A total
in the concentration ranges 0.5–100 ng/mL.
of 5 μL of sample was directly injected into ABSciex 2000 LCMS/MS
For sample preparation, dried whole-plant powder (1 gm) was
system. Separation was achieved on C18 column operated in positive
defatted by refluxing with hexane at 60–70◦ C for 10 min on water
ion and negative ion electrospray mode using ionizing voltage 5500
bath in triplicate. Refluxed material was then filtered, and dried marc
at 350◦ C with mobile phase consisting of 90 : 10 acetonitrile: 5 mm
was then extracted with ethanol. For HPLC analysis, ethanol extracts
ammonium acetate buffer at pH 3. Sample was analyzed in TIC mode
(200 μL) were filtered through membrane filter (0.22 μ) and again
by Analyst software.
evaporated till dryness under nitrogen evaporator. The dried samples
were reconstituted in 50 μL of mobile phase and subjected to analysis
using automatic sample injector. Chemometric analysis
PCA is among the most frequently used methods for chemometric
Standard curve preparation and method validation analysis that help reduce the dimensionality of multivariate data and
HPLC method validation was carried out measuring linearity, inter- generate patterns (14). Peak areas of all the standards were taken
and intra-day precision, accuracy, repeatability, LOD and LOQ. LOD as variables, and PCA analysis of all the samples from different
and LOQ value was calculated based on a signal-to-noise ratio of locations was conducted using PAST 3 (15). Hierarchical clustering
3 : 1 and 10 : 1, respectively. Repeatability of the method was assured was also performed using same software. Euclidian similarity matrix
by relative standard deviations of multiple repeat injection (n = 6). was prepared using Ward’s method (16).
HPLC Coupled with Chemometric Analysis and LC-MS Studies 5

Downloaded from [Link] by guest on 05 September 2019

Figure 3. PCA score (A) and loading plots (B). Ahm: Ahmedabad; Gan: Gandhinagar; Bar: Baroda; Val: Valsad; Pun: Pune; Raj: Rajkot; Dee: Deesa; AP: Andhra
Pradesh.

Results
Optimization of HPLC method and method validation
HPLC parameters as column types and temperatures, mobile phase
combination and flow rate were examined to optimize the resolution
of the three-marker compounds. Among the various columns assessed
for method development (Lichrosphere-60-CN, YMC, Kromasil-CN,
Kromasil C18, Kinetex-C18), Kinetex-C18 was found suitable to
give optimum resolution (Figure 2). Isocratic mobile phases were
not giving good results so a gradient method was developed con-
sisting of acetonitrile, methanol and 1% formic acid on Kinetex
(4.6 mmi.d × 250 mm. 5 μ) column. The UV wavelength in the
range 190–400 nm was scanned to find the maximum absorption
for each marker, which was 270 nm for rutin and 361 nm for
rotenone and deguelin. The HPLC method was validated and used
to quantify rutin, rotenone and deguelin. Upon HPLC analysis of T.
purpurea plant extracts, peaks of marker compounds were success-
fully detected at retention time (tR) of 3.2 (rutin), 16.2 (rotenone) and
17 (deguelin). Calibration curve and method validation parameters
are given in Table II. The method thus optimized for simultaneous Figure 4. Hierarchical clustering of T. purpurea plant samples from Ahm:
estimation of rutin, rotenone and deguelin and was found linear, Ahmedabad; Gan: Gandhinagar; Bar: Baroda; Val: Valsad; Pun: Pune; Raj:
repeatable and highly reproducible. Rajkot; Dee: Deesa; AP: Andhra Pradesh.

Qualitative and quantitative estimation of marker

compounds in plant samples and LC–MS detection are affecting the production of targeted metabolites in the plant.
Quantitation of these marker compounds in various samples of Results obtained in HPLC were validated using more sensitive LC–
T. purpurea collected from different ecogeographical conditions MS analysis. Extracts were analyzed in LC–MS in positive ionization
revealed prevalence of chemovariation. Rutin and rotenone was and negative ionization mode to resolve any doubt pertaining to the
present in plants from all locations with yield ranging from 0.15 w/w presence of these chemical markers in T. purpurea. The results of
% to 0.98 w/w % and from 0.01 w/w % to 0.05 w/w %, respectively LC–MS of alcohol extracts confirmed the presence of all the three
(Table III). However, in the case of deguelin, the HPLC detection markers (rutin, rotenone and deguelin) compounds resolved in HPLC
was possible only in three locations in the range of 0.02 w/w % analysis.
to 0.06 w/w % (Table III). Here, we would like to mention that
initially, quercetin was also included in the study, and HPLC method Chemometric analysis
was found to be efficient in resolving quercetin along with the Chemometric analysis refers to the analysis of chemical data
other three markers (Figure 2). Although quercetin is reported in using mathematical expressions. This is done to extract maximum
T. purpurea in the literature, we were unable to found quercetin information from the chemical data and is very well documented for
above the detection limit in any of samples of the plant, so we identification of chemical markers for medicinal plants (11–13).
eliminated it from further analysis. This chemodiversity of plants HPLC data was subjected to chemometric analysis in order to
may be due to genetic, epigenetic and environmental factors that test chemical-relevant variance statistically (17). Peak areas of
6 Katakam et al.

each marker compound were plotted on PCA plots. PCA is a locations. We have found heterogeneous presence of deguelin. Appli-
pattern recognition technique that demonstrated the distribution cation of LC–MS boosts the sensitivity of analysis and successfully
of T. purpurea samples in three different clusters (Figure 3). PC1 determined deguelin in all samples along with rutin and rotenone.
contributed 99% in the distribution and contained most of the The combined approach including validated HPLC-PDA method and
information. As per the PC1, coefficient value and loading plot LC–MS developed in this paper for rapid and simultaneous deter-
displayed the degree of significance of marker compounds as mination and quantitation of three major flavonoids in T. purpurea
rutin>rotenone>deguelin (13). could provide a platform for future analytical and quality control

Downloaded from [Link] by guest on 05 September 2019

In HCA analysis similar to PCA, the three samples from Deesa, studies of the plant.
Pune and Rajkot clustered together and the other samples classified
in second cluster containing two sub clusters (Figure 4). Acknowledgments
The authors are thankful to B. V. Patel Pharmaceutical Education and Research
Discussion Development (PERD) Centre, Ahmedabad and NIPER, Ahmedabad for pro-
viding facilities to conduct the present study. Conflicts of Interest: None
The lack of standardization is a major barrier for commercialization Declared.
of any bioactive plant material (18). Any crude plant raw material
needs to satisfy certain strict quality control criteria related to
efficacy, reliability, authenticity and toxicity whether it is a botanical References
pesticide for plant healthcare or a medicinal drug for human/animal 1. Gora, R.H., Baxla, S.L., Kerketta, P.S., Patnaik, S., Roy, B.K.; Hepatopro-
healthcare. Tephrosia purpurea is an important medicinal plant, but tective activity of Tephrosia purpurea against arsenic induced toxicity in
reports regarding quality control methods are not sufficient. For, rats; Indian Journal of Pharmacology, (2014); 46: 197–200.
standardization of T. purpurea–related herbal material and final 2. Patel, A., Patel, A., Patel, A., Patel, N.M.; Estimation of flavonoid, polyphe-
products, we analyzed the suitability of HPLC-PDA for analysis of nolic content and in-vitro antioxidant capacity of leaves of Tephrosia
three major flavonoids, rutin, deguelin and rotenone. Resolution purpurea Linn. (Leguminosae); International Journal of Pharma Sciences
capacity of HPLC method was found to be high enough to precisely and Research, (2010); 1: 66–77.
3. Hosseinzadeh, H., Nassiri-Asl, M.; Review of the protective effects of rutin
detect these three compounds (deguelin, rotenone and rutin). The
on the metabolic function as an important dietary flavonoid; Journal of
highest amount of these marker compounds was obtained in samples
Endocrinological Investigation, (2014); 37: 783–788.
from Andhra Pradesh, whereas samples from Pune have the lowest
4. Lu, W., Yan, C.; Molecular mechanism of deguelin in anti-tumor effect;
concentration, and deguelin exhibited disparity in its occurrence. Current Pharmaceutical Analysis, (2012); 8: 14–19.
This indicates existence of chemovariations in plant that may be 5. Sharma, R., Mehan, S., Kalra, S., Khanna, D.; Tephrosia purpurea—
due to intra-species genetic variations or the environmental effects. a magical herb with blessings in human biological system; Interna-
Similarly, chemometric analysis also attested the chemical relevant tional Journal of Recent Advances in Pharmaceutical Research, (2013);
variation in the plant samples. This variation was mapped using 3: 12–22.
pattern recognition and hierarchial clustering tools. In both types of 6. Houghton, Mukherjee, P.K. (eds); Evaluation of herbal medicinal prod-
analysis, distributed clustering of the samples was obtained. In PCA ucts. Pharmaceutical Press, London, UK, (2009).
7. Bansal, S.K., Singh, K.V., Sharma, S., Sherwani, M.R.; Laboratory obser-
analysis, loading plot was prepared to demonstrate contribution of
vations on the larvicidal efficacy of three plant species against mosquito
each variable to its principal component (14). It defined significance
vectors of malaria, dengue/dengue hemorrhagic fever (DF/DHF) and lym-
of each marker compound in the analysis. It can be concluded that
phatic filariasis in the semi-arid desert; Journal of Environmental Biology,
rutin contributed the most in classification and clustering. Rutin (2012); 33: 617–621.
can be used as the primary chemical marker for quality control 8. Belmain, S.R., Amoah, B.A., Nyirenda, S.P., Kamanula, J.F., Stevenson,
of the plant along with the other two compounds, as multimarker P.C.; Highly variable insect control efficacy of Tephrosia vogelii
approach has more potential. Likewise, rutin is present commonly chemotypes; Journal of Agriculture and Food Chemistry, (2012); 60:
in other plants also, so incorporation of rotenone and deguelin will 10055–10063.
reduce the ambiguity in analysis and can give more specific chemical 9. Miresmailli, S., Isman, M.B.; Botanical insecticides inspired by plant–
identity of the medicinal material. Simultaneously, quantitation of herbivore chemical interactions; Trends in Plant Science, (2014); 19:
29–35.
rotenone and deguelin can be used to designate the bioactive efficacy
10. Naumovski, V.R., Tongkao-on, W., Kimble, B., Qiao, V.L., Beilun, L.,
of the plant. Furthermore, extracts were subjected to LC–MS analysis
Li, K.M. et al.; Multiple chromatographic and chemometric methods
as we observed inconsistency in detection of deguelin in samples for quality standardisation of chinese herbal medicines; Moderniza-
from diverse habitats during HPLC analysis. LC–MS displayed more tion of Traditional Chinese Medicine and Materia Medica, (2010); 12:
sensitivity and provided unambiguous confirmation of presence of 99–106.
all these metabolites in all the samples. These results profoundly 11. Dhanani, T., Shah, S., Kumar, S.; A validated high-performance liquid
imply the use of multichromatorgraphic analysis as the prevalent chromatography method for determination of tannin-related marker con-
approach for quality control when chemical makers have been stituents gallic acid, corilagin, chebulagic acid, ellagic acid and chebulinic
used. acid in four Terminalia species from India; Journal of Chromatographic
Science, (2015); 53: 625–632.
12. Sethiya, N.K., Mishra, S.; Simultaneous HPTLC analysis of ursolic acid,
Conclusion betulinic acid, stigmasterol and lupeol for the identification of four medic-
inal plants commonly available in the Indian market as shankhpushpi;
In the present study, a reversed-phase HPLC method has been devel-
Journal of Chromatographic Science, (2015); 53: 816–823.
oped for simultaneous identification of rutin, rotenone and deguelin 13. Du, W., Li, X., Yang, Y., Yue, X., Jiang, D., Ge, W. et al.; Quantitative
for quality control of T. purpurea. Data were analyzed using chemo- determination, principal component analysis and discriminant analysis of
metric tools. Results revealed variation in content of bioactive com- eight marker compounds in crude and sweated Dipsaci radix by HPLC-
pounds in T. purpurea collected from different eco-geographical DAD; Pharmaceutical Biology, (2017); 55: 2129–2135.
HPLC Coupled with Chemometric Analysis and LC-MS Studies 7

14. Cordella, C.B.Y.; PCA: the basic building block of chemometrics. In raphy. B, Analytical Technologies in the Biomedical and Life Sciences,
Analytical chemistry, chapter 1. Intech Open, London, UK, (2012), (2016); 15: 82–91.
pp. 1–46. 17. Kim, J.H., Seo, C.S., Kim, S.S., Shin, H.; Quality assessment of Ojeok-san,
15. Hammer, O., Harper, D.A.T., Ryan, P.D.; PAST: Paleontological statistics a traditional herbal formula, using high-performance liquid chromatogra-
software package for education and data analysis; Palaeontologia Elec- phy combined with chemometric analysis; Journal of Analytical Methods
tronica, (2001); 4: 9. in Chemistry, (2015); 2015: 607252.
16. Liu, S., Liang, Y.Z., Liu, H.T.; Chemometrics applied to quality control and 18. Isman, M.B.; Neem and other botanical insecticides: barriers to commer-

Downloaded from [Link] by guest on 05 September 2019

metabolomics for traditional Chinese medicines; Journal of chromatog- cialization; Phytoparasitica, (1997); 25: 339–344.