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
MC1MC2MAl1MAl2MAm1MAm2GA1GA2 GU1GU2HA1HA2HC-15 -10 -5 0 5 10 15 20Factor 1: 32.66%-14-12-10-8-6-4-2024681012Factor 2: 19.70%(a) 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60 61 62 63 64 65 66 67 68 69 70 71 72 73 74 75 76 77 78 79 80 81 82 83 84 85 86 87 88 89 90 91 92 93 94 95 96 97 98 99 100 101 102-1.0 -0.5 0.0 0.5 1.0Factor 1 : 32.66%-1.0-0.50.00.51.0Factor 2 : 19.70%(b)
(37)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
23 Propyl acetate 976 61 A 1,3,7,8,9,13,17 24 Methyl butanoate 989 87 A - 25 Methyl 2-methylbutanoate 1010 88 A 17 26 Isobutyl acetate 1013 43 A 8,17 27 Methyl 3-methylbutanoate 1018 74 A - 28 Ethyl butanoate 1033 71 A 7,8,16,17 29 Ethyl 2-methylbutanoate 1048 102 A 7,8,9,11,14,17 30 Ethyl 3-methylbutanoate 1064 88 A 7,8,9,14,17 31 Isoamyl acetate 1118 43 A 7,8,9,11,14,17 32 Methyl hexanoate 1185 74 A 7,11,12,17 33 Ethyl hexanoate 1231 88 A 7,8,9,15,16,17 34 Methyl (E)-3-hexenoate 1259 128 A - 35 Hexyl acetate 1268 56 A 1,2,7,8,9,10,11,12,17 36 Ethyl (E)-3-hexenoate 1301 142 A 8 37 Methyl lactate 1322 45 A - 38 Ethyl lactate 1345 75 A 6,7,16 39 Methyl octanoate 1387 74 A 7,9,12 40 Ethyl octanoate 1432 88 A 1,7,8,9,16,17 41 Isoamyl lactate 1566 45 A - 42 Methyl decanoate 1592 87 A 9 43 Ethyl decanoate 1635 88 A 9 44 Ethyl benzoate 1654 105 A 9 45 Benzyl acetate 1721 108 A 6,8,9 46 Methyl salicylate 1758 120 A 6,8,9,10,17 47 Ethyl salicylate 1792 120 A - 48 Methyl hydrocinnamate 1834 104 A 6 49 Ethyl 3-cyclohexenecarboxylate 2182 81 C -
(40)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
72 (+)-Cycloisosativene 1459 161 B 9, 73 Copaene 1477 119 B 2,9,10,11,12 74 Dihydroedulan 1501 179 B - 75 Linalool 1551 93 A 1,2,8,9,10,15 76 Caryophyllene 1575 133 A 2,10,11 77 α-Terpineol 1690 93 A 2,6,8,9,14 78 α-Muurolene 1707 161 C 9,11,12 79 β-Damascenone 1805 121 A 9,10,14 80 Isogeraniol 1811 121 B - 81 Geraniol 1847 69 A - 82 Iridomyrmecin 2129 95 B - Phenols 83 p-Creosol 1949 138 A 9,16,17 84 Phenol 2006 94 A 6,9,17 85 p-Ethyl guaiacol 2022 137 A 9,14,16,17 86 p-Cresol 2083 107 A 9,16 87 p-Propyl guaiacol 2100 137 A - 88 Eugenol 2158 164 A - 89 4-Ethyl phenol 2175 107 A 1,2,6,9,16,17 90 Vanillin 2541 151 A 6,12,14 91 Tyrosol 2804 107 A - Hydrocarbons 92 Octane 807 85 A 1,2,8,9,10,17 93 Decane 1001 57 A 2 94 o-Xylene 1170 91 A 12 95 Styrene 1249 104 A 2,6,7,8,9,14
(42)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.