Effect of Saliva Esterase Activity on Ester

1 EFFECT OF SALIVA ESTERASE ACTIVITY ON ESTER

2 SOLUTIONS AND POSSIBLE CONSEQUENCES FOR THE IN-

3 MOUTH ESTER RELEASE DURING WINE INTAKE

5 María-Pérez-Jiménez, Nuria Rocha-Alcubilla, Maria Ángeles Pozo-Bayón*

7 Instituto de Investigación en Ciencias de la Alimentación (CIAL) CSIC-UAM,

8 C/Nicolás Cabrera, 20049, Madrid, Spain.

*Corresponding author: [email protected] Tel: 34 91 0017961; Fax: 34 910017905

11 Short title: Effect of saliva esterase activity on wine esters

13 Abstract:

14 The aim of the present study was to investigate the role of saliva esterase activity on

15 carboxylic esters typically associated with pleasant and fruity aromas in wine. For this,

16 ex-vivo experiments using the same fresh and inactivated (without enzymatic activity)

17 human saliva with a mixture of carboxylic esters with different aliphatic chain length

18 (ethyl butanoate, ethyl pentanoate, ethyl hexanoate, ethyl octanoate, ethyl decanoate and

19 isoamyl acetate) were prepared. Liquid-liquid extraction with dichloromethane and GC-

20 MS analysis were applied to the saliva systems in order to determine the reduction in

21 ester content and the formation of their corresponding metabolic products (carboxylic

22 acids) in the saliva systems before and after incubation at 37ºC. In addition, to check if

23 there was a relationship between the susceptibility of esters to saliva hydrolysis and the

24 amount of -in mouth ester release during wine intake, the remaining oral amount of each

25 ester was determined by comparing the intraoral amount immediately after spitting out

26 the wine and four minutes later. Ex -vivo experiments showed ester degradation by

27 saliva esterase enzymes mainly acted on long chain esters (ethyl octanoate and ethyl

28 decanoate), which gave rise to the formation of their corresponding carboxylic acids.

29 Nonetheless, in spite of their higher susceptibility to saliva enzymes, -in vivo

30 experiments showed that long chain carboxylic esters remained in the oral cavity long

31 after swallowing. This confirmed that ester hydrophobicity is closely related to the –in

32 mouth temporal release of these odorants and therefore, behind wine aroma persistence.

34 Key words: saliva esterase activity, carboxylic esters, wine, in-mouth aroma release,

35 aroma persistence

37 Practical applications

38 In wines, esters represent a group of aromatic compounds of great interest since they are

39 linked to pleasant fruity aroma nuances. Today wine consumers are demanding fresh

40 and long persistent fruity aromatic wines. The present research contributes to better

41 understanding the relationship between ester content in the wine and oral aroma release

42 experienced during wine tasting, considering the changes in these compounds during

43 oral processing. This is a necessary step when trying to unravel the factors involved in

44 wine aroma perception and in consumer preferences, and it represents a necessary

45 knowledge in promoting winemaking practices (e.g. the use of selected

46 microorganisms) for improving the type and amount of these aroma compounds in the

47 wine.

52 1. INTRODUCTION

53 Ethyl esters of volatile acids and higher alcohols acetates are typical wine aroma

54 compounds produced during alcoholic fermentation. Different studies have associated

55 these compounds to the pleasant and fruity character of many types of wines (Escudero,

56 Campo, Fariña, Cacho, & Ferreira, 2007; Francis & Newton, 2005; Rapp & Mandery,

57 1986). In addition, through perceptual interactions, these compounds enhance the

58 perception of fruity aroma even at concentrations below their individual olfactory

59 thresholds (Lytra, Tempere, Le Floch, de Revel, & Barbe, 2013) The fact that most

60 esters are present in concentrations around their threshold values, implies that minor

61 concentration changes might have a dramatic effect on wine flavor. Therefore, a better

62 understanding of ester hydrolysis/synthesis of these compounds in the wine matrix is

63 essential to aid the winemaker in achieving the best possible winemaking outcome

64 (Sumby, Grbin, & Jiranek, 2010).

65 Aditionally, besides their changes in the wine itself, it is interesting to know what

66 happens with these compounds during oral processing and what type of chemical or

67 biochemical transformations might occur before reaching, via the retronasal route, the

68 olfactory receptors during wine tasting. Recently, it was shown, that some wine esters

69 can be adsorbed into the oral mucosa in a relatively large amount (23% to 44% for

70 isoamyl acetate and ethyl hexanoate) after 30 seconds of oral exposure to wine

71 (Esteban-Fernández, Rocha-Alcubilla, Muñoz-González, Moreno-Arribas, & Pozo-

72 Bayón, 2016). However, they are quickly released from oral mucosa and they do not

73 remain in the oral cavity for a long time (Esteban-Fernández, et al., 2016), thus, having

74 a moderated contribution to aroma persistence. One possible explanation could be the

75 weak interactions between these compounds and saliva proteins from the saliva pellicle

76 (Ployon, Morzel, & Canon, 2017). Although there are not -in vivo studies confirming

77 this, results from -in vitro studies have shown that hydrophobic interactions can be

78 behind these interactions (Pagès-Hélary, Andriot, Guichard, & Canon, 2014). The

79 presence of other wine matrix compounds such as polyphenols, have also been proven

80 to have a role on the adsorption capacity of some esters to oral mucosa (Esteban-

81 Fernández, Muñoz-González, Jiménez-Girón, Pérez-Jiménez, & Pozo-Bayón, 2018;

82 Esteban-Fernández, et al., 2016). Another plausible explanation could be the oral

83 degradation of esters by some saliva enzymes.

84 Salivary aroma converting enzymes might originate from the salivary glands, oral

85 tissues or even microorganisms (Ployon, et al., 2017). The effect of saliva enzymes on

86 the metabolism of different types of aroma compounds such as thiols, aldehydes,

87 ketones and esters has been shown in different works (Buettner, 2002a, 2002b; Muñoz-

88 González, Feron, Brulé, & Canon, 2018; Pagès-Hélary, et al., 2014). In the case of

89 esters, hydrolysis can be seen as the most probable mechanism as many esterolytic

90 enzymes can be found in human saliva (Buettner, 2002b). Apart from carboxylesterases,

91 other enzymes such as acetylcholinesterase, trypsin, chymotrypsin, carbonic anhydrase,

92 and pseudocholinesterase can exhibit esterase activity (Krisch, 1971; Ployon, et al.,

93 2017). The esterolytic activity of saliva was investigated in a previous work (Buettner,

94 2002b). In this work, a reduction in the content of some carboxylic esters (ethyl

95 butanoate, ethyl hexanoate, ethyl octanoate) was observed after ten minutes of

96 incubation in the presence of human saliva, which did not happen when saliva was

97 thermally inactivated. Nonetheless, the corresponding degradation products (carboxylic

98 acids) were not found in the saliva, which the authors explained by the low aroma

99 concentration used in the experiments. In another work, using static headspace

100 conditions, Pages-Hèlary and co-workers (Pagès-Hélary, et al., 2014), also investigated

101 the effect of human saliva on the concentration of a series of carboxylic esters (from

102 ethyl butanoate to ethyl heptanoate). They observed a decrease in the headspace

103 concentrations of all of them, although it was not clear whether ester reduction was due

104 to interaction with salivary proteins or to saliva esterase activity. Moreover, the

105 consequence of this effect for -in vivo ester release during food consumption remains

106 uncertain.

107 Because of the outstanding role of esters for many fermented beverages such as wine,

108 and the lack of a clear relationship between the role of esterase enzymes and their effect

109 on wine aroma, this work has two main objectives. Firstly, to evaluate the esterolytic

110 capacity of human saliva on wine aromatic esters associated to fruity aromas through -

111 ex vivo experiments, in which five ethyl esters of volatile acids (ethyl butanoate,

112 pentanoate, hexanoate, octanoate and decanoate) and one higher alcohol acetate

113 (isoamyl acetate), were incubated (at 37ºC) or not in enzymatic (without any treatment)

114 and non-enzymatic pooled saliva (after enzymatic inhibition with CaCl2) from ten

115 individuals. The reduction in esters content and the formation of the corresponding

116 metabolic products (carboxylic acids) was followed by liquid-liquid extraction with

117 dichloromethane and GC-MS analysis. The second objective was to check the -in mouth

118 ester release after the exposure of the oral cavity to an aromatized wine with the same

119 six esters previously tested in the ex-vivo experiments. The rationale for this objective,

120 was to probe if the higher or lower susceptibility of some esters to saliva esterase, might

121 also affect in-mouth ester release. In this case, oral aroma release was monitored

122 immediately after rinsing the mouth and spitting off the wine and then, four minutes

123 later, using the same ten volunteers who donated the saliva samples for the -ex vivo

124 experiments.

125 MATERIALS AND METHODS

126 1.1.Saliva samples

127 Unstimulated saliva samples were collected from 10 healthy subjects (4 men and 6

128 women), aged between 21 and 36 years old. All subjects were nonsmokers and had not

129 taken any antibiotics or other medical treatments during at least three months prior to

130 the sampling. Participants were asked not to consume any food or drink two hours

131 before the saliva was collected. They let the saliva naturally accumulate in the mouth

132 and then spat it directly into a collection tube. This fresh saliva collected from all the

133 individuals was pooled together and centrifuged at 2600 g for 15 min at 4 ºC. This

134 mixture of saliva had a protein concentration determined using the Bradford protein

135 assay (Pierce Thermo Scientific, Illinois, USA) of 1,04 mg/mL and a pH of 7,13.

136 From this, half saliva (enzymatic saliva: ES) was separated and the other half was

137 enzymatically inactivated (Non-enzymatic saliva: NE) by adding 33 mg CaCl2 (Panreac,

138 Barcelona, Spain) / mL of saliva. Therefore, two saliva types (whole and inactivated

139 saliva) were employed for this study. Saliva samples were aliquoted and stored at -80 ºC

140 until use. All saliva experiments were performed one week after saliva was collected.

141 1.2.Ex-vivo experiment of ester degradation by saliva

142 From the same mixture of saliva, eight types of saliva systems were prepared (table 1).

143 Four of them corresponded to the enzymatic saliva systems (ES) (without enzymatic

144 inactivation). Two ES systems were aromatized and one was incubated at 37 ºC for 120

145 minutes (ES-120), while the other was not (ES-0). The other two ES systems were not

146 aromatized (control saliva), and one was incubated in the same conditions described

147 above (ES-C120), while the other was not (ES-C0). The other four saliva systems

148 corresponded to non-enzymatic saliva (NES). Following the same experimental

149 procedure already explained, two of them were aromatized and one was incubated at 37

150 ºC for 120 minutes (NES-120), while the other was not (NES-0). In addition, as shown

151 in table 1, the other two NES samples without aroma were also prepared (NES

152 controls), and one was incubated at 37 ºC 120 min (NES-C120) while the other was not

153 (NES-C0).

154 [Table 1 here]

155

156 For the aromatization of the saliva systems (ES-0, ES-120, NES-0 and NES-120), six

157 independent stock solution of six esters associated to fruity aromas in wines (Sumby, et

158 al., 2010) and with different physicochemical properties were employed (table 2). The

159 tested esters: ethyl butanoate (105-54-4) from Aldrich (Steinheim, Germany), ethyl

160 pentanoate (539-82-2), ethyl octanoate (106-32-1) and isoamyl acetate (123-92-2) from

161 Fluka (Buchs, Switzerland), ethyl hexanoate (123-66-0) and ethyl decanoate (110-38-3)

162 from Merck (Darmstadt, Germany) were prepared in absolute ethanol (Merck,

163 Darmstadt, Germany) at 10000 mg/L. From this stock, a second solution of each aroma

164 compound in absolute ethanol (1000 mg/L) was also prepared. 100 µL of this solution

165 was added to the saliva systems at a final concentration of 50 mg/L. This concentration

166 was below the solubility values estimated for all the esters used in this study except for

167 ethyl decanoate. In the case of this compound, the selected concentration might have

168 been higher than its estimated solubility value in water. Nonetheless, it is important to

169 bear in mind that these values are mostly estimated in water at 25 ºC. The conditions

170 used in this work, in which esters are firstly completely solubilized in absolute ethanol,

171 and secondly added to the saliva (water) system and even incubated at 37 ºC, might

172 have also increased its solubility. To avoid aroma interactions only one aroma was

173 assayed in each saliva system.

174 [Table 2 here]

175

176 1.3.Analysis of esters and carboxylic acids in the saliva systems by GC-MS

177

178 For the extraction of aroma compounds, two mL of the saliva systems were spiked with

179 100 µL of the internal standard, methyl nonanoate (1 mg/L) and extracted twice with 1

180 mL of dichloromethane (RCI Labscan Ltd, Bangkok, Thailand) then, they were

181 submitted to ultra-sonication (15 min) in an ice bath and finally centrifuged (5000 rpm,

182 4 ºC, 15 min) to separate the two phases. The combined organic extracts were dried

183 over anhydrous Na2SO4, then concentrated to a total volume of 500 µl and subsequently

184 analyzed by gas chromatography-mass spectrometry (GC-MS).

185 For the GC-MS analysis, 2 µL of the concentrated extract from saliva were injected in

186 split mode in the injector port of the GC coupled to an Agilent 5973N Mass Detector.

187 GC-MS conditions are described in a previous work (Esteban-Fernandez et al., 2016).

188 Briefly, the injection temperature was set at 250ºC and volatile compounds were

189 separated on a DB-Wax polar capillary column (60 m × 0.25 mm i.d. × 0.50 μm film

190 thickness) from Agilent (J&W Scientific, Folsom, USA). Helium was the carrier gas at

191 a flow rate of 1 mL/min. The oven temperature was initially held at 50 ºC for 2 min,

192 increased at 8 ºC/min to 100 ºC and held for 1 min, increased 1ºC/min to 120ºC held for

193 1 min, increased 2ºC/min to 164ºC held for 1 min and finally increased at 8ºC/min till

194 240 ºC and held for 5 min. For the MS system, the temperature of the transfer line,

195 quadrupole and ion source was 270, 150 and 230 ºC respectively. Electron impact mass

196 spectra were recorded at 70 eV ionization voltages and the ionization current was 10

197 µA. The acquisitions were performed in Scan (from 35 to 350 amu) and SIM modes,

198 looking for the characteristic ions of the compounds of interest. The identification of

199 compounds was based on the comparison of retention times and mass spectra with those

200 from NIST 2.0 database and with the corresponding reference compounds analyzed in

201 the same conditions as the samples. Relative peak areas (RPAs) were obtained by

202 calculating the relative peak area in relation to that of the internal standard (methyl

203 nonanoate), and they were used to compare among different saliva systems.

204

205 1.4.Intra-oral SPME sampling of odorants after wine rinsing

206

207 2.4.1 Wine

208 A rosé wine with 10 % ethanol, pH of 2, 98 and total polyphenol content of 391mg Eq

209 gallic acid/L, was used for this study. The endogenous presence of esters was previously

210 checked by applying a headspace SPME sampling procedure (Rodríguez‐Bencomo et

211 al., 2011), confirming a very low original ester concentration: ethyl butanoate (89,3

212 µg/L), ethyl octanoate (3,8 µg/L), ethyl decanoate (9,4 µg/L), and isoamyl acetate (1,7

213 µg/L).

214 To reinforce the aroma profile, this wine was aromatized with a mixture of six food

215 grade esters (Sigma-Aldrich, Steinheim, Germany): ethyl butanoate, ethyl pentanoate,

216 ethyl hexanoate, ethyl octanoate, ethyl decanoate and isoamyl acetate. For the

217 aromatization, six independent aroma stock solutions in food grade ethanol (Panreac

218 Química S.A., Barcelona, Spain) were prepared and from there, each aroma compound

219 was added to the wine immediately before the assay to obtain a final concentration of 2

220 mg/L, which is slightly higher than the concentration of these compounds naturally

221 found in wine, allowing us to increase the sensitivity of the oral aroma monitoring.

222 2.4.2. Intra-oral aroma monitoring

223

224 Fifteen mL of aromatized wine was taken into the oral cavity for 30 s and then,

225 expectorated. During rinsing, special care was taken to keep the lips closed, not to

226 swallow and not to open the velum – tongue border prior to expectoration. Immediately

227 after expectoration (t=0 min), and four minutes later (t= 4 min), a DVB/CAR/PDMS

228 (Divinylbenzene/Carboxen/Polydimethylsiloxane 50/30 µm thickness -2 cm length-)

229 coated SPME fiber (Supelco, Bellefonte, PA) contained in a manual holder (Supelco)

230 was placed into the oral cavity of the panelist. This methodology was previously

231 validated (Esteban-Fernández, et al., 2018; Esteban-Fernández, et al., 2016). Briefly, the

232 SPME fiber was placed inside a plastic tube firmly held by the lips of the volunteer and

233 during the in-mouth extraction, the lips were kept closed around the plastic tube

234 containing the SPME fiber. During the extraction, the fiber does not touch the mouth

235 surface and the aroma extraction is done in the free space of the mouth. Swallowing was

236 not allowed throughout this process. After 2 min of extraction, the fiber was removed

237 from the oral cavity, and immediately placed into the split/splitless injector. Four

238 minutes after the first expectoration, a second sampling of the oral cavity was carried

239 out. To do so, a new SPME fiber was used to perform the second oral aroma extraction

240 (t=4 min) in the same conditions previously described. The two SPME fibers were

241 selected before starting the experiment considering their similarity in volatile recovery

242 rates, considering that differences between them could not be higher than 5%. Each

243 assay was performed three times by the ten panelists

244 The SPME fiber with the oral aroma extract from the second oral sampling (t=4

245 min) was immediately desorbed in the injector of the GC system (Agilent 6890N)

246 (Agilent Technologies, California, USA) in splitless mode for 1.5 min at 250 ºC. The

247 SPME fiber with the breath extract corresponding to t=0, was stored in the fridge (4ºC)

248 in a sealed glass tube until the first GC run (corresponding to t=4 min) finished.

249 Preliminary experiments were performed in order to ensure that there were no

250 significant losses of aroma during the storage of the fiber, which was about 1 hour. The

251 chromatographic and MS conditions for the breath analysis were similar to those

252 explained before for the analysis of aroma compounds in the saliva systems. In the case

253 of intraoral-SPME data, since no internal standard was used, absolute peak areas

254 (APAs) were obtained to express aroma release. APAs were used to compare the % of

255 ester remaining in the oral cavity four minutes after wine expectoration considering the

256 amount of ester released at t=0 as 100%. The repeatability of the method expressed as

257 relative standard deviation (% RSD) for each ester considering the average values

258 obtained with the 10 volunteers testing the same wine was: ethyl butanoate: 8,38 %,

259 Isoamyl acetate: 3,61 %, ethyl pentanoate: 8,76%, ethyl hexanoate : 6,64 %, ethyl

260 octanoate: 6,81%, ethyl decanoate: 6,69 %).

261 2. RESULTS AND DISCUSSION

262

263 2.1.Ester degradation by saliva

264

265 First of all, the GC-MS analysis from the saliva extracts did not show the presence of

266 any of the tested esters in the control saliva systems without aroma (NES-C or ES-C

267 after 0 or 120 minutes of incubation). Nonetheless, some traces of hexanoic and

268 decanoic were found in all the saliva systems without incubation. Different primary

269 aliphatic acid volatiles have been already described in the saliva volatolome (de Lacy

270 Costello et al., 2014), thus these compounds could naturally occur in the saliva samples.

271 Figure 1 shows an example of the chromatograms obtained for the control ES system

272 without added aroma (Figure 1a) and the two ES systems with aroma and no incubation

273 (Figure 1b) and the same system after 120 min of incubation (Figure 1c).

274

275 [Figure 1 here]

276

277 To check the effect of the saliva system (inactivated or not) and the incubation time (0

278 or 120 minutes) on the amount of ester recovered, results corresponding to relative peak

279 areas of the six tested esters in each saliva system were submitted to a two-way

280 ANOVA. Results confirmed that both factors significantly affected the recovery of

281 esters. As can be seen in table 3, significant differences in ester recovery between NES

282 and ES were found for all compounds except isoamyl acetate and ethyl hexanoate.

283 Small esters (ethyl butanoate and ethyl pentanoate) were slightly but significantly less

284 recovered in ES. Besides the effect of enzymatic degradation of these compounds by

285 saliva enzymes, the higher recovery of small esters in the NES could be due to the

286 modification of the ionic strength in the saliva system by adding a salt like CaCl2. Many

287 salts have been related with a “salting out” effect enhancing the extraction of aroma

288 compounds to the organic phase and therefore the recovery of these compounds

289 (Leggett, Jenkins, & Miyares, 1990). On the contrary, the long chain esters, ethyl

290 octanoate and ethyl decanoate, were more recovered in ES than in NES. In the case of

291 ethyl hexanoate, the recovery was almost identical in both saliva systems, which did not

292 prove an effect of CaCl2 in the extraction of this compound in the saliva systems. Table

293 3 also shows that except ethyl decanoate, the recovery of all tested esters was lower

294 after 120 minutes of incubation time.

295 [Table 3 here]

296

297 In order to better understand these results, the impact of incubation time in the amount

298 of ester recovered (expressed as relative peak areas) considering separately the two

299 types of saliva systems (NES and ES) is shown in Figures 2a and 2b.

300

301 [Figure 2 here]

302

303 As it can be seen, a significant reduction in the amount of all esters after 120 minutes of

304 incubation time was observed in the ES system. This reduction was higher in the case of

305 the long chain acids ethyl octanoate and ethyl decanote (14-26 %) compared to the

306 small esters, ethyl butanoate, isoamyl acetate and ethyl pentanoate (9-10%) (figure 2a).

307 The reduction in ester recovery could be due to enzymatic degradation by saliva

308 enzymes, however this reduction was also found in saliva without enzymatic activity

309 (figure 2 b). Here a reduction in ethyl butanoate, isoamyl acetate, ethyl pentanoate and

310 ethyl hexanoate was also observed, in spite that it had already been found that esterase

311 activity is inhibited by CaCl2 (data not shown). However, in these saliva systems, a

312 significant increase of the two largest esters, ethyl octanoate and ethyl decanoate was

313 observed after 120 minutes of incubation at 37 ºC (figure 2b). In NES, the observed

314 reduction in ester recovery during incubation, could be due to hydrophobic interactions

315 of aroma compounds with saliva proteins as previously reported (Pagès-Hélary, et al.,

316 2014). The long incubation time (120 min) might have favored these interactions. In

317 addition, previous works have also suggested that some salts might change the

318 conformational state of saliva proteins (such as mucin) and therefore, the interactions

319 between saliva proteins and esters (Friel & Taylor, 2001), which could have also

320 consequences on the recovery of esters from the saliva systems. Thus, the addition of

321 CaCl2 might have induced conformational changes in saliva proteins, which might have

322 affected aroma-protein interactions, and especially, the interactions with the most

323 hydrophobic esters. This effect should be more evident at longer incubation times, and

324 could be the reason for the higher recovery of ethyl octanoate and ethyl decanoate found

325 in the NES after 120 minutes of incubation time. However, another hypothesis, such as

326 the formation of these compounds by transesterification reactions cannot be discarded

327 and it has already been suggested (Buettner, 2002b). Future works will be necessary to

328 validate this hypothesis.

329

330 Even though a significant reduction in all the esters after 120 minutes was determined in

331 the ES system, it was uncertain if the esterolytic activity of saliva could be the reason.

332 The fact that in the NES system, a similar or even higher reduction in ester recovery

333 after 120 minutes of incubation was found, it did not prove the involvement of saliva

334 esterase activity on these results. Previously, hydrolysis had been assumed to be the

335 most probable mechanism to explain reduction of esters in saliva, since many esterolytic

336 enzymes can be found in human saliva (Ployon, et al., 2017). In this case, carboxylic

337 acids should be the metabolic degradation products. Then, a careful analysis of the

338 metabolic degradation products was performed in each saliva system. These results are

339 shown in Figure 3. From the six acids analyzed (butanoic, propanoic, pentanoic,

340 hexanoic, octanoic and decanoic acids), only hexanoic, octanoic and decanoic acids

341 were in sufficient amounts in the saliva systems to be unambiguously identified. As it

342 can be seen in figure 3, the three of them were found in the ES system after 120 minutes

343 of incubation. Also, a small amount of hexanoic and decanoic acids was also detected in

344 the ES without incubation, but their concentration significantly increased after

345 incubation. Contrarily, in the NES saliva systems, only decanoic acid was present at

346 very low concentrations and it did not increase, but in fact it decreased after 120

347 minutes, whereas the concentration of this acid significantly increased in the ES system.

348 These results clearly show the presence of higher amounts of metabolic degradation

349 products from some carboxylic esters, such as ethyl hexanoate, ethyl octanoate and

350 ethyl decanoate only in the enzymatic saliva samples. These acids were not generated in

351 the non-enzymatic saliva systems. As far as the authors know, this is the first time that

352 the production of these three carboxylic acids from ester degradation has been

353 confirmed, comparing in the same experiment saliva enzymatically inhibited or not.

354 These results also show a preference of saliva esterase by long chain carboxylic esters

355 which confirmed previous results (Buettner, 2002b; Pagès-Hélary, et al., 2014).

356

357 [Figure 3 here]

358

359 2.2.In-mouth ester release after wine rinsing

360 The above results confirmed the hydrolysis of esters by saliva enzymes using an ex-vivo

361 approach. If esters, and mainly long chain esters could be degraded by saliva, this could

362 affect their -in mouth release. To check this, the oral release of the 6 target esters was

363 monitored in 10 volunteers immediately after rinsing and spitting off a wine, and then,

364 four minutes later by using intra-oral SPME (Esteban-Fernández, et al., 2018; Esteban-

365 Fernández, et al., 2016).Aroma release values (absolute peak areas) obtained for each

366 volunteer and sampling point (t=0 min and t=4 min) were averaged in order to see the

367 effect of monitoring time on each ester release independently of individual variations. A

368 significant decrease (p<0.05) in ester release from the first oral sampling (immediately

369 after wine expectoration) to the second one (four minutes after the first wine

370 expectoration) was observed. However, this decrease was different depending on ester

371 type. Figure 4 shows these differences. Here, the remaining amount of each ester still

372 present in the oral cavity four minutes after spitting off the wine was calculated by

373 comparing ester release at t=0 and t=4, and considering aroma release at t=0 as 100%.

374 As it can be seen, the remaining oral amount of the smallest esters, like ethyl butanoate,

375 isoamyl acetate and ethyl pentanoate was very low four minutes after spit off. Above

376 9% of the initial amount of ethyl hexanoate was still present in the oral cavity four

377 minutes after wine expectoration. However, the largest differences were found for the

378 two long chain esters ethyl octanoate and ethyl decanote. These compounds were still

379 released in the oral cavity in a relatively large amount (27 % and 65 % compared to the

380 release determined at t=0 minutes) four minutes after wine expectoration. These

381 differences on temporal aroma release could be related to differences in the long lasting

382 aroma perception of the corresponding aromatic notes associated to these chemical

383 compounds, which is usually called aroma persistence (Buffo, Rapp, Krick, &

384 Reineccius, 2005; Linforth & Taylor, 2000; Ployon, et al., 2017). Differences on

385 temporal release depending on ester type could be caused, not only by saliva

386 degradation, but also by interaction of esters with saliva proteins. The fact that in spite

387 of their higher susceptibility to saliva degradation, long chain esters remained in large

388 proportion in the oral cavity supports the hypothesis that differences in -in mouth ester

389 release are largely determined by the strong interaction of these aroma compounds to

390 saliva proteins. Previous -in vitro studies have proven that retention of esters by saliva

391 mucin increases as a function of the aliphatic chain length, thus suggesting the

392 involvement of hydrophobic interactions (Pagès-Hélary, et al., 2014). The conclusion of

393 this work is in agreement with the linear relationship found between the remaining

394 amounts of esters in the oral cavity four minutes after wine expectoration and the

395 hydrophobicity of the aroma compound (characterized by the log P value) (Voilley &

396 Souchon, 2006), as it is shown in Figure 5. A direct relationship between compound

397 hydrophobicity and its presence in the oral cavity four minutes after wine expectoration

398 was found. These results can imply higher aroma persistence of the corresponding

399 aromatic notes associated to these chemical odorants (Table 2), which will be necessary

400 to confirm in sensory studies.

401 In spite of these results, the role of saliva esterase cannot be ignored as it might have

402 important sensory consequences. The formation of new odorant molecules (carboxylic

403 acids) with different aromatic notes and odor thresholds, together with the reduction of

404 esters (associated to pleasant fruity notes) might alter the intensity and perception of

405 wine aroma. In addition, saliva esterase activity could be very different among

406 individuals giving rise to differences in aroma perception. In addition, the specific wine

407 chemical composition (a relatively large ethanol content, acidic pH) and the presence of

408 certain compounds such as polyphenols, might also affect the esterolytic saliva activity..

409 None of these factors have been considered in this study, and should be the objective of

410 future works.

411

412

413 [Figure 4 here]

414 [Figure 5 here]

415

416 3. CONCLUSIONS

417 The incubation (120 minutes at 37 ºC) of human saliva spiked with typical wine

418 carboxylic esters produces a reduction in ester content and the formation of the

419 corresponding metabolic degradation products (carboxylic acids). Although a reduction

420 in ester content has also been proven in the same saliva without enzymatic activity, no

421 carboxylic acid formation have been found in the latter, confirming a saliva esterolytic

422 activity on typical wine carboxylic esters. This activity has shown preference for long

423 carboxylic esters (ethyl octanoate, ethyl decanoate) over small ones (ethyl butanoate,

424 ethyl pentanoate, ethyl hexanote). However, in spite of the higher susceptibility of long

425 chain esters for esterase enzymes, -in vivo results showed that long chain esters

426 remained in the oral cavity longer compared to the smallest ones. These results

427 confirmed that ester hydrophobicity is the main driving force governing -in mouth ester

428 release, and likely, the long lasting perception of these aroma compounds during wine

429 consumption. Nonetheless, the role of saliva esterase on ester degradation and the

430 formation of new odorant molecules might also have important sensory consequences

431 that need to be explored in future sensory studies.

432

433 ACKNOWLEGMENTS

434 This study was funded by Project AGL2016-78936-R from the Spanish Ministry of

435 Economy and Competitiveness (MINECO). Authors greatly thank the volunteers for

436 their participation in this study.

437

438 ETHICAL STATEMENTS

439

440 Conflict of interest: The authors declare that they do not have any conflict of interest.

441 Ethical Review: This study has been approved by the Bioethical Commission of the

442 Spanish National Council of Research (CSIC).

443 Informed Consent: Written informed consent was obtained from all the volunteers who

444 participated in this study.

445

446 REFERENCES:

447 Buettner, A. (2002a). Influence of human saliva on odorant concentrations. 2. 448 Aldehydes, alcohols, 3-alkyl-2-methoxypyrazines, methoxyphenols, and 3- 449 hydroxy-4, 5-dimethyl-2 (5 H)-furanone. Journal of Agricultural and Food 450 Chemistry, 50(24), 7105-7110. 451 Buettner, A. (2002b). Influence of human salivary enzymes on odorant concentration 452 changes occurring in vivo. 1. Esters and thiols. Journal of Agricultural and Food 453 Chemistry, 50(11), 3283-3289. 454 Buffo, R., Rapp, J., Krick, T., & Reineccius, G. (2005). Persistence of aroma 455 compounds in human breath after consuming an aqueous model aroma mixture. 456 Food Chemistry, 89(1), 103-108. 457 de Lacy Costello, B., Amann, A., Al-Kateb, H., Flynn, C., Filipiak, W., Khalid, T., . . . 458 Ratcliffe, N. M. (2014). A review of the volatiles from the healthy human body. 459 Journal of breath research, 8(1), 014001. 460 Escudero, A., Campo, E., Fariña, L., Cacho, J., & Ferreira, V. (2007). Analytical 461 characterization of the aroma of five premium red wines. Insights into the role of 462 odor families and the concept of fruitiness of wines. Journal of Agricultural and 463 Food Chemistry, 55(11), 4501-4510. 464 Esteban-Fernández, A., Muñoz-González, C., Jiménez-Girón, A., Pérez-Jiménez, M., & 465 Pozo-Bayón, M. Á. (2018). Aroma release in the oral cavity after wine intake is 466 influenced by wine matrix composition. Food Chemistry, 243, 125-133. 467 Esteban-Fernández, A., Rocha-Alcubilla, N., Muñoz-González, C., Moreno-Arribas, M. 468 V., & Pozo-Bayón, M. Á. (2016). Intra-oral adsorption and release of aroma 469 compounds following in-mouth wine exposure. Food Chemistry, 205, 280-288. 470 Francis, I., & Newton, J. (2005). Determining wine aroma from compositional data. 471 Australian Journal of Grape and Wine Research, 11(2), 114-126. 472 Friel, E., & Taylor, A. (2001). Effect of salivary components on volatile partitioning 473 from solutions. Journal of Agricultural and Food Chemistry, 49(8), 3898-3905. 474 Krisch, K. (1971). 3 Carboxylic Ester Hydrolases The enzymes (Vol. 5, pp. 43-69): 475 Elsevier. 476 Leggett, D. C., Jenkins, T. F., & Miyares, P. H. (1990). Salting-out solvent extraction 477 for preconcentration of neutral polar organic solutes from water. Analytical 478 Chemistry, 62(13), 1355-1356. 479 Linforth, R., & Taylor, A. J. (2000). Persistence of volatile compounds in the breath 480 after their consumption in aqueous solutions. Journal of Agricultural and Food 481 Chemistry, 48(11), 5419-5423. 482 Lytra, G., Tempere, S., Le Floch, A., de Revel, G., & Barbe, J.-C. (2013). Study of 483 sensory interactions among red wine fruity esters in a model solution. Journal of 484 Agricultural and Food Chemistry, 61(36), 8504-8513. 485 Muñoz-González, C., Feron, G., Brulé, M., & Canon, F. (2018). Understanding the 486 release and metabolism of aroma compounds using micro-volume saliva 487 samples by ex vivo approaches. Food Chemistry, 240, 275-285. 488 Pagès-Hélary, S., Andriot, I., Guichard, E., & Canon, F. (2014). Retention effect of 489 human saliva on aroma release and respective contribution of salivary mucin and 490 α-amylase. Food Research International, 64, 424-431. 491 Ployon, S., Morzel, M., & Canon, F. (2017). The role of saliva in aroma release and 492 perception. Food Chemistry, 226, 212-220. 493 Rapp, A., & Mandery, H. (1986). Wine aroma. Experientia, 42(8), 873-884.

494 Rodríguez‐Bencomo, J. J., Muñoz‐González, C., Andújar‐Ortiz, I., Martín‐Álvarez, P. 495 J., Moreno‐Arribas, M. V., & Pozo‐Bayón, M. Á. (2011). Assessment of the 496 effect of the non‐volatile wine matrix on the volatility of typical wine aroma 497 compounds by headspace solid phase microextraction/gas chromatography 498 analysis. Journal of the Science of Food and Agriculture, 91(13), 2484-2494. 499 Sumby, K. M., Grbin, P. R., & Jiranek, V. (2010). Microbial modulation of aromatic 500 esters in wine: Current knowledge and future prospects. Food Chemistry, 501 121(1), 1-16. 502 Voilley, A., & Souchon, I. (2006). Flavour retention and release from the food matrix: 503 an overview. Flavour in foods, 117-132.

504

505

506

507 FIGURE CAPTIONS:

508 Figure 1. Chromatograms corresponding to: (a) control saliva systems without added

509 aroma, (b) with added aroma (six esters) and without incubation and (c) with added

510 aroma and after 120 minutes of incubation at 37 ºC. Numbers in the chromatogram

511 correspond to the compounds: 1: ethyl butanoate, 2: isoamyl acetate, 3: 2-ethyl

512 pentanoate, 4: ethyl hexanoate, 5: ethyl octanoate, 6: ethyl decanoate, 7:hexanoic acid,

513 8: octanoic acid, 9: decanoic acid.

514

515 Figure 2. Effect on incubation time (120 minutes at 37 ºC) on esters recovered from (a)

516 enzymatic saliva (ES); (b) non enzymatic saliva (NES). Different letters denote

517 statistical significant differences (p<0.05) from Tukey test. Results are expressed as

518 relative peak area (compound peak area / internal standard peak area).

519 Figure 3. Carboxylic acids determined in NES (no enzymatic saliva) and ES (enzymatic

520 saliva) systems before and after incubation (120 min, 37ºC) in presence of the six tested

521 esters.

522 Figure 4. Aroma remaining (%) in the oral cavity four minutes after spitting out the

523 wine

524 Figure 5. Relationship between esters remaining in the oral cavity (4 minutes after

525 spitting out the wine) and logP values (hydrophobicity).

526

527 TABLES:

528 Table 1. Experimental saliva systems employed in this study.

529 Table 2: Physicochemical properties of the six aroma compounds tested in the study.

530 Table 3. Results from two-way ANOVA (saliva system vs incubation time) and Tukey 531 test for mean comparison (relative peak area of each ester).

532

533 Table 1. Experimental saliva systems employed in this study 534 Sample code Type of saliva Ester aromatization(a) Incubation (37ºC) ES-C0 Enzymatic saliva control No No ES-C120 Enzymatic saliva control No 120 min ES-0 Enzymatic saliva 50 mg/L No ES-120 Enzymatic saliva 50 mg/L 120 min NES-C0 Non enzymatic saliva control No No NES-C120 Non enzymatic saliva control No 120 min NES-0 Non enzymatic saliva 50 mg/L No NES-120 Non enzymatic saliva 50 mg/L 120 min (a)535 Each system was aromatized with one single aroma compound. All the samples were 536prepared in triplicated. 537

538 Table 2: Physicochemical properties of the six aroma compounds tested in the study.

Aroma Odor Molecular

MW (g molˉ¹) Log P BP (ºC) OT (µg/L) description compound structure Pinneaple Ethyl butanoate 116 1.9 121.5 20

Banana Isoamyl acetate 130 2.3 142.5 30

Fruity Ethyl pentanoate 130 1.7 145 3

Apple Ethyl hexanoate 136 2.8 167 5-14

Peach Ethyl octanoate 172 3.8 208.5 2-5

Grape Ethyl decanoate 200 4.8 241.5 200

539 MW: Molecular weight, Log P value: logarithm of the octanol/water partition 540 coefficient and BP: boiling point values are from PubChem Open Chemistry database.

541 OT: odor threshold from Francis and Newton, J. L. (2005). Odor description obtained 542 from Flavornet database. (http://www.flavornet.org/)

543

544 Table 3. Results from two-way ANOVA (saliva system vs incubation time) and Tukey test for 545 mean comparison (relative peak area of each ester).

Ethyl Isoamyl Ethyl Ethyl Ethyl Ethyl butanoate acetate pentanoate hexanoate octanoate decanoate Saliva system NES 0,175 a 0,124 a 0,128 a 0,155 a 0,231 b 0,228 b ES 0,149 b 0,114 a 0,108 b 0,154 a 0,306 a 0,432 a Pr > F 0,002 0,002 0,001 0,001 0,000 < 0,0001 Time t=0 min 0,192 a 0,135 a 0,137 a 0,172 a 0,277 a 0,291 b t=120 min 0,132 b 0,103 b 0,099 b 0,138 b 0,259 b 0,369 a Pr > F 0,002 0,002 0,001 0,001 0,000 < 0,0001 546 Different letters within the same column denote significant differences p<0.05 among saliva 547 systems (NES vs ES) or incubation times (t=0 and t=120 min).

548

549

550

551

552

553

554

555

556

557

558

559

560

561 Figure 1

562

563

564

565

566

567

568 Figure 2

569

570

571

Hexanoic acid 0,003 0,0025 0,002

0,0015

RPAs NES 0,001 ES 0,0005 0 0 120 Incubation time (min) 572 Octanoic acid 0,007 0,006 0,005

0,004

RPAs 0,003 NES 0,002 ES 0,001 0 0 120 Incubation time (min) 573 Decanoic acid 0,01

0,008

0,006

RPAS 0,004 NES ES 0,002

0 0 120 Incubation time (min) 574

575 Figure 3

576

70 65,0%

30 27,7%

% aroma remaining in the oral cavity oral the in remaining % aroma 9,3 % four minutes after wine expectoration wine after minutes four 10 4,4 % 5,2 % 4,5%

0 Ethyl Isoamyl Ethyl Ethyl Ethyl Ethyl butanoate acetate pentanoate hexanoate octanoate decanoate

577

578 Figure 4

579

580

5 Etdec 4 Etoct

3 IsoAcet log P values P log 2 Ethex Etpent y = 0,0521x + 1,7579 1 Etbut R² = 0,8872

0 0 10 20 30 40 50 60 70 % aroma remaining in the oral cavity 4 minutes after wine expectoration 581

582

583 Figure 5

584

585

586