`
`The Natural Products Journal, 2013, 3, 15-25
`
`15
`
`Send Orders of Reprints at reprints@benthamscience.net
`
`Deciphering the Inactivation of Human Pancreatic (cid:1)-Amylase, an Anti-
`diabetic Target, by Bisdemethoxycurcumin, a Small Molecule Inhibitor,
`Isolated from Curcuma longa
`
`Sudha Ponnusamya, Smita Zinjardea, Shobha Bhargavab, Urmila Kulkarni-Kalec, Sangeeta Sawantc,
`and Ameeta Ravikumara,*
`
`a
`
`Institute of Bioinformatics and Biotechnology, University of Pune 411007, India;
`c
`Department of Zoology, University of Pune 411007, India;
`
`Bioinformatics Centre University of Pune 411 007, India
`
`b
`
`Molecular Embryology Laboratory,
`
`Abstract: Natural products from plants are an excellent source of Human pancreatic (cid:1)-amylase (HPA) inhibitors which
`have therapeutic application as oral agents to control blood glucose levels. The mechanism of action by
`Bisdemethoxycurcumin (BDMC) has been reported, isolated from Curcuma longa rhizomes, which inactivates HPA, a
`target for type-2 diabetes. This study validates its mode of action and its target which has to date remained largely
`unknown. The cytotoxicity and bioactivity of crude extract and BDMC on the pancreatic acinar AR42J secretory model
`cell line were evaluated with LD50 value of 16.25 μg ml-1 and 63.53 μM, and secretory (cid:1)-amylase inhibition of 41% and
`30%, respectively. BDMC uncompetitively inhibits HPA (Ki(cid:3) of 10.1μM) and a binding affinity (Ka) of 8.5 x 104 M-1 with
`the involvement of surface exposed aromatic residues. The thermodynamic parameters suggest that binding is both
`enthalpically and entropically driven with (cid:2)Gº of - 28.13 kJ mol-1. Computational ligand docking showed that inactivation
`depends on hydrogen bonding and (cid:2)-(cid:2) interactions. Thus, BDMC, a natural product could be lowering post-prandial
`glycemia via a novel mode of binding and inactivation of HPA and may proved to be a good drug candidate to
`reduce/control post-prandial hyperglycemia.
`
`Keywords: Human pancreatic (cid:1)-amylase, bisdemethoxycurcumin, ligand binding, AR42J cell line, docking.
`
`INTRODUCTION
`
`Diabetes mellitus is a carbohydrate metabolic disorder
`affecting 346 million people worldwide, with the most
`prevalent form being Non-insulin dependent Diabetes mellitus
`(NIDDM) associated with post prandial hyperglycemia [1].
`Current oral hypoglycemic agents i.e., acarbose, voglibose,
`miglitol have side effects and the demand for safer
`biomedicine from Natural products (NP) with lesser side
`effects is on the rise. Pancreatic (cid:1)-amylase or (cid:1)-1,4-glucan-
`4-glucanohydrolase (E.C. 3.2.1.1), is a key enzyme in the
`digestive system and catalyses the initial step in hydrolysis
`of starch to maltose which is eventually degraded to glucose.
`Hence retardation of starch digestion by the inhibition of
`enzymes such as (cid:1)-amylase plays a key role in the control of
`diabetes [2, 3]. NPs from plants are known to have a wide
`variety of pancreatic (cid:1)-amylase inhibitors which are of
`therapeutic importance as oral hypoglycemic agents in
`diabetes mellitus with lesser side effects [4]. The chemical
`diversity of these components present in NPs has tremendous
`potential for the discovery of molecules with pharmacological
`activity. The challenge lies in identifying the biological
`targets for chemical inhibitors and validating their mode of
`
`
`
`
`*Address correspondence to this author at the Institute of Bioinformatics
`and Biotechnology, University of Pune, Pune- 411007, Maharashtra, India;
`Tel: +91202569133; Fax: 912025690087; E-mail: ameeta@unipune.ac.in
`
`action [5]. The NP extracts are likely to possess numerous
`primary and secondary metabolites, many of which have not
`been previously characterized and offer a good alternative
`for obtaining lead compounds which can lower post-prandial
`hyperglycemia with lesser side effects. Currently, very few
`NPs exhibiting anti-diabetic properties are in different stages
`of clinical trials. Acarbose, voglibose and miglitol derived
`from microbial sources are currently in use. Dapagliflozin,
`an analog of phlorizin, a polyphenolic glucoside from root
`bark of apple tree, is undergoing phase II while trodusquemine,
`a sulfated amino sterol from dog-fish shark is in phase I
`clinical trials for type 2 diabetes [6]. Recently, potent inhibitors
`for human, rat, hog pancreas and intestinal (cid:1)-amylase and
`(cid:1)-glucosidase have been reported from the crude plant
`extracts but however very few of them have been isolated,
`characterized and validated with respect to their structure or
`their target [4, 7, 8].
`
`In our previous study of the 28 plants screened for
`pancreatic (cid:1)-amylase inhibitors, the IPE extract of C. longa
`rhizomes exhibited a strong inhibitory effect on HPA [7, 8].
`C. longa, rhizomes are used in the traditional Indian
`Ayurvedic medicine for its anti-diabetic, anti-oxidative and
`anti-inflammatory properties. The efficacy of turmeric to
`reduce blood glucose levels in experimentally induced
`diabetic rats has also been reported [9, 10]. Preliminary
`phytochemical and GC-MS analysis of this crude C. longa
`IPE suggested the presence of curcuminoids, sesquiterpenes
`
`
`
` 2210-3163/13 $58.00+.00
`
`© 2013 Bentham Science Publishers
`
`
`NOVARTIS EXHIBIT 2107
`Par v. Novartis, IPR 2016-01479
`Page 1 of 11
`
`

`

`16 The Natural Products Journal, 2013, Vol. 3, No. 1
`
`Ponnusamy et al.
`
`and other phenolics [7]. However, current understanding
`regarding the bioactive principle responsible for hypo-
`glycemia as well as the mechanism of action is rudimentary
`and is limited to the anti-oxidant, anti-cancer and anti-
`inflammatory effects of curcuminoids. The safety of C.
`longa has been studied in various animal models and it is
`clear that turmeric is non-toxic even at high doses in
`laboratory animals [11].
`
` In this study, the isolated curcuminoid responsible for
`the inhibitory action along with the crude extract were
`evaluated for its bioactivity and cytotoxicity on AR42J (rat
`pancreatic acinar) cell line, which produces pancreatic (cid:3)-
`amylase on induction with glucocorticoids [12]. Enzyme
`inhibitor kinetics, ligand binding thermodyanamics, in silico
`docking studies were carried out to reveal the mode of lead
`inhibitor action on HPA.
`
`MATERIALS AND METHODS
`
`HPA and maltopentaose were purchased from Sigma
`Aldrich, USA. AR42J rat pancreatic cell line was from
`ATCC-CRL142. Ham’s F-12K with L-glutamine culture
`media, fetal bovine serum, antibiotics streptomycin, penicillin,
`3, 5-Dinitrosalicylic acid (DNSA) and (3-(4,5-Dimethylthiazol-
`2-yl)-2,5-diphenyltetrazolium bromide (MTT) were procured
`from Himedia Laboratories, Mumbai, India. All other
`chemicals from local manufacturer were of Analytical grade.
`
`Bioactivity Guided Isolation
`
`In order to isolate the principle bioactive component from
`the IPE of C. longa rhizome, a series of bioassay-guided
`purification steps was performed as per the protocol
`mentioned by Revathy et al., 2011 [13] using silica
`chromatography and HPLC. The
`identification of
`the
`compound exhibiting bioactivity in dexamethasone induces
`AR42J celline was confirmed by the complete 1H and 13C
`Nuclear Magnetic resonance (NMR) spectra in dimethyl
`sulfoxide (DMSO)-d6 and was identified as BDMC based on
`the assignments of the chemical shifts (ppm): 1H NMR
`(DMSO): (cid:4) H-1 7.64, H-2 6.8, H-4 6.13, H-6 6.8, H-7 7.64,
`H-9 7.66, H-10 6.91, H-12 6.91, H-13 7.66 and 13C NMR
`(DMSO): (cid:4) C-1 140.8, C-2 121.2, C-3 183.6, C-4 101.4, C-5
`183.6, C-6 121.2, C-7 140.8 C-8,8’ 126.2, C-9,9’ 130.8, C-
`10,10’ 116.3, C-11 11’ 160.2, C-12,12’ 116.3, C-13,13’
`130.8. This was found to be in agreement with the spectral
`analysis and chemical shift data reported in literature [14].
`
`Culturing of AR42-J Cell Line
`
`AR42J cells were cultured in Ham’s F-12K with
`L-glutamine culture media supplemented with nonessential
`amino acids, sodium pyruvate, 7.5 % sodium bicarbonate, 20%
`fetal bovine serum and antibiotics (100 μg/ml streptomycin,
`100 units/ml penicillin) at 37oC under a humidified condition
`of 95 % air and 5 % CO2.
`
`Cytotoxicity and Bioactivity in AR42J Cell Line
`
`Cells were routinely plated at 1(cid:2)105 cells/ml/well onto 12
`wells culture dish. After overnight attachment of the cells,
`the culture medium was replaced by fresh medium with
`dexamethasone at a concentration of 10nM, [12] crude
`
`extract and individual curcuminoids predisolved in serum at
`a concentration of 5-30 μgml-1 and 10-100 μMml-1 for 24 h.
`The amount of (cid:3)-amylase released by cells was determined
`in an aliquot of medium by DNSA method with starch as the
`substrate with appropriate controls [7]. At the end of the 24-h
`incubation period and aspiration of culture media needed
`from all samples for determination of amylase activity, 20 μl
`of MTT tetrazolium bromide 5 mg ml-1 was added to each
`sample well containing 500 μl of fresh culture media and
`AR42J. Cells were incubated for 3 h in the presence of MTT
`tetrazolium bromide at 37°C. Metabolically active cells
`reduce MTT to insoluble purple formazan dye crystals,
`which were dissolved by addition of 500 μl DMSO solution
`at the end of the incubation period. Obtaining absorption
`readings at 550 nm using the microplate spectrophotometer
`(Molecular device, M5, Germany) performed quantitation of
`solubilized formazan [15].
`
`To determine the efficiency of the inhibitors in the
`presence of starch, the substrate for (cid:3)-amylase, at the
`concentration of both crude and BDMC exhibiting maximum
`secretary pancreatic (cid:3)-amylase
`inhibition, starch
`load
`varying from 0.25-1% (w/v) were subjected to dexamethasone
`induced AR42J cell line.
`
`Mechanism of Inhibition
`
`To reveal the mechanism of BDMC inhibition, kinetics,
`ligand binding, thermodynamics of binding in silico docking
`studies with HPA were performed.
`
`Kinetics of Enzyme Inhibition
`
`The mode of inhibition of HPA by BDMC was
`determined by using Michaelis-Menton and Lineweaver-
`Burk (LB) equations. Maltopentaose (0.05-0.4 mM) was
`incubated with BDMC-HPA for 2.5 min and the residual
`enzyme activity was determined by Somogyi’s method [16].
`Secondary Bowden plots of [S]/V vs [I] were used to
`determine the dissociation constants (Ki(cid:2)) of lead inhibitor
`[17].
`
`Rate Constant of Reaction
`
`Rate constants of the reaction were determined by the
`incubation of HPA (0.2-0.5 U) with differing concentrations
`of BDMC (0.022-0.032 mM) for varying time points (2.5-15
`min) and assayed with starch by DNSA method [7].
`
`One unit of enzyme activity is defined as the amount of
`enzyme required to release one micromole of maltose from
`starch/maltopentaose per min under the assay conditions.
`
`Ligand Binding Studies
`
`Fluorescence Measurements
`
`Fluorescence measurements of HPA were carried out in
`0.02 M sodium phosphate buffer, pH 6.9 (containing 6 mM
`NaCl). Fluorescence measurements were performed on a
`Spectra Max M5 using a (cid:5)ex at 280 nm [18-20]. The
`quenching was performed by titrating 350 (cid:1)l, 0.29 mM HPA
`with BDMC (0.001-0.053 mM) (3-45 (cid:1)l aliquots) followed
`by monitoring the change in fluorescence at 350 nm.
`
`NOVARTIS EXHIBIT 2107
`Par v. Novartis, IPR 2016-01479
`Page 2 of 11
`
`

`

`HPA Inhibition by BDMC
`
`The Natural Products Journal, 2013, Vol. 3, No. 1 17
`
`The data was analyzed by linear fits according to Hegde
`et al. [19] as well as by non-linear regression analysis [20]
`using the equations
`
`(F0-F) = (F0-F(cid:1)) · [I0] / Kd + [I0]
`
`
`
`
`
` (1)
`
`where F and F0 are the measured fluorescence emission
`intensity of the protein solution in presence and absence of
`the ligand, F(cid:1) when protein is saturated with ligand and I0
`the total concentration of the bound and unbound ligand. The
`decrease in fluorescence ((cid:1)F) in the presence of different
`concentrations of the ligand is measured, and the dissociation
`constant (Kd) calculated.
`The calculation of the Gibbs free energy ((cid:4)G0) was
`achieved by application of the following equation
`(cid:4)G0 = -RT lnKa
`
`
`
`
`
`
`
`
`
` (2)
`
`where R is the gas constant, T the temperature in Kelvin and
`Ka the association constant.
`
`The enthalpy change ((cid:2)H) and entropy change ((cid:2)S) were
`calculated by van’t Hoff analysis from the temperature
`30 by using the equation
`dependence of Ka
`
`lnKa = - (cid:4)H/RT + (cid:4)S/R
`
`
`
`
`
`
`
` (3)
`
`(106 x126 x 126 points), spacing of 0.5 A° and grid box
`center set at x = 8.499, y = 61.167, and z = 13.444. Based on
`the outputs of blind docking, refined docking simulation was
`done with grid parameters that scored high in blind docking.
`Grid box of 80 x 70 x 70 points was used with a spacing of
`0.5 A° and the grid box center set at x = 6.448, y = 46.002,
`and z = 35.697. The interactions of BDMC-HPA were
`analyzed and visualized using Insight II software (ver 2000).
`
`Statistical Analysis
`
`Analyses were repeated at least three times and evaluated
`by their means and standard deviations. The best-fit values
`were achieved by applying either linear fit or non-linear least
`square regression using the software, Microcal Origin 6.0
`(Micarocal Software Inc., Northampton, USA). Analyses of
`variance were carried out followed by F-test using the SPSS
`statistical package (SPSS 11.5).
`
`The entropy change was then obtained from the equation
`(cid:4)G0 = (cid:4)H-T(cid:4)S
`
` (4)
`
`
`
`
`
`
`
`Circular Dichroism Spectral Analysis
`
`Circular dichroism (CD) spectra, in the near-uv (250–320
`nm) and far–uv (195 to 250 nm) regions were recorded with a
`Jasco J-815 spectrometer at 100 nm min-1, with a 1-s response
`time and a 1-nm bandwidth. The quartz cuvette (0.1-cm)
`contained 50-100 (cid:2)g of HPA in 0.02M sodium phosphate
`buffer (pH 6.9). For ligand binding analysis, BDMC at its Ki(cid:2)
`was pre-incubated along with HPA for 15 min and CD spectral
`scans were taken. Appropriate controls of buffer blanks,
`BDMC in buffer were also performed and the scans corrected.
`
`Docking Simulations
`
`For docking simulations, using AutoDock Vina [21],
`surface pocket identification of HPA (PDB ID: 1HNY) [22]
`was carried out on servers CASTp, Pocket-Finder and Q-
`SiteFinder [23, 24]. Docking protocol and parameters were
`standardized by performing docking simulation of acarbose
`pseudopentasaccharide, using Gaussian-09 [25], with HPA,
`for which a cocrystal structure is available (PDB ID: 1B2Y)
`[26]. The uncomplexed structure of HPA (1HNY) was first
`processed to set protonation states of amino acids with polar
`side chains to neutral pH. Grid box of 80 x 70 x70 points
`was used with three grid spacings: 0.375, 0.5, and 1.0 A°.
`The grid box center was set at x = 6.448, y = 46.002, and z =
`35.697 A° and gasteiger charges assigned to protein and
`BDMC. Exhaustiveness level was set on 8 and a computer
`with four processors was utilized for the computations. A
`total of 270 docked poses of acarbose pseudopentasaccharide
`were generated and compared with crystal structure of the
`complex (1B2Y).
`
`Docking simulations of BDMC with HPA were carried
`out using the standardized docking parameters obtained.
`Blind docking with BDMC was performed with the grid box
`
`
`
`Fig. (1). Cytotoxicity and panceratic (cid:1)-amylase inhibition in
`AR42J cell line. A) Effect of crude extract on cell viability and
`pancreatic (cid:3)-amylase activity of dexamethasone induced AR42J
`cell line. B) Effect of BDMC on cell viability and pancreatic (cid:3)-
`amylase activity of dexamethasone induced AR42J cell line. Error
`bars represent ±SE of the mean of triplicates and p values < 0.05
`were considered significant.
`
`NOVARTIS EXHIBIT 2107
`Par v. Novartis, IPR 2016-01479
`Page 3 of 11
`
`

`

`18 The Natural Products Journal, 2013, Vol. 3, No. 1
`
`Ponnusamy et al.
`
`RESULTS
`
`Cytotoxicity and Bioactivity
`
`The effect of serum soluble crude and BDMC, on 24 h
`dexamethaxone induced AR42J cells was determined by
`MTT assay and the secretory pancreatic (cid:2)-amylase inhibition
`is shown in Figs. 1 (A & B). The resulting growth curve
`shows that both crude and the purified inhibitor, BDMC
`have a sigmoidal dose-dependent effect on the rate of cell
`proliferation. By the MTT technique, crude extract and
`BDMC exhibit an LD50 value of 16.25 μg ml-1 and 63.53 μM,
`respectively. There was no morphological change observed
`in crude and BDMC treated dexamethasone induced AR42J
`cell line as compared to the control (Fig. 2). The maximal
`secretory pancreatic (cid:2)-amylase inhibition of 41 and 30 % is
`observed at 15 μgml-1 and 30 μM for crude extract and
`BDMC, respectively. At these concentrations of both crude
`and BDMC, varying starch loads exhibited a maximum
`inhibition of 55 % and 37 % at 0.5 % (w/v) starch
`respectively.
`
`Inhibition Kinetics of HPA by BDMC
`
`The effect of BDMC on the kinetics of HPA catalyzed
`hydrolysis of maltopentaose was studied at differing
`inhibitor concentrations. The double reciprocal LB plots
`
`revealed that the mode of BDMC inhibition is uncompetitive
`for maltopentaose (Fig. 3A) with decrease in both the
`apparent Km and Vmax. Since the mode of inhibition obtained
`was uncompetitive, the Bowden plot was subsequently
`drawn and a Ki(cid:4) of 0.01mM was obtained for BDMC using
`maltopentaose as substrate (Fig. 3B).
`
`Incubation of HPA with BDMC resulted in a time-
`dependent inactivation (Fig. 3C) with inactivation following
`pseudo-first order kinetic behavior using starch as substrate.
`A reciprocal plot of the pseudo-first order rate constants
`(kobs) versus inhibitor concentration (Fig. 3D) yielded a value
`of k i/Ki (cid:3) of 1.72 min-1 mM -1 for HPA.
`
`BINDING OF BDMC
`
`Fluorescence Studies
`
`Fig. (4) shows the dependence of decrease in fluorescence
`intensity at 350 nm on the binding of BDMC to HPA while
`(cid:3)F is the decrease in fluorescence intensity relative to the
`fluorescence intensity of the free enzyme on ligand binding.
`HPA (16.6 (cid:1)g) bound to maltose exhibited a fluorescence
`intensity of 210 a.u. with a (cid:4)max emission of 350 nm.
`Titration of this enzyme-maltose complex with BDMC
`(0.001-0.053 mM) resulted in quenching of the intrinsic
`fluorescence to 68 a.u. with maximal quenching of 68 % and
`
`Fig. (2). Effect on Morphology of AR42J cell line. Inverted phase contrast image under magnification of 200 X of AR42J cell lines treated
`under following conditions. A) Control B) 10nM Dexamethasone induced C) Crude extract at 15(cid:1)gml-1 D) BDMC at 30 (cid:1)Mml-1 The scale
`bar corresponds to 20 (cid:1)M.
`
`
`
`NOVARTIS EXHIBIT 2107
`Par v. Novartis, IPR 2016-01479
`Page 4 of 11
`
`

`

`HPA Inhibition by BDMC
`
`The Natural Products Journal, 2013, Vol. 3, No. 1 19
`
`
`
`Fig. (3). Kinetic Analysis of HPA Inhibition by BDMC. A) LB plot for HPA inhibition at varying BDMC concentrations (cid:6), 0 mM; (cid:5),
`0.016 mM; (cid:2), 0.032 mM;
`, 0.064 mM; (cid:3); 0.097 mM with maltopentaose as substrate. B) Bowden plot with varying concentrations of
`maltopentaose (cid:6), 0.05 mM; (cid:5), 0.1 mM; (cid:2), 0.2 mM;
`, 0.4 mM. C) Semi-logarithmic plot of activity versus time (cid:6), 0.022 mM; (cid:5), 0.025 mM;
`(cid:2) 0.029 mM;
`, 0.032 mM). D) Replot of first-order rate constants from Fig. (3C). Error bars represent ±SE of mean of triplicates and
`p<0.05 were considered significant.
`
`no shift being observed in (cid:7)max emission (Fig. 4A). The
`unbound HPA (16.6 (cid:1)g) with an activity of 0.25 U ml-1
`exhibited a fluorescence intensity of 225 a.u. with a (cid:7)max
`emission also at 350 nm. Titration resulted in quenching of
`the intrinsic fluorescence to 67 a.u. with maximal quenching
`of 70 % and no change in (cid:7)max emission (Fig. 4B). No
`emission was noted in the same region for BDMC alone
`under any of
`the abovementioned conditions. The
`fluorescence spectra and quenching pattern of BDMC
`binding to both unbound HPA and maltose bound HPA were
`similar. The calculation of binding parameters such as the
`dissociation and association constants (Kd and Ka) was
`performed by the linear as well as non-linear regression
`analysis [18, 19]. Fig. (4A), B show the non-linear fits for
`titration at 25ºC of BDMC with HPA-maltose complex,
`unbound HPA and with R2 values of 0.997, 0.997
`respectively. These high R2 values suggest the one-site
`binding model. The Kd values ascertained from the fits
`exhibited values of 20.05 x 106 and 11.77 x 106, respectively
`for the maltose complexed, and unbound HPA at 25ºC.
`
`Circular Dichroism Studies
`
`The effect of binding on secondary and tertiary structure
`of HPA was performed using CD spectroscopy and the molar
`
`ellipticity plotted. The far uv-CD spectra (195-250 nm) for
`HPA in this study were in accordance with the reported
`spectra for mammalian amylases [27, 28]. However, no
`significant change in the secondary structural elements was
`observed on binding of BDMC to HPA (Fig. 5A). The CD
`spectrum of a protein in the near-UV spectral region (250–
`350 nm) is sensitive to perturbations of tertiary structure. At
`these wavelengths, the chromophores involved are the
`aromatic amino acids which can reflect changes in tertiary
`structure on binding. Fig. (5B) shows the effect of addition
`of BDMC on the CD spectra of HPA in the region 250–320
`nm. The major changes in spectra of HPA can be attributed
`to the change in environment of phenylalanine at 260 nm and
`tryptophan at 290-310 nm, on binding of BDMC. No
`significant structure was noted either in the far-or near-uv
`region for BDMC alone under the same conditions. In an
`earlier study, the binding of Cys3glc to Porcine pancreatic (cid:4)-
`amylase (PPA) resulted in an alteration in the environment of
`phenylalanine residues [18].
`
`Thermodynamic Parameters
`
`Since Ka varies with temperature, the binding parameters
`of BDMC to HPA were carried out at different temperatures
`from 22ºC to 30ºC by fluorescence quenching and analyzed
`
`NOVARTIS EXHIBIT 2107
`Par v. Novartis, IPR 2016-01479
`Page 5 of 11
`
`

`

`20 The Natural Products Journal, 2013, Vol. 3, No. 1
`
`Ponnusamy et al.
`
`
`
`Fig. (4). Fluorescence measurements of BDMC binding to HPA. Quenching of intrinsic fluorescence. A) BDMC and maltose bound HPA.
`B) BDMC and native holo-HPA. Insets: Determination of dissociation constant (Kd) using one site binding analysis. C) Linear van't Hoff plot.
`
`by non-linear regression fits as mentioned above. The
`maximal quenching observed were 71, 68, 62 and 60 % at
`22, 25, 28, and 30ºC, respectively. The association constants
`(Ka) and Gibbs free energy ((cid:1)Gº) of binding were calculated
`for all the four temperatures and are given in Table 1. The Ka
`indicates that the affinity of the ligand for the enzyme
`decreases on increase in temperature with maximal affinity
`at 295 K (22ºC). Though BDMC-HPA complex was found
`to be more stable at 22ºC with a ~2.0 fold lower Kd of
`8.01x10-6 M-1 as compared to 16.6 x10-6 M-1 at 303 K (30ºC),
`most of the studies were carried out at 298 K (25ºC) in
`comparison with other HPA inhibitors reported in literature.
`The free energy ((cid:1)Gº) values calculated from the binding
`constants are negative and range from (cid:2)27.7 to (cid:2)28.8 kJ
`mol(cid:2)1 indicating favorable interactions occurring between
`HPA and BDMC (Table 1). The HPA-maltose complex
`exhibited a Ka of 5.0 x 104 M-1, at 298 K with a (cid:1)Gº of 26.8
`kJ mol(cid:2)1. The linear van’t Hoff plot of Ka as a function of
`temperature
`(Fig. 4C)
`allowed
`calculation of
`the
`2
` of 0.953. The enthalpy
`thermodynamic parameters with R
`((cid:1)HvH) and entropy (-(cid:1)S).were determined to be -66 kJ mol-
`1and 126.6 J mol-1 K-1, respectively, suggesting that binding
`may be both enthalpically and entropically driven. An earlier
`study indicated that the binding of PAMI to PPA is favored
`by enthalpic ((cid:1)H = -10.8 kcal mol-1) and disfavored by
`entropic (-T(cid:1)S= 3.7 kcal/mol-1) contributions [29].
`
`In Silico Docking Studies
`
`In the absence of crystal structure on HPA complexed
`with flavonoids or polyphenols, docking simulations were
`carried out to get an insight into the possible mode of
`binding of BDMC with HPA. The consensus surface pocket
`identified by analyzing the results of three pocket identification
`servers was used for docking analysis. In the docking
`simulations of acarbose pseudopentasaccharide modeled by
`Gaussian 09 [25] with HPA, of the three different grid
`spacings that were evaluated, grid spacing of 0.5Aº showed
`that in 70 % of the total 270 poses generated, acarbose
`pseudopentasaccharide docked at the active site pocket as
`seen in the crystal structure (1B2Y) [26]. Hence, the grid
`spacing of 0.5 Aº was used in the docking simulations of
`BDMC with HPA. Blind docking simulations with BDMC
`generated 90 poses with the minimum interaction energy of -
`24.26 kJ mol-1. In the lowest energy pose obtained from
`blind docking BDMC was docked at a site away from active
`site and was subsequently used for refined docking
`simulation. A total of 90 poses were generated in refined
`docking and two poses were found to have the least binding
`energy (-26.4 kJ mol-1) In both the docked poses, the
`principal interactions of BDMC with the enzyme include (cid:3) -
`(cid:3) interactions between the aromatic rings of BDMC and
`Phe55 as well as hydrogen bonds. In pose 1 Fig. (6A), the
`
`NOVARTIS EXHIBIT 2107
`Par v. Novartis, IPR 2016-01479
`Page 6 of 11
`
`

`

`HPA Inhibition by BDMC
`
`The Natural Products Journal, 2013, Vol. 3, No. 1 21
`
`Fig. (5). CD spectra of HPA bound with BDMC. A) Far –uv spectra (195-250 nm) of BDMC bound and unbound HPA. B) Near –uv
`spectra (250-320 nm) of BDMC.bound and unbound HPA.
`
`
`
`Table 1. Thermodynamic Parameters of BDMC-HPA Binding from Fluorescence Titrations
`
`Conditions
`
`Ka (x 104
`
` M-1)a
`
`Kd (x 10-6
`
` M-1)b
`
`-(cid:1)Go (k J mol-1)b
`
`295 K
`
`298 K
`
`301 K
`
`303 K
`
`HPA- maltose complex
`
`12.4
`
`8.5
`
`6.5
`
`6.0
`
`5.0
`
`8.01±0.43
`
`11.77±0.38
`
`15.33±0.93
`
`16.6±0.49
`
`20.05±0.96
`
`28.78±0.9
`
`28.13±0.3
`
`27.74±0.1
`
`27.72±0.5
`
`26.80±0.3
`
`aThe values were determined using non-linear regression analysis. Error bars represent ±SE of the mean of triplicates.
`bThe values for Ka and -(cid:2)Go were calculated as mentioned in Materials and Methods.
`
`hydrogen bond interactions occur between the C11(cid:1) OH
`group of BDMC and Asn362 and H2O 47, whereas in pose 2
`Fig. 6B) C11(cid:1) OH-Asn355 and C3-Asn362 are involved.
`
`Trp357 is at a distance favorable for quenching in both these
`poses with C(cid:1) of Trp357 at a distance of 9.8 Aº and 8.8 Aº
`from C8(cid:1) of BDMC in pose 1 and 2, respectively (Fig. 6).
`
`NOVARTIS EXHIBIT 2107
`Par v. Novartis, IPR 2016-01479
`Page 7 of 11
`
`

`

`22 The Natural Products Journal, 2013, Vol. 3, No. 1
`
`Ponnusamy et al.
`
`
`
`Fig. (6). Putative mode of interaction between BDMC (represented as sticks) and HPA. A) Pose 1: Two pi–pi interactions of two
`aromatic rings of BDMC with F55 and two hydrogen bond interactions with N362 and H2O 47 are seen. B) Pose 2: A single pi–pi interaction
`of BDMC with F55 and two hydrogen bond interactions with N362 and N355 are seen.
`
`DISCUSSION
`
`The aim of the present study was to investigate the
`principle component
`from C.
`longa
`responsible
`for
`pancreatic amylase inhibition on AR42J rat pancreatic acinar
`cell line followed by elucidation of its mode of inactivation
`of the target, HPA. Thus, enzyme kinetics, ligand binding
`thermodynamics and in silico docking studies were carried
`out to gain insight into the mechanism of inhibition. The
`medicinal property of C. longa is majorly attributed to the
`content of curcuminoids which is about 4-6% and of
`immense value with its pharmaceutical and nutraceutical
`potentials [13]. BDMC, one of the curcuminoid has been
`reported for its anti-oxidant and anti-inflammatory activities
`[30] was identified as the principle component exhibiting
`in vitro cell line secretory pancreatic amylase inhibition.
`
`AR42J cell line is characterized by the presence of
`digestive enzyme-containing dense core vesicles, which are
`released in response to cholecystokinin. In addition, these
`cells also contain small neuroendocrine-specific vesicles, as
`evidenced by the expression of the neuroendocrine-specific
`
`vesicle proteins synaptophysin and S.V.2 [31]. Much more,
`dexamethasone converts pluripotent pancreatic AR42J cells
`into exocrine cells expressing digestive enzymes by
`increasing the intracellular and secretory amylase contents at
`its mRNA level [12]. These properties make AR42J cell line
`an ideal system to work on pancreatic (cid:1)-amylase inhibitors.
`Crude extract exhibits a better bioactivity as compared to
`pure isolated lead inhibitor (BDMC). This could be either
`due to the lower solubility / bioavailability of BDMC or the
`synergistic effect of other components present in the crude
`extract. The former possibility is more likely as neither
`curcumin nor demethoxycurcumin, the other components
`present in the isopropanol extract exhibited HPA inhibition
`on isolation. On assaying the bioactivity in the presence of
`varying starch load, which would mimic the postprandial
`conditions, crude and purified lead inhibitor, both BDMC
`were found to inhibit the secretory amylase activity under
`simulated diabetic conditions.
`
`Enzyme inhibition kinetics have revealed the unusual
`uncompetitive mode of inhibition, since to the best of our
`
`NOVARTIS EXHIBIT 2107
`Par v. Novartis, IPR 2016-01479
`Page 8 of 11
`
`

`

`HPA Inhibition by BDMC
`
`The Natural Products Journal, 2013, Vol. 3, No. 1 23
`
`report by Nagaraj and
`knowledge, only an earlier
`Pattabiraman (1985) on a proteinaceous inhibitor of HPA
`from proso millet (Panicium miliaceum) seeds exhibiting
`uncompetitive inhibition exists [32]. Generally, other plant
`derived polyphenols and flavonoids such as myricetin,
`luteolin, fisetin and quercetin, known to be effective
`inhibitors of mammalian (cid:3)-amylase
`exhibit
`either
`competitive or mixed inhibition [33]. An inhibition constant
`of Ki = 2.5 mM was determined for the synthetic peptide
`PAMI using PPA as target and was indicative of a mixed-
`type, non-competitive inhibition [29], while trans-chalcone, a
`flavonoid intermediate in plants competitively inhibited it
`with a Ki of 48 (cid:1)M [34]. Kinetic analysis of montbretin A
`isolated
`from
`extract of Crocosmia
`crocosmiiflora
`demonstrated it to be a tight-binding competitive inhibitor of
`HPA with Ki of 8.1 nM [5]. Acarbose exhibiting mixed
`inhibition had a Ki(cid:3)of 2.71 and 0.866 (cid:1)M for HSA and PPA,
`respectively [35]. Thus, the mode of inhibition of HPA by
`BDMC is different as compared to the other small molecules
`reported from natural extracts. In an earlier study, Braun et al
`(1995) have shown that 2-deoxy-2, 2-difluoro-D–maltose,
`inactivated HPA with ki/Ki of 0.0073 min-1mM-1 [36].
`
`Since, curcuminoids are well known to possess both
`antioxidant and prooxidative properties, it was possible that
`BDMC might be
`inactivating HPA
`through
`redox
`modifications. In order to eliminate such a possibility, the
`inhibition of HPA by BDMC was measured in the presence
`of 1 and 5 mM dithiothreitol (DTT). As no effect on HPA
`inhibition was observed under reducing conditions, it was
`likely that BDMC inactivated HPA by binding to it.
`
`Intrinsic fluorescence quenching has often been applied
`to study the binding of small organic molecules such as
`phenolics to proteins. We applied this approach to determine
`the interaction of BDMC with HPA. HPA is composed of
`three structural domains viz., domain A (residues 1-99, 169-
`404), in which are located the active site residues, domain B
`(residues 100-168) serves to form a calcium binding site, the
`presence of which is essential for catalytic activity and
`domain C forming a structurally independent antiparallel (cid:4)-
`barrel. Chloride (Cl-) induces an allosteric activation in HPA
`and removal of Cl- by dialysis results in total inactivation of
`the enzyme, which is fully recovered by the addition of Cl-
`[22, 37]. As the mode of inhibition is uncompetitive, the
`binding of BDMC to HPA was studied in the presence of
`maltose; an end-product of HPA catalyzed starch hydrolysis.
`Maltose
`is known
`to bind at
`the enzyme subsite
`competitively to form a nonproductive enzyme-substrate
`complex [38].
`
`The degree of fluorescence quenching for both the HPA
`and HPA-maltose complex suggests the involvement of
`aromatic residues in the enzyme-inhibitor complex formation
`and that the probable binding site of BDMC could be in the
`vicinity of a surface exposed trytophan residue. Ear