Assessment of the Minor Component Transformations in Fat during the

Subscriber access provided by UNIV OF SCIENCES PHILADELPHIA

Food and Beverage Chemistry/Biochemistry

Assessment of the Minor Component Transformations in Fat during the Green Spanish-Style Table Olives Processing Antonio López-López, Amparo Cortés-Delgado, and Antonio Garrido-Fernández J. Agric. Food Chem., Just Accepted Manuscript • DOI: 10.1021/acs.jafc.8b00927 • Publication Date (Web): 13 Apr 2018 Downloaded from http://pubs.acs.org on April 13, 2018

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 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 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.

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 33

Journal of Agricultural and Food Chemistry

1 2

Assessment of the Minor Component Transformations in Fat during the Green Spanish-Style Table Olives Processing

3 4 5 6

Antonio López-López*, Amparo Cortés-Delgado, and Antonio Garrido-Fernández

7 8 9

Instituto de la Grasa (CSIC), Food Biotechnology Department, Campus Universitario Pablo de Olavide, Building 46, Ctra. Utrera km 1, 41013 Sevilla, España

10 11 12 13

Running title: Transformations in green olive fat minor components during processing

14 15 16 17 18 19

*Corresponding author:

20

Dr. Antonio López-López

21

e-mail: [email protected]

1 ACS Paragon Plus Environment


Page 2 of 33

22

ABSTRACT

23

There is an increasing interest of consumers for natural and healthy products. This work

24

asses the transformations that green Spanish-style processing of Manzanilla and

25

Hojiblanca table olives produce on the minor components of its fat. Discriminant

26

Analysis showed that most of the variability was not due to processing (24.4%) but to

27

differences between cultivars (59.2%). Therefore, the final products have a similar

28

quality than the original olive fat; that is, the quality of the fat was scarcely affected.

29

The only systematic trends observed were the decrease in hexacosanol, tetracosanol and

30

octacosanol (fatty alcohols) and C46 (wax), after lye treatment, and the high levels of

31

alkyl esters in the packaged product. Thus, minor component levels in green table olives

32

are, in general, within the limits established for extra virgin olive oil since the alkyl

33

esters should be considered habitual products of fermentation and not as an alteration as

34

in olive oil.

35 36 37 38 39 40

KEYWORDS: green Spanish-style table olives, processing, Manzanilla cultivar, Hojiblanca cultivar, olive fat, minor olive fat components, chemometrics.


Page 3 of 33


41

INTRODUCTION

42

The production of table olives was 2,650,000 tonnes in the 2015/2016 season.1 The

43

European Union produced 868,000 tonnes, with Spain being the main contributor (̴

44

550,000 tonnes). The so-called green Spanish-style table olive is the most popular and

45

accounts for about 60% of the total consumption. Processing always includes several

46

treatments to reduce the natural bitterness of olives. The green Spanish-style consists of

47

an immersion in lye to degrade the bitter compounds, followed by washing, brining,

48

fermentation, conditioning and packaging, all of them performed using aqueous

49

solutions.2

50

Fat is the primary nutrient in table olives.2 The study of numerous table olive

51

commercial presentations showed important differences in the concentrations of their oil

52

components, with triacylglycerols and their associated fatty acids constituting most of

53

the olive lipids.3,4 However, many other minor components (sterols, fatty alcohols,

54

triterpenic dialcohols, waxes, fatty acid esters and products of degradation) play

55

essential roles in oil quality characteristics.5 In this way, several sterols, triterpenic

56

dialcohols, waxes, and alkyl esters are subjected to constraints in olive oil, according to

57

the European Union Legislation and the Standard issued by the International Olive

58

Council.6,7 Furthermore, their levels are determinant for olive oil classification.

59

Therefore, when investigating the effect of processing on the table olive fat, the study of

60

the changes in the minor component profiles is essential.

61

Due to the low concentration of NaOH used for debittering, the exclusive application

62

of aqueous solutions during the elaboration, and the localisation of the oil droplets in the

63

interior of the cell, it was considered for decades that the olive fat was not affected by

64

the green Spanish-style processing. However, recent studies have shown that ripe olive

65

preparation may affect the olive fat composition,8 including the minor components.9 3 ACS Paragon Plus Environment


Page 4 of 33

66

Most of such transformations were produced during the previous storage/fermentation

67

process and were associated with the presence of a lipolytic microflora during this

68

phase; among the species isolated, Candida boidinii showed the strongest activity.10

69

Despite it, individual yeasts have shown a limited effect on the oil quality parameters of

70

green Spanish-style table olives.11 On the contrary, the pitting and pitting and stuffing

71

processes of ripe and green Spanish-style table olives, respectively, produced significant

72

modifications on the minor components of the released oils.12,13 Other cultivars and

73

processing have also been studied by different authors. Sakouhi et al.14 reported the

74

changes in α-tocopherol and fatty acids after processing as green Spanish-style Tunisian

75

autochthonous cultivars, finding important differences among them. Interestly, the

76

presence of α-tocopherol was positively related to the content of unsaturated fatty acids.

77

No changes in tocopherols during different processing conditions (Spanish-style, short

78

process, and “Picholin” style) whereas their contents decreased in the packaged

79

products, with significant losses after 12 months storage.15 Bianchi4 has published a

80

review on the changes in phenols and lipids during processing. However, in mature

81

Oinotria table olive processing (blanching, salting, and drying), biophenols decreased

82

while the fatty acid composition remained without deterioration.16 Also, Lanza et al.17

83

has related the presence of high proportions of alkyls esters in Kalamata and Moresca

84

cultivars with the use of low concentrations of brines in Greek style olives while

85

Giarraffa and Nocellara del Belice cultivars produced only ethyl esters.

86

The combination of analytical techniques and chemometrics can be an excellent

87

strategy for detecting changes in the fat and oil compositions.18 The chemometric

88

analysis was useful to detect changes in the oil characteristics of ripe olives during

89

processing8,9 and to segregate the various oils released during the table olive

90

conditioning.12,13 Furthermore, the same techniques were also applied to study the 4 ACS Paragon Plus Environment

Page 5 of 33


91

effects of green Spanish-style processing on the quality parameters and fatty acid

92

compositions of Manzanilla and Hojiblanca olive fat;19 however, the effects on the

93

minor components still require research.

94

This work aimed to study the minor component (polar compounds, sterols, fatty

95

alcohols, triterpenic dialcohols, waxes, and alkyl esters) profiles of the green Manzanilla

96

and Hojiblanca fat throughout Spanish-style processing. General Linear Model,

97

unsupervised (Principal Component Analysis and Hierarchical Clustering), and

98

supervised (Discriminant Analysis) chemometric techniques were used for the

99

characterisation, identification of transformations, and grouping the corresponding fat

100

101

samples according to cultivars and production phases. MATERIAL AND METHODS

102

Olives. Fruits were of the Manzanilla and Hojiblanca cultivars harvested at the so-

103

called green maturation stage. They were provided by a local processor (JOLCA S.L.,

104

Huevar del Aljarafe, Sevilla, Spain).

105

Processing. Olives of each cultivar were processed in duplicate according to the

106

green Spanish-style. It consisted of treating, in a 50 L container, 30 kg Manzanilla and

107

Hojiblanca olives with 20 L of NaOH solutions at 25 and 30 g L-1 concentrations,

108

respectively, until the alkali reached 2/3 of the flesh. Then the olives were washed with

109

tap water for 18 h and brined in 35 L PVC containers (20 kg olives, 15 L brine), using a

110

90 g L-1 NaCl solution where they followed a spontaneous lactic fermentation. After

111

equilibrium olive-flesh/brine, the NaCl concentration was ∼55 g L-1 and increased

112

progressively to reach ∼60 g L-1 NaCl due to the evaporated brine replacement during

113

the fermentation period. At the end of the fermentation (7 months), 10 kg olives from

114

the two cultivars were packed (separately for each replicate) in glass containers. The



Page 6 of 33

115

physicochemical conditions of brines were fixed to reach the equilibrium at 5 g L-1

116

lactic acid and 55 g L-1 NaCl. Both processing and packaging were intended to closely

117

mimic the standard procedures usually followed for the elaboration of these cultivars at

118

industrial scale.

119

Samples and Fat Extraction. Olive samples from each cultivar and replicate were

120

withdrawn from the fresh fruits, lye-treated olives, the fermented fruits, and the final

121

product (after two months of packaging). The samples were coded as M (Manzanilla)

122

and H (Hojiblanca), followed by T0, T1, T2, or T3, respectively, to identify the

123

successive processing phases.

124

The samples were extracted using the ABENCOR system (Abengoa, Madrid, Spain).

125

In short, the fruits were pitted and mixed with a homogeniser Ultraturrax T25 (IKA-

126

Labortecnik, Staufen, Deutschland); then, hot water was added to the paste to reach

127

about 30 °C in the suspension. The resulting mixture was subjected to malaxation for 40

128

min at room temperature (22 ±2 °C), and the liquid was removed by centrifugation

129

using an ABENCOR equipment. The liquid phase was allowed to decant, and the oil

130

was obtained, filtered and subjected to analysis. The process was similar to that used for

131

the estimation of olive oil yield 20 and the efficient extraction of table olive fat.3,5,8,9,10

132 133

Separation of Polar and non-Polar Compounds. The oils were fractioned using silica gel columns, according to the procedure developed by Dobarganes et al.21

134

Determination of Polar Compounds. The total polar compounds (PC) and their

135

components (polymerized triacylglycerols (PTG), oxidised triacylglycerols (OxTG),

136

diacylglycerols (DG), and free fatty acids plus traces of unsaponifiable matter (FFA))

137

were analysed according to the method developed by Dobarganes et al. 21


Page 7 of 33


138 139

140 141

142 143

Determination of Sterols and Triterpenic Dialcohols. This analysis was performed

according to the method described by the EU 1348/2013.22 Determination of Fatty Alcohols. This analysis was carried out according to the method outlined by the EU 2015/1833.23 Determination of Waxes and Alkyl Esters Contents. This analysis was performed according to the method described by EU 1348/2013.22

144

Apparatus and Reagents. Apparatus. The determinations were carried out using an

145

equipment HP1050 system (HPLC) equipped with a refractive index detector (Hewlett-

146

Packard, CA, USA) and an HP 5890 Series II gas chromatograph (Hewlett-Packard,

147

Minnesota, USA) fitted with a flame ionization detector. Reagents.All reagents were of

148

analytical grade and chromatographic grade.

149

Statistical analysis. For statistical analysis, data were structured in a matrix array

150

where rows were cases (treatments), and columns were variables (the contents of the

151

minor components). The effect of processing phases on the minor components was

152

studied by GLM and chemometric techniques (after data standardisation according to

153

variables).13

154

The chemometric analysis comprised the first survey of correlation, followed by

155

Principal Component Analysis (PCA), Hierarchical Clustering (HC), and Discriminant

156

Analysis (DA). For PCA only eigenvalues ≥1 (according to the Kaiser criterion) were

157

retained.24 DA was achieved by using the complete matrix of components, but for the

158

deduction of the canonical and discriminant functions, only those variables containing

159

the most powerful information were retained. The selection of variables was made by

160

the backwards stepwise procedure, with 0.05 and 0.10 probability values for entering



Page 8 of 33

161

and removing them. The minimum tolerance was fixed at 0.00001 and the number of

162

steps at 100. A leaving-one-out cross-validation procedure was used for assessing the

163

performance of the classification rule.12

164

Statistica software version 7.0 (StatSoft Inc., Tulsa, USA), XLSTAT (Addinsoft,

165

2017), and IBM SPSS Statistics Version 22 (IBM Corporation, Spain) was used for data

166

processing.

167

RESULTS AND DISCUSSION

168

The samples of this study were subjected to the usual green Spanish-style table olive

169

processing and packaging.2 Therefore, the transformations observed in their olive fat

170

mimic those expected at industrial scale. It should be pointed out that this work

171

complements the previous research carried out on Manzanilla and Hojiblanca to study

172

the effects of green Spanish-style processing on the quality parameters and fatty acid

173

and triacylglycerols of olive fat. Also, green Spanish-style processing may introduce

174

important changes on other fat/olive characteristics like α-tocopherols or pigment

175

(chlorophylls and carotenoids).14,25 However, this work is exclusively focussed on the

176

changes that this processing style may produce on the olive fat minor components

177

included in the olive oil legislations. 6,7

178

Polar Compound Profiles. A previous description of the diverse polar compounds

179

in diverse table olive processing was reported by Bianchi.4 Most of the polar

180

compounds (PC) correspond to degradation products. The absence of polymerized

181

triacylglycerol in the fats from all samples may indicate a limited secondary oxidation.26

182

The concentrations of total PC were always higher in Manzanilla than in Hojiblanca,

183

but processing did not produce systematic effects within cultivar (Table 1). The levels

184

in Hojiblanca were reduced; however in Manzanilla were comparable to those observed 8 ACS Paragon Plus Environment

Page 9 of 33


185

in recently extracted extra virgin olive oil (EVOO) (∼25 mg/g) but lower than those

186

found after 12 month olive oil storage (33 mg/g).27 The lowest values observed during

187

ripe olive processing (18-50 mg/g) were similar to those reported here for Manzanilla

188

but higher than those for Hojiblanca.8 Therefore, green Spanish-style processing is,

189

apparently, less aggressive with the fat than Californian style processing (ripe olive);

190

furthermore, its packaged products may have total polar compounds comparable to

191

EVOO. In Italian cultivars, the initial contents of PC in the fresh fruit fats were of the

192

same order than in this work; but, after fermentation as natural green olives, the contents

193

were higher, regardless of cultivar.26 Similarly, the levels (27-37 mg/g) in the oils

194

released from pitting of green Spanish-style or ripe olives were higher than those found

195

in this work (Table 1).12

196

Oxidized triacylglycerols (OxTG) constituted the major fraction of the polar compounds

197

and were higher in Manzanilla than in Hojiblanca (Table 1). Due to the high stability

198

and low volatility of OxTG, these compounds are considered as an index of the

199

secondary oxidation level of oils, usually correlated with K270.26 However, the

200

progression of these compounds into further oxidation products, like polymerised

201

triacylglycerols, during green Spanish-style Manzanilla and Hojiblanca processing was

202

not observed due to the reduced absolute changes observed in OxTG. This result is in

203

contrast to the changes observed in K270 which had a marked decrease after lye

204

treatment, followed by an increase in the fermented product but without influence of

205

packaging.19 In ripe olives, the OxTG increased progressively throughout processing,

206

leading to the similar concentrations in the packaged fruits of both cultivars (5.8 and 7.3

207

mg/g oil in Hojiblanca and Manzanilla, respectively),8 with levels below most of the

208

values in Table 1. In the table olives conditioning operations, the lowest proportions of

209

OxTG were observed in oils from the ripe olive pitting operation while the highest was 9 ACS Paragon Plus Environment


Page 10 of 33

210

observed in oils from green olive pitting.12 The OxTG contents in green Spanish-style

211

processing are higher (Table 1) than those found in green, directly brined, Italian

212

cultivars, possibly because the fat oxidation was stronger and progressed further into

213

PTG.26

214

Diacylglycerols (DG) are usually considered as a useful index for oil quality and are

215

produced by the hydrolysis of triacylglycerols.28,29 In this work, Manzanilla and

216

Hojiblanca had low and similar contents at the end of processing (3.75-4.15 mg/g oil)

217

(Table 1). The results are in contrast with the higher concentrations found in the fats

218

from ripe olives (17-24 mg/g oil) or in oils released during the pitting and stuffing green

219

table olives with vegetable products (15-30 mg/g oil).8,12 The difference may be

220

attributed to the higher lipolytic activity in ripe olives or the environmental conditions,

221

in the case of the released oils during the conditioning operations. The presence of DG

222

was elevated in the fresh fruits of Italian cultivars which content significantly increased

223

as the natural processing progressed, possibly because of the lack of inactivation of the

224

lipolytic enzymes in this elaboration.26

225

The free fatty acids (FFA) are also produced by hydrolysis of triacylglycerols.26,29

226

Their concentrations in this work did not have a clear trend during processing and,

227

overall, the changes were limited (Table 1), although their highest concentrations were

228

found in the packaged products in agreement with the acidity observed when studying

229

the quality parameters of these olives.19 FFA was particularly high (∼45 mg/g oil) in oils

230

released from the pitting and stuffing with vegetable origin material green Spanish style

231

table olives.12 Furthermore, there was also an evident increase of FFA during the natural

232

processing of Italian cultivars, in agreement with the DG formation and the apparent

233

higher enzymatic lipolytic activity in the non-lye treated olives.26


Page 11 of 33


234

Therefore, the overall hydrolytic degradation during processing green Spanish-style

235

olives was limited and moderate in comparison with the changes in DG observed in

236

natural fermentation.26 This reduced activity could be due to the partial inactivation of

237

the enzymatic hydrolytic process in the lye-treated fruits and the low contribution of

238

lipolytic yeasts to this degradation during fermentation since, at the moment, their

239

effects are questionable as recently demonstrated in individually yeast inoculated

240

processed olives.11 Shimizu et al. also found a good correlation between FFA and DG

241

and demonstrated that the hydrolytic degradation in olive oil occurred during the

242

previous to extraction storage.30

243

Unsaponifiable Fraction Profiles. Sterols. Sterols are important non-glyceridic

244

compounds usually related to the olive oil authenticity in the EU.22 Approximately, 10-

245

15% are esterified, but most of them (85-90%) are free.31 Regardless of cultivar, sterol

246

concentrations are remarkable (Table 2) and, therefore, green Spanish-style table olives

247

may contribute to the beneficial effects of phytosterols as blood cholesterol-lowering

248

agent.32 Total sterols, β-sitosterol, and campesterol concentrations were always higher

249

in Hojiblanca than in Manzanilla (Table 2), and processing hardly affected their

250

concentrations. Among components, β-sitosterol was the most abundant followed by

251

∆5-avenasterol, campesterol and clerosterol (Table 2). The cholesterol content,

252

regardless of cultivar, was higher in the fresh fruits than after processing but no

253

systematic effects were detected in the other sterols (Table 2). Sterols contributed to

254

discrimination among oils from diverse geographic regions in Apulian.33 Sterolic

255

composition, with sitostanol, campestanol, and campesterol as the major contributors

256

played an essential role in the characterisation of oils from some Tunisian olive

257

cultivars.34 Oils from the same cultivar but different geographical areas were correctly

258

classified using their sterol profiles.35 In Spanish commercial table olives, the sterol 11 ACS Paragon Plus Environment


Page 12 of 33

259

contents were significantly different between cultivars but not between processing

260

styles.5 In ripe olive, subjected to more degrading processing than the green Spanish-

261

style products, β-sitosterol, ∆5-avenasterol and cholesterol significantly contributed to

262

segregation according to elaboration phases.9

263

Some of the sterols (in percentages) are subjected to constraints in the EU22 and the

264

IOC7 for olive oil. Their proportions are relevant concerning olive oil classification with

265

thresholds shown in parenthesis in the last row of Table 2, when appropriate. As

266

observed, the percentages found for them in this work were mostly below their limits

267

(Table 2, in brackets).

268

Triterpene Dialcohols.

These compounds are triterpenic pentacyclic dialcohols

269

identified by Fedeli36 and are currently considered for the classification of olive oils

270

according to the EU22. Usually, the proportion of erythrodiol is higher in oils extracted

271

by solvents than in those obtained by the press.22 In this work, the levels of erythrodiol

272

were greater than those of uvaol and this, in turn, was higher in Hojiblanca than in

273

Manzanilla (except for MTO). However, processing did not produce any systematic

274

trend (Table 3). Also, the erythrodiol contents were close to the upper range of those

275

observed in ripe olives (6-40 mg/kg oil in Hojiblanca; 54-87 mg/kg, in Manzanilla), but

276

higher than in the oils released during the conditioning process of table olives (not

277

detected-15 mg/kg oil) (Table 3).9,13 Uvaol was absent in oils from several steps of ripe

278

olive processing and in oils from table olives conditioning operations.9,13 The

279

erythrodiol+uvaol percentages (EU regulation) were higher in Manzanilla than in

280

Hojiblanca but were not affected by treatments, regardless of cultivar (Table 3). Their

281

levels were always below the limit established in the EU legislation for EVOO (4.5%).22


Page 13 of 33


282

Fatty Alcohols. In olive oil, fatty alcohols with an even number of carbon atoms are

283

abundant while those with odd are only found as traces.37 The total concentration of

284

these compounds in fresh fruits was higher in Manzanilla (305±7 mg/kg oil) than in

285

Hojiblanca (200±7 mg/kg oil). Tetracosanol (C24) (except in Hojiblanca), hexacosanol

286

(C26) and octacosanol (C28) markedly decreased after lye treatment (Figure 1).

287

Docosanol (C22) increased during processing in Hojiblanca but was stable in Manzanilla

288

(Figure 1). Fatty alcohol levels increased with the olive oil extraction by solvent.38 In

289

oils from seven autochthonous cultivars (Calabria, South of Italy) was found significant

290

differences among cultivars and a general decline as harvesting time progressed.39

291

Wax and Alkyl Ester Profiles. The waxes are esters of fatty alcohols and fatty

292

acids. Most waxes in olive oils have an even number of carbons (C36-C46). Hydrolysis

293

of triacylglycerols favours the formation of waxes. Wax content in pomace oil is usually

294

higher than in EVOO; therefore, the wax level in oils is used to detect fraudulent

295

mixtures.6 Waxes can also be formed spontaneously during olive oil storage, at rates

296

depending on the proportions of reactants and environmental conditions.40 A detailed

297

description of the wax composition, alkyl esters and methyl phenyl esters and their

298

changes in table olives were reported by Bianchi.4 In this work, total waxes

299

(C40+C42+C44+C46) and sum waxes (C42+C44+C46) were always higher in oils from

300

Manzanilla than in Hojiblanca but were not affected by processing (Table 3). In the EU

301

regulation, the sum waxes for EVOO and virgin olive oil (VOO) should be ≤150 mg/kg;

302

however, the threshold of total waxes affects only to lampante (≤300 mg/kg) or lower

303

categories of olive oils (≤350 mg/kg).6 In this study, the sum waxes (Table 3) was

304

always below the limit established in the EU for EVOO; furthermore, the total waxes

305

were also far below the limits for lampante oil.6 Therefore, the wax content of the green

306

Spanish-style table olives fat is comparable to those in EVOO. Concerning the 13 ACS Paragon Plus Environment


Page 14 of 33

307

individual components, the levels of C40 and C42 were always significantly higher in

308

Manzanilla (Figure 2, panel a and b) but C44 and C46 had greater contents in Manzanilla

309

only after fermentation and packaging (Figure 2, panels c and d). The trends of C40, C42,

310

and C44 were scarcely, or not, affected by processing, but C46 significantly decreased

311

after lye treatment. The low proportion of waxes in oils from green Spanish-style olives

312

(very far below the limits for EVOO) is a good indicator of the soft processing of this

313

product.6 On the contrary, these compounds suffer a sharp increase in hard extraction

314

conditions or careless oil storage.40

315

Alkyl esters are produced by the reaction of free fatty acids and ethanol or methanol

316

and are associated with fermentative deterioration of olives before extraction. Low

317

levels indicate oils from recently picked olives and, indirectly, of good quality since

318

their presence leads to sensory defects in olive oil.41 In olive processing, the formation

319

of methanol during lye treatments is well known.42 Also, production of ethanol,

320

methanol and other volatile during all table olive processing is common.2,43 However,

321

excessively high levels (450-550 mg kg-1 oil) in directly brined olives have been related

322

to abnormal fermentation and poorly conducted technological treatments.17 In this work,

323

lye treatment produced a significant increase of methyl esters regardless of cultivar but

324

formed ethyl esters only in Manzanilla (Figure 3). However, fermentation significantly

325

increased the ethyl esters in Hojiblanca and only ethyl oleate contents in Manzanilla

326

(Fig. 3, panel b and c).

327

The limit for total ethyl esters established in the EU regulation for EVOO is ≤35

328

mg/kg.6 The ethyl ester concentrations in oils from fermented and packaged Manzanilla

329

were below it and, on the contrary, those from Hojiblanca were above (Table 3);

330

consequently, the fats from the final products of the green Spanish-style Hojiblanca

331

table olives should not be considered as EVOO according to this criterion. However, the 14 ACS Paragon Plus Environment

Page 15 of 33


332

direct application of the olive oil legislation6,7 to these fats, in this case, can be

333

misleading because ethyl alcohols have not any unfavourable connotation in table olive

334

storage/fermentation.43 Consequently, the use of ethyl ester as a quality criterion for

335

table olive fat will be questionable. Furthermore, only oils with very high levels of ethyl

336

esters (which is not the case) have been sensory assessed as defective (with fermentative

337

notes like fusty, winery, vinegary, muddy sediment, or musty).44 Therefore, the levels

338

reached in green Spanish-style table olives could hardly produce the unpleasant

339

sensations found in oils from non-brined fusty/muddy olives.41

340

Nevertheless, the formation of these volatile compounds during processing should

341

not be underestimated. The activity of yeasts during fermentation should eventually be

342

controlled by inoculation of lactic acid bacteria or by the use of yeast starter cultures

343

with limited production of methyl or ethyl esters. They could prevent excessive

344

formation of esters and preserve the processed olive fat quality in levels comparable to

345

EVOO.

346

In general, ethyl esters and waxes were the minor components which showed the

347

most consistent trend during processing, although their roles in the interpretation of the

348

processing effects on table olive fat or sensory characteristics will still require further

349

clarification.

350

Segregation of Overall Minor Component Profiles throughout Processing by

351

Multivariate Analysis. The presence of several significant correlations among

352

compounds qualified the data for multivariate analysis.

353

The PCA of data extracted five factors with eigenvalues >1. The overall variance

354

explained by the two first Factors was 57.58% (Figure S1, a). Factor 1 was responsible

355

for differentiation between cultivars (variance explained 36.69%) while Factor 2 was 15 ACS Paragon Plus Environment


Page 16 of 33

356

expected to segregate according to treatments (21.09%). Oils from the fresh fruits of

357

both cultivars were in the 2nd quadrant well separated between them and from the

358

processed olives. Oils from the different processing phases of Hojiblanca were situated

359

on the right of the graph with lye treated olives being on the upper-left but without any

360

trend among fermented and packaged olives (Figure S1, panel a). All Manzanilla

361

samples, except the fresh fruit, were grouped on the 3rd quadrant with scarce differences

362

among them. Therefore, apparently, processing had a lower effect on the minor

363

components of Manzanilla than on the Hojiblanca fat (somewhat dispersed in the

364

graph).

365

Another approach for studying dissimilarities among treatments is cluster analysis,

366

using the interaction cultivar·treatment as the grouping variable. Three clusters were

367

obtained (Figure S1, panel b). The most considerable dissimilarity was also observed

368

between cultivars. The technique segregated Hojiblanca from Manzanilla. However, the

369

Hojiblanca cluster did not distinguish between fresh fruit and processing phases,

370

grouping all the samples in the same cluster. Samples from Manzanilla were grouped

371

into two closely related clusters, but significantly different. One consisted of fresh and

372

lye treated fruits (plus one fermented sample) while the other was composed of

373

fermented and packaged olives. Therefore, clustering was not efficient in segregating

374

the fresh fruits but confirmed that the highest dissimilarity corresponded to cultivars; on

375

the contrary, PCA was more consistent and did not lead to any wrong assignation.

376

A further study to differentiate among treatments was attempted by DA. The

377

preliminary ANOVA showed that there were significant differences among treatments

378

for most of the variables, except for ∆5-avenasterol and ∆7-stigmastenol (data not

379

shown) and, therefore, DA application was viable. The most powerful discriminating

380

variables were chosen by applying the backwards elimination (Table S1). The major 16 ACS Paragon Plus Environment

Page 17 of 33


381

contributions to discrimination, as assessed by their standardized coefficients, were FFA

382

(1.674, Function 1), ethyl oleate (1.790, Function 2), methyl oleate (-1.347, Function 1)

383

and OxTG (-1.265, Function 1), C24 (-0.700, Factor 5), C26 (0.918, Function 3), C22

384

(0.947, Function 4), and β-sitosterol (0.931, Function 1) (Table S1). In this case, the

385

first two functions accounted for high cumulative variance (83.6% vs only 57.58% in

386

the PCA analysis), making the DA analysis reliable. In the plane of the two first

387

function axes, the treatments are represented by their centroids while the samples within

388

them are indicated by circles (Figure S2). All samples were correctly grouped. The

389

overall situation was similar to that observed in the PCA but clearer. Segregation

390

between cultivars was based on Function 1 (Manzanilla treatments are situated on the

391

left of the similar Hojiblanca counterparts) and explained most of the variance (59.2%).

392

Regardless of cultivars, the fresh fruits are situated at the bottom of the graph while the

393

samples from the processed olives are placed above them. Therefore, Function 2

394

segregates among treatments, and the various processing phases are well differentiated

395

within cultivar. However, Function 2 only accounts for 24.4% of the total variance; that

396

is, for less than half of the cultivar variance (Function 1). Therefore, the DA has

397

confirmed the differences between cultivars above commented and has also disclosed

398

scarce but noticeable differences between processing phases, only partially revealed by

399

PCA and clustering. The higher overall segregation capacity of DA (based on the high

400

variance explained) was also reaffirmed by the confusion matrix which showed, even in

401

the validation process, 100% correct assignations of samples to treatments.

402

In summary, the most affected compounds during processing were fatty alcohols and

403

alkyl esters. In general, the contents of minor components were below or close to the

404

limits established for EVOO by EU and IOC.6,7 Only the ethyl esters in fermented and

405

packaged Hojiblanca were above them, due to their generation during fermentation. 17 ACS Paragon Plus Environment


Page 18 of 33

406

Therefore, based on their composition, the fats from Manzanilla and Hojiblanca green

407

Spanish-style table olives may have fat qualities comparable to EVOO. However, these

408

transformations were scarce since, overall, the DA assigned most of the variability to

409

cultivars (~59.2%) and only a limited proportion (~24.4%) to processing. That is, the

410

differences in minor components between cultivars are more relevant than the changes

411

introduced by processing. This result is in agreement with the results obtained for

412

qualtiy parameters and fatty acids and triacylglecerols, whose chemometric analysis

413

also showed that most of the variance was due to differences between cultivars and only

414

a reduced proportion to processing. In addition, the parameters with restrictions on the

415

European or IOC legislation were within the limits stablished for them.19 On the

416

contrary, Pasqualone et al.26 reported significant increase of all the indixes of oxidative

417

and hydrolytic degradation of lipids; mainly the hydrolitic degradation was found higher

418

than that reported in the literature for Californian-style (ripe) olives. Therefore, in

419

general, the green Spanish-style table olive fat in Manzanilla and Hojiblanca cultivars

420

preserves their original natural nutritive and healthy characteristics.

421

ASSOCIATED CONTENT

422

Supporting Information

423

Segregation as projections of the cultivar·treatment interactions on the plane of the first

424

two Principal Component Factors and cluster dendrogram, standardized canonical

425

discriminant function coefficients, and the corresponding projection of the

426

cultivar·treatment interactions on the plane of the first two canonical discriminant

427

functions (PDF).

428 429

AUTHOR INFORMATION Corresponding Author 18 ACS Paragon Plus Environment

Page 19 of 33


430

*Phone.: +34 954 692 516. Fax: +34 954 691 262. E-mail: [email protected]

431

ORCID

432

Antonio López-López: 0000-0002-1938-9410

433

Funding

434

This work was supported by the Spanish Government (projects AGL2009-07436/ALI

435

and AGL2010-15494/ALI, partially financed by European regional development funds,

436

ERDF) and Junta de Andalucía (through financial support to group AGR-125).

437

Notes

438

The authors declare no competing financial interest.

439 440 441

ACKNOWLEDGEMENTS The technical assistance of Elena Nogales is acknowledged.

442

443

444

REFERENCES (1) IOC (International Olive Council). On line reference included in World table olives

445

figures: production. IOC: Madrid, Spain, 2017.

446

(http://www.internationaloliveoil.org/estaticos/view/132-world-table-olive-figures)

447

(accessed November 2017).

448 449

(2) Garrido-Fernández, A.; Fernández-Díez, M.J.; Adams, R.M. Table olives. Production and Processing. Chapman & Halls: London, UK, 1997.

450

(3) López, A.; Montaño, A.; García, P.; Garrido, A. Fatty acid profile of table olives and

451

its multivariate characterization using unsupervised (PCA) and supervised (DA)

452

chemometrics. J. Agric. Food Chem. 2006, 54, 6747-6753.



453 454

Page 20 of 33

(4) Bianchi, G. Lipids and phenols in table olives. Eur.J. Lipid Sci.Technol. 2003, 105, 229-242.

455

(5) López-López, A.; Montaño, A.; Ruíz-Méndez, M.V.; Garrido Fernández, A. Sterols,

456

fatty alcohols, and triterpenic alcohols, in commercial table olives. J. Am. Oil Chem.

457

Soc. 2008, 83, 253-262.

458

(6) EU 2016/2095. Commission Delegated Regulation (EU) No 2016/2095 of 26

459

September 2016 amending Regulation (EEC) No 2568/91 on the characteristics of

460

olive oil and olive-residue oil and on the relevant methods of analysis. OJEU

461

(Official Journal of the European Union), 2016, L326, 1-6.

462

(7) IOC (International Olive Council). Trade standard applying to olive oils and olive-

463

pomace oils. Document COI/T.15/NC nº 3/Rev.8. IOC: Madrid, Spain, 2015.

464

(http://www.internationaloliveoil.org/estaticos/view/222-standards)

465

November 2017).

(accessed

466

(8) López-López, A.; Rodríguez-Gómez, F.; Cortés-Delgado, A.; Montaño, A.: Garrido-

467

Fernández, A. Influence of ripe table olive processing on oil characteristics and

468

composition as determined by chemometrics. J. Agric. Food Chem. 2009, 57, 8973-

469

8981.

470

(9) López-López, A.; Rodríguez-Gómez, F.; Cortés-Delgado, A.; Ruíz-Méndez, M.V.;

471

Garrido-Fernández, A. Sterols, fatty alcohol and triterpenic alcohol changes during

472

ripe table olive processing. Food Chem. 2009, 117, 127-134.

473

(10) López-López, A.; Rodríguez-Gómez, F.; Cortés-Delgado, A.; García-García, P.;

474

Garrido-Fernández, A. Effect of the previous storage of ripe olives on the oil

475

composition of fruits. J. Am. Oil Chem. Soc. 2010, 87, 705-714.


Page 21 of 33


476

(11) Rodríguez-Gómez, F.; Bautista- Gallego, J.; Romero- Gil, V.; Arroyo-López, F.N.;

477

Garrido-Fernández, A. Influence of yeasts on the oil quality indexes of table olives.

478

J. Food Sci. 2013, 78, M1208-M1216.

479

(12) López-López, A.; Cortés-Delgado, A.; Garrido-Fernández, A. Chemometric

480

characterisation of the fats released during the conditioning process of table olives.

481

Food Chem. 2011, 126, 1620-1628.

482

(13) Cortés-Delgado, A.; Garrido-Fernández, A.; López-López, A. Chemometric

483

classification of the fat residues from the conditioning operations of table olive

484

processing, based on their minor components. J. Agric. Food Chem. 2011, 59, 8280-

485

8288.

486

(14) Sakouhi, F.; Harrabi, S.; Absalon, C.; Boukhchina, S.; Kallel, H. α-Tocopherol and

487

fatty acids contents of some Tunisian table olives (Olea europea L.): Changes in

488

their composition during ripening and processing. Food Chem. 2008, 108, 833-839.

489

(15) Montaño, A.; Casado, F.J.; de Castro, A.; Sánchez, A.H.; Rejano, L. Influence of

490

processing, storage time, and pasteurisation upon the tocopherol and amino-acid

491

contents of treated green table olives. Eur. Food Res. Technol. 2005, 220, 255-260.

492

(16) Borzillo, A.; Iannotta, N.; Uccella, N. Oinotria table olives: Quality evaluation

493

during ripening and processing by biomolecular components. Eur. Food Res.

494

Technol., 2000, 212, 113-121.

495

(17) Lanza, B.; Di Serio, M.G.; Di Giacinto, L. Fatty-acid alkyl esters in table olives in

496

relation to abnormal fermentation and poorly conducted technological treatments.

497

Grasas Aceites, 2016, 2, 1-10.



Page 22 of 33

498

(18) Bosque-Sendra, J.M.; Cuadros Rodríguez, L.; Ruiz-Samblás, C.; Paulina de la

499

Mata, A. Combining chromatography and chemometrics for the characterization

500

and authentication of fats and oils from triacylglycerols compositional data-A

501

review. Anal. Chim. Acta 2012, 724, 1-11.

502

(19) López-López, A.; Cortés Delgado, A.; Garrido Fernández, A. Effect of green

503

Spanish-style processing (Manzanilla and Hojiblanca) on the quality parameters

504

and fatty acid and triacylglycerol compositions of olive fat. Food Chem. 2015,

505

188, 37-45.

506 507

(20) Martínez, J.M.; Muñoz, E.; Alba, J.; Lanzón, A. Informe sobre la utilización del analizador de rendimientos “Abencor”. Grasas Aceites 1975, 26, 379-385.

508

(21) Dobarganes, M.C.; Velasco, J.; Dieffenbacher, A. Determination of polar

509

compounds, polymerized and oxidized triacylglycerols, and diacylglycerols in oils

510

and fats. Pure Appl. Chem. 2000, 72, 1563-1575.

511

(22) EU 1348/2013. Commission Implementing Regulation (EU) No 1348/2013 of 16

512

December 2013 amending Regulation (EEC) No 2568/91 on the characteristics of

513

olive oil and olive-residue oil and on the relevant methods of analysis. OJEU 2013,

514

L338, 31–67.

515

(23) EU 2015/1833. Commission Implementing Regulation (EU) No 2015/1833 of 26

516

October 2015 amending Regulation (EEC) No 2568/91 on the characteristics of

517

olive oil and olive-residue oil and on the relevant methods of analysis. OJEU 2015,

518

L266, 29-52.

519

(24) Jolliffe, I.T. Principal components analysis. Springer-Verlag: New York, 1986.


Page 23 of 33


520

(25) Gandul-Rojas, B.; Roca, M.; Gallardo-Guerrero, L. Chlorophyls and carotenoids in

521

food products from olive tree. In Products from olive tree. Boskou, D., Clodoveo,

522

M.L. Eds.; InTech: Open Access book publisher, 2016. Available from:

523

https://mts.intechopen.com/books/products-from-olive-tree/chlorophylls-and-

524

carotenoids-in-food-products-from-olive-tree

525

(26) Pasqualone, A.; Nasti, R.; Montemurro, C.; and Gomes, T. Effect of natural style

526

processing on the oxidative and hydrolytic degradation of the lipid fraction of table

527

olives. Food Control 2014, 37, 99-103.

528

(27) Pasqualone, A.; Caponio, F.; Gomes, T.; Motemurro, C.; Summo, C.; Simeone, R.;

529

Bihmidine, S.; Kalaitzis, P. Use of innovative analytical methodologies to better

530

assess the quality of virgin olive oil. Riv. Ital. Sostanze Grasse 2005, 82, 173-176.

531 532

(28) Mariani, C.; Fedeli, E. Determination of glyceridic structures present in edible oils. Note 1. The case of olive oil. Riv. Ital. Sostanze Grasse 1985, 62, 3-7.

533

(29) Pérez-Camino, M.C.; Moreda, W.; Cert, A. Determination of diacylglycerol

534

isomers in vegetable oils by solid-phase extraction followed by gas chromatography

535

on a polar phase. J. Chromatography A 1996, 721, 305-314.

536

(30) Shimizu, M.; Kudo, N.; Nasuragi, Y.; Tokimitsu, I.; Barceló, I.; Mateu, C.;

537

Barceló, F. Acidity and DAG content of olive oils recently produced on the island

538

of Mallorca. J. Am. Oil Chem. Soc. 2008, 85, 1051-1056.

539

(31) Grobb, K.; Lanfranchi, M.; Mariani, C. Evaluation of olive oils through the fatty

540

alcohols, the sterols and their esters by coupled LG-GC. J. Am. Oil. Chem. Soc.

541

1990, 67, 629-634.



542 543

Page 24 of 33

(32) Kritchevski, D.; Chen, S. Phytosterols-health benefits and potential concerns. Nutr. Res. 2005, 25, 413-428.

544

(33) Longobardi, F.; Ventrella, A.; Casiello, G.; Sacco, D.; Catucci, L.; Agostiano, A.;

545

Kontominas, M.G. Instrumental and multivariate statistical analysis for the

546

characterization of the geographical origin of Apulian virgin olive oils. Food Chem.

547

2012, 133, 597-584.

548

(34) Kammoun, N.G.; Zarrouk, W. Exploratory chemometric analysis for the

549

characterization of Tunisian olive cultivars according to their lipid and sterolic

550

profiles. Int. J. Food Sci. Technol. 2012, 47, 1496-1504.

551

(35) Ouni, Y.; Flamini, G.; Youssef, N.B.; Guerfel, M.; Zarrouk, M. Sterolic

552

composition and triacylglycerols of Oueslati virgin olive oil: Comparison among

553

different geographic areas. Int. J. Food Sci. Technol. 2011, 46, 1747-1754.

554 555 556 557

558 559

560 561

562 563

(36) Fedeli, E. Lipids of olives. In Progress in the chemistry of fats and other lipids. Ralph, R., Holman, T. Eds.; Pergamon Press: París, 1977; 15, pp. 57-74. (37) Aparicio, R.; Harwood, J. Manual del aceite de oliva. AMV Ediciones y Mundi Prensa: Madrid, Spain, 2003. (38) Tacchino, C.E.; Borgogni, C. Research on the content of aliphatic alcohols on pressed and extraction olive oil. Riv Ital Sostanze Grasse 1983, 60, 575-581. (39) Giuffrè, A.M. Evolution of fatty alcohols in olive oils produced in Calabria (South Italy) during fruit ripening. J Oleo Sci. 2014, 63, 458-468. (40) Mariani, C.; Venturini, S. Sull’aumento delle cere durante la conservazione degli olive di oliva. Riv Ital Sostanze Grasse 1996, 73, 486-498.


Page 25 of 33


564 565

566 567

(41) Moreda, W. El significado de los ésteres alquílicos de los ácidos grasos en el aceite de oliva. Oleo 2010, 139 (Sept.), 36-37. (42) Borbolla y Alcalá, J.M R. de la. Sobre la preparación de aceitunas estilo sevillano. El tratamiento con lejía. Grasas Aceites 1981, 32, 181-189.

568

(43) Sabatini, N.; Mucciarella, M.R.; Marsilio, V. Volatile compounds in uninoculated

569

and inoculated table olives with Lactobacillus plantarum (Olea europaea, L. cv.

570

Moresca and Kalamata). LWT-Food Sci. Technol. 2008, 41, 2017-2022.

571

(44) Gómez-Coca, R.B.; Moreda, W.; Pérez-Camino, M.C. Fatty acid alkyl esters

572

presence in olive oils vs. organoleptic assessment. Food Chem. 2012, 135, 1205-

573

1209.



Page 26 of 33

574

FIGURE CAPTIONS

575

Figure 1. Fatty alcohol profiles (unweighted means) of Manzanilla and Hojiblanca olive

576

fat throughout green Spanish-style processing. Vertical bars denote 0.95 confidence

577

intervals.

578

Figure 2. Waxe profiles (unweighted means) of Manzanilla and Hojiblanca olive fat

579

throughout green Spanish-style processing. Vertical bars denote 0.95 confidence

580

intervals.

581

Figure 3. Alkyl ester profiles (unweighted means) of Manzanilla and Hojiblanca olive

582

fat throughout green Spanish-style processing. Vertical bars denote 0.95 confidence

583

intervals.


Page 27 of 33


Table 1. Polar compound profiles of Manzanilla and Hojiblanca olive fat throughout green Spanish-style processing.

Total polar Polar compound components (mg/g oil) compounds (mg/g oil) OxTG DG FFA a a a Manzanilla Fresh fruit 20.91 14.14 6.56 0.21d After lye 19.36ab 13.37a 3.15d 2.84a c b c Fermentation 17.85 11.61 3.72 2.52ab Packaging 18.27bc 11.42bc 3.75c 3.10a d c d Hojiblanca Fresh fruit 15.63 10.18 2.71 2.74ab After lye 9.11e 5.74e 1.44e 1.93bc d d b Fermentation 14.38 8.55 4.22 1.61c Packaging 14.44d 7.71d 4.15bc 2.57ab Pooled Standard Error 0.51 0.41 0.17 0.27 Cultivar

Treatment

Notes: OxTG, oxidised triacylglycerols; DG, diacylglycerols; FFA, free fatty acids+residual unsaponifiable matter. Data in each treatment are the average (unweighted mean) of two samples per replicate, except fresh fruits (two samples per cultivar). Values within columns followed by different letters are different at p≤0.05.



Page 28 of 33

Table 2. Sterol profiles (mg/kg oil or percentages, in parenthesis) of Manzanilla and Hojiblanca olive fat throughout green Spanish-style processing.

Cultivar

Treatment

Manzanilla Fresh fruit After lye

Cholesterol

Campesterol

Stigmasterol

Clerosterol

β-sitosterol

β-sitosterol (apparent)

Sitostanol

∆5-Avenasterol ∆5,24-Stigmastadienol

27.6 cd (2.5)

14.78 bc (1.4)

19.9 bc

907 de

(94.0)

22.8 a

52.8 a

18.6 a

7.0 abc (0.5)

7.95 abc

1087 cd

5.74 cd (0.6)

28.3 cd (2.8)

13.07 c (1.3)

14.5 c

864 e

(94.0)

8.9 bc

43.4 bc

9.0 c

4.8 c (0.4)

6.25 bcd

998 d

cd

b

b

a

bc

5.83

(0.5)

34.9 (2.9)

16.82 (1.4)

15.7

Packaging

5.64 cd (0.5)

31.5 bc (2.6)

19.16 a (1.6)

16.8 c

1079 c

(94.6)

15.3 b

46.8 ab

8.6 bc

a

11.52 (0.6)

a

47.3 (2.4)

b

15.93 (0.8)

24.4

ab

a

(95.5)

8.7

bc

c

bc

6.69 c (0.4)

41.7 a (2.3)

12.49 c (0.7)

26.4 a

1660 b

(95.7)

7.9 c

44.5 abc

Fermentation

d

4.79 (0.2)

a

45.3 (2.3)

b

15.62 (0.8)

24.6

ab

a

(95.7)

5.5

c

abc

Packaging

5.84 cd (0.2)

45.6 a (2.3)

16.18 b (0.8)

24.1 ab

1817 a

(95.7)

7.1 c

45.0 abc

6.9 bc

1.8

45

2.1

2.6

1.3

After lye

Pooled Standard Error Limits EU and IOC

Total

8.21 b (0.8)

Fermentation

Hojiblanca Fresh fruit

∆7-Stigmastenol ∆7-Avenasterol

0.45

2.4

0.84

(≤0.5)

(≤4.0)

(

Assessment of the Minor Component Transformations in Fat during the

Recommend Documents