Research Article

Received: 13 May 2011, Revised: 30 September 2011, Accepted: 14 October 2011 Published online in Wiley Online Library

(wileyonlinelibrary.com) DOI 10.1002/pca.1375

Bio-assay Guided Isolation of a-GlucosidaseInhibitory Constituents from HibiscusMutabilis LeavesDeepak Kumar, Hemanth Kumar, J. R. Vedasiromoni and Bikas C. Pal*

Introduction – The increasing demand for natural-product-based medicines and health-care products for the management ofdiabetes encouraged investigation of this commonly available Indian plant.Objectives – To establish the anti-diabetic (a-glucosidase inhibitory) activity of H. mutabilis leaf extract, isolate and identifythe constituents responsible for the activity, and validate a HPLC method for quantification of the active constituents forstandardisation of the extract.Material and methods – The methanolic extract of leaves was partitioned between water, n-butanol and ethyl acetate.Bio-assay guided fractionation, based on inhibition of a-glucosidase, allowed isolation and identification of the activecomponents. The active components were quantified using RP-HPLC-DAD validated for linearity, limit of detection, limit ofquantification, precision, accuracy and robustness for this plant extract and the partitioned fractions.Results – Ferulic acid and caffeic acid were identified as the a-glucosidase inhibitors present in H. mutabilis. They were partitionedinto an ethyl acetate fraction. The HPLC-DAD calibration curve showed good linearity (r2> 0.99). For the recovery studies the%RSD was less than 2%. The interday and intraday variations were found to be less than 4% RSD for retention time and response.Conclusion – The identification of a-glucosidase inhibition activity in H. mutabilis supports further investigations into thepossible use of the plant for the management of diabetes. The HPLC method validated for these extracts will be useful infuture research with the plant. Copyright © 2011 John Wiley & Sons, Ltd.

Supporting information can be found in the online version of this article.

Keywords: HPLC method validation; a-glucosidase; caffeic acid; ferulic acid; Hibiscus mutabilis

* Correspondence to: Bikas C. Pal, National Institute of PharmaceuticalEducation and Research, 4, Raja S. C. Mullick Road, Jadavpur, Kolkata700032, India. Email: drbcpal@gmail.com

National Institute of Pharmaceutical Education and Research (NIPER), 4,Raja S. C. Mullick Road, Jadavpur, Kolkata 700032, India

INTRODUCTIONDiabetes is a major health-related problem and the treatment oftype-2 diabetes mellitus often requires combined regimens in-cluding diet and medication, which act through different mecha-nisms. One of the approaches to control diabetes is to inhibit intes-tinal a-glucosidases, which controls postprandial hyperglycaemiaby inhibiting the digestion and absorption of carbohydrates (Pulset al., 1977; Jong-Anurakkun et al., 2007). In addition the a-glucosidaseinhibitors are also reported to reduce triglyceride and postpran-dial insulin levels and to possess anti-HIV activity (Bridges et al.,1994; Johnston et al., 1994; Lebowitz, 1998).

H. mutabilis L. (Family: Malvaceae) is commonly known as cot-ton rosemallow or confederate rose. Traditionally the plant isused as an emollient, a stimulant and for pulmonary complaints.The plant is reported to contain flavones, flavone glycosides,anthocyanins and lectin in different parts. The reported bioactivi-ties of the plant extract and isolated constituents include anti-allergic, nitric oxide scavenging, anti-proliferative and HIV-1reverse transcriptase inhibitory activities (Lowry, 1976; Chauhanet al., 1979; Iwaoka et al., 2009; Lam and Ng, 2009). This current in-vestigation focused on examining any inhibition of a-glucosidasepresent in H. mutabilis.

EXPERIMENTALChemicals and reagents

a-Glucosidase (Maltase) and p-nitrophenyl a-D-glucopyranosidewere purchased from SRL Mumbai, India. Acarbose was pur-chased from Sigma-Aldrich. DMSO, NaH2PO4, Na2HPO4, Na2CO3,

Phytochem. Anal. 2011 Copyright © 2011 John

HPLC grade methanol and acetic acid, other chemicals and sol-vents were of highest purity grade and purchased from MerckIndia. Milli-Q water (Milli-Q Academic with 0.22mm Millipak(R) 40)was used for the assays and HPLC analysis.

Plant material

Fresh leaves of the plant were collected from around Kolkata, In-dia and authenticated by Botanical Gardens, Howrah, India. Avoucher specimen has been deposited at the herbarium of theNational Institute of Pharmaceutical Education and Research,Kolkata for future reference (NIP-K/BCP/002).

Extraction and fractionation

Fresh leaves were air dried for 1week in shade at room tempera-ture and ground to a coarse power (40mesh) using a mechanicalgrinder. The leaf powder (2 kg) was extracted with methanol (4 L)at room temperature for 48h. The whole extract was filtered andevaporated under reduced pressure at 40–45 �C using a rotaryevaporator (Eyela). The step was executed five times, and theextracts from each were pooled together and lyophilised to yieldthe crude methanolic extract (142.48 g). A part (100 g) of themethanolic extract was suspended inmilli-Qwater and fractionatedsuccessively with ethyl acetate and n-butanol (Kumar et al., 2010).

Wiley & Sons, Ltd.

D. Kumar et al.

All three fractions were evaporated under reduced pressure andlyophilised to yield residues of ethyl acetate fraction (30.14 g),n-butanol fraction (18.43 g) and aqueous fraction (51.4 g). Themethanolic extract and the three fractions were stored at 4 �C.

Isolation of a-glucosidase inhibitory constituentsand characterisation

A part (5 g) of the ethyl acetate fraction was chromatographedover a bed of silica gel (50 g, 100–200mesh, Merck India) with in-creasing polarity of solvents in the order of petroleum ether(500mL, Fr-1), petroleum ether:chloroform 1:1 (500mL, Fr-2), pe-troleum ether:chloroform 1:3 (500mL, Fr-3), chloroform (500mL,Fr-4), chloroform:methanol 95:5 (500mL, Fr-5), chloroform:metha-nol 90:10 (500mL, Fr-6), chloroform:methanol 80:20 (500mL, Fr-7)and chloroform:methanol 70:30 (500mL, Fr-8). The most activefraction, Fr-6 (1.796g), was chromatographed over a bed of silicagel (20 g, 100–200mesh, Merck India) eluting successively with pe-troleum ether:chloroform 1:1 (200mL, Fr-6-1), petroleum ether:chloroform 1:3 (200mL, Fr-6-2), chloroform (200mL, Fr-6-3), chlo-roform:methanol 95:5 (200mL, Fr-6-4), chloroform:methanol90:10 (200mL, Fr-6-5), chloroform:methanol 85:15 (200mL, Fr-6-6),chloroform:methanol 80:20 (200mL, Fr-6-7), chloroform:methanol75:25 (200mL, Fr-6-8) and chloroform:methanol 70:30 (200mL,Fr-6-9). Among these, Fr-6-4 (635mg) and Fr-6-7 (382mg) showedmaximum inhibition of the enzyme. The active fractions werefurther purified with semi-preparative HPLC (XTerraTM Prep RPC18, 7.8� 300mm, 10mm particle size) and elution was carriedout using methanol:water:acetic acid (34:65:1 v/v/v) at a flow rateof 2.5mL/min, monitoring the eluate at 254nm. Collection of themajor peak from these fractions and evaporation under reducedpressure yielded two amorphous solids, which were characterisedas ferulic acid (from FR-6-4) and caffeic acid (from Fr-6-7) by com-parison of their spectroscopic data (mass, 1H- and 13C-NMR) withthose reported previously (Lee et al., 2001; Tan et al., 2004).

Preparation of enzyme, sample and standard solution

a-Glucosidase enzyme was stored at �20�C. A stock solution of1mg/mL (1mg= 63U) in 0.1 M phosphate buffer, pH 6.8 was pre-pared and diluted to 0.5 U/mL at the time of assay with the samebuffer. The extract and fractions were dissolved in DMSO andfurther diluted to the desired concentrations with buffer. The fi-nal DMSO concentration was maintained below 2% v/v, whichwas found not to have any effect on enzyme activity. For HPLCanalysis the methanolic extract and fractions were accuratelyweighed and dissolved in methanol to obtain a concentrationof 10mg/mL. The isolated constituents were more than 99%pure as determined by HPLC and their standard solutions wereprepared in methanol (1mg/mL).

a-Glucosidase inhibition assay

The a-glucosidase inhibitory assay was performed in a 96-well plateaccording to Li et al. (2009) with some modifications. Briefly, 25mLof sample solution in 0.1 M phosphate buffer were mixed with25mL of enzyme solution containing 0.5U/mL and incubatedat 37� 1 �C for 10min. After the incubation 25mL of the substratep-nitrophenyl a-D-glucopyranoside (0.5mM concentration in 0.1M

phosphate buffer, pH6.8) was added to the mixture, allowed to in-cubate at 37� 1 �C for 30min, and then the reaction was termi-nated by the addition of 100mL of 0.2M sodium carbonate solution.The absorbance of the solution produced was measured at 405nm.Acarbose was used as the reference inhibitor of the enzyme. Theuninhibited enzyme was taken as control; an appropriate DMSO

Copyright © 2011wileyonlinelibrary.com/journal/pca

control was used, wherever applicable. Appropriate blank was usedfor all the samples. The assaywas performed in duplicate and in fourindependent experiments. Percentage inhibition of the enzymewas calculated in comparison with the control and expressed asmean� SD. Inhibition concentration (IC50) values were calculatedfrom concentration versus percentage inhibition curves.

Quantitative analysis by HPLC-DAD

HPLC analyses were performed at 25� 1 �C using sample solu-tions filtered through 0.45mm membrane (Whatman’s syringe fil-ter) and analysed (10mL injected volume) using a Shimadzu(Japan) system equipped with SCL-10A VP Shimadzu system con-troller, SPD-M10A VP Shimadzu diode array detector, LC-10AT VPShimadzu liquid chromatography pump, Class VP software, and aRheodyne injector with 20mL loop. Separation was achieved usingWater’s XTerraTM RP C18-column, 4.6� 250mm, 5mm particlesize; isocratic elution was carried out using the mobile phase of1% acetic acid in methanol:water (30:70 v/v) at a flow rate of1mL/min. The eluate was monitored at 254 nm. Quantitative esti-mation of the constituents present in the methanolic extract andfractions was carried out by using the calibration curve of the stan-dard solution (Microsoft Office Excel 2007).

Method validation

The HPLC method validation parameters such as linearity, limitof detection (LOD), limit of quantification (LOQ), accuracy, preci-sion and repeatability, were studied (ICH, 2005; Chen et al.,2011). The mixed standard solution was diluted serially to obtainsolutions of five different concentrations, which were injected intriplicate and their regression equations were calculated. LODand LOQ were calculated by determining standard deviation(SD) of the response of a number of blank runs and slope ofthe linear equation. The accuracy of the HPLC method was stud-ied by the standard addition technique to calculate percentagerecovery of ferulic acid and caffeic acid in the methanolic extractand fractions. Interday (n=3) and intraday (n= 6) variabilitieswere studied to analyse the precision in three different concen-trations. The robustness of the method was analysed at differentcolumn temperatures and flow rates, mobile phases and col-umns of the same configuration (Waters Spherisorb, 5mmODS2, 4.6� 250mm) to determine the effect on retention time.

Statistical analysis

All the statistical analyses were performed using Graph PadPrism and expressed in terms of mean� SD/SEM. Statistical anal-ysis was done by one-way ANOVA followed by the Bonferronipost-test (Graph pad prism); p< 0.05 and p< 0.001 were consid-ered significant as compared to control.

RESULTS AND DISCUSSION

a-Glucosidase inhibitory potential

In the present investigation H. mutabilis leaf extract was evalu-ated for its a-glucosidase inhibitory potential for the first time.The crude methanolic extract demonstrated considerable inhibi-tion of the enzyme a-glucosidase in a concentration-dependentmanner (2.5–10mg/mL, Fig. 1). The extract was fractionated intothree fractions among which the ethyl acetate fraction wasfound to be most active, while the n-butanol fraction showedsome inhibition of the enzyme. The aqueous fraction showed

Phytochem. Anal. 2011John Wiley & Sons, Ltd.

Figure 1. Inhibitory effect of (A) methanolic extract and (B) ethyl acetate fraction on a-glucosidase (0.5 U/mL, using substrate p-nitrophenyl a-D-glucopyranoside 0.5mM) determined at concentrations ranging from 2.5 to 10mg/mL for methanolic extract and 0.125 to 1mg/mL for ethyl acetatefraction. * and ** represent significant difference compared with control at p< 0.05 and p< 0.001, respectively.

Table 1. Effect of H. mutabilis leaf extract, its fractions, andof ferulic acid and caffeic acid on the enzyme a-glucosidase

Sample IC50*

Methanolic Extract 7.534� 0.177mg/mLEthyl acetate fraction 0.704� 0.070mg/mLn-butanol fraction > 10mg/mLAqueous fraction NIFr-6 0.630� 0.052mg/mLFerulic acid 0.335� 0.031mg/mLCaffeic acid 0.283� 0.036mg/mLAcarbose 9.675� 0.089mg/mL

* Mean� SD; n=4 and NI = no inhibition

Table 2. Concentration of ferulic acid and caffeic acid in H.mutabilis leaf extract and fractions as determined by HPLC

Sample Concentration (% w/w)*

Ferulic acid Caffeic acid

Methanolic extract 0.295� 0.041 0.096� 0.013Ethyl acetate fraction 7.346� 0.687 3.056� 0.433n-butanol fraction 0.288� 0.025 0.058� 0.011Aqueous fraction ND 0.015� 0.002

* Mean� SD; n=3, ND=not detected

Figure 2. HPLC chromatogram obtained for (A) methanolic extract, (B) ethyl acetate fraction, (C) n-butanol fraction and (D) aqueous fraction using theisocratic mobile phase 1% acetic acid in methanol:water (30:70 v/v) at a flow rate of 1mL/min and eluate monitoring at 254 nm. Arrows indicate peaksfor ferulic acid (retention time 28.942min) and caffeic acid (retention time 13.257min).

Isolation and Quantification of a-Glucosidase Inhibitors of H. Mutabilis

no inhibition of the enzyme. The bioactivity-guided approachled to the isolation of ferulic acid and caffeic acid from the ethylacetate fraction for the first time from the plant. These constitu-ents, present in very low concentrations in the n-butanol fraction

Phytochem. Anal. 2011 Copyright © 2011 John Wile

and negligible in the aqueous fraction, showed potent inhibitionof the enzyme (Table 1). Both the constituents, from othersources, have been previously reported for their potential to in-hibit a-glucosidase (Adisakwattana et al., 2009). The inhibition of

y & Sons, Ltd. wileyonlinelibrary.com/journal/pca

n=3) Re

spon

se

Mean

%RS

D

3202

841.23

1674

313.61

4990

770.72

D. Kumar et al.

the enzyme by the extract/fractions and isolated constituents isless compared with the reference inhibitor acarbose, but keep-ing in mind the multiple health benefits of herbal medicines thisreport may provide further scope for evaluation and/or utilisa-tion of the plant alone or in combinations for the treatment ormanagement of diabetic complications.

Interday

(

tiontim

e

%RS

D

0.18

02

0.33

51

0.30

7

Method validation

The method developed for the quantification of ferulic acid andcaffeic acid (Table 2; Fig. 2) was validated and showed good

Table 3. Limit of detection (LOD) and limit of quantification(LOQ) of ferulic acid and caffeic acid obtained by the RP-HPLC-DAD method

Samples Linearityrange(mg/mL)

LOD(mg/mL)

LOQ(mg/mL)

Correlationcoefficient

(r2)

Linearityequation

Ferulicacid

3–337 0.314 1.047 0.996 y=15196x+12216

Caffeicacid

3–337 0.290 0.968 0.999 y=13212x+22225

Table 4. Recovery study of ferulic acid and caffeic acid fromH. mutabilis leaf extract and fractions by the RP-HPLC-DADmethod

Sample Amountpresent(mg)

Amountadded(mg)

Totalamount(mg)

Amountfound(mg)

Percentagerecovery

%RSD

FA+ME 29.5 33 62.5 61.21 96.09 1.8129.5 66 95.5 95.26 99.6329.5 167 196.5 193.05 97.93

FA+ EAF 734.6 33 767.6 766.31 96.09 0.72734.6 66 800.6 798.15 96.29734.6 167 901.6 897.24 97.39

FA+NBF 28.8 33 61.8 61.24 98.30 0.6028.8 66 94.8 93.27 97.6828.8 167 195.8 193.9 98.86

FA+AF 0 33 33 32.59 98.76 0.290 66 66 65.26 98.880 167 167 165.85 99.31

CA+ME 9.6 33 42.6 42.02 98.24 0.999.6 66 75.6 73.17 96.329.6 167 176.6 171.94 97.21

CA+ EAF 305.6 33 338.6 338.29 99.06 0.55305.6 66 371.6 370.92 98.97305.6 167 472.6 472.54 99.96

CA+NBF 5.8 33 38.8 38.17 98.09 1.185.8 66 71.8 71.46 99.485.8 167 172.8 168.09 97.17

CA+AF 1.5 33 34.5 34.01 98.52 0.461.5 66 67.5 66.63 98.681.5 167 168.5 164.88 97.83

FA= ferulic acid, CA = caffeic acid, ME=methanolic extract,EAF = ethyl acetate fraction, NBF = n-butanol fraction, AF =aqueous fraction. Ta

ble

5.Intrad

ayan

dinterday

precisionof

ferulic

acid

andcaffeicacid

determ

inationby

RP-HPLC-DADmetho

d

Caffeicacid

Ferulic

acid

Intrad

ay(n=6)

Interday

(n=3)

Intrad

ay(n=6)

Retentiontim

eRe

spon

seRe

tentiontim

eRe

spon

seRe

tentiontim

eRe

spon

seRe

ten

Mean

%RS

DMean

%RS

DMean

%RS

DMean

%RS

DMean

%RS

DMean

%RS

DMean

167mg

/mL

13.252

0.02

321

9271

51.84

13.259

0.42

324

0954

40.88

28.952

0.15

923

4828

31.38

28.939

66mg

/mL

13.247

0.21

197

6358

0.64

13.269

0.24

987

6243

0.69

28.947

0.09

710

9559

23.29

28.926

33mg

/mL

13.252

0.07

542

7635

1.25

13.262

0.26

442

9882

1.86

28.944

0.12

149

7941

1.51

28.946

Copyright © 2011wileyonlinelibrary.com/journal/pca

Phytochem. Anal. 201John Wiley & Sons, Ltd. 1

Isolation and Quantification of a-Glucosidase Inhibitors of H. Mutabilis

linearity (r2> 0.99) in a concentration range wide enough toquantify the constituents in the extract. Table 3 shows the line-arity range, LOD, LOQ and linearity equation for both the constit-uents determined by injecting a number of concentrations ofthe standard solution and establishing the standard deviationof response of a number of blank runs (n= 9). The accuracy ofthe method was proved by percentage recovery on standardaddition of ferulic acid and caffeic acid to methanolic extractand the fractions. Table 4 shows the percentage recovery and%RSD of both the constituents when determined in the metha-nolic extract and the fractions at three different concentrations.The %RSD was found to be less than 2% for all the sets, which sug-gests that themethod is accurate enough to analyse both the con-stituents in the leaf extract as well as the fractions. The interday(n=3) and intraday (n=6) precision study was performed atthree different concentrations for the retention time and response(Table 5). The variations were less than 4%, which may be consid-ered to be in the acceptable range. The robustness of the methodwas also in the acceptable range as there is not much variation onthe retention time and separation of the analysed constituents.Therefore the HPLC-DAD method is suitable for quantitativeanalysis of caffeic and ferulic acids in extracts (and subsequentfractions) of H. mutabilis.

Acknowledgements

Authors are thankful to the Council of Scientific and IndustrialResearch (CSIR), New Delhi, India for financial support (21(0775)/09/EMR-II) and Director IICB Kolkata for providing thenecessary facilities.

Supporting informationSupporting information can be found in the online version ofthis article.

REFERENCESAdisakwattana S, Chantarasinlapin P, Thammarat H, Yibchok-Anun S.

2009. A series of cinnamic acid derivatives and their inhibitory

Phytochem. Anal. 2011 Copyright © 2011 John Wile

activity on intestinal alpha-glucosidase. J Enzyme Inhib Med Chem24: 1194–1200.

Bridges CG, Brennan TM, Taylor DL, McPherson M, Tyms AS. 1994. Theprevention of cell adhesion and the cell-to-cell spread of HIV-1 in-vitroby the a-glucosidase 1 inhibitor, 6-O-butanoyl castanospermine (MDL28574). Antiviral Res 25: 169–175.

Chauhan JS, Vidyapati TJ, Gupta AK. 1979. A new flavanone glycosidefrom the stem of Hibiscus mutabilis. Phytochemistry 18: 1766–1767.

Chen H, Zhang Q, Wang X, Yang J, Wang Q. 2011. Qualitative analysis andsimultaneous quantification of phenolic compounds in the aerialparts of Salvia miltiorrhiza by HPLC-DAD and ESI/MS(n). PhytochemAnal 22: 247–257.

ICH. 2005. Q2(R1). Validation of Analytical Procedures: Text and Methodol-ogy. International Conference on Harmonization: Geneva.

Iwaoka E, Oku H, Takahashi Y, Ishiguro K. 2009. Allergy-preventive effectsof Hibiscus mutabilis ‘versicolor’ and a novel allergy-preventive flavo-noid glycoside. Biol Pharm Bull 32: 509–512.

Johnston PS, Coniff RF, Hoogwerf BJ, Santiago JV, Pi-Sunyer, FX, Krol A.1994. Effects of the carbohydrase inhibitor miglitol in sulfonylurea-treated NIDDM patients. Diabetes Care 17: 20–29.

Jong-Anurakkun N, Bhandari MR, Kawabata J. 2007. a-Glucosidaseinhibitors from Devil tree (Alstonia scholaris). Food Chem 103:1319–1323.

Kumar D, Mallick S, Vedasiromoni JR, Pal BC. 2010. Anti-leukemic activityof Dillenia indica L. fruit extract and quantification of betulinic acid byHPLC. Phytomedicine 17: 431–435.

Lam SK, Ng TB. 2009. Novel galactonic acid-binding hexameric lectinfrom Hibiscus mutabilis seeds with antiproliferative and potentHIV-1 reverse transcriptase inhibitory activities. Acta Biochim Pol 56:649–654.

Lebowitz HE. 1998. a–Glucosidase inhibitors as agents in the treatmentof diabetes. Diabetes Rev 6: 132–145.

Lee HS, Beon MS, Kim MK. 2001. Selective growth inhibitor toward hu-man intestinal bacteria derived from Pulsatilla cernua root. J AgricFood Chem 49: 4656–4661.

Li HL, Song FR, Xing JP, Tsao R, Liu ZQ, Liu SY. 2009. Screening and struc-tural characterization of a-glucosidase inhibitors from hawthorn leafflavonoids extract by ultrafiltration LC-DAD-MSn and SORI-CID FTICRMS. J Am Soc Mass Spectrom 20: 1496–1503.

Lowry JB. 1976. Floral anthocyanins of some malesian Hibiscus species.Phytochemistry 15: 1395–1396.

Puls W, Keup U, Krause HP, Thomas G, Hofmeister F. 1977. Gluco-sidase inhibition: a new approach to the treatment of diabe-tes, obesity, and hyperlipoproteinaemia. Naturwissenschaften 64:536–537.

Tan J, Bednarek P, Liu J, Schneider B, Svatos A, Hahlbrock K. 2004. Univer-sally occurring phenylpropanoid and species-specific indolic metabo-lites in infected and uninfected Arabidopsis thaliana roots and leaves.Phytochemistry 65: 691–699.

y & Sons, Ltd. wileyonlinelibrary.com/journal/pca