Docking and Molecular Dynamics of Steviol GlycosideâHuman Bitter

Subscriber access provided by Northern Illinois University

Article

Docking and Molecular Dynamics Simulations of Steviol Glycosides - Human Bitter Receptors Interactions Waldo Andres Acevedo, Fernando Danilo Gonzalez-Nilo, and Eduardo Agosin J. Agric. Food Chem., Just Accepted Manuscript • DOI: 10.1021/acs.jafc.6b02840 • Publication Date (Web): 17 Sep 2016 Downloaded from http://pubs.acs.org on September 18, 2016

Just Accepted “Just Accepted” manuscripts have been peer-reviewed and accepted for publication. They are posted online prior to technical editing, formatting for publication and author proofing. The American Chemical Society provides “Just Accepted” as a free service to the research community to expedite the dissemination of scientific material as soon as possible after acceptance. “Just Accepted” manuscripts appear in full in PDF format accompanied by an HTML abstract. “Just Accepted” manuscripts have been fully peer reviewed, but should not be considered the official version of record. They are accessible to all readers and citable by the Digital Object Identifier (DOI®). “Just Accepted” is an optional service offered to authors. Therefore, the “Just Accepted” Web site may not include all articles that will be published in the journal. After a manuscript is technically edited and formatted, it will be removed from the “Just Accepted” Web site and published as an ASAP article. Note that technical editing may introduce minor changes to the manuscript text and/or graphics which could affect content, and all legal disclaimers and ethical guidelines that apply to the journal pertain. ACS cannot be held responsible for errors or consequences arising from the use of information contained in these “Just Accepted” manuscripts.

Journal of Agricultural and Food Chemistry is published by the American Chemical Society. 1155 Sixteenth Street N.W., Washington, DC 20036 Published by American Chemical Society. Copyright © American Chemical Society. However, no copyright claim is made to original U.S. Government works, or works produced by employees of any Commonwealth realm Crown government in the course of their duties.

Page 1 of 50

Journal of Agricultural and Food Chemistry

Docking and Molecular Dynamics Simulations of Steviol Glycosides - Human Bitter Receptors Interactions Waldo Acevedo,1 Fernando González-Nilo,2 and Eduardo Agosin1,∗ 1

Department of Chemical and Bioprocess Engineering, School of Engineering,

Pontificia Universidad Católica de Chile, Avenida Vicuña Mackenna 4860, Santiago, Chile. 2Universidad

Andrés Bello, Center for Bioinformatics and Integrative Biology, Faculty

of Biological Sciences, Avenida República 239, Santiago, Chile. ∗Phone:

+56 2 2354 4253. Fax: +56 2 2354 5803. E-mail: [email protected].

ACS Paragon Plus Environment


1

ABSTRACT: Stevia is one of the most sought after sweeteners by consumers due to its

2

natural origin and minimum calorie content. Steviol glycosides (SG) are the main

3

active compounds present in the leaves of Stevia rebaudiana, and are responsible for

4

its sweetness. However, recent in vitro studies in HEK 293 cells revealed that SG

5

specifically activate the hT2R4 and hT2R14 bitter taste receptors, triggering this mouth

6

feel. The objective of the present study was to characterize the interaction of SG with

7

these two receptors at molecular level. The results showed that SG have only one

8

orthosteric binding site to these receptors. The binding free energy (∆Gbinding) between

9

receptor and SG was negatively correlated with SG bitterness intensity, for both hT2R4

10

(r=-0.95) and hT2R14 (r=-0.89). We also determined, by steered molecular dynamics

11

(SMD) simulations, that the force required to extract stevioside from the receptors

12

was greater than that required for rebaudioside A, in accordance with the ∆Gs

13

obtained by molecular docking. Finally, we identified the loop responsible for the

14

activation by SG of both receptors. As a whole, these results contribute to a better

15

understanding of the resulting off-flavor perception of these natural sweeteners in

16

foods and beverages, allowing for better prediction – and control – of the resulting

17

bitterness.

18

KEYWORDS: stevia, steviol glycosides, bitter taste receptor, molecular docking,

19

molecular dynamics simulations


Page 2 of 50

Page 3 of 50


20

INTRODUCTION

21

Sugar has historically played a leading role as a tabletop sweetener and in the

22

preparation of various food products. However, at present it has been determined that

23

a significant number of chronic, noncommunicable diseases are a consequence of the

24

excessive consumption of sugar.1 Among these diseases are cardiovascular accidents,

25

diabetes, and obesity, which claim the lives of 38 million people each year.2

26

One way to prevent these diseases is by consuming noncaloric sweeteners to replace

27

sucrose.3,4 There are numerous natural compounds derived from plants that perform

28

this function,5 where steviol glycosides (SG), extracted from Stevia rebaudiana, are the

29

most popular. Among these glycosides, the most important are stevioside, various

30

rebaudiosides, dulcosides, steviobioside, and rubusoside, which give the extract a

31

sweetening activity about 300 times that of sucrose (Figure 1).6 The sweetening

32

capacity – similar to aspartame – and stability of these active compounds transform

33

them into an excellent alternative for the – at least partial – substitution of sugar in

34

processed foods.7

35

Unfortunately, these compounds generate, in addition to their sweetening effect,

36

unwanted sensory attributes such as bitterness, metallic and licorice tastes in foods

37

and beverages sweetened with Stevia.8 Recent in vitro studies in HEK 293 cells revealed

38

that SG specifically activate the hT2R4 and hT2R14 bitter taste receptors, triggering



39

this mouth feel.9 It is worthy to note that other artificial sweeteners, like saccharin and

40

acesulfame-K, also induce a bitter taste, but this is due to the activation of hT2R43 and

41

hT2R44 receptors instead.10 Humans have 25 taste receptors (hT2R type) responsible

42

for the sensation of bitterness, which belong to a superfamily of G-protein coupled

43

receptors (GPCR).

44

Specifically, these receptors share partial sequence similarity with the GPCR Class A

45

subfamily. However, the activation mechanisms of this subfamily are poorly

46

understood.11 In structural terms, hT2R receptors have between 290 and 333 amino

47

acids and are composed by 7 transmembrane segments, a short extracellular N-

48

terminal domain, and an intracellular C-terminal domain.12,13

49

Functional studies in humans showed that hT2R receptors exhibit high selectivity for

50

their ligands.14 On the one hand, the hT2R4 receptor has a limited degree of

51

promiscuity and,15 on the other hand, hT2R14 responds to a wide variety of natural and

52

synthetic, structurally different bitter compounds, including alkaloids and medicinal

53

drugs;16 while other receptors – such as hT2R16 and hT2R38 – are activated by a family

54

of structurally related ligands that contain β-glucopyranoside and isothiocyanate.17,18

55

Thus, the architecture of the ligand-binding site, as well as the amino acid residues

56

implicated in the interaction, are required for the activation of these bitter taste

57

receptors.19


Page 4 of 50

Page 5 of 50


58

GLY28 and SER285 residues are highly conserved in transmembrane domains 1 and 7

59

of hT2R receptors.20 These residues could play an important role in receptor activation,

60

mediated by the formation of hydrogen bonds between residues. However, only GLY28

61

is conserved in Class A GPCRs.21 Moreover, most hT2R structural models have been

62

constructed by comparative modeling, using the crystal structure of rhodopsin (a

63

member of the GPCR Class A family) from different species as a template.20,22,23 Despite

64

the limited sequence similarity, the conserved residues act as binding sites for various

65

compounds. The first modeled structure developed to study the activation mechanism

66

of hT2R receptors, was hT2R1.22 Opsin (PDB ID: 3DQB), which has 22% sequence

67

identity with hT2R1, was used as template. Subsequently, the hT2R4 and hT2R38

68

receptors were modeled, using bovine (PDB ID: 1U19) and squid (PDB ID: 2Z73)

69

rhodopsin as templates, respectively, which have about 20% sequence identity with

70

each receptor.20,23

71

Kubota and Kubo were the first to propose that the chemical basis of bitterness, based

72

on the structure of diterpenes isolated from the Rabdosia plant, could be represented

73

by a three-point attachment model, resulting from the presence of a proton donor

74

group AH-, a proton acceptor group B- and a third binding site X- that interacts with

75

the hydrophobic site on the receptor.24 More recently, Nofre and Tinti developed the

76

Multipoint Attachment Model,25 a more elaborated theory for the taste of any



77

molecule, including simple sugars and artificial sweeteners. This model assumes the

78

presence of at least eight fundamental recognition points or sites, namely, B-, AH-,

79

XH-, G1- ,G2- ,G3- , G4- and D- that interact with a receptor through three types of

80

elementary interactions, i.e. ionic, H-bonding and steric interactions.

81

The objective of the present study was to develop a model capable of predicting the

82

bitterness intensity of the main sweetening compounds extracted from Stevia

83

rebaudiana, by molecular simulation. To this end, we built a model for hT2R4 and

84

hT2R14 bitter taste receptors, using as reference structures, bovine and squid

85

rhodopsin, respectively. Then, we identified the corresponding binding sites by both

86

hydrophobic interactions and hydrogen bonding and we calculated the binding free

87

energy associated with SG. The calculated binding free energies were correlated with

88

the SG relative bitterness previously reported in the literature.3 Finally, we simulated,

89

by steered molecular dynamics (SMD), the unbinding process of the ligands and the

90

conformational changes of the binding sites of the modeled bitter taste receptors.

91

MATERIALS AND METHODS

92

Template selection. The amino acid sequences of hT2R4, hT2R14 and hT2R1 bitter

93

taste receptors were retrieved from the NCBI (Accession: EAW83983, EAW96222, and

94

EAX08075, respectively). Squid rhodopsin (PDB ID: 2Z73) was selected as a template

95

to model hT2R14, since the latter does not have homolog proteins with a known


Page 6 of 50

Page 7 of 50


96

structure. To this end, we used the PGenTHREADER method,26 based on the technique

97

of protein fold recognition, which allows for the detection of templates for three-

98

dimensional (3D) structural prediction of this protein family with more certainty.

99

Bovine rhodopsin (PDB ID: 1U19) and opsin (PDB ID: 3DQB) were used as templates

100

to model hT2R4 and hT2R1, respectively, in accordance to that reported in the

101

literature.22,20

102

Modeling of bitter taste receptors. We built comparative models for hT2R4, hT2R14

103

and hT2R1 using the crystal structure of the protein templates mentioned above. The

104

hT2R1 receptor was used as a negative control. Models were built using the SWISS-

105

MODEL server.27-28 The resulting models were subjected to cycles of energy

106

minimization until convergence and equilibration for 20 ps, in order to relax the

107

conformation of side chains and avoid conformational tension generated by the

108

homology model. All calculations were performed using the programs NAMD

109

(NAnoscaled Molecular Dynamics; v 2.9) and CHARMM force field.29,30 Subsequently,

110

we assessed the stereochemical quality of the models using Ramachadran plot

111

(RAMPAGE), PROCHECK (Overall quality factor) and ERRAT (Overall quality factor).

112

The models were visualized and rendered using Visual Molecular Dynamics (VMD

113

1.9).31,32



Page 8 of 50

114

Ligand preparation. The 3D structure of the active ingredients of stevia were

115

obtained from the PubChem database33 as sdf file format (Figure 1). The program

116

OpenBabel 2.3.1 was employed to convert ligand representations to mol2 format and

117

visually checked to correct structural errors.

118

Molecular docking of “sweeteners – bitter taste receptors”. Binding sites, as well

119

as the associated free energies (∆Gbinding) of the different sweeteners with both, the

120

hT2R4 and hT2R14 receptors, were predicted using the program Autodock Vina.34,35

121

Protein and ligand preparation was carried out using Autodock Tools 1.5.6.36 Gasteiger

122

partial charges were assigned to the atoms of ligands. The AutoTors option was used to

123

define rotatable bonds in the sweeteners. The Lamarckian genetic algorithm was

124

employed for all docking calculations. The visual inspection of the results was carried

125

out using the MGL Tools package. We selected a volume of 60x60x60 grid points,

126

enough to cover each receptor. In Autodock Vina,34 the binding free energies - ∆Gbinding

127

- were estimated by the sum of inter- and intra-molecular contributions that

128

approximate the thermodynamic stability of the ligand with the receptor.

129

∆

130

Where ∆Ggaussian is the attractive term for dispersion, ∆Grepulsion is the repulsive term,

131

∆Ghbond describes interactions by hydrogen bonds, ∆Ghydrophobic depicts hydrophobic

132

interactions and ∆Gtorsional corresponds to the number of rotatable bonds. The value of

=

∆

+

∆

+

∆

+

∆


+

∆

(1)

Page 9 of 50


133

weights was calculated by minimizing the difference between interaction affinities

134

(pKd), predicted and measured, using a set of 1300 complexes from the ligand-protein

135

database (LPDB), and a nonlinear optimization algorithm.35 LPDB contains data of

136

high-resolution, ligand-protein complexes and known experimental binding

137

affinities.37

138

Subsequently, we carried out redocking simulations on the previously predicted

139

binding sites, both to validate binding free energies (∆Gbinding) and the bioactive

140

conformation of the ligands. Finally, we correlated the ∆Gbinding calculated with the

141

relative bitterness of the evaluated sweeteners. The relative bitterness values for SG

142

were obtained from Hellfritsch et al.9 It is worthy to note that the SwissDock docking

143

server38 yielded significantly lower correlations than those obtained by Autodock

144

Vina.

145

Mathematical modeling of the relationship between relative bitterness and the

146

interaction parameter. The experimental ligand-receptor binding free energy is

147

described by:

148

∆ =

149

Where R is the gas constant equivalent to 1.98721 [cal/K*mol], T is the temperature in

150

Kelvin (K), and Kd is the dissociation constant. The relative bitterness (RB) of the

× ln(

(2)

)



Page 10 of 50

151

different active compounds was obtained from the literature based on 1 mM of

152

rubusoside.9 Thus, we express the RB values as relative Kd values: =

153

(

)

154

From equation 2, we obtain a linear relationship for RB:

155

ln(

156

Analysis of protein-ligand interaction. The residues responsible for hydrogen

157

bonding and hydrophobic interactions with SG were identified and visualized with the

158

program LigPlot+.39 LigPlot+ is a graphic system that generates multiple two-

159

dimensional (2D) diagrams of protein-ligand interactions. These analyses were carried

160

out from 3D coordinates from lower energy poses of receptor-ligand complexes.

161

Hydrogen bonds were identified by geometric criteria between hydrogen (H), the

162

donor (D) and the acceptor (A). The H-A and D-A distances should be less than 2.7 Å

163

and 3.3 Å, respectively. The D-H-A and H-A-AA angles should be greater than 90°,

164

where AA is the atom bound to the acceptor from the amino acid. The distance and

165

angle geometric criteria were established based on crystal structures of protein-ligand

166

complexes.40 We estimated the hydrophobic interactions of ligands at the binding site

167

between pairs of carbon atoms whose minimum and maximum contact distance was

168

2.9 and 3.9 Å, respectively. Finally, intramolecular interactions are represented

169

schematically in 2D from 3D coordinates of the protein-ligand complex.

) = ln(

) − 1

(3)

×∆


Page 11 of 50


170

Molecular dynamics simulations. Based on receptor docked conformations with

171

stevioside and rebaudioside A, we run Conventional Molecular Dynamics (CMD)

172

simulations using NAMD (NAnoscaled Molecular Dynamics; v 2.9).29 The latter

173

allowed to relax the collision between atoms to obtain a more stable complex

174

conformation. These SG were employed as references, considering that they are the

175

most quantitatively important SG present in Stevia rebaudiana.41 We applied periodic

176

boundary conditions to the system to obtain consistent behavior. We assigned a

177

distance cutoff of electrostatic interaction of 12, which is sufficient to achieve

178

reasonable results from the molecular dynamics calculations.42 The particle mesh

179

Ewald method was employed for electrostatic interactions in periodic boundary

180

conditions with grid dimensions of 1.0 Å. The files generated for the ligand-protein

181

complexes using the VMD 1.9 software were first inserted into a POPC lipid bilayer

182

using VMD’s “membrane” plugin.43

183

Then, simulations were carried out using the parameter file CHARMM 22 (Chemistry

184

at HARvard Macromolecular Mechanics) for proteins and lipids. The energy of each

185

system was minimized, using the Powell algorithm at constant temperature (310 K),

186

followed by equilibration with the Langevin dynamics44 to control the kinetic energy,

187

temperature, and/or pressure of the system. The Langevin dynamics method follows

188

the equation:



189

=

( ) −

+

( )

Page 12 of 50

(4)

190

Where F is the force, γ is the damping coefficient, m is the mass and R(t) is the frictional

191

force.

192

The predicted models were refined by minimization for 2000 steps, keeping the

193

backbone fixed the first 1000 steps, and equilibration for 18 ps. Later, ligand

194

parameterization was accomplished using the ACPYPE web server.45,46 Parameterized

195

ligands were inserted into the binding site of the protein and saved as a ligand-protein

196

complex using VMD. Finally, the different complexes inserted in lipid membrane were

197

minimized for 1000 steps (Supplementary material S2) and then equilibrated for 10 ns

198

in order to relax the conformation of the ligand, as well as the lateral chains of the

199

residues of the receptor. Each equilibration trajectory was saved at 5 ps for further

200

analysis. After the simulations, the RMSD values of alpha carbon atoms (Cα) in the

201

receptors were calculated, using the initial respective structures as reference points in

202

the MD trajectories; and the root mean square fluctuation (RMSF) was executed to

203

compare the conformation of free and SG-bound proteins. The analyzed trajectories

204

were plotted using GNUPLOT.

205

Steered molecular dynamics simulations. After the different complexes were

206

minimized and equilibrated by CMD, we performed SMD simulations to study the

207

unbinding process of the ligands, as well as the conformational changes of the binding


Page 13 of 50


208

sites of the modeled bitter taste receptors. To do this, the equilibrated ligand-receptor

209

complexes were driven at constant velocity (cv), using the approach implemented in

210

NAMD.29 The ligands were extracted from the binding site at a constant velocity of

211

0.005 Å/ps by applying an external force variable to its center of mass. The force exerted

212

on the ligand was defined as:

213

= (

(5)

− )

214

Where k is the spring constant (k = 7 Kcal/mol*Å) of the contraint, v is the pulling

215

velocity (v = 0.005 Å/ps), t represents time (in ps), and x is the position of the ligand at

216

time t. Force was applied to carbon 1 (C-1) of the aglycone part of stevioside and

217

rebaudioside A. SMD simulations were run for 20 ps at 300 K. Unbinding directions

218

were calculated using Tcl/Tk scripts and the information of residues was stored as a

219

reference file. Trajectories were saved every 10 steps. Each simulation was repeated 6

220

times. All graphical analyses were performed using VMD 1.9 software.

221

RESULTS

222

Modeling of taste receptors responsible for the bitterness of stevia. The template

223

for the comparative modeling of hT2R14 was selected using PGenTHREADER, based

224

on the percentage of identity and p-value. Squid rhodopsin (PDB ID: 2Z73) showed a

225

sequence identity of 18.7% and a p-value of 2e-04 with respect to hT2R14. Meanwhile,

226

bovine rhodopsin (PDB ID: 1U19) and opsin (PDB ID: 3DQB) showed a sequence



227

identity of 21.6% and 20.3% with respect to hT2R4 and hT2R1, respectively; and a p-

228

value of 6e-04, similar for both receptors. These results indicate that these

229

transmembrane proteins - 2Z73, 1U19 and 3DQB - which belong to the class A GPCR

230

family, are appropriate templates for predicting the 3D structure of hT2R4, hT2R14 and

231

hT2R1 (Figure 2). Figure 2 (a to c) illustrates the loops in the extracellular regions and

232

seven transmembrane domains, which were relaxed by molecular simulations to refine

233

their structures.

234

We compared the structures of our homology model, generated using a single

235

template, with the fragment modeling approach (GPCR-SSFE Database), using a suite

236

of structure alignment techniques (TopMatch webserver).47,48 Results indicated that

237

the structures of transmembranal regions were similar, with a root-mean-square error

238

superposition of 2.86 and 2.19 Å for hT2R4 and hT2R14, respectively.

239

Ramachandran plots (Supplementary material S1) evidenced that 92.2% of the residues

240

are in favoured regions, 4.8% in allowed regions, and 3.0% in outlier regions for hT2R4;

241

83.8% of the residues are in favoured regions, 12.4% in allowed regions, and 3.8% in

242

outlier regions for hT2R14 and 83.3%, 9.0%, and 7.6%, respectively for hT2R1.

243

PROCHECK analyses of the hT2R4, hT2R4 and hT2R1 models yielded overall average

244

G factors of -1.08, -1.22 and -1.27, respectively. Overall quality factors were 81.395,

245

66.809 and 69.585 for hT2R4, hT2R4 and hT2R1, respectively.


Page 14 of 50

Page 15 of 50


246

Molecular docking and binding site characterization. Molecular docking showed

247

that the calculated binding free energies (∆Gbinding) negatively correlated with SG

248

bitterness intensity 9 (Figure 3). The correlation coefficient for the 10 evaluated SG with

249

hT2R4 and hT2R14 was -0.95 and -0.89, respectively. On the other hand, we found a

250

poor correlation with hT2R1 (R = -0.50).

251

Rubusoside, the bitterest SG, has a lower ∆Gbinding than rebaudioside D, for both

252

receptors (Figure 3a and 3b); this could result from their different structural

253

characteristics, as rubusoside and rebaudioside D contain 2 and 5 glucose residues,

254

respectively. On the other hand, those SG that contain rhamnose, e.g. dulcoside A and

255

rebaudioside C, have lower ∆Gs for hT2R14.

256

Table 2 identifies those amino acids responsible for hydrogen bonding and

257

hydrophobic interactions in hT2R4 and hT2R14 with SGs. The fact that 43 and 75%, of

258

the aminoacidic residues responsible for these interactions in the receptors are polar

259

suggests that the hydrophilic nature of the pocket favors the binding of these

260

sweeteners with the bitter taste receptors.

261

Figures 4 and 5 depict the potential binding sites of SG and the pose of docked

262

stevioside in the receptors, respectively. The binding cavity in these receptors was

263

mainly composed by transmembrane regions – with moderate participation from the

264

extracellular region – which interact with monosacharides bound to the C-13 and C-19



265

of SG (Figure 1). The surface of the hT2R4 and hT2R14 binding cavity is 2017 and 3430

266

Å 2, respectively.

267

Docking of different complexes allowed to identify the most representative modes of

268

SG binding to the bitter taste receptors (Figures 6 and 7). SG differ both in their

269

orientation and conformation, selectively interacting with specific residues at the

270

binding site. In addition, the interactions of SG with the binding pocket of bitter taste

271

receptors are governed by both hydrogen bonding and hydrophobic interactions.49 On

272

the one hand, oxygen and nitrogen atoms from residues of the binding pocket interact

273

with the OH groups of monosaccharides bound at positions C-13 and C-19 of SG by

274

hydrogen bonds. On the other, the carbons of the aglycone part of SG and of apolar

275

residues at the binding site interact with each other through hydrophobic interactions.

276

Ligand-receptor molecular dynamics simulations. Figure 8 shows all the

277

conformations of the hT2R4 and hT2R14 receptors bound to stevioside or to

278

rebaudioside A after 10 ns of equilibration. RMSD values of the receptor bound to the

279

ligand were between 3.0 and 4.0 Å (Supplementary material S3). On the other hand,

280

RMSD values of the ligand bound to the receptor were between 2.0 and 3.0 Å

281

(Supplementary material S4). Accordingly, it can be inferred that the overall

282

conformations of receptors, as well as of ligands, were stable.


Page 16 of 50

Page 17 of 50


283

RMSD calculations allowed to estimate the global stability of the conformations, but

284

do not consider the local region of proteins, particularly the binding site. Therefore,

285

we decided to analyze some physical parameters for this purpose, including the

286

number of hydrogen bonds and RMSF. The results evidenced that, on average, the

287

number of hydrogen bonds of both ligands with binding site residues is between 3 and

288

4 (Supplementary material S5), and fluctuations are less than 1.5% of RMSF (Figure 9).

289

In general, the conformations of the different protein-ligand complexes were stable

290

(Figures 10a and 10b).

291

Steered molecular dynamics simulations. After inspection of the SG binding sites

292

on the receptors, and a comparative analysis of SG-hT2R interactions, we run SMD

293

simulations at constant velocity to extract SG – stevioside and rebaudioside A – from

294

the binding site in order to better understand the different affinities of both ligands.

295

Stevioside and rebaudioside A were stabilized, during the unbinding process, at 3 and

296

1 ps for hT2R4 and hT2R14, respectively. Moreover, the force peak employed to extract

297

stevioside from the binding site was greater than the one required for rebaudioside A,

298

with values of 3747 and 2057 pN; and 1647 and 1279 pN for hT2R4 and hT2R14,

299

respectively (Figure 10).

300

Figure 11 illustrates the different phases of the unbinding process. After equilibration

301

inside the binding pocket, the D-glucopyranosyl C-19 hydroxyl group of stevioside and



302

rebaudioside A disentangles the Tyr 68 – Ser 81 extracellular loop of hT2R4, and the

303

His 151 – Thr 184 of hT2R14, making its way to the binding pocket entrance. RMSF

304

values indicate that these extracellular loops are highly flexible during ligand

305

unbinding. In particular, Ser 71 and Ser 72 residues for hT2R4 - and Ser 155 and Ser 170

306

for hT2R14 - correspond to peaks of RMSF fluctuations. Furthermore, hT2R4

307

experiences bigger conformational changes than hT2R14 in the dissociation of the

308

ligand (Supplementary material S6).

309

DISCUSSION

310

Hellfritch et al. 3 proposed that the bitter taste of stevia is promoted by the structural

311

characteristics of SG - e.g. the sugar content in their structures – and the nature of the

312

SG binding cavity of hT2R4 and hT2R14 receptors.9 In the present study, we predicted

313

the 3D structure of the hT2R4 and hT2R14 bitter taste receptors, and elucidated the

314

interaction patterns with SG.

315

Even though Worth et al. suggested that building comparative models of GPCR using

316

transmembrane fragments from multiple templates might lead to more accurate

317

results than homology modeling, using a single template,

318

approaches were very similar (Supplementary material S7). The latter confirms that

319

homology modelling using a single template is an appropiate tecnique to build this

320

type of receptors.


47

we found out that both

Page 18 of 50

Page 19 of 50


321

The resulting models share the same structural pattern – seven transmembrane helices

322

and the presence of loops in both, the extracellular and intracellular regions – similar

323

to other previously modeled bitter taste receptors. 22,20,23 Unlike sweet taste receptors,

324

these proteins have a small extracellular region – around 60 amino acid residues; thus,

325

it is likely that these sweeteners bind to both, the extracellular and the transmembrane

326

regions.

327

Previous studies have shown that the most divergent sequences of hT2Rs are the

328

extracellular loops, which affect the selectivity of various bitter taste receptors, such as

329

hT2R3, hT2R7, hT2R10, hT2R13 and hT2R61.14,50 Moreover, the intracellular loops play

330

a critical role in the activation of these receptors,11 as the conformational change in

331

these regions allows for the formation of bridges between transmembrane domains,

332

stabilizing the inactive state of the receptor.

333

We determined, by molecular docking, that the calculated binding free energies

334

(∆Gbinding) negatively correlate with the bitterness intensity of different SG. This strong

335

correlation confirms the potential binding sites predicted for the SG to hT2R4 and

336

hT2R14. Furthermore, the poor correlation obtained for hT2R1 suggests that SG only

337

bind to hT2R4 and hT2R14. Indeed, the binding site identified on hT2R1 is not

338

appropriate to accommodate this type of sweeteners.



339

It is possible that the SG binding site identified on hT2R1 was mainly composed by

340

transmembrane regions, while the binding pocket of hT2R4 and hT2R14 included

341

both,

342

regions; thus, receptor activation by these sweeteners – similar to other water soluble

343

molecules, such as phenylthiocarbamide, quinine and caffeine – might start with a

344

marked intervention of the extracellular regions.51

345

In addition, our docking results confirm previous sensory studies reporting that

346

rubusoside is the bitterest SG, and rebaudioside D is the least bitter9. Therefore, this

347

suggests that the affinity of SG to bitter taste receptors, as well as the differences in

348

bitterness intensity and the reported activation thresholds,9 could be due to the steric

349

characteristics of SG and their binding site architecture.19

350

The SG with more sugars have less affinity for bitter taste receptors, since the size of

351

the receptor binding cavity is limited. In fact, the surface of the hT2R14 binding cavity

352

is greater than that of hT2R4, allowing larger SG to adequately bind. This characteristic

353

could explain the capacity of hT2R14, as well as hT2R10 and hT2R46, to respond to a

354

wide variety of different bitter compounds.16,52

355

Furthermore, the high affinity of rhamnose-containing SG, rebaudioside C and

356

dulcoside A, to hT2R14 does not result only from the total amount of monosaccharides,

357

but also from the type of sugars present in their structure. This confirms the results of

extracellular regions and a moderate participation of the transmembrane


Page 20 of 50

Page 21 of 50


358

Hellfritsch et al. in that the activation threshold of hT2R14 by rhamnose-containing SG

359

are lower than for other SG.9

360

Regarding potential SG binding sites on these receptors, besides those recently

361

identified in hT2R4 by Singla and Jaitak53, i.e. Phe 69, Phe 168, Phe 169, Ile 170, Glu 172,

362

Met 259, Gly 260 and Lys 262, we showed that SG lies in the same binding cavity sharing

363

a set of amino acids, including Asn 65, Thr 66, Val 78, Phe 84 for hT2R4; and Ser 167,

364

Asp 168, Ser 246 for hT2R14. Activation of these receptors by SG may result from the

365

stimulation of residues, such as Val 78 and Phe 84 for hT2R4; and Ser 167 and Asp 168

366

for hT2R14, at a first, extracellular binding site by monosaccharides at the C-13

367

position; and the allosteric modulation of a second, transmembraneous binding site

368

by the monosaccharidic moieties of SG. This multipoint stimulation model has also

369

been suggested for sweet taste receptors but, unlike bitter taste receptors, the sugars at

370

the C-13 and C-19 positions only stimulate residues at the extracellular region.49

371

In addition, SG interactions with the binding pockets of the hT2R4 and hT2R14 bitter

372

receptors are governed by hydrogen bonding and hydrophobic interactions. The

373

binding mechanism of SG mediated by hydrogen bonds has also been identified on

374

sweet taste receptors, 49,54 due to the sugars present in these sweeteners. On the other

375

hand, the hydrophobic interactions between aminoacidic residues and both

376

components of SG could further promote their stability at the binding site. Moreover,



377

SG could simultaneously trigger sweetness and bitterness due to the polarization of

378

their sugars and their proximity – probably 3 - 4 Å – to the binding sites of sweet and

379

bitter taste receptors.55

380

The analysis of the SG unbinding process by SMD unravelled the difference in ligand

381

affinity between both receptors. Indeed, the peak applied force for stevioside was

382

greater than for rebaudioside A, as stevioside forms more hydrogen bonds with the

383

residues of both receptors (Figures 6 and 7). Moreover, we propose that the

384

extracellular loops Tyr 68 – Ser 81 on hT2R4, and His 151 – Thr 184 on hT2R14 are

385

responsible for anchoring stevioside and rebaudioside A to their respective binding

386

sites. Furthermore, the SG coupling process possibly induces some changes in the

387

structure of both receptors, allowing for the formation of hydrogen bonds between the

388

transmembrane regions and, thus, stabilizing the active conformations of the

389

receptors.56,57

390

In conclusion, in this work we constructed a comparative model for the hT2R4 and

391

hT2R14 receptors using the 3D structure of 1U19 and 2Z73, respectively, which was

392

subsequently validated using a Ramachandran plot. Furthermore, we predicted the

393

intensity of bitterness of the active compounds in stevia, from their binding free energy

394

(∆Gbinding) with the modeled bitter taste receptors. Calculated ∆Gs negatively


Page 22 of 50

Page 23 of 50


395

correlated with reported bitterness of SG, that is, the more negative the ∆Gbinding

396

between the sweetener and the receptor, the greater the bitterness intensity.

397

Finally, bitterness intensity of SGs resulted from the following features: 1) the presence

398

of transmembrane regions at the binding cavity of hT2R4, and hT2R14, which favors

399

the interactions with SG by hydrophobic contact; 2) the ability to form hydrogen bonds

400

with amino acids at the binding pocket due to the presence of sugars at the C-13 and

401

C19 positions; and 3) the difference in the total number and type of monosaccharides

402

present in their structures.

403

Therefore, the binding model of SG with hT2R4 and hT2R14 taste receptors developed

404

here could be useful to propose strategies to reduce unwanted attributes, such as

405

bitterness, from this type of sweeteners.

406

NOMENCLATURE

407

CHARMM: Chemistry at HARvard Macromolecular Mechanics

408

CMD: Conventional Molecular Dynamics

409

Kd: Dissociation constant

410

LPDB: Ligand-Protein Database

411

NAMD: NAnoscale Molecular Dynamics

412

PDB: Protein Data Bank

413

POPC: 1-Palmitoyl-2-Oleyl-PhosphatidylCholine



414

RB: Relative Bitterness

415

RMSD: Root Mean Square Deviation

416

RMSF: Root Mean Square Fluctuations at Cα carbon of the aminoacidic residues

417

SG: Steviol Glycosides

418

SMD: Steered Molecular Dynamics

419

VMD: Visual Molecular Dynamics

420

ASSOCIATED CONTENT

421

Supporting Information

422

Supplementary material supporting the results (S1, S2, S3, S4, S5, S6 and S7) is available

423

free of charge via the Internet at http://pubs.acs.org.

424

AUTHOR INFORMATION

425

Corresponding Author

426

∗Phone: +56 2 2354 4253. Fax: +56 2 2354 5803. E-mail: [email protected].

427

Notes

428

The authors declare no competing financial interests.

429

ACKNOWLEDGMENTS

430

We are grateful to Javier Sainz, Prodalysa Inc. for recommendations and suggestions

431

during the course of this work; to Francisco Saitúa and Martín Cárcamo for their

432

constructive comments on the manuscript; and to the anonymous referees who


Page 24 of 50

Page 25 of 50


433

provided valuable feedback to improve the final manuscript. Waldo Acevedo was

434

supported by a doctoral fellowship from the Chilean National Council of Scientific and

435

Technological Research (CONICYT).

436 437 438 439 440 441 442 443 444 445 446 447 448 449 450 451 452 453 454 455



456

REFERENCES

457

(1)

Malik, V. S.; Popkin, B. M.; Bray, G. A. Sugar-Sweetened Beverages and Risk of

458

Metabolic Syndrome and Type 2 Diabetes. Rev. Statements 2010, 33 (11), 2477–

459

2481.

460

(2)

World Health Organization. About Noncommunicable Diseases and Mental

461

Health http://www.who.int/mediacentre/factsheets/fs355/en/ (accessed Nov 7,

462

2013).

463

(3)

Update. Nutrition 2013, 29 (11-12), 1293–1299.

464 465

(4)

(5)

(6)

Prakash, I.; Markosyan, A.; Bunders, C. Development of Next Generation Stevia Sweetener: Rebaudioside M. Foods 2014, 3 (1), 162–175.

470 471

Kinghora, A. D.; Soejarto, D. D.; Inglett, G. E. Sweetening Agents of Plant Origin. CRC. Crit. Rev. Plant Sci. 1986, 4 (2), 79–120.

468 469

Hu, F. B.; Malik, V. S. Sugar-Sweetened Beverages and Risk of Obesity and Type 2 Diabetes: Epidemiologic Evidence. Physiol. Behav. 2010, 100 (1), 47–54.

466 467

Shankar, P.; Ahuja, S.; Sriram, K. Non-Nutritive Sweeteners: Review and

(7)

Kroyer, G. Stevioside and Stevia-Sweetener in Food: Application, Stability and

472

Interaction with Food Ingredients. J. Consum. Prot. Food Saf. 2010, 5 (2), 225–

473

229.

474

(8)

Rocha, I. F. de O.; Bolini, H. M. A. Passion Fruit Juice with Different Sweeteners:


Page 26 of 50

Page 27 of 50


475

Sensory Profile by Descriptive Analysis and Acceptance. Food Sci. Nutr. 2015, 3

476

(2), 129–139.

477

(9)

Hellfritsch, C.; Brockhoff, A.; Stähler, F.; Meyerhof, W.; Hofmann, T. Human

478

Psychometric and Taste Receptor Responses to Steviol Glycosides. J. Agric. Food

479

Chem. 2012, 60 (27), 6782–6793.

480

(10)

Kuhn, C.; Bufe, B.; Winnig, M.; Hofmann, T.; Frank, O.; Behrens, M.;

481

Lewtschenko, T.; Slack, J. P.; Ward, C. D.; Meyerhof, W. Bitter Taste Receptors

482

for Saccharin and Acesulfame K. J. Neurosci. 2004, 24 (45), 10260–10265.

483

(11)

Pydi, S. P.; Singh, N.; Upadhyaya, J.; Bhullar, R. P.; Chelikani, P. The Third

484

Intracellular Loop Plays a Critical Role in Bitter Taste Receptor Activation.

485

Biochim. Biophys. Acta 2014, 1838 (1), 231–236.

486

(12)

S. A Novel Family of Mammalian Taste Receptors. Cell 2000, 100 (6), 693–702.

487 488

(13)

Meyerhof, W. Elucidation of Mammalian Bitter Taste. Rev. Physiol. Biochem. Pharmacol. 2005, 154 (July), 37–72.

489 490

Adler, E.; Hoon, M. A.; Mueller, K. L.; Chandrashekar, J.; Ryba, N. J.; Zuker, C.

(14)

Ji, M.; Su, X.; Su, X.; Chen, Y.; Huang, W.; Zhang, J.; Gao, Z.; Li, C.; Lu, X.

491

Identification of Novel Compounds for Human Bitter Taste Receptors. Chem.

492

Biol. Drug Des. 2014, 84 (1), 63–74.

493

(15)

Meyerhof, W.; Born, S.; Brockhoff, A.; Behrens, M. Molecular Biology of



494

Mammalian Bitter Taste Receptors. A Review. Flavour Fragr. J. 2011, 26 (4), 260–

495

268.

496

(16)

Behrens, M.; Brockhoff, A.; Kuhn, C.; Bufe, B.; Winnig, M.; Meyerhof, W. The

497

Human Taste Receptor hTAS2R14 Responds to a Variety of Different Bitter

498

Compounds. Biochem. Biophys. Res. Commun. 2004, 319 (2), 479–485.

499

(17)

Bufe, B.; Hofmann, T.; Krautwurst, D.; Raguse, J. D.; Meyerhof, W. The Human

500

TAS2R16 Receptor Mediates Bitter Taste in Response to Beta-Glucopyranosides.

501

Nat Genet 2002, 32 (3), 397–401.

502

(18)

Bufe, B.; Breslin, P. A. S.; Reed, D. R.; Tharp, C. D.; Slack, J. P.; Kim, U.-K.;

503

Drayna, D.; Meyerhof, W. The Molecular Basis of Individual Differences in

504

Phenylthiocarbamide and Propylthiouracil Bitterness Perception. Can. Field-

505

Naturalist 2014, 128 (2), 189–190.

506

(19)

Brockhoff, A.; Behrens, M.; Niv, M. Y.; Meyerhof, W. Structural Requirements

507

of Bitter Taste Receptor Activation. Proc. Natl. Acad. Sci. U. S. A. 2010, 107 (24),

508

11110–11115.

509

(20) Pydi, S. P.; Bhullar, R. P.; Chelikani, P. Constitutively Active Mutant Gives Novel

510

Insights into the Mechanism of Bitter Taste Receptor Activation. J. Neurochem.

511

2012, 122 (3), 537–544.

512

(21)

Smith, S. O. Structure and Activation of the Visual Pigment Rhodopsin. Annu.


Page 28 of 50

Page 29 of 50

513


Rev. Biophys. 2010, 39, 309–328.

514

(22) Upadhyaya, J.; Pydi, S. P.; Singh, N.; Aluko, R. E.; Chelikani, P. Bitter Taste

515

Receptor T2R1 Is Activated by Dipeptides and Tripeptides. Biochem. Biophys.

516

Res. Commun. 2010, 398 (2), 331–335.

517

(23) Marchiori, A.; Capece, L.; Giorgetti, A.; Gasparini, P.; Behrens, M.; Carloni, P.;

518

Meyerhof, W. Coarse-Grained/molecular Mechanics of the TAS2R38 Bitter

519

Taste Receptor: Experimentally-Validated Detailed Structural Prediction of

520

Agonist Binding. PLoS One 2013, 8 (5), 1–12.

521 522 523 524 525

(24) Kubota, A.; Kubo, I. Bitterness and Chemical Structure. Nature 1969, 223, 97– 99. (25) Nofre, C.; Tinti, J. M. Sweetness Reception in Man: The Multipoint Attachment Theory. Food Chem. 1996, 56 (3), 263–274. (26) Lobley,

A.;

Sadowski,

M.

I.;

Jones,

D.

T.

pGenTHREADER

and

526

pDomTHREADER: New Methods for Improved Protein Fold Recognition and

527

Superfamily Discrimination. Bioinformatics 2009, 25 (14), 1761–1767.

528

(27) SIB Swiss Institute of Bioinformatics and the Biozentrum; der Universität Basel,

529

S. SWISS-MODEL http://swissmodel.expasy.org/ (accessed Nov 26, 2015).

530

(28) Biasini, M.; Bienert, S.; Waterhouse, A.; Arnold, K.; Studer, G.; Schmidt, T.;

531

Kiefer, F.; Cassarino, T. G.; Bertoni, M.; Bordoli, L.; Schwede, T. SWISS-MODEL:



532

Modelling Protein Tertiary and Quaternary Structure Using Evolutionary

533

Information. Nucleic Acids Res. 2014, 42 (W1), W252–W258.

534

(29) Phillips, J. C.; Braun, R.; Wang, W.; Gumbart, J.; Tajkhorshid, E.; Villa, E.;

535

Chipot, C.; Skeel, R. D.; Kalé, L.; Schulten, K. Scalable Molecular Dynamics with

536

NAMD. J. Comput. Chem. 2005, 26 (16), 1781–1802.

537

(30) Vanommeslaeghe, K.; Hatcher, E.; Acharya, C.; Kundu, S.; Zhong, S.; Shim, J.;

538

Darian, E.; Guvench, O.; Lopes, P.; Vorobyov, I.; Jr, A. D. M. CHARMM General

539

Force Field: A Force Field for Drug-Like Molecules Compatible with the

540

CHARMM All-Atom Additive Biological Force Fields. J. Comput. Chem. 2010, 31

541

(4), 671–690.

542

(31)

Lovell, S. C.; Davis, I. W.; Arendall, W. B.; de Bakker, P. I. W.; Word, J. M.;

543

Prisant, M. G.; Richardson, J. S.; Richardson, D. C. Structure Validation by

544

Calpha Geometry: Phi,psi and Cbeta Deviation. Proteins 2003, 50 (3), 437–450.

545

(32) Humphrey, W.; Dalke, A.; Schulten, K. VMD: Visual Molecular Dynamics. J. Mol. Graph. 1996, 14, 33–38.

546 547 548

(33)

National Center for Biotechnology Information. The PubChem Project http://pubchem.ncbi.nlm.nih.gov/ (accessed Nov 26, 2015).

549

(34) Trott, O.; Olson, A. J. AutoDock Vina: Improving the Speed and Accuracy of

550

Docking with a New Scoring Function , Efficient Optimization , and


Page 30 of 50

Page 31 of 50


Multithreading. J. Comput. Chem. 2009, 31 (2), 455–461.

551 552

(35)

Serio, C. Di; Lio, P.; Nonis, A.; Tagliaferri, R. Computational Intelligence Methods for Bioinformatics and Biostatictic; 2014.

553 554

(36) Morris, G. M.; Huey, R.; Lindstrom, W.; Sanner, M. F.; Belew, R. K.; Goodsell, D.

555

S.; Olson, A. J. AutoDock4 and AutoDockTools4: Automated Docking with

556

Selective Receptor Flexibility. J. Comput. Chem. 2010, 31, 2967–2970.

557

(37) Brooks, C. L. Ligand Protein Database http://lpdb.chem.lsa.umich.edu/ (accessed May 30, 2015).

558 559

(38) Grosdidier, A.; Zoete, V.; Michielin, O. SwissDock, a Protein-Small Molecule

560

Docking Web Service Based on EADock DSS. Nucleic Acids Res. 2011, 39 (SUPPL.

561

2), 270–277.

562

(39) Wallace, A. C.; Laskowski, R. A.; Thornton, J. M. LIGPLOT: A Program to

563

Generate Schematic Diagrams of Protein-Ligand Interactions. Protein Eng. 1995,

564

8 (2), 127–134.

565

(40) Rezácová, P.; Borek, D.; Moy, S. F.; Joachimiak, A.; Otwinowski, Z. Crystal

566

Structure and Putative Function of Small Toprim Domain-Containing Protein

567

from Bacillus Stearothermophilus. Proteins 2008, 70 (2), 311–319.

568 569

(41)

Espinoza, M. I.; Vincken, J. P.; Sanders, M.; Castro, C.; Stieger, M.; Agosin, E. Identification, Quantification, and Sensory Characterization of Steviol



570

Glycosides from Differently Processed Stevia Rebaudiana Commercial Extracts.

571

J. Agric. Food Chem. 2014, 62 (49), 11797–11804.

572 573

(42) Loncharich, R. J.; Brooks, B. R. The Effects of Truncating Long-Range Forces on Protein Dynamics. Proteins Struct. Funct. Genet. 1989, 6 (1), 32–45.

574

(43) Sommer, B. Membrane Packing Problems: A Short Review on Computational

575

Membrane Modeling Methods and Tools. Comput. Struct. Biotechnol. J. 2013, 5

576

(6), e201302014.

577 578

(44) Adcock, S. A.; McCammon, J. A. Molecular Dynamics: Survey of Methods for Simulating the Activity of Proteins. Chem. Rev. 2006, 106 (5), 1589–1615.

579

(45) Fogh, R. H.; Boucher, W.; Vranken, W. F.; Pajon, A.; Stevens, T. J.; Bhat, T. N.;

580

Westbrook, J.; Ionides, J. M. C.; Laue, E. D. A Framework for Scientific Data

581

Modeling and Automated Software Development. Bioinformatics 2005, 21 (8),

582

1678–1684.

583

(46) Nilges, M.; Bernard, A.; Bardiaux, B.; Malliavin, T.; Habeck, M.; Rieping, W.

584

Accurate NMR Structures Through Minimization of an Extended Hybrid

585

Energy. Structure 2008, 16 (9), 1305–1312.

586

(47) Worth, C. L.; Kreuchwig, A.; Kleinau, G.; Krause, G. GPCR-SSFE: A

587

Comprehensive Database of G-Protein-Coupled Receptor Template Predictions

588

and Homology Models. BMC Bioinformatics 2011, 12 (1), 185.


Page 32 of 50

Page 33 of 50


589

(48) Sippl, M. J.; Wiederstein, M. Structural Bioinformatics A Note on Difficult

590

Structure Alignment Problems. Struct. Bioinforma. 2008, 24 (3), 426–427.

591

(49) Mayank; Vikas, J. Interaction Model of Steviol Glycosides from Stevia

592

Rebaudiana (Bertoni) with Sweet Taste Receptors: A Computational Approach.

593

Phytochemistry 2015, 116, 12–20.

594

(50) Pronin, A. N.; Tang, H.; Connor, J.; Keung, W. Identification of Ligands for Two Human Bitter T2R Receptors. Chem. Senses 2004, 29 (7), 583–593.

595 596

(51)

Floriano, W. B.; Hall, S.; Vaidehi, N.; Kim, U.; Drayna, D.; Goddard, W. a.

597

Modeling the Human PTC Bitter-Taste Receptor Interactions with Bitter

598

Tastants. J. Mol. Model. 2006, 12 (6), 931–941.

599

(52) Meyerhof, W.; Batram, C.; Kuhn, C.; Brockhoff, A.; Chudoba, E.; Bufe, B.;

600

Appendino, G.; Behrens, M. The Molecular Receptive Ranges of Human TAS2R

601

Bitter Taste Receptors. Chem. Senses 2010, 35 (2), 157–170.

602

(53)

Singla, R.; Jaitak, V. Synthesis of Rebaudioside A from Stevioside and Their

603

Interaction Model with hTAS2R4 Bitter Taste Receptor. Phytochemistry 2016,

604

125, 106–111.

605 606 607

(54) Shallenberger, R. S.; Acre, T. E. Molecular Theory of Sweet Taste. Nature 1967, 216, 480–482. (55) Birch, G. G.; Mylvaganam, A. R. Evidence for the Proximity of Sweet and Bitter



608

Receptor Sites. Nature 1976, 260 (5552), 632–634.

609

(56) Singh, N.; Pydi, S. P.; Upadhyaya, J.; Chelikani, P. Structural Basis of Activation

610

of Bitter Taste Receptor T2R1 and Comparison with Class A G-Protein-Coupled

611

Receptors (GPCRs). J. Biol. Chem. 2011, 286 (41), 36032–36041.

612

(57) Tan, J.; Abrol, R.; Trzaskowski, B.; Goddard, W. A. 3D Structure Prediction of

613

TAS2R38 Bitter Receptors Bound to Agonists Phenylthiocarbamide (PTC) and

614

6-N-Propylthiouracil (PROP). J. Chem. Inf. Model. 2012, 52 (7), 1875–1885.

615 616 617 618 619 620 621 622 623 624 625 626 627


Page 34 of 50

Page 35 of 50


628

FIGURE CAPTIONS

629

Figure 1. Chemical structure of steviol glycosides (SG) 1 - 10 (R1 and R2 referred to in

630

Table 1).

631

Figure 2. Comparative models of hT2R4 (a), hT2R14 (b) receptors, and hT2R1 (c) used

632

as a negative control, generated by SWISS-MODEL.

633

Figure 3. Relationship between relative bitterness and binding free energy of SG with

634

bitter taste receptors. Steviol glycosides (SG) were inserted into the binding sites of

635

hT2R4 (a), hT2R14 (b) receptors, and hT2R1 (c), used as a negative control. Data

636

correspond to: (a) stevioside; (b) reb A; (c) reb B; (d) reb C; (e) reb D; (f ) steviolbioside;

637

(g) dulcoside A; and (h) rubusoside. The relative bitterness data for SG were obtained

638

from the study by Hellfritsch et al 3.

639

Figure 4. Two-dimensional representation of hT2R4 (a) and hT2R14 (b) amino acid

640

sequence. The potential binding sites of SG are indicated by red circles.

641

Figure 5. Snapshots of VMD showing the potential binding site and pose of docked

642

stevioside on hT2R4 (a) and hT2R14 (b).

643

Figure 6. Binding pattern of SG with the hT2R4 bitter taste receptor. The amino acids

644

responsible for the hydrogen bonds and hydrophobic interactions are represented by

645

three letter codes in green and black, respectively. Carbon, oxygen and nitrogen atoms

646

are indicated by closed black, red and blue circles, respectively.



647

Figure 7. Binding pattern of SG with the hT2R14 bitter taste receptor. The amino acids

648

responsible for the hydrogen bonds and hydrophobic interactions are represented by

649

three letter codes in green and black, respectively. Carbon, oxygen and nitrogen atoms

650

are indicated by closed black, red and blue circles, respectively.

651

Figure 8. Snapshots from Visual Molecular Dynamics showing the stevioside (c and

652

d) and rebaudioside A (e and f) poses on hT2R4 (red color) and hT2R14 (green color)

653

bitter taste receptors.

654

Figure 9. Root mean square fluctuation value at Cα carbon of hT2R4 (a) and hT2R14

655

(b) residues bound to stevioside (black line) and rebaudioside A (red line).

656

Figure 10. Evolution of the pulling force exerted on stevioside (black line) and

657

rebaudioside A (red line) during unbinding from hT2R4 and hT2R14 pockets.

658

Figure 11. Images of the stevioside (a and b) and rebaudioside A (c and d) unbinding

659

process from the hT2R4 (red color) and hT2R14 (green color) binding sites during 8

660

ps simulations.


Page 36 of 50

Page 37 of 50


Table 1. Chemical structure of steviol glycosidesa Compound

R1

R2

Stevioside (1)

β − glc −

β − glc − β − glc −

Rebaudioside A (2)

β − glc −

(β − glc) − β − glc −

Rebaudioside B (3)

H−

(β − glc) − β − glc −

Rebaudioside C (4)

β − glc −

(β − glc) − α − rha −

Rebaudioside D (5)


(β − glc) − β − glc −

Rebaudioside E (6)



Rebaudioside F (7)

β − glc −

(β − glc) − α − xyl −

Steviolbioside (8)

H−


Dulcoside A (9)

β − glc −

β − glc − α − rha −

Rubusoside (10)

β − glc −

β − glc −

aR 1

and R2 referred to in Figure 1. Glc, D-glucopyranosyl; rha, Lrhamnopyranosyl; xyl, D-xylopyranosyl



Page 38 of 50

Table 2. Binding free energy and binding sites of steviol glycosides (SG) for hT2R4 and hT2R14 bitter taste receptors. Compounda

∆Gbinding for hT2R4 (Kcal/mol)

∆Gbinding for hT2R14 (Kcal/mol)

Binding site in hT2R4

Binding site in hT2R14

1

-6.7

-6.6

Leu 61, Asn 65, Thr 66, Phe 69, Asn 73, Glu 75, Val 78, Phe 84

-6.3

-6.5

Asn 65, Thr 66, Leu 80, Phe 84

3

-6.3

-6.3

4

-6.3

-6.6

Asn 65, Thr 66, Phe 84, Ser 77 Asn 65, Thr 66, Leu 80, Phe 84

5

-5.3

-5.2

6

-6.3

-7.2

7

-6.5

-6.5

8

-6.0

-6.8

9

-6.6

-7.1

Ser 155, Ile 156, Asn 157, Ser 167, Asp 168, Ser 170, Arg 174, Tyr 240, Ser 246, Phe 247, Glu 258 Ser 167, Asp 168, Ser 170, Tyr 240, Ser 246, Phe 247 Ser 155, Ile 156, Lys 163, Asp 168, Arg 174 Ser 167, Asp 168, Ser 170, Tyr 240, Ser 246, Phe 247 Arg 160, Ser 169, Thr 173 Ser 155, Asn 157, Gly 158, Ser 246, Glu 258 Ser 167, Asp 168, Ser 170, Tyr 240, Ser 246, Phe 247 Asn 86, Ile 152, Ile 156, Arg 160, Asp 168, Arg 174 Asn 157, Ser 167, Asp 168, Ser 169, Asn 171, Tyr 240 Ser 155, Asn 157, Arg 160, Ser 167, Asp 168, Ser 169, Tyr 240

2

Val 78, Leu 80, Ser 81 Thr 66, Glu 75, Val 78, Leu 80 Asn 65, Thr 66, Asn 73, Glu 75, Ser 77, Val 78 Thr 66, Glu 75, Ser 77

Relative bitterness at 1 mMb

xlogPc

1.4

-1.2

1.3

-2.8

1.1

-1.0

1.4

-2.3

0.6

-4.4

-

-2.8

-

-2.7

0.9

0.6

Asn 65, Thr 66, 1.7 -0.7 Asn 73, Glu 75, Val 78, Leu 80 10 -7.0 -7.6 Leu 61, Phe 69, 2.7 0.4 Asn 65, Thr 66, Asn 73, Ser 77, Val 78, Phe 83, Phe 84 aCompound number refers to structures given in Figure 1 and Table 1. bRelative bitterness values extracted from Hellfritch et al. (2012) 9. c Values extracted from the PubChem database 33


Page 39 of 50


Figure 1



Figure 2


Page 40 of 50

Page 41 of 50


Figure 3



Figure 4


Page 42 of 50

Page 43 of 50


Figure 5



Figure 6


Page 44 of 50

Page 45 of 50


Figure 7



Figure 8


Page 46 of 50

Page 47 of 50


Figure 9



Figure 10


Page 48 of 50

Page 49 of 50


Figure 11



TOC graphic


Page 50 of 50

Docking and Molecular Dynamics of Steviol GlycosideâHuman Bitter

Recommend Documents

Docking and Molecular Dynamics of Steviol GlycosideâHuman Bitter