Volatile profile of Spanish-style green table olives prepared from different cultivars grown at different locations (2024)

1

Volatile profile of Spanish-style green table olives prepared from different 1

cultivars grown at different locations 2

3

Amparo Cortés-Delgado, Antonio Higinio Sánchez, Antonio de Castro, Antonio López-4

López, Víctor Manuel Beato, Alfredo Montaño* 5

6

Food Biotechnology Department, Instituto de la Grasa-CSIC, Pablo Olavide University 7

Campus, building 46, Utrera road, km 1, 41013 Seville, Spain 8

9

*Tel.: +34 95 4611550, fax: +34 95 4616790, e-mail corresponding author (A. 10

Montaño): [emailprotected] 11

12

E-mail addresses for co-authors: 13

Amparo Cortés-Delgado: [emailprotected] 14

Antonio Higinio Sánchez: [emailprotected] 15

Antonio de Castro: [emailprotected] 16

Antonio López-López: [emailprotected] 17

Víctor Manuel Beato: [emailprotected] 18

19

Running title: Volatile profile of Spanish-style green table olives 20

Postprint of Food Research International Volume 83, May 2016, Pages 131-142 DOI: https://doi.org/10.1016/j.foodres.2016.03.005

2 Abstract

21

The volatile profiles of Spanish-style green table olives elaborated with Manzanilla, 22

Gordal and Hojiblanca cultivars grown at different locations in Spain were established 23

by solid phase micro-extraction (SPME) and gas chromatography coupled to mass 24

spectrometry (GC-MS). A total of 102 volatile compounds were identified, belonging to 25

distinct chemical classes, and 20 of them are reported for the first time in table olives. 26

The headspace profile was predominated by alcohols and phenols, followed by acids 27

and esters, whereas the relative amounts of the remaining classes were quite lower (< 28

5% in general). The principal compounds characterizing the headspace for most samples 29

were p-creosol, phenylethyl alcohol, acetic acid, ethanol, benzyl alcohol, ethyl acetate, 30

and (Z)-3-hexen-1-ol. Significant differences in the proportions of volatile compounds 31

between samples from the Gordal cultivar and those from Manzanilla and Hojiblanca 32

cultivars were detected and statistically visualized by principal component analysis 33

(PCA). Among all the identified compounds, only (E)-2-decenal showed significant 34

differences between the three cultivars without being significantly affected by locations 35

where the fruits were grown. 36

37

Keywords: table olives, volatile composition, olive cultivar, SPME, GC-MS, PCA 38

3 Highlights

39

● More than 100 headspace compounds were identified in Spanish-style green table 40

olives. 41

● Headspace profile of product was predominated by alcohols and phenols. 42

● PCA discriminated samples according to olive cultivar. 43

4 1. Introduction

45

Spain is the main producer of table olives in the world, with 573,371 t in the 46

season 2013/2014, and over 50% of its production corresponds to Spanish-style green 47

table olives (ASEMESA, 2015). This type of table olive is considered the main 48

fermented vegetable product in western countries. Its processing consists of a treatment 49

with alkaline lye (1.8-2.5%, w/v NaOH) to hydrolyze the bitter glucoside oleuropein, 50

followed by a washing step to remove the excess alkali. A solution of NaCl (10-13%, 51

w/v) is then added, and a lactic acid fermentation takes place (Rejano, Montaño, 52

Casado, Sánchez, & de Castro, 2010). After this step, which can last a few months, the 53

fruits are kept in the fermenter until they are marketed either in bulk with their own 54

fermenting brine or packed in small containers with an acidified cover brine. The unique 55

and pleasant flavor of this product is probably the most appreciated characteristic for 56

consumers. The flavor of table olives is closely related to both the qualitative and 57

quantitative composition of volatile compounds and can be influenced by a number of 58

factors, including olive cultivar, fruit ripeness stage, and processing method (Sabatini 59

and Marsilio, 2008). Optimal processing conditions and microbial spoilage have been 60

extensively studied for Spanish-style green table olives, yet the literature on the volatile 61

composition of this product is rather limited. Previous studies regarding this subject 62

were carried out to evaluate the major headspace compounds of olive brine (Montaño, 63

Sánchez, & Rejano, 1990), to identify the volatile compound responsible of the 64

unpleasant odor of zapatera olives (Montaño, de Castro, Rejano, & Sánchez, 1992), to 65

screen for key odor compounds in Moroccan green table olives (Iraqi, Vermeulen, 66

Benzekri, Bouseta, & Collin, 2005), to compare the volatile compounds in Spanish-67

style, Greek-style and Castelvetrano-style green olives (Sabatini and Marsilio, 2008), 68

5

and to evaluate the effects of regulated deficit irrigations on the profile of volatile 69

compounds (Cano-Lamadrid et al., 2015). 70

Today, solid-phase microextraction (SPME) followed by gas chromatography 71

coupled to mass spectrometry (GC-MS) is one of the most often used techniques for 72

analysis of volatile compounds in foods (Merkle, Kleeberg, & Fritsche, 2015). SPME-73

GG-MS has been applied to study the volatile composition of raw olives as well as of 74

different types of table olives. A total of 34 volatile compounds were identified in intact 75

raw olives from three Portuguese olive cultivars (Cobrançosa, Madural, and Verdeal 76

Transmontana), with the main contributors being hexen-1-ol, hexanal, and (Z)-3-77

hexen-1-ol acetate (Malheiro, Casal, Cunha, Baptista, & Pereira, 2015). These authors 78

demonstrated that volatile composition of olives is dependent on the olive cultivar, and 79

is highly influenced by olives maturation. In unfermented “Campo Real” table olives, 80

the main aroma compounds identified were ethanol, 2-butanol, 3-hexen-1-ol, ethyl 81

hexanoate, benzaldehyde, eucalyptol, γ-terpinene, fenchone, linalool, and terpinen-4-ol 82

(Navarro, de Lorenzo, & Pérez, 2004). In green Sicilian table olives from five different 83

cultivars (Brandofino, Castriciana, Nocellara del belice, Passalunara, and Manzanilla), a 84

total of 52 compounds were identified after 60 days of fermentation (Aponte et al., 85

2010). This study evidenced several differences in the volatile profiles among cultivars 86

and considerable changes in their profiles during storage. In the Portuguese preparation 87

known as “alcaparras” table olives, 42 volatile compounds consisting manly of 88

aldehydes were identified (Malheiro, de Pinho, Casal, Bento, & Pereira, 2011). Again, it 89

was demonstrated that the volatile profile was influenced by the olive cultivar used. In 90

Greek-style green table olives, analyses of volatile compounds by SPME-GC-MS have 91

been more numerous. Using olives from Nocellara del Belice cultivar, Martorana et al. 92

(2015) identified 49 volatile compounds, with acids, alcohols and aldehydes being 93

6

detected at the highest concentrations. A more complex volatile profile (82 volatiles) 94

was found with olives from Bella di Cerignola cultivar (De Angelis et al., 2015). A 95

comparative study between Greek-style green table olives from Giarraffa and Grossa di 96

Spagna cultivars was conducted by Randazzo et al. (2014). Notable differences among 97

volatile compounds were detected (35 compounds in Giarraffa samples vs. 24 in Grossa 98

di Spagna ones), indicating that cultivar strongly influenced the final product. Another 99

comparative study was carried by Bleve et al. (2015) in Greek-style black table olives 100

from Conservolea and Kalamàta cultivars. Forty-six compounds were identified and 101

principal component analysis (PCA) was carried out at three different fermentation 102

times. Aldehydes were closely associated with the first stage of fermentation (30 days), 103

isoamylalcohols and styrene with the middle stage (30-90 days) and ethyl esters and 104

fatty acids with the final stage (180 days). Finally, in Spanish-style table olives from 105

Manzanilla cultivar, a total of 43 volatile compounds have been identified (Cano-106

Lamadrid et al., 2015). The five most abundant volatile compounds by these authors 107

were: acetic acid, 2-decenal, tetrahydrogeraniol, 1,4-dimethoxybenzene, and 4,8-108

dimethyl-1,3,7-nonatriene. The main objective of the present work was to 109

comparatively study the volatile profile of Spanish-style green table olives produced 110

from the cultivars Manzanilla, Gordal, and Hojiblanca using the HS-SPME-GC-MS 111

technique. These three cultivars are the most prominent cultivars dedicated to table 112

olives in Spain (Hojiblanca, 51% of the total exports in 2014; Manzanilla, 33%; and 113

Gordal, 8%) (ASEMESA, 2015). Manzanilla olive is a fleshy olive with a fine texture, 114

spherical shape and medium size. Gordal olive has a very low oil content and is larger 115

than most. Hojiblanca variety is a dual-purpose olive, that is, it can be used either for 116

making oil or for table olives. In order to choose the most adequate HS-SPME 117

7

procedure based on extraction efficiency, different sample preparation procedures were 118

previously assessed. 119

120

2. Materials and Methods 121

2.1. Samples and chemicals 122

Manzanilla, Gordal, and Hojiblanca cultivars, grown at different locations with 123

ample tradition in Spanish-style table olives processing, were selected. The growing 124

locations were: Manzanilla cv: Alcalá de Guadaira (Seville), Posadas (Córdoba), and 125

Almendralejo (Badajoz), the corresponding samples were denoted by the codes MAl, 126

MC, and MAm, respectively; Gordal cv.: Utrera (Seville) and Arahal (Seville), the 127

corresponding samples were denoted by the codes GU and GA, respectively; and 128

Hojiblanca cv.: Alameda (Málaga), Estepa (Seville), and Casariche (Seville), with the 129

corresponding samples being denoted by the codes HA, HE, and HC, respectively. The 130

olives were harvested between September 23rd and October 26th, 2013, at their mature-131

green stage and transported to our laboratories for processing. At the laboratory, the 132

olives were placed in polyethylene vessels (5.2 kg fruits plus 3.4 L liquid capacity) and 133

the typical steps of Spanish-style method were carried out. An alkaline treatment was 134

carried out using a lye solution of 1.90-2.10% w/v NaOH. The olives remained in this 135

solution until the lye had penetrated two-thirds of the way through the flesh. Then, a 136

long-period water washing (11-17 h duration) was applied. The only exception was the 137

sample HE, which was subjected to two washings of 1h and 1.5 h. This change in 138

washing stage was decided in view of the rapid evolution of alkaline treatment, in order 139

to prevent possible damage of fruits due to the effect of NaOH on texture and, at the 140

same time, an excessive loss of sugars, which would affect the lactic acid fermentation.141

8

Finally, the olives were covered with brine (11.4% NaCl) and kept at room temperature 142

for fermentation. The experiments were conducted in duplicate (denoted with the 143

numbers 1 and 2 after each sample code) except for sample HC which was processed 144

only once (due to the supplied amount of olives was not sufficient to make duplicate 145

elaborations). Corrections to prevent any microbial spoilage were not necessary in the 146

case of samples from Manzanilla or Gordal cultivars, but lactic acid was added to 147

samples from the Hojiblanca cultivar at the end of fermentation in order to reach final 148

pH values lower than 4 units. After 5 months of brining, once olives were totally 149

fermented as indicated by the absence of reducing sugars according to the Fehling´s test, 150

sampling was performed for the determination of chemical and microbiological 151

characteristics, and analysis of volatile compounds. 152

Isopropyl alcohol, ethanol, 2-butanol, 1-propanol, isobutanol, 1-butanol, 153

isopentanol, 3-methyl-3-buten-1-ol, 1-pentanol, 3-methyl-2-buten-1-ol, 3-methyl-1-154

pentanol, hexanol, (Z)-3-hexen-ol, octen-3-ol, heptanol, 2-ethyl-hexanol, 1-155

octanol, 1-nonanol, benzyl alcohol, phenylethyl alcohol, ethyl acetate, methyl 156

propanoate, propyl acetate, methyl butanoate, methyl 2-methylbutanoate, isobutyl 157

acetate, methyl methylbutanoate, ethyl butanoate, ethyl 2-methylbutanoate, ethyl 3-158

methylbutanoate, isoamyl acetate, methyl hexanoate, ethyl hexanoate, hexyl acetate, 159

ethyl (E)-3-hexenoate, methyl lactate, ethyl lactate, methyl octanoate, ethyl octanoate, 160

methyl decanoate, ethyl decanoate, ethyl benzoate, benzyl acetate, methyl salicylate, 161

ethyl salicylate, propanoic acid, isobutanoic acid, butanoic acid, 2-methylbutanoic acid, 162

hexanoic acid, heptanoic acid, (E)-3-hexenoic acid, octanoic acid, decanoic acid, 163

benzoic acid, pentanal, heptanal, octanal, 2-heptenal, nonanal, (E)-2-octenal, 164

benzaldehyde, (E)- 2-decenal, limonene, 6-methyl-5-hepten-2-one, linalool oxide, 165

linalool, α-terpineol, β-damascenone, geraniol, creosol, phenol, ethyl guaiacol, p-166

9

cresol, p-propyl guaiacol, eugenol, 4-ethyl phenol, vanillin, tyrosol, octane, decane, o-167

xylene, styrene, dimethyl sulfide, theaspirane (mixture of theaspirane A and theaspirane 168

B), dimethyl sulfoxide, butyrolactone, 1,4-dimethoxybenzene, and n-alkane standards 169

(C7-C30) were supplied by Sigma-Aldrich (St. Louis, MO). Methyl (E)-3-hexenoate 170

was purchased from Across Organics (Thermo Fisher Scientific, Madrid, Spain). 171

Isoamyl lactate, methyl hydrocinnamate, and β-caryophyllene were purchased from TCI 172

Chemicals (Cymit Química SL, Barcelona, Spain). Acetic acid was purchased from 173

Panreac (Barcelona, Spain). Ultra-pure water from Milli-Q system (Millipore, Bedford, 174

MA) was used throughout. De Man, Rogosa, Sharpe (MRS) agar and oxytetracycline-175

glucose-yeast extract (OGYE) agar were from Oxoid (Basingstoke, UK). All other 176

chemicals and solvents were of analytical or chromatographic grade from various 177

suppliers. 178

179

2.2. Selection of sample preparation procedure for HS-SPME analysis 180

In order to assess the effect of the sample preparation procedure, three different 181

sample preparations were tested: (1) the extraction of 5 g of hom*ogenized olive pulp 182

plus 5 mL of ultra-pure Milli-Q water, (2) the extraction of 5 g of hom*ogenized olive 183

pulp plus 5 mL of 30% (w/v) NaCl, and (3) the extraction of a 10 g aliquot of a 184

hom*ogenized sample obtained by mixing 20 g of pulp with 20 mL of a solution 185

containing 30% (w/v) NaCl, 0.3% (w/v) ascorbic acid and 0.3% (w/v) citric acid. In all 186

cases, the experimental conditions were adjusted so that the same amount of pulp was 187

extracted. To assess the effect of sample dilution, the preparation mode 1 (dilution 1:1) 188

was compared with (a) the extraction of 3.5 g of hom*ogenized pulp plus 7 mL of ultra-189

pure Milli-Q water (dilution 1:2), and (b) the extraction of 2.5 g of hom*ogenized pulp 190

plus 7.5 mL of ultra-pure Milli-Q water (dilution 1:3). Three replicates per sample were 191

10

prepared and analyzed. All measurements were performed under constant stirring (600 192

rpm) using the following extraction conditions: equilibration time, 30 min; extraction 193

temperature, 60 ºC; and extraction time, 30 min. The volume of the sample phase (10 194

mL) in the 15 mL vial was kept constant in all assays. This minimizes the headspace 195

volume and improves extraction efficiency according to the operating instructions for 196

SPME sampling supplied by Supelco. The experiments were carried out with Spanish-197

style green olives from the Manzanilla cultivar. 198 199 2.3. HS-SPME-GC-MS analyses 200 A 1 cm, 50/30 µm divinylbenzene/carboxen/polydimethylsiloxane (DVB/CAR/PDMS) 201

StableFlex fiber (Supelco, Bellefonte, PA) was used. This triple fiber was chosen in the 202

present work in order to obtain the highest recoveries and a wider profile, according to 203

previous studies from the literature for samples of fermented table olives (Aponte et al., 204

2010) or other fermented foods (Riu-Aumatell, Miró, Serra-Cayuela, Buxaderas, 205

&López-Tamames, 2014). It combines the absorption properties of the liquid polymer 206

with the adsorption properties of porous particles and has bipolar properties. Before the 207

first use, the fiber was conditioned at 270 ºC for 1 h according to the supplier´s 208

instructions. Olives (approximately 200 g), which were separated from brine and dried 209

with a tissue, were pitted and then hom*ogenized in a blender. Aliquots of 2.5 g of 210

hom*ogenized olive pulp were placed in a 15 mL glass vial, and 7.5 mL of 30% (w/v) 211

NaCl were added. After the addition of a stirring bar (cross shaped PTFE bar of 5 mm 212

long and 10 mm diameter, for stirring at 600 rpm), the vial was closed and placed in a 213

water bath adjusted to 60 ºC. We used this relatively high temperature in order to 214

improve extraction of semi-volatile compounds, as reported in other foods with high fat 215

content such as cocoa products (Ducki, Miralles-García, Zumbé, Tornero, & Storey, 216

11

2008). After this step, fiber was manually inserted into the sample vial headspace during 217

60 min. After completion of the extraction process, the fiber was retracted prior to 218

removal from the sample vial and immediately inserted into the injection port of the GC 219

for desorption at 250 ºC for 15 min. All measurements were made in triplicate using 220

different vials. 221

All GC-MS analyses were performed on an Agilent 7890A gas chromatograph 222

coupled to an Agilent 5975C mass selective detector and GC/MSD ChemStation 223

software (version E.02.01.1177) (Agilent Technologies, Santa Clara, CA). A VF-WAX 224

MS capillary column (30 m x 0.25 mm x 0.25 μm film thickness) from Agilent was 225

used. The GC/MS conditions used were slightly modified from those described by 226

Aprea et al. (2012). The injector port (equipped with a glass liner of 0.75 mm I.D.) was 227

heated to 250 ºC. The injections were performed in the splitless mode. The carrier gas 228

was helium at a constant flow of 1 mL min-1. The initial oven temperature was 40 ºC (5 229

min), which was ramped up at 3 ºC/min to 195 ºC, and then at 10 ºC/min to 240 ºC and 230

held there for 15 min. For the mass selective detector conditions, the quadrupole, ion 231

source and transfer line temperatures were maintained at 150, 230, and 250 ºC, 232

respectively. Electron ionization mass spectra in the full-scan mode were recorded at 70 233

eV in the range 40-400 amu. Peaks were identified by comparing their mass spectra, 234

retention times and linear retention indices (RI) against those obtained from authentic 235

standards. The compounds for which it was not possible to find authentic standards 236

were tentatively identified by comparing their mass spectra with spectral data from the 237

NIST 08 MS library as well as retention indices sourced from NIST Standard Reference 238

Database. For the determination of the RI, a C7-C30 n-alkanes series was used, and the 239

values were compared, when available, with values reported in the literature for similar 240

chromatographic columns. The GC peak area of each compound was obtained from the 241

12

ion extraction chromatogram (IEC) by selecting target ions for each one. These ions 242

corresponded to base ion (m/z 100% intensity), molecular ion (M+) or another 243

characteristic ion for each molecule. Hence, some peaks that could be co-eluted in scan 244

mode can be integrated with a value of resolution greater than 1. Results were expressed 245

as percentages of the total area represented by each one of the volatile compounds. 246

247

2.4. Physico-chemical and microbiological analyses of olive brines 248

The pH, free acidity, and combined acidity of samples were measured using a 249

Metrohm 670 Titroprocessor (Herisau, Switzerland). Free acidity was determined by 250

titrating up to pH 8.3 with 0.2N NaOH and expressed as percent (w/v) of lactic acid. 251

Combined acidity was determined with 2N HCl until the pH value reached 2.6 and 252

expressed as the equivalent of sodium hydroxide per liter. Sodium chloride by titration 253

with AgNO3 and reducing sugars by the Fehling´s test were determined as described by

254

Fernández-Díez et al. (1985). Total polyphenols were measured with the Folin-255

Ciocalteau reagent following the procedure described by Casado, Sánchez, Rejano, and 256

Montaño (2007). 257

The microbial population was determined by plating the brines on the 258

appropriate solid media, both by spreading 0.1 mL onto the surface and plating their 259

decimal dilutions (in 0.1% peptone water) with a Spiral Plater (Don Whitley Sci. Ltd., 260

Shipley, England). De Man, Rogosa, Sharpe (MRS) agar with and without 0.02% 261

sodium azide was used for the lactic acid bacteria (LAB) determination, and 262

oxytetracycline-glucose-yeast extract (OGYE) agar was used for yeasts. Plates were 263

incubated at 32 ºC (MRS) or 26 ºC (OGYE) for up to 5 days, and the colony numbers 264

were recorded. 265

13 2.5. HPLC analyses

267

Organic acids (lactic, acetic and succinic acids) and ethanol were analyzed by 268

HPLC using a C18 column and deionized water (pH adjusted to 2.2 using concentrated 269

H3PO4) as the mobile phase (Sánchez, de Castro, Rejano, & Montaño, 2000).

270

Carbohydrates (sucrose, glucose, fructose, and mannitol) were determined by HPLC 271

using a Rezex RCM Monosaccharide column (Phenomenex, Torrance, CA) and 272

deionized water as the mobile phase (Casado and Montaño, 2008). 273

274

2.6. Statistical analyses 275

All the data were compiled and calculated using a combination of Microsoft 276

Excel 2010 (Microsoft Corporation, Redmond, WA) and Statistica software version 7.0 277

(Statsoft Inc., Tulsa, OK). The ANOVA test of Student-Newman-Keuls of multiple 278

comparisons of mean values was applied to the results to ascertain possible significant 279

differences among the samples studied. Significant differences were determined at the 280

p<0.05 level. In order to reveal any grouping of the table olives based on the 281

composition of volatile compounds, as well as to identify the main components 282

contained within each group, the data were subjected to principal component analysis 283

(PCA). 284

3. Results and Discussion 285

3.1. Selection of sample preparation methodology for HS-SPME analysis 286

As expected, the addition of salting-out agents such as NaCl improved extraction 287

efficiency (Fig. 1a). This can be attributed to a lower solubility of analytes in solution, 288

thus increasing the amount of sorbed analytes on the fiber (Balasubramanian and 289

14

Panigrahi, 2011). The preparation mode 3 did not have a significant effect on the 290

extraction efficiency of volatile compounds from olive pulp in comparison with 291

preparation mode 2, although a better precision was apparent. The number of detected 292

compounds for preparations 1 to 3 did not differ significantly (average ± SD was 156 ± 293

6, 154 ± 6, and 161 ± 5, respectively). Since enzymatic reactions and consequent 294

oxidation due to polyphenol-oxidases are supposedly absent in preparation mode 3 due 295

to the presence of ascorbic and citric acids (Aprea et al., 2012), the above result 296

suggests that these types of reactions (assuming these reactions occur during preparation 297

mode 2) do not contribute significantly to the volatile profile. In other words, this means 298

that the formation of volatile compounds did not occur during the hom*ogenization step 299

of olive pulp, prior to SPME. 300

Sample dilution affected the extraction efficiency (Fig. 1b), which can be due to 301

improved agitation conditions at higher sample dilutions with a noticeable influence on 302

repeatability (RSD were 15.4, 18.7, and 4.3% for dilutions 1:1, 1:2, and 1:3, 303

respectively). It is well known that, in SPME, extraction increases with the stirring rate 304

of the aqueous phase (Zhang & Pawliszyn, 1993). Taking into account these results, a 305

simple sample dilution 1:3 with 30% NaCl was chosen as the optimum sample 306

preparation for the analysis of volatile compounds from Spanish-style green table olives 307

by SPME-GC-MS. 308

309

3.2. Chemical and microbiological characteristics of table olive samples 310

311

Physico-chemical and microbiological characteristics along with concentrations 312

of free sugars, organic acids, and ethanol in the different samples of Spanish-style green 313

olives after 5 months of brining are shown in Table 1. As expected, the main 314

15

fermentation substrates (glucose, fructose, sucrose) were totally metabolized. Mannitol, 315

another free carbohydrate present in raw olives (Montaño, Sánchez, López-López, de 316

Castro, & Rejano, 2010), was detected in small amounts (<0.1% in general). Final 317

values of physicochemical parameters and fermentation end-products in all samples 318

were within the normal ranges found in brines of Spanish-style green olives in bulk 319

(Montaño, Sánchez, Casado, de Castro, & Rejano, 2003). The only exception was the 320

sample HE, which contained a relatively high content of ethanol. This is consistent with 321

a greater growth of yeasts in detriment of the LAB, which could be partly inhibited by a 322

greater content of phenolic compounds (5.2 g/L expressed as gallic acid in HE versus 323

2.5-4.0 g/L in the remaining samples). In turn, this could be a consequence of applying 324

a less efficient washing step after lye treatment of olives. 325

326

3.3. Volatile compounds of table olive samples 327

328

A total of 102 individual aroma compounds were identified using the HS-SPME-329

GC-MS technique (Table 2). The compounds were grouped into the following chemical 330

classes: alcohols, esters, acids, aldehydes, terpenes/terpenoids, phenols, hydrocarbons, 331

and other compounds. Most of the identified compounds (82 out of 102) had been 332

previously found as volatile compounds in one or various types of table olives, 333

including Spanish-style (Montaño et al., 1992; Sabatini and Marsilio, 2008; Iraqi et al., 334

2005; Cano-Lamadrid et al., 2015), Greek-style (Sabatini, Mucciarella, & Marsilio, 335

2008; Randazzo et al., 2014; Bleve et al., 2014, 2015; De Angelis et al., 2015; 336

Martorana et al., 2015), Tunisian-style (Dabbou et al., 2012), Californian-style 337

(Sansone-Land, Takeoka, & Shoemaker, 2014), “alcaparras” stoned olives from 338

Portugal (Malheiro et al., 2011), “Greek-style” Moroccon black olives (Collin et al., 339

16

2008) and “Campo Real” unfermented olives (Navarro et al., 2004). Twenty compounds 340

(isopropyl alcohol, methyl-1-pentanol, 1-octen-ol, methyl butanoate, methyl 3-341

methylbutanoate, methyl (E)-3-hexenoate, methyl lactate, isoamyl lactate, ethyl 342

salicylate, ethyl 3-cyclohexenecarboxylate, (E)-3-hexenoic acid, decanoic acid, benzoic 343

acid, dihydroedulan, isogeraniol, geraniol, iridomyrmecin, p-propyl guaiacol, eugenol, 344

and tyrosol) were reported for the first time as volatile compounds in table olives. 345

Although it is known that the last compound (tyrosol) is normally present in table olives 346

as a result of the hydrolysis of ligstroside (a heterosidic ester of tyrosol and elenolic 347

acid) (Brenes, Rejano, García, Sánchez, & Garrido, 1995), to the best of our knowledge, 348

its detection as a headspace volatile component has not been previously reported. 349

The total amounts of the identified chemical classes in the different samples are 350

shown in Fig. 2. The headspace profile was predominated by alcohols and phenols, 351

followed by acids and esters, whereas the relative amounts of the remaining classes 352

were quite lower (< 5% in general). Cano-Lamadrid et al. (2015), using an SPME 353

method similar to ours, reported that aldehydes were one of the most abundant families 354

in the volatile profiles of Spanish-style green table olives from Manzanilla cultivar. This 355

discrepancy with our study may be due to the different origins of olives and differences 356

in microbial growth during fermentation. Climatic and agronomic conditions of olive 357

growing can affect volatile composition in case of virgin olive oils obtained by the same 358

cultivar (Angerosa et al., 2004). Formation of volatile compounds in Spanish-style (or 359

Greek-style) table olives is a dynamic process that develops mainly by indigenous lactic 360

acid bacteria and yeasts, together with a variety of contaminating microorganisms 361

(Sabatini & Marsilio, 2008). However, the fermentation process is not fully predictable. 362

It has been reported that differences in the fermentation process affect the 363

concentrations of volatile compounds in Greek-style green table olives (De Angelis et 364

17

al., 2015). The ANOVA study showed that the effect of cultivar was significant 365

(p<0.05) for all chemical classes, with the exception of alcohols (data not shown). In 366

samples from the Manzanilla cultivar, there were significant differences between the 367

samples for all chemical classes, with the exception of esters and terpenes/terpenoids (). 368

In case of Gordal cultivar, significant differences between the samples were only found 369

for terpenes and “other compounds” However, we must mention that HS-SPME 370

analyses of duplicate fermenters of sample GU were coincident with the SPME fiber 371

death, which forced us to change the fiber. This could explain the high standard errors 372

for different chemical classes in this sample (Fig. 2). It is known that the change of fiber 373

in a study can negatively affect the reproducibility especially for fibers from different 374

batches (Kalua, Bedgood, & Prenzler, 2006).The most pronounced effect occurred in 375

the Hojiblanca cultivar, where significant differences between the samples occurred for 376

all chemical classes. Sample HE was characterized by a higher content of alcohols 377

compared to samples HA and HC, which agrees with the higher content of ethanol in 378

HE mentioned in the previous section. In addition, clear differences in other chemical 379

classes were found in HE in comparison with HA and HC. It appears that differences in 380

the fermentation process significantly affect the volatile profile of product. 381

Regarding individual volatile compounds, the relative amounts of 102 volatile 382

compounds, expressed as percentage of the total peak area, for the different samples are 383

shown in Table 3. Compounds are ordered according to their chemical class. 384

385

3.3.1. Alcohols 386

387

Alcohols are compounds formed from enzymatic reactions during fruit ripening 388

and from heterolactic and alcoholic fermentation during olive processing. In our study, 389

18

20 alcohols were identified, with phenylethyl alcohol (representing 8-19% of all volatile 390

compounds in the headspace), benzyl alcohol (3.1-8%), (Z)-3-hexen-1-ol (2.7-5.8%), 391

and ethanol (1.2-13.1%) being the major ones in all the samples. Phenylethyl alcohol is 392

an aromatic alcohol with a rose-like odor and occurs in many essential oils and 393

fermented foods. It is likely that this alcohol in Spanish-style green olives is formed, at 394

least in part, as a result of yeast fermentation, as yeast species such as Saccharomyces 395

cerevisiae could produce phenylethyl alcohol from L-phenylalanine (Eshkol, 396

Sendovski, Bahalul, Katz-Ezov, Kashi, & Fishman, 2009). Benzyl alcohol is naturally 397

synthesized by many plants, notably accumulating in edible fruits and tea leaves (CoE, 398

1992). The presence of 1-hexanol and (Z)-3-hexen-1-ol, which are higher alcohols from 399

the lipoxygenase pathway (Siegmund, 2015), may be due to a lipoxygenase-like 400

metabolism of polyunsaturated fatty acids, affected by enzymes produced in the brine 401

medium by lactic acid bacteria and yeasts together with other different microorganisms 402

(Sabatini & Marsilio, 2008). Ethanol can be classified as a fermentation-derived 403

compound, which is produced in table olives via yeasts and hetero-fermentative lactic 404

acid bacteria from sugars (Sabatini and Marsilio, 2008). The relatively high content of 405

ethanol in sample HE agrees with results obtained by HPLC and microbiological 406

analysis, which indicates that fermentation process is mostly produced by yeasts. As a 407

consequence, the number of alcohols showing significant differences between the 408

samples was higher in the Hojiblanca cultivar (a significant effect was found for 19 out 409

of 20 alcohols) compared to Manzanilla (12 out of 20) and Gordal (6 out of 20) 410 cultivars (Table 3). 411 412 3.3.2. Esters 413 414

19

The largest group of volatile compounds identified in our study was esters of 415

which there were 29 compounds. Volatile esters are major components of the aroma of 416

all fruits, and are sometimes mainly responsible for the pleasant flavor appreciated by 417

consumers (Sabatini and Marsilio, 2008). Their formation and content mainly depend 418

on the number of alcohols and acids. Acetate esters and propanoate esters could be 419

synthesized by the esterification of volatile alcohols with acetyl-CoA and propionyl-420

CoA, respectively (Sabatini and Marsilio, 2008). Ethyl and methyl esters were the most 421

numerous esters, with ethyl acetate being the dominant compound in all samples 422

(representing 0.8-8% of all volatile compounds). Ethyl lactate was relatively important 423

in samples from the Gordal cultivar and sample HE (representing more than 1% of all 424

volatile compounds). In the latter sample, this could be explained by its high content of 425

ethanol, as shown in Table 1. The presence in sample HE of relatively high amounts of 426

ethyl octanoate (6%), ethyl hexanoate (2%), ethyl decanoate (1.2%), and methyl 427

octanoate (1%) is noteworthy. As occurred with alcohols, the number of alcohols 428

showing significant differences between the samples was higher in olives from 429

Hojiblanca and Manzanilla cultivars compared to Gordal cultivar (Table 3). 430

431

3.3.3. Volatile acids 432

433

Within the family of volatile acids, 11 compounds were identified. Acetic acid 434

was the dominant acid in all cases, representing 8-14% of all volatile compounds in the 435

headspace. It is known that this acid is formed in olives during the lye treatment step, 436

presumably from fragmentation by alkali from other compounds, and during the 437

fermentation step (Sánchez et al., 2000). The content of propanoic acid was relatively 438

high in samples MC, MAm, GU, HA, and HC (6-12%); in the remaining samples its 439

20

content was low (0-1%). The formation and content of this acid depends on the growth 440

of Propionibacterium species, characteristic of the “fourth stage” of fermentation in 441

Spanish-style table olives (Montaño et al., 2003). Among the remaining acids identified, 442

it is worth mentioning that 2-methylbutanoic acid was present at a 0.6-2.5% level. For 443

each cultivar, its content was significantly different between the samples studied (Table 444

3). On the contrary, the content of benzoic acid was not significantly different between 445

the samples in any cultivar. However, since no significant differences in benzoic acid 446

were found among cultivars according to ANOVA (data not shown), this acid is not 447

considered a good candidate for marker of olive cultivar in Spanish-style green olives. 448

Hexanoic, octanoic and decanoic acids were present at relatively high amounts in HE 449

compared to the other samples, which is consistent with the high contents of the 450

corresponding ethyl esters, as mentioned above. 451

452

3.3.4. Aldehydes 453

454

Among 8 aldehydes identified, benzaldehyde was the most abundant, 455

representing 0.5-1.2% of all volatile compounds. This aldehyde may result from 456

enzymatic reactions during fruit ripening, and is present in intact raw olives (Malheiro, 457

Casal, Cunha, Baptista, & Pereira, 2015), but benzaldehyde formation during the 458

fermentation phase of Spanish-style olives should not be ruled out. Lactobacillus 459

plantarum, the main species of LAB during fermentation, has been reported to convert 460

phenylalanine to benzaldehyde (Nierop Groot and De Bont, 1998). For each olive 461

cultivar, the benzaldehyde content in Spanish-style olives was significantly different 462

between the samples studied (Table 3). The contrary occurred in the case of octanal and 463

(E)-2-decenal. In addition, the mean content of the latter compound was significantly 464

21

different among cultivars (Manzanilla > Gordal > Hojiblanca). It suggests that (E)-2-465

decenal could be used as a potential marker of olive cultivar in Spanish-style green 466

olives. However, a greater amount of data is necessary to corroborate this hypothesis. 467

468

3.3.5. Terpene/terpenoids 469

470

Terpene compounds consisted of oxygenated as well as non-oxygenated 471

monoterpenes, sesquiterpenes and irregular terpenes, which all occurred in relatively 472

low amounts in the headspace of the samples. All of these compounds can be classified 473

as olive-derived compounds. The oxygenated monoterpenes detected included the 474

alcohols linalool, linalool oxide, α-terpineol, geraniol, and isogeraniol; and the iridoid 475

monoterpene iridomyrmecin. The non-oxygenated terpenes consisted of the common 476

monoterpene limonene and the sesquiterpenes copaene, caryophyllene, 477

cycloisosativene, and α-muurolene. The irregular terpenes detected included 6-methyl-478

5-hepten-2-one, dihydroedulan, and β-damascenone, which are most likely formed from 479

carotenoids. Of the 14 terpene compounds identified, dihydroedulan, geraniol, 480

isogeraniol, and iridomyrmecin were identified for the first time in table olives. As 481

found for other chemical classes, the contents of most terpene compounds were not 482

significantly different between the samples from Gordal cultivar (Table 3). However, 483

Manzanilla and Hojiblanca cultivars showed significant changes in most of the terpene 484

compounds. In particular, copaene and α-muurolene showed significant changes 485

between the samplesfor all three cultivars. Although the size of our data set is too small 486

and the stability of these compounds during table olive processing has not been 487

evaluated, it suggests that these sesquiterpenes could be considered as potential 488

molecular marker candidates or play a role in determining the authenticity and 489

22

protection of regional produce. In fact, copaene and muurolene, along with α-490

farnesene, have been proposed as markers of extra virgin olive oil origin (Damascelli 491 and Palmisano, 2013). 492 493 3.3.6. Volatile phenols 494 495

Within the volatile phenols, 9 compounds were identified including 5 guaiacol 496

derivatives (p-creosol, p-ethyl guaiacol, p-propyl guaiacol, eugenol, vanillin) and 4 497

phenol derivatives (phenol, p-cresol, 4-ethyl phenol, tyrosol). The most abundant 498

compound in all the samples was p-creosol (26-37%, except in sample HE). The content 499

of 4-ethyl phenol was also relatively important in all samples (1-5%). Most of these 500

compounds are likely formed during the fermentation process as a result of the activity 501

of microorganisms. Thus, the presence of volatile phenols in olive oils with strong 502

fusty, musty, and muddy defects as well as in stored olive paste has been attributed to 503

microbial activity (Vichi, Romero, Gallardo-Chacón, Tous, López-Tamames, & 504

Buxaderas, 2009). It is known that certain strains of LAB, L. plantarum among them, 505

are able to produce volatile phenols from the metabolism of phenolic acids (Silva, 506

Campos, Hogg, & Couto, 2011). Changes in individual phenols were particularly 507

important in the case of the Manzanilla cultivar (significant differences between the 508

samples were found for 7 out of 9 phenols, Table 3). For each cultivar, the contents of 509

tyrosol were not significantly different between the samples studied. 510

511

3.3.7. Hydrocarbons 512

23

Five hydrocarbons (octane, decane, o-xylene, styrene, and 2-bornene), all of 514

them previously detected in table olives, were identified in all samples. In general, 2-515

bornene, a bridge cyclic hydrocarbon previously detected in Spanish-style olives (Iraqi 516

et al., 2005) was the most abundant (0.5-1.6%), with contents significantly different 517

between the samples from Manzanillaor Hojiblanca cultivar (Table 3). 518

519

3.3.8. Other volatile compounds 520

521

Finally, other volatile compounds identified in our study were dimethyl sulfide, 522

dimethyl sulfoxide, butyrolactone, 1,4-dimethoxybenzene, and the stereoisomeric 523

compounds theaspirane A and B. In general, the major compounds were dimethyl 524

sulfide and theaspirane. On the other hand, 1,4-dimethoxybenzene was only found in 525

samples from the Hojiblanca cultivar and at low concentrations (0.04-0.05%), but 526

further research is needed to know if this compound could be considered as a potential

527

marker candidate of Spanish-style green olives elaborated with the Hojiblanca cultivar.

528

Dimethyl sulfide contents were significantly different between the samples from 529

Manzanilla or Hojiblanca cultivar. Mean contents of theaspirane A and B were 530

significantly different among cultivars (Hojiblanca > Manzanilla > Gordal) while the 531

differences between the samples studied for each cultivar were small or not significant. 532

533

3.3.9. Principal component analysis (PCA) of volatile compounds 534

535

PCA was performed using the contents of individual volatile compounds as the 536

variables. For this study, the sample HE was not considered due to its distinct 537

processing operations and final characteristics in comparison with the other samples. 538

24

The first two principal components accounted for 52.36% of the variation in the data. 539

The score plot showed that three separate groups were clearly visible (Fig 3a): all 540

samples from the Gordal cultivar (GU1, GU2, GA1, GA2) formed one group; all 541

samples from the Manzanilla cultivar (MC1, MC2, MAl1, MAl2, MAm1, MAm2) 542

formed a second group, and the third group was composed of samples from the 543

Hojiblanca cultivar (HA1, HA2, HC). The locations where the fruits were grown for a 544

given cultivar were not clearly distinguished, indicating that the fruit growing 545

environment had a minor influence on the volatile composition of Spanish-style green 546

table olives. Similarly, in virgin olive oil, it has been reported that cultivar is the 547

dominant factor in the formation of the aroma whereas the fruit grown environment has 548

little effect (Angerosa et al., 2004). The loading plot (Fig. 3b) showed that the volatile 549

compounds mainly associated with the first group were the esters ethyl acetate (21), 550

ethyl lactate (38) and ethyl benzoate (44). The second group was particularly related 551

to1-octanol (17), phenylethyl alcohol (20) and (E)-2-decenal (68), while the third group 552

was mainly related to propyl acetate (23) and 1,4-dimethoxybenzene (102). 553

554

4. Conclusions 555

In this study, the volatile profiles of Spanish-style green olives prepared from 556

Manzanilla, Gordal and Hojiblanca cultivars each grown at different locations in Spain 557

were evaluated using HS-SPME-GC-MS. All samples presented complex aroma 558

profiles rich in different families of aroma compounds, mainly alcohols and phenols. 559

More than 100 volatile compounds distributed over different chemical groups were 560

identified in the pulp of olives. The major volatile compounds characterizing the 561

headspace for most samples were: p-creosol, phenylethyl alcohol, acetic acid, ethanol, 562

benzyl alcohol, ethyl acetate, and (Z)-3-hexen-1-ol. Based on the content of individual 563

25

volatile compounds and PCA, the samples were clearly separated according to their 564

olive cultivar. However, the different locations of samples for each cultivar were poorly 565

distinguished. The contents of benzoic acid, octanal, (E)-2-decenal, and tyrosol were 566

not significantly different between the samples studied for each cultivar, but only (E)-2-567

decenal showed significant differences among the three cultivars. Therefore, this 568

aldehyde would be a promising candidate as marker of olive cultivar in Spanish-style 569

green table olives. However, further studies are needed to support the results obtained 570

by this first screening. Apart from this, new experiments are in progress in our 571

laboratories to determine the contribution of each volatile compound to the 572

characteristic aroma of Spanish-style green table olive and to elucidate the relationship 573

between aroma compounds and sensory attributes. 574

575

Acknowledgements 576

This work was supported in part by the European Union (FEDER funds) and the 577

Spanish government through Project AGL2014-54048-R. 578

579

References 580

Angerosa, F., Servili, M., Selvaggini, R., Taticchi, A., Esposto, S., & Montedoro, G. 581

(2004). Volatile compounds in virgin olive oil: occurrence and their relationship 582

with the quality. Journal of Chromatography A, 1054, 17-31. 583

Aponte, M., Ventorino, V., Blaiotta, G., Volpe, G., Farina, V., Avellone, G., Lanza, 584

C.M., & Moschetti, G. (2010). Study of green Sicilian table olive fermentations 585

26

through microbiological, chemical and sensory analyses. Food Microbiology, 586

27, 162-170. 587

Aprea, E., Corollaro, M.L., Betta, E., Endrizzi, I., Demattè, M.L., Biasioli, F., & 588

Gasperi, F. (2012). Sensory and instrumental profiling of 18 apple cultivars to 589

investigate the relation between perceived quality and odour and flavor. Food 590

Research International, 49, 677-686. 591

ASEMESA (Spanish Association of Producers and Exporters of Table Olives)). Datos 592

generales del sector, 2015. URL 593

http://www.asemesa.es/content/datos_generales_del_sector. Accessed 05.02.16. 594

Balasubramanian, S., & Panagrahi, S. (2011). Solid-phase microextraction (SPME) 595

techniques for quality characterization of food products: a review. Food and 596

Bioprocess Technology, 4, 1-26. 597

Bleve, G., Tufariello, M., Durante, M., Grieco, F., Ramires, F.A., Mita, G., Tasioula-598

Margari, M., & Logrieco, A.F. (2015). Physico-chemical characterization of 599

natural fermentation process of Conservolea and Kalamàta table olives and 600

development of a protocol for the pre-selection of fermentation starters. Food 601

Microbiology, 46, 368-382. 602

Bleve, G., Tufariello, M., Durante, M., Perbellini, E., Ramires, F.A., Grieco, F., 603

Cappello, M.S., De Domenico, S., Mita, G., Tasioula-Margari, M., & Logrieco, 604

A.F. (2014). Physico-chemical and microbiological characterization of 605

spontaneous fermentation of Cellina di Nardò and Leccino table olives. 606

Frontiers in Microbiology, 5, 570. doi: 10.3389/fmicb.2014.00570. 607

Brenes, M., Rejano, L., García, P., Sánchez, A.H., & Garrido, A. (1995). Biochemical 608

changes in phenolic compounds during Spanish-style green olive processing. 609

Journal of Agricultural and Food Chemistry, 43, 2702-2706. 610

27

Cano-Lamadrid, M., Giron, I.F., Pleite, R., Burlo, F., Corell, M., Moriana, A., & 611

Carbonell-Barrachina, A.A. (2015). LWT- Food Science and Technology, 62, 19-612

26. 613

Casado, F. J., & Montaño, A. (2008). Influence of processing conditions on acrylamide 614

content in black ripe olives. Journal of Agricultural and Food Chemistry, 56, 615

2021-2027. 616

Casado, F.J., Sánchez, A.H., Rejano, L., & Montaño, A. (2007). Estudio de nuevos 617

procedimientos de elaboración de aceitunas verdes tratadas con álcali, no 618

fermentadas, conservadas mediante tratamientos térmicos. Grasas y Aceites, 58, 619

275-282. 620

CoE (Council of Europe). Flavouring substances and natural sources of flavourings, 621

1992. 4th Edition, Volume I, Chemically-defined flavouring substances, 622

Strasbourg. 623

Collin, S., Nizet, S., Muls, S., Iraqi, R., & Bouseta, A. (2008). Characterization of odor-624

active compounds in extracts obtained by simultaneous extraction/distillation 625

from Moroccan black olives. Journal of Agricultural and Food Chemistry, 56, 626

3273-3278. 627

Dabbou, S., Issaoui, M., Brahmi, F., Nakbi, A., Chehab, H., Mechri, B., & Hammami, 628

M. (2012). Changes in volatile compounds during processing of Tunisian-Style 629

table olives. Journal of the American Oil Chemists' Society, 89, 347-354. 630

Damascelli, A., & Palmisano, F. (2013). Sesquiterpene fingerprinting by headspace

631

SPME–GC–MS: Preliminary study for a simple and powerful analytical tool for

632

traceability of olive oils. Food Analytical Methods, 6, 900–905.

28

De Angelis, M., Campanella, D., Cosmai, L., Summo, C., Rizzello, C.G., & Caponio, F. 634

(2015). Microbiota and metabolome of un-started and started Greek-type 635

fermentation of Bella di Cerignola table olives. Food Microbiology, 52, 18-30. 636

Ducki, S., Miralles-Garcia, J., Zumbé, A., Tornero, A., & Storey, D.M. (2008). 637

Evaluation of solid-phase micro-extraction coupled to gas chromatography–mass 638

spectrometry for the headspace analysis of volatile compounds in cocoa 639

products. Talanta, 74, 1166-1174. 640

Eshkol, N., Sendovski, M., Bahalul, M., Katz-Ezov, T., Kashi, Y., & Fishman, A. 641

(2009). Production of 2-phenylethanol from L-phenylalanine by a stress tolerant 642

Saccharomyces cerevisiae strain. Journal of Applied Microbiology, 106, 534-643

542. 644

Fernández-Díez, M. J., de Castro, R., Garrido, A., Cancho, F., González-645

Pellissó, F., Vega, M.N., Moreno, A.H., Mosquera, I.M., Rejano L., Durán, M. 646

C., Roldán, F.S., García, P., & de Castro, A. (1985). Biotecnología de la 647

Aceituna de Mesa. Madrid: Consejo Superior de Investigaciones Científicas. 648

Iraqi, R., Vermeulen, C., Benzekri, A., Bouseta, A., & Collin, S. (2005). Screening for 649

key odorants in Moroccan green olives by gas chromatography-650

olfactometry/aroma extract dilution analysis. Journal of Agricultural and Food 651

Chemistry, 53, 1179-1184. 652

Kalua, C.M., Bedgood, D.R., & Prezler, P.D. (2006). Development of a headspace solid 653

phase microextraction-gas chromatography method for monitoring volatile 654

compounds in extended time-course experiments of olive oil. Analytica Chimica 655

Acta, 556, 407-414. 656

Malheiro, R., Casal, S., Cunha, S.C., Baptista, P., & Pereira, J.A. Olive volatiles from 657

Portuguese cultivars Cobrançosa, madural and Verdeal Transmontana: role in 658

29

oviposition preference of Bactrocera oleae (Rossi) (Diptera: Tephritidae). PLoS 659

ONE 10(5): e0125070, doi: 10.1371/journal.pone.0125070. May 18, 2015. 660

Malheiro, R., de Pinho, P.G., Casal, S., Bento, A., & Pereira, J.A. (2011). 661

Determination of the volatile profile of stoned table olives from different 662

varieties by using HS-SPME and GC/IT-MS. Journal of the Science of Food and 663

Agriculture, 91, 1693-1701. 664

Martorana, A., Alfonzo, A., Settanni, L., Corona, O., La Croce, F., Caruso, T., 665

Moschetti, G., & Francesca, N. (2015). An innovative method to produce Green 666

table olives based on “pied de cuve” technology. Food Microbiology, 50, 126-667

140. 668

Merkle, S., Kleeberg, K.K., & Fritsche, J. (2015). Recent developments and 669

applications of solid phase microextraction (SPME) in food and environmental 670

analysis—A review. Chromatography, 2, 293-381. 671

Montaño, A., de Castro, A., Rejano, L., & Sánchez, A.H. (1992). Analysis of zapatera 672

olives by gas and high-performance liquid chromatography. Journal of 673

Chromatography, 594, 259-267. 674

Montaño, A., Sánchez, A.H., & Rejano, L. (1990). Rapid quantitative analysis of 675

headspace components of green olive brine. Journal of Chromatography, 521, 676

153-157. 677

Montaño, A., Sánchez, A.H., Casado, F.J., de Castro, A., & Rejano, L. (2003). 678

Chemical profile of industrially fermented green olives of different varieties. 679

Food Chemistry, 82, 297-302. 680

Montaño, A., Sánchez, A.H., López-López, A., de Castro, A., & Rejano, L. (2010). 681

Chemical composition of fermented green olives: acidity, salt, moisture, fat, 682

protein, ash, fiber, sugar, and polyphenol. In V.R. Preedy, & R.R. Watson 683

30

(Eds.), Olives and olive oil in health and disease prevention (pp. 291-297). 684

Amsterdam: Elsevier Inc. 685

Navarro, T., de Lorenzo, C., & Pérez, R.A. (2004). SPME analysis of volatile 686

compounds from unfermented olives subjected to termal treatment. Analytical 687

and Bioanalytical Chemistry, 379, 812-817. 688

Nierop Groot, M.N. and De Bont, J.A.M. (1998). Conversion of phenylalanine to 689

benzadehyde initiated by an aminotransferase in Lactobacillus plantarum. 690

Applied and Environmental Microbiology, 64, 3009-3013. 691

Randazzo, C.L., Todaro, A., Pino, A., Pitino, I., Corona, O., Mazzaglia, A., & Caggia, 692

C. (2014). Giarraffa and Grossa di Spagna naturally fermented table olives: 693

Effect of starter and probiotic cultures on chemical, microbiological and sensory 694

traits. Food Research International, 62, 1154-1164. 695

Rejano, L., Montaño, A., Casado, F.J., Sánchez, A.H., & de Castro, A. (2010). Table 696

olives: varieties and variations. In V.R. Preedy, & R.R. Watson (Eds.), Olives 697

and olive oil in health and disease prevention (pp. 5-15). Amsterdam: Elsevier 698

Inc. 699

Riu-Aumatell, M., Miró, P., Serra-Cayuela, A., Buxaderas, S. & López-Tamames, E. 700

(2014). Assessment of the aroma profiles of low-alcohol beers using HS-SPME-701

GC-MS. Food Research International 57,196-202. 702

Sabatini, N., & Marsilio, V. (2008). Volatile compounds in table olives (Olea Europaea 703

L., Nocellara del Belice cultivar). Food Chemistry, 107, 1522-1528. 704

Sabatini, N., Mucciarella, M.R., & Marsilio, V. (2008). Volatile compounds in 705

uninoculated and inoculated table olives with Lactobacillus plantarum (Olea 706

europea L., cv. Moresca and Kalamata). LWT- Food Science and Technology, 707

41, 2017-2022. 708

31

Sánchez, A.H., de Castro, A., Rejano, L., & Montaño, A. (2000). Comparative study on 709

chemical changes in olive juice and brine during green olive fermentation. 710

Journal of Agricultural and Food Chemistry, 48, 5975-5980. 711

Sansone-Land, A., Takeoka, G.T., & Shoemaker, C.F. (2014). Volatile constituents of 712

commercial imported and domestic black-ripe table olives (Olea europaea). 713

Food Chemistry, 149, 285-295. 714

Siegmund, B. (2015). Biogenesis of aroma compounds: flavor formation in fruits and 715

vegetables. In J.K. Parker, J.S. Elmore, & L. Methven (Eds.), Flavor 716

development, analysis and perception in food and beverages (pp. 127-144). 717

Amsterdam: Elsevier Inc. 718

Silva, I., Campos, F.M., Hogg, T., & Couto, J.A. (2011). Wine phenolic compounds 719

influence the production of volatile phenols by wine-related lactic acid bacteria. 720

Journal of Applied Microbiology, 111, 360–370. 721

Vergara, J.V., Blana, V., Mallouchos, A., Stamatiou, A., & Panagou, E.Z. (2013). 722

Evaluating the efficacy of brine acidification as implemented by the Greek table 723

olive industry on the fermentation profile of Conservolea green olives. LWT- 724

Food Science and Technology, 53, 113-119. 725

Vichi, S., Romero, A., Gallardo-Chacón, J., Tous, J., López-Tamames, E., & 726

Buxaderas, S. (2009). Influence of olives’ storage conditions on the formation of 727

volatile phenols and their role in off-odor formation in the oil. Journal of 728

Agricultural and Food Chemistry, 57, 1449–1455. 729

Zhang, Z., & Pawliszyn, J. (1993). Headspace solid-phase microextraction. Analytical 730

Chemistry, 65, 1843-1852. 731

32 FIGURE CAPTIONS

732 733

Figure 1. Influence of (a) sample preparation mode and (b) sample dilution on HS-734

SPME extraction efficiency of volatile compounds in pulp of Spanish-style green table 735

olives (Manzanilla cultivar) under constant stirring (600 rpm). Experiments are 736

described in text. Error bars indicate 95% confidence intervals (n = 3). Sample 737

preparation modes: (1) 5 g of pulp + 5 mL of water, (2) 5 g of pulp + 5 mL of 30% 738

NaCl, and (3) 10 g aliquot of a mix composed of 20 g of pulp + 20 mL of a solution 739

containing 30% (w/v) NaCl, 0.3% (w/v) ascorbic acid and 0.3% (w/v) citric acid. 740

741

Figure 2. Chemical classes of the volatile compounds in Spanish-style green olives 742

obtained with olives from cultivars Manzanilla, Gordal, and Hojiblanca grown at 743

different locations. Sample codes are described in text. Error bars indicate 95% 744

confidence intervals (n = 6). 745

746

Figure 3. Principal component analysis (PCA) performed on individual volatile 747

compounds: (a) distinction between the samples (scores); (b) relationships between the 748

variables (loadings). 749

0.0E+002.0E+074.0E+076.0E+078.0E+071 2 3Total peak area

Sample preparation mode

0.0E+001.0E+072.0E+073.0E+074.0E+07 1:1 1:2 1:3Total peak areaDilution (a) (b)

01020304050Content (%)

Alcohols

051015202530

Esters

051015202530Content (%)

Acids

0.00.51.01.52.02.53.0Content (%)

Aldehydes

0510Content (%)

Terpenes

01020304050

Phenols

0.01.02.03.04.0

Hydrocarbons

0.01.02.03.04.05.0

Others

Table 1. Physico-chemical and microbiological characteristics, and substrates and end-products of fermentation after 5 months of briningSamplesa

MC MAl MAm GA GU HA HE HC

Physico-chemical parameter

pH 4.03± 0.01c 3.80± 0.04b 4.01 ±0.03c 3.69 ±0.02a 3.68 ±0.04a 3.88 ±0.01b 3.85 ±0.04b 3.91 ±0.00b Titratable acidity (%) 0.82 ±0.02a 0.9 ±0.1ab 0.71 ±0.04a 1.05 ±0.01ab 0.98 ±0.00ab 1.2 ±0.1bc 1.3 ±0.1c 1.0 ±0.0ab Combined acidity (N) 0.11 ±0.00c 0.09 ±0.00a 0.10 ±0.00b 0.09 ±0.00a 0.09 ±0.00a 0.13 ±0.00d 0.14 ±0.00e 0.11 ±0.00c Salt (% NaCl) 5.6 ±0.2a 5.95 ±0.01b 6.0 ±0.1b 5.8 ±0.1ab 6.3 ±0.1c 5.59± 0.01a 5.7 ±0.1ab 5.70 ±0.00ab Total phenols (g/L gallic

acid)

3.6±0.3c 4.0±0.3d 3.4±0.5c 2.9±0.3b 2.5±0.2a 2.9±0.5b 5.2±0.4e 3.4 ±0.2c

Microbial population (log cfu/mL)

Lactic acid bacteria 6.8±0.1d 6.4±0.1c 6.6±0.2cd 6.4±0.0c 5.7±0.2b 5.7±0.2b <3.3a 5.9 ± 0.0b

Yeasts 2.2±0.3a 2.7±0.1a 4.0±0.1b 4.1±0.3b 3.7±0.8ab 4.8±0.8bc 5.6±0.0c 3.4 ± 0.2b

Substrates and end-products of fermentation (g/L)b

Mannitol 0.05 ±0.00a 0.09 ±0.01a 0.05 ±0.01a 1 ± 1a 0.7 ±0.5a 0.05 ±0.00a 0.6 ±0.3a 0.07 ±0.01a Lactic acid 13.66± 0.08b 14.3±0.9b 11.4± 0.5a 18.6± 0.1d 16.8± 0.3c 17.2± 0.6c 19.9± 0.3e 15.1 ±0.1b Acetic acid 1.92 ±0.05ab 0.96±0.01b 1.44 ±0.03ab 1.0±0.2a 0.9±0.2a 2.2± 0.2b 1.7 ±0.1ab 2.4 ±0.1b Succinic acid 0.50± 0.02bc 0.11±0.01a 0.30± 0.01ab 0.29 ±0.02ab 0.3±0.1a 0.29 ± 0.00ab 0.6± 0.2c 0.34 ±0.02ab Ethanol 0.61 ± 0.06 a 0.3 ± 0.1a 1.1 ± 0.1a 1.3 ± 0.2a 1.1 ± 0.2a 0.21 ± 0.02a 3.2 ± 0.9b 0.25 ±0.01a

a

Values are means ± SD of two fermenters, each analysed in duplicate, except for sample HC whose values are from one fermenter. b Glucose, fructose and sucrose were not detected in any case. Data in the same row with different letters are significantly different (p < 0.05).

Table 2. Volatile compounds in the headspace of Spanish-style green table olives, processed with olives from different cultivars grown at different locations, identified in the present study.

Code Compound LRI a IEC (m/z)b IDc Ref.dAlcohols 1 Isopropyl alcohol 928 45 A - 2 Ethanol 935 45 A 1,3,4,5,10,13,15,16,17 3 2-Butanol 1025 59 A 3,4,5,7,10,13,15,17 4 1-Propanol 1039 59 A 3,4,5,7,8,13 5 Isobutanol 1107 43 A 3,7,8,13,17 6 1-Butanol 1153 56 A 3,7,13,17 7 Isopentanol 1211 55 A 2,3,6,7,8,9,11,13,14,16,17 8 3-Methyl-3-buten-1-ol 1256 68 A 7,8,17 9 1-Pentanol 1255 55 A 2,3,7,8,9,12,13 10 3-Methyl-2-buten-1-ol 1324 71 A 7,8,12,17 11 3-Methyl-1-pentanol 1329 56 A - 12 1-Hexanol 1356 56 A 1,2,3,6,7,8,9,10,11,13,14,16,17 13 (Z)-3-Hexen-1-ol 1385 67 A 1,2,3,6,7,8,9,11,12,13,14,15,17 14 1-Octen-3-ol 1454 57 A - 15 1-Heptanol 1458 70 A 2,7,8,9,10,12 16 2-Ethyl-1-hexanol 1491 57 A 9,16,17 17 1-Octanol 1560 84 A 1,6,9,12,17 18 1-Nonanol 1661 70 A 2,8,9,12 19 Benzyl alcohol 1871 108 A 2,6,7,8,9,12,14,17 20 Phenylethyl alcohol 1903 91 A 1,2,6,9,12,14,16,17 Esters 21 Ethyl acetate 897 43 A 3,4,5,7,8,12,13,16,17 22 Methyl propanoate 911 57 A 16,17

Acids 50 Acetic acid 1460 60 A 1,3,6,7,8,12,13,16,17 51 Propanoic acid 1549 74 A 1,3,6,7,8,13,16,17 52 Isobutanoic acid 1581 43 A 2,7,8,14,16,17 53 Butanoic acid 1640 73 A 6,16,17 54 2-Methylbutanoic acid 1680 74 A 1,7,8,14,17 55 Hexanoic acid 1854 73 A 6,17 56 Heptanoic acid 1959 73 A 6,17 57 (E)-3-Hexenoic acid 1966 114 A - 58 Octanoic acid 2065 73 A 6,10,12,17 59 Decanoic acid 2277 73 A - 60 Benzoic acid 2436 105 A - Aldehydes 61 Pentanal 980 44 A 9,11 62 Heptanal 1177 70 A 8,9,10,11 63 Octanal 1284 84 A 1,2,6,9,11,12,16,17 64 2-Heptenal 1313 83 A 1,2,9,11,14 65 Nonanal 1388 98 A 1,2,3,6,7,8,9,11,13,14,16,17 66 (E)-2-Octenal 1420 55 A 9,11,12 67 Benzaldehyde 1511 106 A 1,2,6,7,8,9,10,11,14,15,16,17 68 (E)- 2-Decenal 1633 70 A 1,6,8,9,10,11,12,14

Terpenes and terpenoids

69 Limonene 1182 93 A 1,2,9,10,11,14,17

70 6-Methyl-5-hepten-2-one 1334 108 A 1,2,7,9,10,11,17

96 2-Bornene 1505 121 C 2 Other compounds 97 Dimethyl sulfide 765 62 A 1,5,7 98 Theaspirane A 1484 138 A 9 99 Theaspirane B 1524 138 A 9 100 Dimethyl sulfoxide 1556 63 A 2 101 Butyrolactone 1613 42 A 6,7,8 102 1,4-Dimethoxybenzene 1730 123 A 1 a

Linear retention index on VF-Wax column. b

Ion extraction chromatogram, m/z used to obtain the GC peak area of each compound. c

Identification: A, identified, mass spectrum and RI were in accordance with standards; B, tentatively identified, mass spectrum matched in the standard NIST 2008 library and RI matched with the NIST Standard Reference Database (NIST Chemistry WebBook); C, tentatively identified, mass spectrum agreed with the standard NIST 2008.

d

Previously reported as volatile compound in typical Spanish-style green table olives (1, Cano-Lamadrid et al., 2015; 2, Iraqi et al., 2005; 3, Sabatini and Marsilio, 2008; 4, Montaño et al., 1992; 5, Vergara et al., 2013) or other preparations of table olives (6, Martorana et al., 2015; 7, Bleve et al., 2015; 8, Bleve et al., 2014; 9,Sansone-Land et al., 2014; 10, Dabbou et al., 2012; 11, Malheiro et al., 2011; 12, Aponte et al., 2010; 13, Sabatini et al, 2008; 14, Collin et al., 2008; 15, Navarro et al., 2004; 16, Randazzo et al., 2014; 17, De Angelis et al., 2015). -, not reported.