This article appeared in a journal published by Elsevier. The attachedcopy is furnished to the author for internal non-commercial researchand education use, including for instruction at the authors institution

and sharing with colleagues.

Other uses, including reproduction and distribution, or selling orlicensing copies, or posting to personal, institutional or third party

websites are prohibited.

In most cases authors are permitted to post their version of thearticle (e.g. in Word or Tex form) to their personal website orinstitutional repository. Authors requiring further information

regarding Elsevier’s archiving and manuscript policies areencouraged to visit:

http://www.elsevier.com/copyright

Author's personal copy

Journal of Ethnopharmacology 137 (2011) 592– 600

Contents lists available at ScienceDirect

Journal of Ethnopharmacology

jo ur nal homep age : www.elsev ier .com/ locate / je thpharm

Pharmacological evaluation of insulin mimetic novel suppressors of PEPCK genetranscription from Paeoniae Rubra Radix

Yi-Chen Juana,b,1, Chia-Chuan Changc,1, Wei-Jern Tsai c, Yun-Lian Lind, Yi-Shin Hsua, Hui-Kang Liua,∗

a Division of Herbal Drugs and Natural Products, National Research Institute of Chinese Medicine, Taipei, Taiwan, ROCb Institute of Pharmacology, University of Yang Ming, Taipei, Taiwan, ROCc Division of Basic Chinese Medicine, National Research Institute of Chinese Medicine, Taipei, Taiwan, ROCd Division of Medicinal Chemistry, National Research Institute of Chinese Medicine, Taipei, Taiwan, ROC

a r t i c l e i n f o

Article history:Received 30 August 2010Received in revised form 3 June 2011Accepted 9 June 2011Available online 16 June 2011

Keywords:Paeoniae Rubra RadixNon-PGG fraction (NPF)Phosphoenolpyruvate carboxykinase(PEPCK)Glucocorticoids antagonistcAMP-responsive element-binding protein(CREB) inhibitor1,2,3,4,6-Penta-O-galloyl-beta-d-glucose(PGG)

a b s t r a c t

Ethnopharmacological relevance: Paeoniae Rubra Radix (root of Paeonia lactiflora) has been frequentlyemployed in Traditional Chinese Medicine (TCM) as and anti-diabetic therapy to enhance blood circula-tion and dissipate stasis.Aim of the study: Previously, we identified a novel hypoglycemic action of a crude extract from PaeoniaeRubra Radix, which also suppressed phosphoenolpyruvate carboxykinase (PEPCK) gene transcription.Therefore, the current investigation intended to elucidate potential active bio-constituents of this herband mechanisms of action.Materials and methods: Glucocorticoid receptor (GR) nuclear localization, the PEPCK messenger (m)RNAlevel, pregnane X receptor (PXR) mRNA expression, cAMP-responsive element-binding protein (CREB)serine phosphorylation and DNA binding were evaluated in dexamethasone (Dex) and 8-bromo-cAMP(CA)-stimulated H4IIE cells, while efficacy of agents was assessed in a stable cell line containing a greenfluorescent protein (GFP) reporter driven by the PEPCK promoter. HPLC profiling, colorimetric assays,and NMR analysis were employed for chemical characterization purpose.Results: An extract of Paeoniae Rubra Radix lacking the insulin mimetic compound, 1,2,3,4,6-penta-O-galloyl-beta-d-glucose (PGG), and termed the non-PGG fraction (NPF), consisting of tannin polymers,suppressed PEPCK expression in the presence of an insulin receptor antagonist (HNMPA-AM3), suggest-ing the action of this fraction is independent of the insulin receptor. Furthermore, Dex-stimulated GRnuclear localization and transactivation were prevented by the NPF. Similarly, CA-stimulated CREB ser-ine phosphorylation and DNA binding were also inhibited by the NPF in H4IIE cells. Hence NPF antagonizesboth signaling pathways that induce PEPCK gene transcription.Conclusion: In conclusion, the current study proposes that the potent suppressive activity on PEPCKgene transcription observed with Paeoniae Rubra Radix extract, can be attributed to at least two distinctcomponents, namely PGG and NPF.

© 2011 Elsevier Ireland Ltd. All rights reserved.

1. Introduction

Phosphoenolpyruvate carboxykinase (PEPCK) is a key enzymerate-determining hepatic gluconeogenesis and transcriptional reg-ulation plays an important role to regulate the enzyme activity(Yabaluri and Bashyam, 2010). Both cortisol from the adrenal cor-tex and glucagons from islets could increase PEPCK activity viaelevation of PEPCK mRNA level. In contrast, insulin dominantlysuppresses PEPCK mRNA induced by both hormones. However, the

∗ Corresponding author. Tel.: +886 2 28201999x3711; fax: +886 2 28250743.E-mail address: hk.liu@nricm.edu.tw (H.-K. Liu).

1 These authors contributed equally to this work.

inhibitory mechanism of hepatic gluconeogenesis is attenuated dueto insulin resistance or insulin deficiency in diabetes. As a result, up-regulation of PEPCK mRNA and over-activated gluconeogensis andglucose production in liver leads to a promotion of fasting hyper-glycemia. Restoration of abnormal PEPCK expression could benefitdiabetic population.

The model of Dex and CA stimulated PEPCK transcription iswell established and represents a situation in response to gluco-corticoid and glucagons (cAMP as a second messenger) (Magnusonet al., 1987). Several transcriptional factors play important roles inPEPCK transcription via binding to corresponding cis-elements inthe PEPCK promoter region. In response to stress hormones such asglucocorticoid, the interaction of its cytosolic glucocorticoid recep-tors (GRs), promote the translocation into the nuclear compartment

0378-8741/$ – see front matter © 2011 Elsevier Ireland Ltd. All rights reserved.doi:10.1016/j.jep.2011.06.007

Author's personal copy

Y.-C. Juan et al. / Journal of Ethnopharmacology 137 (2011) 592– 600 593

and GR DNA binding to the complex glucocorticoid-responsive unit(GRU). In contrast, in response to glucagons and catecholamines,protein kinase A (PKA) activation by second messenger cAMPresults in the promotion of CREB DNA binding to a cAMP-responsiveelement (CRE) via protein phosphorylation (Nichols et al., 1992;Xing and Quinn, 1993; Vallejo et al., 1995). The presence of insulindominantly suppressed glucocorticoid- and/or glucagon-inducedPEPCK transcription via thymidine-rich insulin response elements(TIREs) at the GRU (Patel et al., 2003), disrupted the protein com-plex at the CRE (He et al., 2009), and influenced the epigenetics ofthe promoter (Hall et al., 2007). In terms of insulin’s action, it wasclearly demonstrated that the PI3 kinase (PI3K)-protein kinase B(PKB)-glycogen synthase kinase (GSK)3 pathway plays an impor-tant role in insulin-mediated suppression of PEPCK (Sutherlandet al., 1996). Alternatively, under an insulin-resistance condition,the activation of AMPK can also facilitate suppression of PEPCKtranscription in an insulin-independent manner (Lochhead et al.,2000).

In history, Paeoniae Rubra Radix (root of Paeonia lactiflora)is frequently employed in TCM formula in diabetes therapy toactivate blood and dissipate stasis. The association among increas-ing arterial stiffness, reducing blood flow, and glucose or insulinmetabolism is well documented (Cameron and Cruickshank, 2007;Perrin et al., 2007).

In our previous studies, we demonstrated that the ethanolextract of Paeoniae Rubra Radix (PR-Et) could suppress phos-phoenolpyruvate carboxykinase (PEPCK) messenger (m)RNAexpression both in vitro and in vivo. In addition, the suppressiveeffect of the PR-Et on the dexamethasone (Dex) and 8-bromo-cAMP(CA) induced PEPCK (m)RNA level appeared to be unaffected by aninsulin-desensitized condition. The possibility that two major com-pounds, paeonol and paeoniflorin, suppressed PEPCK was ruled out(Juan et al., 2010). Therefore, both responsible bioactive compo-nent(s) and potential mechanisms of actions on PEPCK suppressionby this herb remained to be elucidated.

According to the literature, 1,2,3,4,6-penta-O-galloyl-beta-d-glucose (PGG) can bind to the insulin receptor to facilitate glucosetransport in 3T3-L1 adipocytes (Li et al., 2005). By a high-performance liquid chromatographic (HPLC) examination, it wasdetermined that this phytochemical also exists in the PR-Et. How-ever, we also obtained a novel non-PGG fraction (NPF) whichalso possesses an inhibitory effect on PEPCK transcription. There-fore, the purpose of the current investigation was to elucidate thepotential inhibitory actions of this fraction on PEPCK transcrip-tion.

2. Methods

2.1. Materials

Dex and CA were purchased from Sigma (St. Louis, MO, USA).PGG was purchased from International Laboratory (San Bruno,CA, USA). Paeoniflorin and paeonol were purchased from NacalaiTesque (Kyoto, Japan). Insulin was purchased from Novo Nordisk(Princeton, NJ, USA). The TRI reagent was purchased from Invitro-gen (Taipei, Taiwan). LY294002 and HNMP-(AM)3 were purchasedfrom Merck (Darmstadt, Germany). Rat hepatoma H4IIE cells werepurchased from the Bioresource Collection and Research Center(BCRC; Taichung, Taiwan).

2.2. Preparation of the NPF from Paeoniae Rubra Radix

The NPF was prepared from the PR-Et by Prof. Y.L. Lin (Juanet al., 2010). Briefly, slices of the dried commercially availablePaeoniae Rubra Radix (Paeonia lactiflora roots, 2 kg) were extracted

with 80% ethanol/H2O at 60 ◦C overnight. The extract was evap-orated to dryness under a vacuum to yield the crude extract(797 g, ∼40% of raw materials). A portion of the crude extract(500 g) was resuspended in H2O and added to a Diaion HP-20 col-umn (10 cm × 70 cm). Elution with a gradient of methanol in H2Ogave 5 major fractions. The fraction eluted with pure methanolwas combined and concentrated to leave a brown residue whichwas used as the experimental material (NPF, ∼21 g). Thin-layerchromatography (TLC) was performed on a silica gel plate anddeveloped with solvent consisting of ethyl acetate:methanol:H2O(10:0.5:0.5). A high-performance liquid chromatography (HPLC)apparatus (Shimadzu, Kyoto, Japan) assembled with an LC-10ATpump, an SPD-10AV UV–vis spectrophotometric detector, and anODS column was employed for both the PR-Et and NPF chem-ical profiling. The solvent system was set as: solvent A: 20%acetonitrile/H2O; solvent B: acetonitrile; 0–10 min, A: 100–85%;10–30 min, A: 85–30%; at a flow rate of 1 ml/min. The detectionwavelength was set to 230 nm. Paeoniflorin, PGG, and paeonol wereemployed under the same HPLC running conditions as the chem-ical standard. Qualitative characterization of NPF was carried outby colorimetric assays (Makkar and Becker, 1993). The NPF (10 mg)was dissolved in 1 ml of DMSO and diluted to 10 ml by distilledH2O to become the test solution. The test solution (1 ml) was thenadded with one drop of 1% ferric chloride (Alfa Aesar, Ward Hill,USA), one drop of three solutions stepwise [(1) 1% lead tetra-acetate(Sigma–Aldrich, St. Louis, USA); (2) one drop of 4N acetic acid(Mallinckrodt Baker, Phillipsburg, USA); (3) one drop of ammo-nium iron (III) sulfate (Merck KGaA, Darmstadt, Germany)], and3% bromine water (1 ml) (Acros, Geel, Belgium), correspondingly.The 1H NMR and 13C nuclear magnetic resonance (NMR) spectraof the NPF were obtained from a Varian VNMRS NMR (600 MHz)with a DCH Cold Probe spectrometer. Chemical shifts (�) were givenin parts per million (ppm) relative to �H 2.49/�C 39.7 for DMSO-d6.

2.3. The measurement of condensed tannin content byvanillin–HCl test

NPF solution was treated with 4% vanillin (w/v in MeOH) andconcentrated HCl. After 20 min, the absorption at 500 nm was mea-sured for (±)-catechin equivalent content by comparing with thestandard curve in five different concentration of (±)-catechin (0.25,0.5, 1.0, 1.5 and 2.0 mg/ml) (Muchuweti et al., 2005).

2.4. Dialysis of the NPF

40 mg of the NPF was dialyzed by a molecular-porous membranetubing (MWCO 12–14,000) to give the inner-membrane fraction(NPF-DI; 33.2 mg) and outer-membrane fraction (NPF-DO). Theratio of the inner membrane to the total loading amount was 83.1%.Both fractions were lyophilized for later cell based bio-assay.

2.5. Cell culture

H4IIE cells were cultured with Dulbecco’s modified Eagle’smedium (DMEM) (Gibco, Grand Island, NY, USA) containing 1 g/lglucose, 5% (v/v) fetal bovine serum (FBS), and 1% (v/v) antibiotics(100 U/ml penicillin and 0.1 g/l streptomycin) and maintained at37 ◦C in an atmosphere of 5% CO2 and 95% air. For experiments, oncecells became confluent, cells were serum starved overnight andthen stimulated with Dex (500 nM) and/or CA (100 �M) in the pres-ence or absence of test agents described in the figure legends. ForPEPCK expression, RNA and Protein was harvested at 4 h or 8 h post-treatment, respectively. For signal transduction analysis, cells wereharvested for protein analysis after treated with insulin (10 nM) orPGG for 30 min. H4-pPCK-GFP cells containing the PEPCK promoter

Author's personal copy

594 Y.-C. Juan et al. / Journal of Ethnopharmacology 137 (2011) 592– 600

regulating a green fluorescent protein (GFP) reporter were culturedwith DMEM consisting of 5% FBS and G418 (2.2 mg/ml) according toprevious description (Juan et al., 2010). After overnight serum star-vation, cells were treated with inhibitors for 30 min prior to addinginsulin, PGG or NPF. Cells were harvested after 8 h post-treatment.

2.6. Reverse-transcription polymerase chain reaction (RT-PCR)

Gene expression was carried out based on a previousdescription (Chen et al., 2008). In brief, treated H4IIE cellswere directly homogenized in TRI reagent, and RNA wasextracted according to the manufacturer’s instructions. RTwas performed with a RevertAidTM First Strand cDNA Synthe-sis kit (Fermentas, Burlington, ON, Canada). Complementary(c)DNA at 50 ng was employed for a further PCR. Primersequences for �-actin were 5′-CGTAAAGACCTCTATGCCAA-3′ and 5′-AGCCATGCCAAATGTGTCAT-3′, for PEPCKwere 5′-AAGGCCGCACCATGTATGTC-3′ and 5′-AGCAGTGAGTTCCCACCGTAT-3′, and for PXR were5′-ATGTCTGATGCCGCTGTG-3′ and 5′-TGGAGGGAGGTTGGTAGTT-3′. In addition, the annealing temperatures for amplification of�-actin (57 ◦C), PEPCK (57 ◦C), and PXR (55 ◦C) were employedto generate PCR products with sizes of 349, 319, and 352 bp,respectively. PCR products were separated by gel electrophoresis,visualized, photographed with a digital camera, and quantifiedwith Genetools 3.06 (Syngene, Frederick, MD, USA).

2.7. Reporter assays

H4-pPCK-GFP cells were seeded in 6-cm dishes and allowed togrow to confluence. After overnight serum starvation, cells weretreated under various conditions described in the figure legendsfor another 8 h. At the end of treatment, the whole-cell lysate wasplaced in 96-well plates to measure GFP using Flexstation (Molec-ular Device, Sunnyvale, CA, USA), where lysates were excited at475 nm, and the emission was detected at 525 nm. The resultingrelative fluorescent units (RFU) were normalized to the protein con-centration. Finally RFU/protein concentration in a serum-starvedcondition was set to 100% to determine the promoter responsive-ness under each condition.

2.8. Western blot

Western blotting was carried out based on a previous descrip-tion (Chen et al., 2008). In brief, cells were washed with ice-coldphosphate-buffered solution (PBS) twice and scraped into ice-cold lysis buffer (25 mM Tris/HCl (pH 7.4), 50 mM NaF, 100 mMNaCl, 1 mM sodium vanadate, 5 mM EGTA, 1 mM EDTA, 1% (v/v)Triton X-100, 10 mM sodium pyrophosphate, 1 mM benzamidine,0.1 mM PMSF, 0.27 M sucrose, 2 �M microcystin, and 0.1% (v/v)2-mercaptoethanol). For the nuclear protein analysis or later DNA-binding assay, washed cells were collected, and the nuclear proteinwas extracted by employing a nuclear extraction kit (Active Motif,Carlsbad, CA, USA) according to the manufacturer’s instructions.The protein concentration was determined by the Bradford assay(Sigma) using bovine serum albumin (BSA) as the standard basedon the manufacturer’s instructions. Equal amounts of proteinsamples (40 �g) were separated by a 10% sodium dodecylsulfatepolyacrylamide gel electrophoresis (SDS-PAGE) and then trans-ferred to an activated polyvinylidene difluoride (PVDF) membrane.The blot membrane was further blocked with 5% non-fat milk inTBST (Tris-buffered saline containing 0.1% (v/v) Tween 20) priorto replacement with the appropriate primary antibodies at 4 ◦Covernight. The next day, after further incubation with the corre-sponding secondary antibodies, the results were detected by anenhanced chemiluminescence (ECL) kit (Amersham Biosciences,

Taipei, Taiwan). Primary anti-bodies used in the current investiga-tion including anti-GR (anti-mouse; purchased from Calbiochem,Darmstadt, Germany), PKB serine473 (anti-rabbit; purchased fromCell signaling technology, Danvers, MA, USA), PKB total (anti-rabbit; from Cell signaling technology), CREB total (anti-rabbit;from Cell signaling technology), CREB serine133 (anti-rabbit; fromCell signaling technology), GSK3 serine9 (anti-rabbit; from Cell sig-naling technology), GS serine641 (anti-rabbit; from Cell signalingtechnology), PEPCK (anti-rabbit; purchased from Cayman Chemi-cal, Ann Arbor, MI, USA), and actin (anti-mouse; purchased fromChemicon, Billerica, MA, USA).

2.9. ELISA based DNA-binding assay

To evaluate the effect of the NPF on GR and CREB DNA bindingto the consensus GRE or CRE sequences, we used an enzyme-linked immunosorbent assay (ELISA)-based approach according tothe manufacturer’s instructions (TransAMTM, Active Motif). Briefly,H4IIE cells were exposed to Dex or CA for 1 h in the presenceof the NPF at indicated concentrations, and then nuclear extractswere prepared. Extracts were incubated in 96-well plates contain-ing bound oligonucleotides including the GRE or CRE consensusmotif. After the capture of GR or phosphorylated CREB by theoligonucleotide, a primary antibody to GR or CREB was added. Ahorseradish peroxidase (HRP)-conjugated secondary antibody wasused to quantify the DNA-binding activity of GR or CREB from thecolorimetric readouts.

2.10. Statistical analysis

All results are expressed as the mean ± standard error of themean (SEM). Statistical significance was determined by a one-wayanalysis of variance (ANOVA) followed by the Tukey–Kramer test.Differences were considered significant when p < 0.05, p < 0.01, andp < 0.001.

3. Results

3.1. An NPF isolated from Paeoniae Rubra Radix suppressedPEPCK transcription in H4IIE cells in an insulinreceptor-independent manner

As shown in Fig. 1A, a qualitative TLC analysis revealed thatthe reported insulin mimetic, PGG was not detectable in thefractionated Paeoniae Rubra Radix ethanol extract. In addition, acomparative HPLC analysis indicated that paeoniflorin, PGG, orpaeonol, could not be detected in this extract, hereafter referredto as non-PGG fraction (NPF) (Fig. 1B). In addition, judged by theretention times of HPLC, detection peaks of the NPF appeared tobe quicker than those of the PGG standard (Fig. 1B). NPF, at con-centrations of >10 �g/ml dose-dependently suppressed Dex andCA-induced PEPCK gene transcription in H4IE cells (Fig. 1C). Thepotency of the NPF was approximately 10-fold greater than theparent PR-Et preparation. Finally, as shown in Fig. 1D, the NPF sup-pressed PEPCK promoter-regulated GFP activity at 20 �g/ml. Thesuppressive effect of the NPF on the GFP reporter was unaffectedby the presence of the insulin receptor antagonist HNMPA-AM3(50 �M), a treatment that partially blocked the action of insulin onthis PEPCK-reporter (Fig. 1D).

3.2. The NPF reduced Dex-mediated GR nuclear localization andtransactivation

Next we evaluated the effect of the NPF on glucocorticoid recep-tor (GR) localization and activity, a factor previously shown tomediate the glucocorticoid induction of the PEPCK gene promoter.

Author's personal copy

Y.-C. Juan et al. / Journal of Ethnopharmacology 137 (2011) 592– 600 595

Fig. 1. Identification of the non-PGG fraction (NPF) with a suppressive effect on dexamethasone (Dex) and 8-bromo-cAMP (CA)-induced phosphoenolpyruvate carboxykinase(PEPCK) mRNA expression in an insulin receptor-independent manner. (A) A qualitative TLC analysis was carried out to compare the distribution of chemical spots from1,2,3,4,6-penta-O-galloyl-beta-d-glucose (PGG) or the NPF. (B) A qualitative HPLC analysis was carried out to compare the chemical profiling of the ethanol extract of PaeoniaeRubra Radix (PR-Et) and NPF. Paeoniflorine, PGG, and paeonol were the three standards used. (C) A representative picture and quantified results of Dex + CA-induced PEPCKmRNA expression in the presence of the PR-Et or NPF are shown. Insulin (10 nM) served as a positive control. All quantitative data are presented as the mean ± SEM (n ≥ 4).*p < 0.05 and **p < 0.01, significant differences compared to the Dex + CA only group. (D) H4-pPCK-GFP cells were generated from H4IIE cells stably integrated with the PEPCKpromoter-driven green fluorescent protein (GFP) reporter. Full promoter activity (100%) is relative fluorescence units (RFU) per mg total protein under serum-free conditions.All quantitative data are presented as the mean ± SEM (n ≥ 10). ***p < 0.001, compared to the basal condition (none). ��p < 0.01, compared to the response in the absence ofHNMPA-AM3.

Author's personal copy

596 Y.-C. Juan et al. / Journal of Ethnopharmacology 137 (2011) 592– 600

Fig. 2. The non-PGG fraction (NPF) intervenes in dexamethasone (Dex)-mediated glucocorticoid receptor (GR) nuclear localization, glucocorticoid responsive element (GRE)DNA binding, and pregnane X receptor (PXR) mRNA expression. (A) A representative result of Dex-induced GR protein expression in the nuclear compartment in the presenceof different dosages of the NPF. The total GR and protein kinase B (PKB) from whole-cell lysates were used as the control. The experiments were repeated three times. (B)Quantified results of Dex-induced GR DNA binding in the presence of the NPF at different dosages are shown. All quantitative data are presented as the mean ± SEM (n = 3).***p < 0.001 compared to full GR DNA binding to the GRE in the presence of Dex. (C) A representative picture and quantified results of Dex-induced PXR mRNA expressionin the presence of the NPF at different dosages are shown. All quantitative data are presented as the mean ± SEM (n = 3). **p < 0.01 and ***p < 0.001 compared to the fullexpression of PXR mRNA in the presence of Dex.

As shown in Fig. 2A, we first assessed GR nuclear localization byWestern blot of purified nuclear lysates. GR nuclear accumula-tion occurred within 1 h of treatment of the H4IIE cells with thesynthetic glucocorticoid dexamethasone (Dex). The NPF, at con-centrations of >20 �g/ml, dose-dependently decreased nuclear GRlocalization without affecting total GR expression in whole-celllysates (Fig. 2A). This was in stark contrast to insulin which did notaffect GR localization. Next we assessed the amount of active GRin nuclear lysates by quantifying GR after binding to an oligonu-cleotide based on a consensus glucocorticoid response element(GRE). As before, the proportion of GR in nuclear extracts avail-able to bind to a GRE was increased by Dex treatment of cells andthis interaction was prevented by adding competing GRE oligonu-cleotides (Fig. 2B). The promotion of GR DNA binding by Dex wasalso inhibited by the NPF at concentrations of 40 �g/ml (p < 0.001;n = 3).

Others have shown that Dex treatment of cells induces PXRmRNA within 12 h, and this induction is inhibited by RU486, aknown GR antagonist (Pascussi et al., 2000). In Fig. 2C, as expected,the Dex-induced PXR mRNA was dose-dependently suppressed bythe NPF at concentrations of >20 �g/ml (p < 0.01; n = 3).

3.3. CA-stimulated CREB serine phosphorylation and -promotedCREB DNA binding at the CRE were prevented by the NPF

The induction of PEPCK by glucagon is mediated through cAMP(CA) induction of the transcription factor CREB. CREB phosphory-

lation at serine 133 (activating phosphorylation) was observed inH4IIE cells 1 h after exposure to CA (Fig. 3A). In the presence of theNPF at concentrations of >20 �g/ml, the CA-induced phosphory-lation of CREB was inhibited. Consistent with this, CA-enhancedCREB DNA binding at the CRE was also inhibited by the NPF atconcentrations of >10 �g/ml (p < 0.05; n = 3).

3.4. The NPF contains high-molecular-weight tannin polymerspossessing PEPCK mRNA suppressive bioactivity.

Chemical characteristics of the NPF were further evaluated bycolorimetric assays. Firstly, addition of one drop of 1% ferric chlo-ride to the NPF initiated formation of a dark blue color. Secondly,addition of one drop of 1% lead tetra-acetate induced formationof a precipitate in the NPF solution. This precipitate could be dis-solved by addition of one drop of 4N acetic acid. Finally, a precipitatewas also generated by addition of three drops of bromine waterto the NPF. These data are consistent with the presence of bothhydrolysable and condensed tannins in the NPF. The amount ofcondensed tannins in NPF was estimated at 23.6% by the contentof (±)-catechin equivalent in the NPF using the vanillin–HCl test(Data not shown). The 13C NMR spectrum of the NPF (Fig. 4A)showed broad signals in the region between 100 and 160 ppm.Such down-shifted signals were attributed to the aromatic carbonsof the constitutive monomer. The broadening pattern also indi-cates the structural conformation and the relatively high viscosity

Author's personal copy

Y.-C. Juan et al. / Journal of Ethnopharmacology 137 (2011) 592– 600 597

Fig. 3. The non-PGG fraction (NPF) blocks 8-bromo-cAMP (CA)-induced cAMP-responsive element (CRE)-binding protein (CREB) 133 serine phosphorylation andDNA binding at the CRE. (A) A representative result of CA-induced CREB proteinphosphorylation at serine 133 at the nuclear compartment in the presence of differ-ent dosages of the NPF. Total CREB from the whole-cell lysate was used as the control.The experiments were repeated three times. (B) Quantified results of CA-inducedCREB DNA binding in the presence of the NPF at different dosages are shown. Allquantitative data are presented as the mean ± SEM (n = 3). *p < 0.05, **p < 0.01 and***p < 0.001 compared to full CREB DNA binding to the CRE in the presence of CA.

of the polymer chain. In addition, the 13C NMR spectrum containeda typical catechin signal without C-2 and C-3 peaks.

The importance of these potential mixed high-molecular-weight tannin polymers in the NPF was further evaluated. Firstly,the NPF was fractionated into an inner membrane fraction (NPF-DI) containing high-molecular-weight tannin polymers or outermembrane fraction (NPF-DO) using dialysis. Then H4IIE cells wereexposed to these purified extracts as depicted in Fig. 4B. Dex + CAinduced PEPCK mRNA expression was significantly suppressed byNPF, NPF-DI, or insulin, with the NPF and the NPF-DI exhibitingsimilar potency. In contrast, NPF-DI at this same concentration wasunable to suppress PEPCK mRNA leveling the presence of Dex andCA.

3.5. PGG regulated PEPCK mRNA levels in a way similar to insulin

As shown in Fig. 5A, exposure of H4IIE cells to insulin (10 nM)for 1 h stimulated serine phosphorylation of both PKB and GSK3while reducing the serine phosphorylation of glycogen synthase(GS). PGG also dose-dependently stimulated both PKB and GSK3serine phosphorylation in this same time scale. In addition, PGG atconcentrations of >5 �g/ml significantly suppressed Dex and CA-induced PEPCK mRNA expressions in H4IIE cells (Fig. 5B). Moreover,suppression of the PEPCK-GFP reporter by both insulin (10 nM) andPGG (20 �g/ml) was blocked by the PI 3-kinase inhibitor. LY294002(50 �M). and the insulin receptor antagonist. HNMPA-AM3 (50 �M)(Fig. 5B). Finally, induction of PEPCK protein expression by Dex andCA was also inhibited by both insulin and PGG, effects which wereprevented by the presence of HNMPA-AM3 (Fig. 5D).

4. Discussion and conclusions

The PEPCK promoter is regulated by glucocorticoids by acomplex Glucorticoid Response Unit (GRU), containing two corecis-acting GREs which bind GR, and several accessory factor bindingsites (He et al., 2009). Meanwhile the CA induction of PEPCK is medi-ated through a cAMP Response Unit (CRU) containing a core CREthat binds CREB, and additional accessory factors. Binding of CREBor C/EBP � at the CRE also enhances the induction of PEPCK genetranscription by Dex, showing that the CRE is an accessory elementfor the glucocorticoid response (Yamada et al., 1999). Moreover,CREB regulates the transcription of the general co-activator pro-tein peroxisome proliferator-activated receptor gamma coactivatoralpha (PGC1-�). PGC1-� works in concert with the transcriptionfactor, HNF4�, and the forkhead family activator, FoxO1 (Bartheland Schmoll, 2003) to induce the PEPCK gene.

In contrast, insulin negatively regulates PEPCK gene transcrip-tion following induction of the PI 3-kinase-PKB signaling pathway(Lipina et al., 2005; Logie et al., 2007), and probably the FOXOand CBP/p300 transcriptional regulators. Therefore, activation ofinsulin signaling and disruption of key transcriptional regulatorsare potential targets for the suppression of PEPCK gene transcrip-tion.

Suppression of PEPCK mRNA and protein expression by thePaeoniae Rubra Radix extract (PR-Et) has been successfully demon-strated in rats with insulin-deficient streptozotocin-induceddiabetes or insulin-resistant db/db mice (Juan et al., 2010). Thisimplied the existence of insulin mimetic actions of the PR-Et withproperties that could bypass the deficit in insulin signaling foundin these animal models.

In the current investigation, PGG was identified in the PR-Et viaqualitative HPLC. Consistent with previous work, we found thatPGG also mimics the action of insulin in H4IIE cells, where it acti-vates the PI3K-PKB-GSK3 signaling cascade and suppresses PEPCKgene transcription. However, it is quite possible that simple insulinmimetic agents (especially those working at the insulin receptor)may have limited efficacy in insulin resistant states such as type 2diabetes, where there is impaired insulin signal transduction.

In order to establish that the only active ingredient of PR-Et wasthis simple insulin mimetic agent, PGG, we generated a fractionof PR-Et lacking PGG (NPF). Interestingly the NPF retained PEPCK-suppressive activity and in contrast to PGG and insulin, the actionof the NPF on PEPCK was independent of activation of the insulinreceptor. We further showed that the NPF reduces nuclear localiza-tion and GRE DNA binding of the GR, and inhibited the expressionof PXR mRNA, another GR regulated gene (Kliewer et al., 2002).Similarly, the NPF inhibited CREB serine 133 phosphorylation andsubsequent CRE DNA binding. Therefore, the current investigationprovides evidence for disruption of two important PEPCK transacti-vators, the GR and CREB by the NPF, which is achieved independentof activation of the insulin receptor (Fig. 6).

Agents like NPF that antagonize glucocorticoid action may haveadditional benefits in diabetes therapy, over and above simplyrepressing PEPCK and reducing gluconeogenesis. This hypothesisis supported by the beneficial affects observed in diabetic animalstreated with RU486 (Taylor et al., 2009).

Similarly, CREB plays a key role in regulating fasting hyper-glycemia as demonstrated by knockdown in type 2 diabetic animals(Erion et al., 2009). Although insulin probably represses PEPCK tran-scription by interfering with several transcriptional activators, oneof the major targets is likely to be the interaction between CREBand p300 (a CREB-binding protein) (Sutherland et al., 2003; Heet al., 2009). Hence although the NPF appears to function inde-pendently of the proximal part of the insulin receptor signalingpathway it converges at the PEPCK promoter on the GRU and theCRU.

Author's personal copy

598 Y.-C. Juan et al. / Journal of Ethnopharmacology 137 (2011) 592– 600

Fig. 4. NPF facilitates its PEPCK transcriptional suppression via the constituents of high-molecular-weight tannin polymers. (A) A representative result of 1H and 13C NMRspectrums of NPF. (B) NPF was fractionated by molecular weight via dialysis to results in its high-molecular-weight constituents (tannin polymers), NPF-DI, and low molecularconstituents, NPF-DO. Comparison of PEPCK mRNA suppressive activities among NPF, NPF-DI, and NPF-DO, was then carried out. A representative picture and quantifiedresults are shown. Insulin (10 nM) served as a positive control. All quantitative data are presented as the mean ± SEM (n ≥ 3). *p < 0.05, **p < 0.01 and ***p < 0.001, significantdifferences compared to the Dex + CA only group.

Clearly the NPF is an unlikely therapeutic in its crude state, henceestablishing the active component in NPF is an important goal. Ourresults from various colorimetric assays, NMR analysis, dialysis-based fractionation, and cell based bio-assay, have suggested thata mixture of high-molecular-weight tannin polymers are possiblecandidates as the active ingredient of the NPF. To date, the hypo-glycemic activities of hydrolysable and condensed tannins havebeen poorly characterised (Thompson et al., 1984; Suryanarayanaet al., 2004; Serrano et al., 2009; Adisakwattana et al., 2010). There-fore, we propose that there is great potential for these agents to bedeveloped as anti-diabetes therapeutics in the future.

In conclusion, we have isolated a novel fraction from theChinese herbal medicine, Paeoniae Rubra Radix that has suppres-sive effects on PEPCK transcription most likely via disruption ofthe activity of the transcription factors GR and CREB. This dataextends our understanding of the potent anti-hyperglycemic activ-ities observed with the Paeoniae Rubra Radix crude extract (Juanet al., 2010). In particular, we propose that there are at leastfour anti-diabetic constituents in this mixture (the NPF (high-molecular-weight tannins), paeoniflorin, PGG, and paeonol), butthat the NPF has greatest potential to reduce blood glucose in aninsulin resistant state.

Author's personal copy

Y.-C. Juan et al. / Journal of Ethnopharmacology 137 (2011) 592– 600 599

Fig. 5. 1,2,3,4,6-Penta-O-galloyl-beta-d-glucose (PGG) mimics insulin’s actions on phosphoenolpyruvate carboxykinase (PEPCK) transcription at the level of the insulinreceptor. (A) A representative result of insulin and PGG-mediated activation of the protein kinase B (PKB)-GSK3-GS signaling cascades. Phosphorylated PKB (serine 473),GSK3� (serine 9), and GS (serine 641) in the presence of insulin (10 nM) and PGG at different dosages are illustrated. Total PKB served as the control. The experiments wererepeated three times. (B) A representative picture and quantified results of Dex + CA-induced PEPCK mRNA expression in the presence of PGG at different dosages are shown.Insulin (10 nM) served as the positive control. All quantitative data are presented as the mean ± SEM (n ≥ 4). *p < 0.05 and ***p < 0.001, significant differences compared to theDex + CA only group. (C) Full promoter activity (100%) measured from H4-pPCK-GFP cells is in relative fluorescence units (RFU)/mg total protein under serum-free conditions.All quantitative data are presented as the mean ± SEM (n ≥ 10). *p < 0.05 and ***p < 0.001, compared to the basal condition (none). �p < 0.05 and ���p < 0.001, compared tothe response in the absence of LY294002 or HNMPA-AM3. (D) A representative result of Dex + CA-induced PEPCK protein expression under insulin and PGG conditions in thepresence or absence of HNMPA-AM3. Actin was used as the control. The experiments were repeated three times.

GR1 CRE

radixPaeoniae

Insulin/PGG NPF

INSRGR

CREBPEPCK mRNA

GR2AF1 AF2TIRE

GRU

Ser133P

Nucleus translocation

Fig. 6. Proposed model of distinct actions from two phosphoenolpyruvate car-boxykinase (PEPCK) transcriptional suppressors in Paeoniae Rubra Radix. Two typesof PEPCK suppressors are contained in Paeoniae Rubra Radix. 1,2,3,4,6-Penta-O-galloyl-beta-d-glucose (PGG) serves as an insulin mimetic, and its activity is at thelevel of the insulin receptor. Thymidine-rich insulin response elements (TIREs) arethe major DNA element through which insulin carries out its inhibitory action. Incontrast, the non-PGG fraction (NPF) directly intervenes with the active transcrip-tional factors, glucocorticoid receptor (GR) and cAMP-responsive element-bindingprotein (CREB), on the promoter of PEPCK in an insulin receptor-independent man-ner.

Conflict of interest

Authors declared that there is no conflict of interest in this study.

Acknowledgements

This work was supported by grants from the National ScienceCouncil (NSC97-2320-B-077-007-MY3) and the National ResearchInstitute of Chinese Medicine (NRICM-98-DHM6). The PhD stu-dentships for Ms. YC Juan were funded by both the NRICM andNSC. We would like to thank Dr. JJ Cheng, and Dr. MK Lu for fruitfuldiscussions and Prof. C Sutherland for the manuscript editing.

References

Adisakwattana, S., Jiphimai, P., Prutanopajai, P., Chanathong, B., Sapwarobol, S.,Ariyapitipan, T., 2010. Evaluation of alpha-glucosidase, alpha-amylase and pro-tein glycation inhibitory activities of edible plants. International Journal of FoodSciences and Nutrition 61, 295–305.

Barthel, A., Schmoll, D., 2003. Novel concepts in insulin regulation of hepatic glu-coneogenesis. American Journal of Physiology-Endocrinology and Metabolism285, E685–692.

Cameron, J.D., Cruickshank, J.K., 2007. Glucose, insulin, diabetes and mechanismsof arterial dysfunction. Clinical and Experimental Pharmacology and Physiology34, 677–682.

Author's personal copy

600 Y.-C. Juan et al. / Journal of Ethnopharmacology 137 (2011) 592– 600

Chen, C.C., Hsu, C.Y., Chen, C.Y., Liu, H.K., 2008. Fructus Corni suppresses hepaticgluconeogenesis related gene transcription, enhances glucose responsivenessof pancreatic beta-cells, and prevents toxin induced beta-cell death. Journal ofEthnopharmacology 117, 483–490.

Erion, D.M, Ignatova, I.D., Yonemitsu, S., Nagai, Y., Chatterjee, P., Weismann, D., Hsiao,J.J., Zhang, D., Iwasaki, T., Stark, R., Flannery, C., Kahn, M., Carmean, C.M., Yu, X.X.,Murray, S.F., Bhanot, S., Monia, B.P., Cline, G.W., Samuel, V.T., Shulman, G.I., 2009.Prevention of hepatic steatosis and hepatic insulin resistance by knockdown ofcAMP response element-binding protein. Cell Metabolism 10, 499–506.

Hall, R.K., Wang, X.L., George, L., Koch, S.R., Granner, D.K., 2007. Insulin repressesphosphoenolpyruvate carboxykinase gene transcription by causing the rapiddisruption of an active transcription complex: a potential epigenetic effect.Molecular Endocrinology 21, 550–563.

He, L., Sabet, A., Djedjos, S., Miller, R., Sun, X., Hussain, M.A., Radovick, S., Wondisford,F.E., 2009. Metformin and insulin suppress hepatic gluconeogenesis throughphosphorylation of CREB binding protein. Cell 137, 635–646.

Juan, Y.C., Tsai, W.J., Lin, Y.L., Wang, G.J., Cheng, J.J., Yang, H.Y., Hsu, C.Y., Liu,H.K., 2010. The novel anti-hyperglycemic effect of Paeoniae radix via thetranscriptional suppression of phosphoenopyruvate carboxykinase (PEPCK).Phytomedicine 17, 626–634.

Kliewer, S.A., Goodwin, B., Willson, T.M., 2002. The nuclear pregnane X receptor: akey regulator of xenobiotic metabolism. Endocrine Review 23, 687–702.

Li, Y., Kim, J., Li, J., Liu, F., Liu, X., Himmeldirk, K., Ren, Y., Wagner, T.E.,Chen, X., 2005. Natural anti-diabetic compound 1,2,3,4,6-penta-O-galloyl-d-glucopyranose binds to insulin receptor and activates insulin-mediated glucosetransport signaling pathway. Biochemical and Biophysical Research Communi-cations 336, 430–437.

Lipina, C., Huang, X., Finlay, D., McManus, E.J., Alessi, D.R., Sutherland, C., 2005. Anal-ysis of hepatic gene transcription in mice expressing insulin-insensitive GSK3.Biochemical Journal 392, 633–639.

Lochhead, P.A., Salt, I.P., Walker, K.S., Hardie, D.G., Sutherland, C., 2000. 5-aminoimidazole-4-carboxamide riboside mimics the effects of insulin on theexpression of the 2 key gluconeogenic genes PEPCK and glucose-6-phosphatase.Diabetes 49, 896–903.

Logie, L., Ruiz-Alcaraz, A.J., Keane, M., Woods, Y.L., Bain, J., Marquez, R., Alessi, D.R.,Sutherland, C., 2007. Characterization of a protein kinase B inhibitor in vitro andin insulin-treated liver cells. Diabetes 56, 2218–2227.

Magnuson, M.A., Quinn, P.G., Granner, D.K., 1987. Multihormonal regulation ofphosphoenolpyruvate carboxykinase-chloramphenicol acetyltransferase fusiongenes. Insulin’s effects oppose those of cAMP and dexamethasone. The Journalof Biological Chemistry 262, 14917–14920.

Makkar, H.P.S., Becker, K., 1993. Vanillin–HCl method for condensed tannins: effectof organic solvents used for extraction of tannins. Journal of Chemical Ecology19, 613–621.

Nichols, M., Weih, F., Schmid, W., DeVack, C., Kowenz-Leutz, E., Luckow, B., Boshart,M., Schutz, G., 1992. Phosphorylation of CREB affects its binding to high and low

affinity sites: implications for cAMP induced gene transcription. EMBO Journal11, 3337–3346.

Pascussi, J.M., Drocourt, L., Fabre, J.M., Maurel, P., Vilarem, M.J., 2000. Dexametha-sone induces pregnane X receptor and retinoid X receptor-alpha expression inhuman hepatocytes: synergistic increase of CYP3A4 induction by pregnane Xreceptor activators. Molecular Pharmacology 58, 361–372.

Patel, S., Lipina, C., Sutherland, C., 2003. Different mechanisms are used by insulin torepress three genes that contain a homologous thymine-rich insulin responseelement. FEBS Letters 549, 72–76.

Perrin, R.M., Harper, S.J., Bates, D.O., 2007. A role for the endothelial glycocalyx inregulating microvascular permeability in diabetes mellitus. Cell Biochemistryand Biophysics 49, 65–72.

Serrano, J., Puupponen-Pimia, R., Dauer, A., Aura, A.M., Saura-Calixto, F., 2009.Tannins: current knowledge of food sources, intake, bioavailability and bio-logical effects. Molecular Nutrition & Food Research 53 (Suppl. 2), S310–329.

Suryanarayana, P., Kumar, P.A., Saraswat, M., Petrash, J.M., Reddy, G.B., 2004. Inhibi-tion of aldose reductase by tannoid principles of Emblica officinalis: implicationsfor the prevention of sugar cataract. Molecular Vision 10, 148–154.

Sutherland, C., O’Brien, R.M., Granner, D.K., 1996. New connections in the regulationof PEPCK gene expression by insulin. Philosophical Transactions of the RoyalSociety of London 351, 191–199.

Sutherland, C., O’Brien, R.M., Granner, D.K., 2003. Genetic regulation of glucosemetabolism. In: Jefferson, L.S., Cherrington, A.D. (Eds.), Handbook of Physiology.Oxford University Press, UK, Oxford, pp. 707–734.

Taylor, A.I., Frizzell, N., McKillop, A.M., Flatt, P.R., Gault, V.A., 2009. Effect of RU486on hepatic and adipocyte gene expression improves diabetes control in obesity-type 2 diabetes. Hormone and Metabolic Research 41, 899–904.

Thompson, L.U., Yoon, J.H., Jenkins, D.J., Wolever, T.M., Jenkins, A.L., 1984. Rela-tionship between polyphenol intake and blood glucose response of normaland diabetic individuals. The American Journal of Clinical Nutrition 39,745–751.

Vallejo, M., Gosse, M.E., Beckman, W., Habener, J.F., 1995. Impaired cyclicAMP-dependent phosphorylation renders CREB a repressor of C/EBP-inducedtranscription of the somatostatin gene in an insulinoma cell line. Molecular andCellular Biology 15, 415–424.

Xing, L., Quinn, P.G., 1993. Involvement of 3’,5’-cyclic adenosine monophosphateregulatory element binding protein (CREB) in both basal and hormone-mediatedexpression of the phosphoenolpyruvate carboxykinase (PEPCK) gene. MolecularEndocrinology 7, 1484–1494.

Yabaluri, N., Bashyam, M.D., 2010. Hormonal regulation of gluconeogenic gene tran-scription in the liver. Journal of Biosciences 35, 473–484.

Yamada, K., Duong, D.T., Scott, D.K., Wang, J.C., Granner, D.K., 1999. CCAAT/enhancer-binding protein beta is an accessory factor for the glucocorticoid response fromthe cAMP response element in the rat phosphoenolpyruvate carboxykinase genepromoter. The Journal of Biological Chemistry 274, 5880–5887.