International Journal of Food Microbiology 189 (2014) 153–163

Contents lists available at ScienceDirect

International Journal of Food Microbiology

j ourna l homepage: www.e lsev ie r .com/ locate / i j foodmicro

Microbiological spoilage and investigation of volatile profile duringstorage of sea bream fillets under various conditions

Foteini F. Parlapani a, Athanasios Mallouchos b, Serkos A. Haroutounian c, Ioannis S. Boziaris a,⁎a Department of Ichthyology and Aquatic Environment, School of Agricultural Sciences, University of Thessaly, Fitoko Street, 38446 Volos, Greeceb Department of Food Science and Human Nutrition, Agricultural University of Athens, Iera Odos 75, 118 55 Athens, Greecec Department of Animal Science and Aquaculture, Agricultural University of Athens, Iera Odos 75, 118 55 Athens, Greece

⁎ Corresponding author at: School of Agricultural ScieAquatic Environment, Fitoko Street, 38446 N. Ionia, V93153; fax: +30 24210 93157.

E-mail address: boziaris@uth.gr (I.S. Boziaris).

http://dx.doi.org/10.1016/j.ijfoodmicro.2014.08.0060168-1605/© 2014 Published by Elsevier B.V.

a b s t r a c t

a r t i c l e i n f o

Article history:Received 28 March 2014Received in revised form 30 July 2014Accepted 2 August 2014Available online 10 August 2014

Keywords:Sea bream filletsSeafoodSpoilageVolatilesSPME–GC/MSChemical spoilage indices

Volatile organic compound (VOC) profile was determined during storage of sea bream (Sparus aurata) filletsunder air and Modified Atmosphere Packaging (MAP— CO2/O2/N2: 60/10/30) at 0, 5 and 15 °C. Microbiological,TVB-N (Total Volatile Base Nitrogen) and sensory changes were also monitored. Shelf-life of sea bream filletsstored under air was 14, 5 and 2 days (d) at 0, 5 and 15 °C respectively, while under MAP was 18, 8, and 2 d at0, 5 and 15 °C respectively. At the end of shelf life, the total microbial population ranged from 7.5 to8.5 log cfu/g. Pseudomonas spp. were among the dominant spoilagemicroorganisms in all cases, however growthof Brochothrix thermosphacta and Lactic Acid Bacteria (LAB) were favoured under MAP compared to air. TVB-Nproduction was favoured at higher temperatures and under air compared to lower temperatures and MAP.TVB-N increased substantially from the middle of storage and its value never reached concentrations higherthan 30–35 mg N/100 g, which is the legislation limit, making it a poor chemical spoilage index (CSI). A lot ofalcohols, aldehydes, ketones and ethyl esters that were detected in the present study have been reported asbacterial metabolites, others as products of chemical oxidation while others as aroma constituents. VOCs suchas 3-methylbutanal, acetic acid, ethanol, ethyl esters of isovaleric and 2-methylbutyric acids, 1-penten-3-ol,1-octen-3-ol and cis-4-heptenal appeared from the early or middle stages and increased until the end of storage.From those only 3-methylbutanal, acetic acid, ethanol and the ethyl esters have been reported asmicrobial origin,making them potential CSI candidates of sea bream fillets.

© 2014 Published by Elsevier B.V.

1. Introduction

Fish is highly perishable product and spoil due tomicrobiological ac-tivity, chemical oxidation of lipids and autolysis (Gramand Huss, 1996).However, microbial spoilage is the main mechanism affecting fresh fishquality. Under particular storage conditions (e.g., atmosphere, tempera-ture), a consortium of bacteria known as specific spoilage organisms(SSOs) produces metabolites (chemical spoilage indices—CSIs) respon-sible for off-flavours and causes the organoleptic rejection of theproduct (Dalgaard, 2003; Gram and Huss, 1996). Off flavours/offodour production is the major indicator used by consumers to evaluatefish freshness (Oehlenschläger, 2014). Compounds with characteristicsmell such as trimethylamine (TMA), various nitrogenous (TVB-N)and sulphuric compounds, aldehydes, ketones, and esters are producedby various microorganisms during fish spoilage (Dalgaard, 2003; Gramand Huss, 1996; Olafsdottir et al., 1997). Many groups of microorgan-isms contain species and strains that contribute to fish spoilage under

nces, Dept of Ichthyology andolos, Greece. Tel.: +30 24210

particular storage conditions. For example, under aerobic storage condi-tions, Shewanella putrefaciens has been recognised as potential spoilageorganism of chilled fish from northern seas due to the ability to reduceTrimethylamine Oxide (TMAO) to TMA (Gram et al., 1987). In fish orig-inated fromMediterranean Seawaters, Pseudomonas spp. aremainly in-volved in the spoilage of fish at low temperatures (Parlapani et al., 2013;Tryfinopoulou et al., 2002). However, under MAP, the dominant micro-biota of sea bream fillets is influenced by the different storage condi-tions (Parlapani, 2013). It is known that, the succession of spoilagemicroorganisms as well as their metabolic activity is greatly influencedby temperature and type of packaging (Dalgaard, 2003; Gram andHuss,1996). Indeed, the different gaseous atmospheres of MAP not only pro-longs the shelf-life of fishery products but also affects the synthesis ofspoilage microbiota and the profile of themetabolites produced leadingto a different type of spoilage (Gram and Huss, 1996).

Many methods have been used to assess fish and seafood qualitybased on physical, microbiological and chemical changes during storage(Olafsdottir et al., 1997). Sensory evaluation is the most common wayof assessing the freshness of fish and fish products (Howgate, 1982).Various sensory attributes (appearance of the skin, eyes, mucus andgills, colour, odour and texture) have been used to estimate the overallquality of fish (Cakli et al., 2006; Kyrana et al., 1997; Özogul et al., 2007;

154 F.F. Parlapani et al. / International Journal of Food Microbiology 189 (2014) 153–163

Rodriguez et al., 2003). The reliability of sensory analyses is decreased infish products, such as fillets, due to the diminished number of assess-ment attributes (Duflos et al., 2006). Additionally, sensory methodshave to be carried out by trained assessors, hence it is expensive to per-form and difficult to standardize, while microbiological methods usedby now are retrospectively expensive and time consuming (Dainty,1996; Dalgaard, 2003). Thus, there is a need to implement new andrapid methods for fish and fish product freshness/spoilage monitoringbased on determination of microbial metabolites (Dainty, 1996;Dalgaard, 2003; Edirisinghe et al., 2007; Ellis and Goodacre, 2001).

Traditional chemical methods to monitor microbial activity infish include the determination of TVB-N and TMA. However, TVB-Nand TMA increase in fish occurs only at the late stages of storage,hence those two parameters cannot be used as freshness indicator(Oehlenschläger, 2014). Additionally, TMA development in Mediterra-nean fish is not significant, presumably due to the low level of precursorcompound TMAO (Drosinos et al., 1997; Koutsoumanis and Nychas,1999; Kyrana and Lougovois, 2002).

Solid Phase MicroExtraction coupled with gas chromatography/mass spectrometry (SPME–GC/MS) is an analytical method which hasbeen used to study volatile organic compounds (VOCs) in seafood inorder to evaluate the degree of seafood spoilage, by providing us withvaluable information which can be used for the identification of poten-tial CSIs for rapid quality assessment and estimation of the remainingshelf-life (Duflos et al., 2006; Edirisinghe et al., 2007; Joffraud et al.,2001; Jonsdottir et al., 2008; Jorgensen et al., 2001; Leduc et al., 2012;Noseda et al., 2012; Soncin et al., 2008; Wierda et al., 2006). Accordingto Jay (1986) a metabolite that can be used for spoilage assessmenthas to (i) be absent or at least present at only low levels in food, (ii) in-crease during storage, and (iii) be produced by the dominant flora andshow good correlation with sensory score. Obviously, it is essential tofind out which metabolites fulfil those requirements, hence a prelimi-nary investigation has to focus on the VOC profile throughout storageand record which of them appear at the early stages of storage andhave the tendency to increase until the end of shelf-life. Afterwards ina future study, documentation of which VOCs are produced by spoilagemicrobiota has to be done before setting up trials with various productbatches and determine accurately the concentrations of the VOCs of in-terest throughout storage stages and then correlate themwith bacterialcounts and sensory changes.

Gilthead sea bream (Sparus aurata) is one of the main fish speciesfarmed in Greece and other Mediterranean countries. Greece is theleading producer in the world with approximately 45% of the total pro-duction (FAO, 2012). To our knowledge, there is no study regardingmicrobiological spoilage analysis and investigation of VOCs productionof sea bream fillets stored at various temperatures and atmosphericconditions. The aims of this work were to (i) determine the microbio-logical changes and shelf-life of sea bream fillets stored under air andMAP with a commercial gaseous mixture used by Hellenic AquacultureIndustry, at 0, 5 and 15 °C, and (ii) carry out a preliminary investigationof VOC profile using SPME–GC/MS, in order to reveal any potential CSIsof sea breamfillet spoilage/freshness. This studywill give valuable infor-mation regarding spoilage of sea bream fillets which is an importantadded-value product of seafood market.

2. Materials and methods

2.1. Sea bream fillet provision and storage

Packages from two different batches containing two sea breamfilletsof approximately 120 g each were taken from the Fish Processing Plantof Dias Aquaculture SA (Magoula, Attica, Greece). Sea breamwas farmedin thegeographical area designated as FAO37, 3.1 (Aegean Sea) andcap-tured on March of 2011. Fillets were packaged in polystyrene boxes(Sirap Gema S.p.A., Italy) under air or MAP. The MAP gas concentrationswere CO2: 60%, O2: 10%, and N2: 30%, which is one of the commercial gas

composition used by the Hellenic seafood industry for sea bream fillets,while theMAP film was the BDF 8050F (Cryovac-Sealed Air Ltd, Athens,Greece). The samples were transferred to the laboratory within 4 h afterpackaging using insulated boxes with melted ice. The samples werestored in incubators operating at 0, 5 and 15 °C.

2.2. Sensory analysis

Sensory evaluation was carried out by five trained panellists ac-cording to ISO 8586-1 (1993). The sensory attributes that were evalu-ated were appearance of the skin and flesh (translucent, glossy, naturalcolour, opaque, dull, discoloured) and odour of flesh (marine, fresh,neutral, sour, stale, spoiled, putrid). The rating of each sensory attributewas scored using a 1 to 9 descriptive hedonic scale (9 being the highestquality score and 1 the lowest). A score of 5 was taken as the averagescore for minimum acceptability (Tsironi and Taoukis, 2011). Theaim of the sensory evaluation was the determination of shelf-life offish fillets.

2.3. Microbiological analysis

All microbiological media were supplied by LABM (Lancashire, UK),apart from streptomycin sulphate, thallus acetate, cycloheximide(actidione) agar (STAA) which was supplied by Biolife Italiana srl(Milano, Italy). Iron agar (IA) was prepared according to Gram et al.(1987) by mixing the following ingredients: peptone 20 g/L, meatextract 3.0 g/L, yeast extract 3.0 g/L, ferric citrate 3.0 g/L, sodiumthiosulphate 0.3 g/L, NaCl 5 g/L, L-cysteine 0.6 g/L, agar 14 g/L, beforepH was adjusted to 7.4.

At every sampling point 4 packages (2 from each batch) wereopened and one fillet from each package was taken for microbiologicalanalyses. Twenty-five (25) gramsample fromeachfilletwas transferredaseptically to stomacher bags with 225 mL MRD (Maximum RecoveryDiluent, 0.1% w/v peptone, 0.85% w/v NaCl) and homogenized for2 min using a Stomacher (Bug Mixer, Interscience, London, UK). Sam-ples of 0.1 mL of serial dilutions in MRD were spread on the surfaceof dried media in Petri dishes for enumeration of (a) total microbialpopulation as aerobic plate counts (APC) on TSA (Tryptone Soy Agar),incubated for 48–72 h at 25 °C, (b) Pseudomonas spp., on cetrimide–fucidin–cephaloridine agar (CFC), incubated for 48 h at 25 °C and(c) Brochothrix thermosphacta, on STAA, incubated for 48–72 h at25 °C. Samples of 1 mL of serial dilution in MRD were used for thepour plate technique for enumeration of (a) H2S producing bacteria(presumptive S. putrefaciens) on IA by counting only black colonies,after incubation at 25 °C for 72 h., (b) Enterobacteriaceae on VioletRed Bile Glucose agar (VRBGA), incubated at 37 °C for 24 h and(c) Lactic Acid Bacteria (LAB) on Mann, Rogosa, Sharpe agar (MRS)after incubation at 25 °C for 72 h. The results were expressed as meanlog cfu g−1 ± standard deviation of 4 replicates.

2.4. Determination of TVB-N

At every sampling point, portions of 10 g of flesh were taken from4 different packages (2 from each batch) and homogenized in trichloro-acetic acid (TCA) 6% w/v and filtered using Whatman No. 1 filter paperin a 100 mL volumetric flask. Fifty milliliters in duplicates was takenfor TVB-N analysis using the steam-distillation procedure accordingto Vyncke et al. (1987). The results were expressed as mean TVB-N mg N/100 g ± standard deviation of 4 replicates (2 replicates fromeach batch of fish).

2.5. VOC determination by headspace SPME–GC/MS analysis

A slight modification of the method described by Iglesias et al.(2009)was used. At every sampling point, a total amount of 50 g of filletflesh was removed from 4 different packages (2 from each batch) and

Fig. 1. Sensory scores of appearance (a) and odour (b) of sea bream fillets stored at 0 °C(●,○), 5 °C (■,□) and 15 °C (▲,△) under air (closed symbols) and MAP (open symbol).The dashed lines show the time of organoleptic rejection.

155F.F. Parlapani et al. / International Journal of Food Microbiology 189 (2014) 153–163

pooled into one sample. Then 5 g in duplicates of pooled sample wasstored at −20 °C and analysed within a month. For analysis each 5 gportion was transferred with 5 mL of 30% NaCl solution into a 20 mLglass vial. The vial was closed hermetically using a mininert valve(Sigma Aldrich, Greece) and the contents were magnetically stirredfor 15 min at 40 °C. Then, the SPME fibre (DVB/CAR/PDMS 50/30 μm)was exposed to the headspace for another 30 min, under the sameconditions. The length of the fibre in the headspace was kept constant.Before each analysis, the fibre was exposed to the injection port for10 min to remove any volatile contaminants.

Analysis was performed on an Agilent 7890A gas chromatographcoupled to an Agilent 5973C mass spectrometer. Helium was used as acarrier gas at a constant flow rate of 1 mL/min. The injection port wasequipped with a liner (0.75 mm i.d.) suitable for SPME analysis. It wasoperated in splitlessmode for 1min at 250 °C. Separation of compoundswas performed on anHP-5MS column (30m, 0.25mm i.d., 0.25 μm filmthickness, Agilent). Oven temperature was maintained at 40 °C for5 min, subsequently the temperature was raised to 150 °C with a rateof 4 °C/min, and then was raised to 250 °C with a rate of 30 °C/minand held for 5 min. The interface temperature was set at 280 °C. Themass spectrometer was operated in electron impact mode with theelectron energy set at 70 eV and a scan range of 29–350 m/z. The tem-perature of MS source and quadrupole was set at 230 and 150 °C,respectively.

Identification of the compounds was effected by comparing: (i) thelinear retention indices (LRI) based on a homologous series of evennumbered n-alkanes (C8–C24, Niles, Illinois, USA) with those of stan-dard compounds and by comparison with literature data, and (ii) MSdata with those of reference compounds and by MS data obtainedfrom NIST library (NIST/EPA/NIH Mass Spectral Library with SearchProgram, data version NIST 05, software version 2.0d). Amdis software(version 2.62, http://chemdata.nist.gov/mass-spc/amdis/) was usedfor the deconvolution of mass spectra and identification of targetcomponents.

The amount of volatile compounds was expressed in arbitrary unitof the peak area of deconvoluted component multiplied by 10−6.

2.6. Statistical analysis

Differences of means in viable counts, TVB-N and sensory score,were statistically tested by performing Analysis of Variance followedby Tukey's significant difference test, using STATISTICA 6.0. A probabil-ity level of p ≤ 0.05 was considered statistically significant. Statisticalanalysis of VOCs was not applicable, for the reason that the measure-ments were conducted in duplicates and the determination of theirconcentration did not take place, since our aim was to monitor whichof the detected compounds are produced from the early stage of storageand have the tendency to increase until the end of shelf-life, in order toidentify which of them can be realistic candidates for CSI.

3. Results

3.1. Sensory changes & shelf-life

Shelf-life of sea bream fillets decreased with increasing storage tem-perature. Sensory score changes for appearance and odour of filletsstored under air and MAP are shown in Fig. 1. Initially, fish fillet fresh-ness was excellent. For example at 0 °C, appearance and odour werenot statistically different in the first two days of storage for both airand MAP stored fillets (p N 0.05). The fresh characteristics werediminished gradually with time. The differences in appearance andodour between air and MAP stored fillets at 0 °C became apparentfrom the 10th day of storage (p b 0.05). For the fillets stored at 15 °C,no statistically significant differences were found in appearance andodour scores between air and MAP (p N 0.05). At time point of mini-mumacceptability (grade 5 of hedonic scale) the sensory characteristics

of fillets represented opaque and dull appearance and stale odour. Afterthis point fillets were discoloured with putrid odour, for these reasonsthe fillets were graded as unfit (grade b 5 of hedonic scale) and rejected.Finally, the shelf-life of sea bream fillets stored under air determined bysensory assessmentwas 14, 5 and 2 d at 0, 5 and 15 °C respectively. Theshelf-life of MAP fillets was extended up to 18 and 8 d at 0 and 5 °C re-spectively, while at 15 °C shelf life for both air and MAP fillets was thesame (2 d) (Fig. 1).

3.2. Microbiological changes

Initially, the total microbial population of fillets expressed by APCwas at the level of 3.9 log cfu/g. At the end of shelf life, the total micro-bial population ranged from 7.5 to 8.5 log cfu/g (Fig. 2). Initial microbialpopulations of spoilage bacteria were at the level of 3.5, 3.1, 2.7 and2.5 log cfu/g, for LAB, Pseudomonas spp., H2S producing bacteria andEnterobacteriaceae respectively, while B. thermosphactawas below de-tection limit of 2 log cfu/g (Fig. 2). Bacteria grew faster under aerobicconditions and higher temperature. Elevated CO2 and reduced O2

inhibited bacterial growth and changed the microbial spoilage associa-tion at 0 and 5 °C, by suppressingmostly Gram negatives and favouringslightly Gram positives. However at 15 °C the effect of modified atmo-sphere on bacterial growth was not pronounced compared to lower

Fig. 2.Microbiological changes during storage of sea bream fillets at 0 °C (a), 5 °C (b) and 15 °C (c) under air (1) andMAP (2). Each data point and the error bars show themean and±st.dev. of 4 replicates. APC (○), Pseudomonas spp. (●), H2S bacteria (▲), Enterobacteriaceae (△), Brochothrix thermosphacta (□) and LAB (■). The vertical dashed lines indicate the point oforganoleptic rejection.

156 F.F. Parlapani et al. / International Journal of Food Microbiology 189 (2014) 153–163

temperatures as it can be seen from the changes of microbial spoilagepopulations under air and MAP (Fig. 2c1, c2).

Pseudomonas spp. were among the dominant spoilage microorgan-isms in all cases. Pseudomonas spp. reached up to 8 or 8.5 logs cfu/gunder air and between 7 and 8 logs cfu/g under MAP, depending ontemperature. The higher the temperature the higher was Pseudomonasspp. maximum population density. H2S producing bacteria andEnterobacteriaceae were also co-dominants with pseudomonads insome of the storage conditions. Indeed, H2S producing bacterial popula-tion at the time of spoilage of fillets stored at 0 °C under MAP, was ashigh as pseudomonads (p N 0.05), while Enterobacteriaceae reachedPseudomonas population (p N 0.05) in fillets stored at 5 and 15 °Cunder MAP (Fig. 2). MAP also enhanced the growth of B. thermosphacta

and LAB compared to aerobic storage. B. thermosphacta and LAB grewup to 5 or 6 logs cfu/g depending on the temperature and atmosphere(Fig. 2).

3.3. TVB-N changes

Production of TVB-N during the storage of sea bream fillets is shownin Fig. 3. The amount of TVB-N initially was 15.29 ± 0.31 mg N/100 gfish. The higher the temperature, the faster was TVB-N productionand higher its final concentration. Aerobic conditions favoured TVB-Nproduction compared to MAP. Indeed, at 0 °C under MAP, TVB-N wasvirtually unchanged until the 16th d of storage (p N 0.05). At theend of shelf-life at 0 °C, TVB-N concentration was 24.54 ± 0.25 and

Fig. 3. TVB-N changes on sea bream fillets stored under air (a) and MAP (b) at 0 °C (●),5 °C (■) and 15 °C (▲). Each data point and the error bars show the mean and ±st. dev.of 4 replicates. The dashed lines show the time of organoleptic rejection.

157F.F. Parlapani et al. / International Journal of Food Microbiology 189 (2014) 153–163

16.64 ± 0.21 mg N/100 g for fillets stored under air and MAP, respec-tively. Accordingly, at 5 °C TVB-N was 25.30 ± 0.19 and 19.10 ±0.50 mg N/100 g for fillets stored under air and MAP respectively,while at 15 °C the corresponding concentrations were 28.02 ± 0.34and 18.39 ± 0.27 mg N/100 g.

3.4. Headspace SPME–GC/MS analysis of VOCs

A total of 44 VOCs were detected, without taking into account ter-penes, aromatic compounds and hydrocarbons which were found intraces. From those, 18 compounds were detected only in fillets storedunder air, 3 only under MAP, while 23 were detected in both atmo-spheres. The detected compounds comprised of alcohols, aldehydes,ketones, esters and one organic acid (acetic). By comparing the chro-matographic peak areas it is possible to find which compounds havethe tendency to increase, decrease, disappear orfluctuate during storage.The VOC changes during storage of fillets under air and MAP are shownin Tables 1 and 2, respectively.

There are differences between aerobic storage and MAP. Esters andmost of the ketones were detected only under air, while differences in

relative amounts of various compounds were found between air andMAP. Quite few compounds appeared sporadically or fluctuated, duringstorage depending on storage conditions. Most of the compounds werefound to exhibit a more consistent profile during storage dependinghowever on storage conditions and can be classified into three groups.The first group included few compounds that appeared at the earlystages and decreased during storage. The second group included a lotof compounds that found to increase consistently only at the late stagesand/or the end of shelf-life and the third group included fewer com-pounds which appeared initially (d 0) or from the early stages and in-creased during storage until the end of shelf-life.

The first group of compounds included mostly aldehydes likehexenal, heptanal and nonanal, which showed the tendency to decreasegradually under air storage at 0 and 5 °C, while hexenal decreasedand under MAP at 5 °C. In the second group the compounds whichwere appeared at the end of aerobic storage, were esters like ethyl-2-methylbutyrate, ethyl isovalerate, ethyl crotonate and ethyl tiglateat 0 and 5 °C, ethyl isobutyrate at 5 °C, and ethyl propionate andethyl butyrate at 15 °C. From carbonylic compounds, aldehydes like2-methylbutanal appeared at the end of shelf-life at all temperatures,cis-6-nonenal at 0 and 5 °C, while ketones like 3-pentanone, 2-heptanone and 6-methyl-5-hepten-2-one appeared at the end ofshelf life at 0 °C. Among alcohols, 3-methyl-1-butanol was foundat the end of shelf-life at all temperatures, while 1-penten-3-oland 2-methylbutanol only at 15 °C (Table 1). In MAP storage, both3-methylbutanal and 2-methylbutanal appeared at the end of shelf-life at 5 and 15 °C, while trans-2-hexenal only at 15 °C. Alcohols likedecanol appeared at the end of shelf life at all temperatures, whileheptyl and lauryl alcohol at 0 °C and 1-penten-3-ol at 15 °C only(Table 2). From the third group, the compounds which were found toincrease from the early stages were 3-methylbutanal and 1-octen-3-olin fillets stored under air at 0 and 15 °C, cis-6-nonenal at 15 °C, ethanolunder air at 5 and 15 °C, while esters like ethyl-2-methylbutyrate, ethylisobutyrate, ethyl acetate and ethyl isovalerate at 15 °C. In fillets storedunder MAP, cis-4-heptenal increased at 0 and 15 °C, hexanal at 15 °C,acetic acid at 5 °C, 1-penten-3-ol at 0 °C, while 1-octen-3-ol increasedat all temperatures.

The amount changes of 3-methylbutanal, acetic acid and ethanol, inrelation to APC are shown in Figs. 4 and 5. At 0 °C in fillets under air, theamount of 3-methylbutanal was initially low, following a gradual in-crease during early–middle stages of storage, while an abrupt increasewas observed at the end of storage, when APC reached the levelof 8 logs cfu/g and the remaining shelf-life was less than 2 d (Figs. 2and 4a). In fillets under MAP at 0 °C, the amount of 3-methylbutanalwas substantially reduced (Fig. 4a). At 5 °C, 3-methylbutanal increasedunder both air and MAP but only at the end of storage. The amount of3-methylbutanal was higher compared to 0 °C and a rapid increaseoccurred when APC reached again the level of 8 logs cfu/g (Fig. 4b).Finally at 15 °C, 3-methylbutanal increased gradually mostly in filletsunder air (Fig. 4c). Ethanol increased in fillets stored under air at5, while under MAP no changes were observed (Fig. 5a). Acetic acid in-creased only in fillets stored under MAP at 5 °C. Its increase wasgradual at the beginning with an abrupt increase when APC reachedthe level of 8 logs cfu/g at the end of shelf-life (Fig. 5b). Changes ofethyl esters of 2-methylbutyric and isovaleric acids against APC areshown in Fig. 6. Esters appeared only in fillets stored under air andtheir concentration increased substantially at the end of shelf life at5 °C, while a gradual increase was observed at 15 °C.

4. Discussion

Shelf-life of fish is extended underMAP compared to aerobic storage(Noseda et al., 2014). However, chilling temperature has to be main-tained, since the CO2 effect diminishes at elevated temperatures mainlydue to the reduced solubility of CO2 into the water phase of microbialcytoplasm (Daniels et al., 1985; Gill and Molin, 1991). Indeed in our

Table 1Detected VOCs and their relative concentrations (surface × 10−6 under the chromatographic peak) in fillets during their storage (in days) under air at 0, 5 and 15 °C. Each value is themean of duplicate measurement of pooled sample.

Air (area × 10−6)

0 °C 5 °C 15 °C

d0 d4 d8 d12 d14 d2 d4 d5 d6 d1 d2

EstersEthyl acetate – – – – – – – – – 2.65 3.90Ethyl propionate – – – – – – – – 0.36 – 0.93Ethyl isobutyrate – – – – – – – 0.05 0.22 0.12 0.35Ethyl butyrate – – – – – – – – 0.21 – 0.46Ethyl crotonate – – – – 0.18 – – 0.04 0.19 – –

Ethyl-2-methylbutyrate – – – – 0.39 – – 0.05 0.44 0.14 0.20Ethyl isovalerate – – – – 0.65 – – 0.10 0.63 0.09 0.25Ethyl tiglate – – – – 0.48 – – – 0.27 – –

Ethyl hexanoate 0.20 – – 0.05 0.15 – – – 0.27 – 0.13

Aldehydes3-Methylbutanal – 0.10 0.15 0.20 0.63 – – 1.30 0.43 1.64 2.312-Methylbutanal – – – – 0.31 – – 0.39 – – 0.70Hexenal 5.42 5.13 1.85 1.80 1.10 34.4 2.45 0.34 – 0.68 0.40Hexanal – – – – – – – – 0.70 – –

Heptanal 0.36 0.16 0.09 0.13 0.73 0.14 – – – –

Nonanal 0.34 0.18 0.14 0.12 0.10 0.37 0.13 0.14 0.07 – –

Trans-2-decenal 0.05 – 0.05 0.05 – – 0.02 – – – –

Trans-2-hexenal – 0.08 – 0.15 – 0.55 0.08 – – – –

Cis-4-heptenal 0.25 0.91 0.52 0.31 0.36 2.02 0.38 – 0.14 – –

Trans,trans-2,4-heptadienal 0.17 0.13 0.06 0.48 – 0.44 0.20 – – – –

Trans-2-octenal 1.20 1.49 1.89 0.99 1.59 0.13 – – 0.30 – 0.44Cis-6-nonenal – – – 0.15 0.55 – – 0.14 – 0.06 0.12Trans-2-nonenal – – – – – – – – 0.14 – 0.15

Ketones2-Pentanone 0.14 – 0.18 – 0.72 – 0.20 – 0.19 – 0.313-Pentanone – – – – 4.48 – – 0.77 – – –

2-Heptanone – – – – 0.45 – – – – – –

2-Nonanone – – – – – – – – 0.05 – –

2,3 Octanedione 0.27 – 0.16 – – – 0.18 – – – –

2,3-Pentanedione 1.21 – 0.68 1.21 – – 1.69 – 0.38 – 0.336-Methyl-5-hepten-2-one – – – 0.21 0.22 – – – – – –

AlcoholsEthanol 1.51 4.10 4.62 1.41 2.46 8.53 8.88 14.8 16.4 17.6 19.43-Methyl-1-butanol – 0.19 – – 0.60 – – 1.22 0.17 – 4.07Amyl alcohol 0.07 0.13 0.14 0.14 0.14 0.71 0.16 – – – –

2-Penten-1-ol 0.26 0.73 0.30 0.22 0.19 1.59 0.35 0.20 – – 0.161-Penten-3-ol – 13.3 3.73 1.04 6.40 19.5 3.39 5.24 0.66 – 1.231-Octen-3-ol – 0.58 0.75 0.77 0.79 1.35 0.54 0.29 0.14 0.11 0.17Cis-6-nonen-1-ol 0.10 – – – – – – – – – –

2-Methyl-1-butanol – 2.06 – 0.44 – – – – – – 0.37Lauryl alcohol – – 0.04 0.04 – – – – – – –

Hexanol 0.11 – – – – 0.16 0.13 – – – –

Decanol 0.05 – 0.04 0.03 – 0.02 – – 0.04 – 0.042-Ethyl-1-hexanol 1.78 1.92 1.20 0.40 1.52 1.76 5.00 3.3 1.05 – 0.58

AcidsAcetic acid 0.13 – – – – – – – – – 0.19

158 F.F. Parlapani et al. / International Journal of Food Microbiology 189 (2014) 153–163

study no shelf-life extension was observed at 15 °C, while significantshelf-life extension was observed at 0 and 5 °C. According to Tsironiet al. (2009) and Tsironi and Taoukis (2011), shelf-life of sea bream fil-lets stored aerobically at 5 °C varied from 4 to 7 d, while Dalgaard et al.(1997) found that shelf-life of sea bream fillets stored under 50%CO2 and 50% N2 was 14 d at 0 °C. The differences are presumablydue to the different levels of initial microbial population, the type ofpackaging film, the different gaseous composition, gas to product ratioetc. (Noseda et al., 2014).

It is known from scientific literature that primarily Pseudomonas spp.and secondarily H2S producing bacteria are the dominant spoilage mi-croorganisms of chilled stored fish under aerobic condition caughtfrom warm waters (Álvarez et al., 2008; Koutsoumanis et al., 1999;Koutsoumanis and Nychas, 1999, 2000; Tryfinopoulou et al., 2002,2007). Our study agrees with these results. Pseudomonaswas the dom-inant bacterial population from the middle until the end of storage,

while H2S producing bacteria followed Pseudomonas population at lowtemperatures (Koutsoumanis and Nychas, 2000). With increasing tem-peratures, Enterobacteriaceae becomes also co-dominant with pseudo-monads and/or H2S producing bacteria (Gram et al., 1987; Gram andHuss, 1996). H2S producing bacteria are consistedmainly by Shewanellaspp. in chilled fish from both Northern seas (Dalgaard et al., 1993;Dalgaard, 1995; Gram et al., 1987; Jorgensen and Huss, 1989) andMediterranean area (Tryfinopoulou et al., 2007).

MAP favoured the growth of B. thermosphacta and LAB compared toaerobic storage. LAB and B. thermosphacta usually predominate underMAP (Drosinos and Nychas, 1996; Koutsoumanis et al., 2000), due tothe tolerance of these microorganisms to elevated CO2 (Gill and Molin,1991). Indeed LAB and B. thermosphacta have been reported as SSO ofGreek fish stored under MAP. Koutsoumanis et al. (2000) and Drosinosand Nychas (1996) found that B. thermosphacta and S. putrefaciensgrew under ΜΑΡ at low temperatures and outcompeted Pseudomonas

Table 2Detected VOCs and their relative concentrations (surface × 10−6 under the chromatographic peak) in fillets during their storage (in days) under MAP at 0, 5 and 15 °C. Each value is themean of duplicate measurement of pooled sample.

MAP (area × 10−6)

0 °C 5 °C 15 °C

d0 d4 d8 d12 d16 d18 d4 d8 d12 d1 d2

Aldehydes3-Methylbutanal – – – – 0.89 – – 0.29 1.68 – 0.182-Methylbutanal – – – – 0.40 – – 0.22 1.04 – –

Hexenal 5.42 1.00 1.54 3.18 1.18 2.21 2.49 1.87 0.93 – –

Hexanal – – – – – – – – – 0.80 4.12Nonanal 0.34 0.44 0.37 0.41 0.35 0.27 0.25 0.20 0.24 – 0.20N-decanal 0.04 0.08 0.06 0.05 0.04 – – 0.03 – – –

Trans-2-decenal 0.05 0.44 0.27 0.15 0.14 0.11 0.09 0.10 0.07 0.06Cis-4-heptenal 0.25 0.40 0.72 0.98 1.37 1.39 0.63 0.67 0.54 0.26 1.01Trans-2-octenal – – – – – – – 0.13 0.11 – –

Trans-2-hexenal – – – – – – – – – – 0.19Trans-2-nonenal – – – – – – – – 0.12 – –

Trans,trans-2,4-heptadienal 0.17 0.33 0.29 0.17 – 0.20 0.23 0.14 – – 0.23

Ketones3-Hydroxy-2-butanone – – – – – – – – 0.26 – –

2,3-Pentanedione 1.21 – – – – – – – – – 0.66

AlcoholsEthanol 1.51 1.75 1.55 2.05 3.73 0.83 1.52 1.72 1.89 1.86 2.03-Methyl-1-butanol – – – – – – – – 0.61 – –

Amyl alcohol 0.07 – – – – – – 0.12 – – 0.34Heptyl alcohol – – – – – 0.38 – – – – –

Lauryl alcohol – – – – – 0.07 0.05 – – – –

Hexanol 0.11 – – – – – – 0.14 0.20 – –

Decanol – – – – – 0.06 0.05 0.04 1.261-Penten-3-ol – 0.63 1.08 2.16 2.36 2.54 1.17 1.00 1.03 – 1.672-Penten-1-ol 0.26 – – 0.27 0.36 – – 0.32 0.35 – –

1-Octen-3-ol – 0.34 0.81 1.05 1.20 1.23 0.75 0.95 0.96 0.35 1.002-Ethyl-1-hexanol 1.78 – 0.23 0.54 0.15 0.30 0.39 0.27 0.54 0.14 0.15

AcidsAcetic acid 0.13 0.90 0.30 0.28 0.31 – 0.18 0.28 0.47 – –

159F.F. Parlapani et al. / International Journal of Food Microbiology 189 (2014) 153–163

spp., in red mullet (Mullus barbatus) and sea bream (S. aurata), respec-tively. In the present work, LAB and B. thermosphacta reached relativelylow population densities, not higher than 6 logs cfu/g. However,the gaseous composition in our case differed compared to the above-mentioned works. Pournis et al. (2005), using the same gaseouscomposition as in our study, found that Pseudomonas spp. and H2Sproducing bacteria were the dominant spoilage microorganisms inMediterranean mullet (Mullus surmuletus), while B. thermosphactaand LAB reached populations not higher than 5 logs cfu/g. LAB werefound in higher numbers in initial microbiota compared to othermicroorganisms, but they did not reach high numbers during storage.It is known that despite the diversity and abundance of the initialmicrobiota, the composition of spoilage microbiota is selected by stor-age conditions especially temperature and atmospheric composition(Dalgaard, 2003). In our case, the conditions did not favour adequatelythe growth of LAB.

Pseudomonas spp. produces mainly volatile nitrogenous compounds(Dainty, 1996). Indeed TVB-N increased from the middle of storage,when Pseudomonas spp. had reached a considerable population levelof about 6 logs cfu/g (Figs. 2 and 3). The initial value of 15.29 ±0.31 mg N/100 g was similar with the value reported for fresh seabream by Goulas and Kontominas (2007) (15.9 mg N/100 g). At theend of shelf-life TVB-N value never reached amounts higher than30–35 mg N/100 g, which is the legislation limit (EC, 2074/2005). Thishas been also observed by other researchers as well (Castro et al.,2006; Koutsoumanis and Nychas, 2000; Kyrana and Lougovois, 2002).In fillets under MAP, TVB-N development was not pronounced com-pared to air. The reduced development of TVB-N under MAP can beattributed to the suppression of Pseudomonas growth and metabolismand to the enhancement of LAB and B. thermosphacta growth whichproduce organic acids instead of basic nitrogenous compounds as

pseudomonads do (Nychas and Arkoudelos, 1991; Drosinos andNychas, 1997). From our results it can be concluded that TVB-N is apoor CSI of sea bream fillets. The same was also concluded by Castroet al. (2006) studying European sea bass while stored on ice. Accordingto Oehlenschläger (2014), TVB-N can monitor fish quality only on thelast stages of storage making it unsuitable for freshness evaluation.

Most of the VOCs detected in this study have also been reportedfor other fish and seafood (cod fillets, pangasius fillets, king salmon,yellowfin tuna, cold-smoked salmon, sole, sea bass, shellfish) by otherresearchers as well (Duflos et al., 2006; Edirisinghe et al., 2007; Fratiniet al., 2012; Jonsdottir et al., 2008; Jorgensen et al., 2001; Leduc et al.,2012; Moreira et al., 2013; Noseda et al., 2012; Tuckey et al., 2013;Wierda et al., 2006). A lot of them were also detected by Alasalvaret al. (2005) and Soncin et al. (2008), during the storage of whole seabream while stored on ice. Most of the VOCs detected in the presentstudy have been reported in the literature as bacterial metabolites,others as products of chemical oxidation of fatty acids, while others asconstituents of aroma.

Compounds such as, hexanal, hexenal, nonanal, heptanal, anddecanal have been reported as sea bream aroma compounds (Selliand Cayhan, 2009). Various C6–C9 aldehydes that were also found inEuropean sea basswhile stored on ice had great impact on fresh fish fla-vour (Leduc et al., 2012). Those compounds mostly come from enzymiccatabolism of unsaturated fatty acids (Yasuhura and Shibamoto, 1995;Kawai, 1996). A lot of C6–C9 aldehydes, ketones and alcohols arearoma compounds characterizing fresh fish (Kawai, 1996). However, al-dehydes and ketones contribute more to the aroma due to their lowodour and flavour thresholds in contrast to alcohols (Alasalvar et al.,2005; Kawai, 1996).

Chemical activity is the second most important mechanism, afterbacterial action, affecting quality of fresh fish. Compounds such as

Fig. 5. Changes of ethanol (a) and acetic acid (b) in relation to APC in fillets stored at 5 °Cunder air (●) and MAP (○).

Fig. 4. Changes of 3-methylbutanal in relation to APC in fillets stored at 0 °C (a), 5 °C(b) and 15 °C (c) under air (●) and MAP (○).

160 F.F. Parlapani et al. / International Journal of Food Microbiology 189 (2014) 153–163

1-octen-3-ol, cis-4-heptenal and 1-penten-3-ol, which were found inour study, are involved with n−6 polyunsaturated fatty acid oxida-tion and/or other chemical reactions (Dalgaard, 2003; Duflos et al.,2006; Iglesias and Medina, 2008; Fratini et al., 2012). Despite thefact that they exhibited good profile in many storage conditions,these compounds are not related to bacterial action, which is theprimer mechanism of sea bream fillet quality deterioration and conse-quently, according to Jay (1986), are not suitable for CSIs.

Production of various alcohols, aldehydes and ketones (e.g. 2-methyl-1-butanol, ethanol, 2-methylbutanal, 3-methylbutanal, 2,3-pentanedione,2-heptanone, 3-hydroxy-2-butanone) has been reported as a resultof metabolic activity by microorganisms, such as Pseudomonas spp.,Shewanella spp., Enterobacteriaceae, B. thermosphacta, LAB andPhotobacterium phosphoreum, during fish and meat spoilage (Casaburiet al., 2014; Edirisinghe et al., 2007; Joffraud et al., 2001; Jonsdottiret al., 2008; Jorgensen et al., 2001; Olafsdottir et al., 2005; Wierdaet al., 2006). A number of ethyl esters and some ketones that were de-tected in this study presumably are involvedwith Pseudomonas spp. ac-tivity (Casaburi et al., 2014; Edwards et al., 1987). Indeed, in this study,esters and the majority of ketones were found only in fillets storedunder air where the aerobic conditions enhance growth and metabolicactivity of pseudomonads. However, esters like ethyl acetate can alsobe produced by Lactobacillus spp. (Joffraud et al., 2001). Productionof acetic acid has mainly attributed to the metabolic activity ofB. thermosphacta and LAB (Borch and Molin, 1989; Laursen et al.,

2006; Tsigarida et al., 2003). Indeed, in our study, acetic acid increasedin fillets under MAP where B. thermosphacta and LAB growth wasmore pronounced compared to storage under air. Ethanol is producedboth by LAB under reduced oxygen packages and pseudomonadsunder air storage (Casaburi et al., 2014; Olafsdottir et al., 2005). Acetoin(3-hydroxy-2-butanone) production, which has been attributed toP. phosphoreum (Olafsdottir et al., 2005) and LAB (Jonsdottir et al.,2008) when they reach high population densities, did not play anyimportant role in our case. One explanation is that in chilled fishfrom Greek waters P. phosphoreum consists only a negligible part of thespoilage microbiota (Koutsoumanis and Nychas, 1999) and LAB inour case never reached numbers higher than 6 logs cfu/g. Alcohols like3-methylbutanol, 2-methylbutanol, and the aldehyde, 3-methylbutanalhave been detected in products such as vacuum-packed smokedsalmon (Jorgensen et al., 2001)where LAB often predominates. The alde-hydes like 3-methylbutanal, 2-methylbutanal, 2-methylpropanal, and 2-methylbutanal, the alcohols such as 3-methylbutanol, 2-methylbutanol,2-methylpropanol, and 2-methylbutanol and various ketones like 2,3pentanedione, and 2-heptanone, are also produced by Carnobacteriumspecies (Joffraud et al., 2001; Laursen et al., 2006). However, 2-methylbutanal and 3-methylbutanal have also been reported to be pro-duced by other bacteria as well such as Pseudomonas spp., Shewanellaspp. and Brochothrix spp. in spoiled chicken carcasses and meat(Nychas et al., 1998, 2007).

It is known that different temperatures and atmospheric conditionsnot only select different dominant microorganisms with different me-tabolism but also affect both their growth rate and metabolic activity(Nychas et al., 2007). This can explain the difference on VOC profile be-tween air andMAP stored fillets. Regardless of the diversity of the initialmicrobiota, which is influenced by various sources of contamination(seawater, processing surfaces etc.), specific microorganisms finallypredominate. Indeed, low temperatures select psychrotrophic specieswhile further selection is introduced by the availability of oxygenwhich affects not only the microbial growth but also the microbial me-tabolism, hence only few species or strains are able to predominate andconsequently affect the production of spoilagemetabolites that degradethe product quality (Nychas et al., 2007, 2008). The different initial mi-crobiota of various product batches presumably affects the diversity ofthe detected VOCs, but the VOCs associated with the spoilage wouldbe detected throughout storage and undoubtedly they are produced

Fig. 6. Changes of ethyl-2-methylbutyrate (1) and ethyl isovalerate (2) in relation to APC in fillets stored at 5 °C (a) and 15 °C (b) under air (●).

161F.F. Parlapani et al. / International Journal of Food Microbiology 189 (2014) 153–163

by the dominant spoilage microbiota. The volatile metabolites of ourinterest detected in this study are metabolic products of dominantspoilage microorganisms which indicate the minor effect of the initialdiversity to the VOC profile.

Various compounds are utilized by microorganisms for the pro-duction of volatiles. Organic acids, ethanol and ethyl esters are pro-duced mainly by glucose, while leucine and iso-leucine have beenreported as precursors for 2-methyl-1-butanol, 2-methylbutanal and3-methylbutanal (Nychas et al., 2008; Olafsdottir et al., 2005; Soncinet al., 2008; Wierda et al., 2006). In many cases metabolites producedduring spoilage can also be further metabolised by other microorgan-isms, such as Shewanella which can use ethanol or organic acids afterdepletion of other key metabolic molecules (Nychas et al., 2008).Additionally, other compounds such as aroma constituents are depletedthroughout storage (Leduc et al., 2012; Wierda et al., 2006). Those as-pects presumably explain the fluctuation or depletion of VOCs observedin the present and/or other studies.

A lot of VOCs of microbial origin have been reported as potentialCSIs such as ethanol, acetic acid, 2-methylbutanal, 3-methylbutanal,1-propanol, 2-methyl-1-butanol, 3-methyl-1-butanol, 3-hydroxy-2-butanone, and 2-butanone, in fish such as sea bream, sole, whiting,mackerel, tuna, cod and vacuum-packed cold-smoked salmon (Alasalvaret al., 2005; Duflos et al., 2006; Edirisinghe et al., 2007; Jonsdottiret al., 2008; Jorgensen et al., 2001; Moreira et al., 2013; Olafsdottiret al., 2005; Soncin et al., 2008; Wierda et al., 2006). However, to iden-tify an effective metabolite for determining and/or monitoring fishquality, a lot of factors have to be taken into account, such as tempera-ture and its fluctuation, gaseous atmosphere and film permeability ofthe package, bacterial competition etc. (Tsigarida et al., 2003). In ourstudy the compounds that attributed to microbial origin and exhibiteddesirable profiles were the 3-methylbutanal, ethanol, acetic acid andthe ethyl esters of isovaleric and 3-methylbutyric acid. Acetic acid and3-methylbutanal were among the quality indicators that Alasalvaret al. (2005) and Soncin et al. (2008) respectively suggested for wholesea bream while stored on ice. In our case, only 3-methylbutanal ex-hibited satisfactory profile in almost all storage conditions. Ethyl estersalso showed good profile for aerobically stored fillets at elevated tem-peratures only. Finally, ethanol and acetic acid were suitable for filletsstored at 5 °C, with the latter only for MAP fillets.

5. Conclusions

Different temperatures and atmospheric conditions affected notonly the shelf-life but also the spoilage microbiota population and com-position and consequently the volatile profile. Quite few volatile com-pounds of microbial origin can be considered from this study aspotential CSIs of sea bream fillets. Further work is required to confirmby which microorganisms those compounds are produced in order tofully document them as CSIs of sea bream fillets. Then, accurate deter-mination of the concentrations of VOCs documented as CSI candidatesin sea bream fillets from different batches stored under various condi-tions and their relation with remaining shelf-life will be the first stepfor designing bio-sensors for on-pack freshness assessment and shelflife determination. The spoilage process, focused on themicroorganismsthat dominate and produce the VOCs considered as CSI candidates hasto be studied in depth. Various new species and strains of spoilage mi-croorganisms which might contribute to the spoilage of sea breamhave been revealed recently using molecular techniques (Parlapani,2013). The spoilage potential and activity of those microorganismsto produce VOCs have to be studied in order to give us insights on seabream spoilage mechanism.

Acknowledgements

This research has been co-financed by the European Union(European Social Fund — ESF) and Greek national funds through theOperational Program “Education and Lifelong Learning” of the NationalStrategic Reference Framework (NSRF) — Research Funding Program:Heracleitus II. Investing in knowledge society through the European So-cial Fund. Finally, we would like to thank Dias Aquaculture SA (Greece)for the provision of sea bream fillets.

References

Alasalvar, C.,Taylor, K.D.A.,Shahidi, F., 2005. Comparison of volatiles of cultured and wildsea bream (Sparus aurata) during storage in ice by dynamic headspace analysis/gaschromatography–mass spectrometry. J. Agric. Food Chem. 53, 2616–2622.

Álvarez, A.,García García, B.,Garrido, M.D.,Hernández, M.D., 2008. The influence of starva-tion time prior to slaughter on the quality of commercial-sized gilthead seabream(Sparus aurata) during ice storage. Aquaculture 284, 106–114.

162 F.F. Parlapani et al. / International Journal of Food Microbiology 189 (2014) 153–163

Borch, E.,Molin, G., 1989. The aerobic growth and product formation of Lactobacillus,Leuconostoc, Brochothrix, and Carnobacterium in batch cultures. Appl. Microbiol.Biotechnol. 30, 81–88.

Cakli, S.,Kilinc, B.,Cadun, A.,Dincer, T.,Tolasa, S., 2006. Effects of gutting and ungutting onmicrobiological, chemical, and sensory properties of aquacultured sea bream (Sparusaurata) and sea bass (Dicentrarchus labrax) stored in ice. Crit. Rev. Food Sci. Nutr. 46,519–527.

Casaburi, A., Piombino, P.,Nychas, G.-J.,Villani, F., Ercolini, D., 2014. Bacterial populationsand the volatilome associated to meat spoilage. Food Microbiol. http://dx.doi.org/10.1016/j.fm.2014.02.002.

Castro, P.,Padron, J.C.P.,Cansino, M.J.C.,Velazquez, E.S.,De Larriva, R.M., 2006. Total volatilebase nitrogen and its use to assess freshness in European sea bass stored in ice. FoodControl 17, 245–248.

Commission Regulation (EC) No 2074/2005 of 5 December 2005, 2005. Off. J. Eur. Union L338/27 (22.12.2005).

Dainty, R.H., 1996. Chemical/biochemical detection of spoilage. Int. J. Food Microbiol. 33,19–33.

Dalgaard, P., 1995. Qualitative and quantitative characterization of spoilage bacteria frompacked fish. Int. J. Food Microbiol. 26, 319–333.

Dalgaard, P., 2003. Spoilage of seafood. In: Caballero, B., Trugo, L., Finglas, P. (Eds.), Ency-clopedia of Food Science and Nutrition. Academic Press, London, pp. 2462–2472.

Dalgaard, P., Gram, L., Huss, H.H., 1993. Spoilage and shelf life of cod fillets packed invacuum or modified atmospheres. Int. J. Food Microbiol. 19, 283–294.

Dalgaard, P., Mejlholm, O., Christiansen, T.J., Huss, H.H., 1997. Importance ofPhotobacterium phosphoreum in relation to spoilage of modified atmosphere packedfish products. Lett. Appl. Microbiol. 24, 373–378.

Daniels, J.A.,Krishnamurthi, R.,Rizvi, S.S.H., 1985. A review of effects of carbon dioxide onmicrobial growth and food quality. J. Food Prot. 48 (6), 532–537.

Drosinos, E.H.,Nychas, G.-J.E., 1996. Brochothrix thermosphacta, a dominant microorgan-ism in Mediterranean fresh fish (Sparus aurata) stored under modified atmosphere.Ital. J. Food Sci. Technol. 4, 323–330.

Drosinos, E.H.,Nychas, G.-J.E., 1997. Production of acetic acid in relation to the contentof glucose during storage of gilt-head seabream (Sparus aurata) under modified at0 ± 1 °C. Food Res. Int. 30, 711–717.

Drosinos, E.H., Lambropoulou, K., Mitre, E., Nychas, G.-J.E., 1997. Attributes of freshgilt seabream (Sparus aurata) fillets treated with potassium sorbate, sodium gluco-nate and stored under a modified atmosphere at 0 ± 1 °C. J. Appl. Microbiol. 83,569–575.

Duflos, G., Coin, V.M., Cornu, M.,Antinelli, J.F.,Malle, P., 2006. Determination of volatilecompounds to characterize fish spoilage using headspace/mass spectrometry andsolid-phase microextraction/gas chromatography/mass spectrometry. J. Sci. FoodAgric. 86, 600–611.

Edirisinghe, R.K.B., Graham, A.J., Taylor, S.J., 2007. Characterisation of the volatiles ofyellowfin tuna (Thunnus albacares) during storage by solid phase microextractionand GC–MS and their relationship to fish quality parameters. Int. J. Food Sci. Technol.42, 1139–1147.

Edwards, R.A.,Dainty, R.H.,Hibbard, C.M., 1987. Volatile compounds produced by meatpseudomonads and related reference strains during growth on beef stored in air atchill temperatures. J. Appl. Bacteriol. 62, 403–412.

Ellis, D.I.,Goodacre, R., 2001. Rapid and quantitative detection of the microbial spoilageof muscle foods: current status and future trends. Trends Food Sci. Technol. 12,414–424.

FAO, 2012. FishStatJ — software for fishery statistical time series. FAO Fisheries Depart-ment, Fishery Information, Data and Statistics Unit, Version 2.0.0, Release Date:March 2012.

Fratini, G.,Lois, S.,Pazos, M.,Parisi, G.,Medina, I., 2012. Volatile profile of Atlantic shellfishspecies by HS-SPME GC/MS. Food Res. Int. 48, 856–865.

Gill, C.D.,Molin, G., 1991. Modified atmospheres and vacuum packaging. In: Russell, N.S.,Gould, G.W. (Eds.), Food Preservatives. Blackie and Son Limited, Glasgow, London,pp. 172–199.

Goulas, A.E., Kontominas, M.G., 2007. Combined effect of light salting, modified atmo-sphere packaging and oregano essential oil on the shelf-life of sea bream (Sparusaurata): biochemical and sensory attributes. Food Chem. 100, 287–296.

Gram, L.,Huss, H.H., 1996. Microbiological spoilage of fish and fish products. Int. J. FoodMicrobiol. 33, 121–137.

Gram, L.,Trolle, G.,Huss, H.H., 1987. Detection of specific spoilage bacteria from fish storedat low (0 °C) and high (20 °C) temperatures. Int. J. Food Microbiol. 4, 65–72.

Howgate, P.F., 1982. Quality assessment and quality control. In: Aitken, A., Mackie, I.M.,Merrit, J.H., Windsor Crown, M.L. (Eds.), Fish Handling and Processing. Edinburgh,Scotland.

Iglesias, J.,Medina, I., 2008. Solid-phase microextraction method for the determinationof volatile compounds associated to oxidation of fish muscle. J. Chromatogr. A 1192,9–16.

Iglesias, J.,Medina, I.,Bianchi, F.,Careri, M.,Mangia, A.,Musci, M., 2009. Study of the volatilecompounds useful for the characterisation of fresh and frozen–thawed culturedgilthead sea bream fish by solid-phase microextraction gas chromatography–massspectrometry. Food Chem. 115, 1473–1478.

International Organization for Standardization (ISO) 8586-1, 1993. Sensory analysis—general guidance for the selection, training and monitoring of assessors, part 1:selected assessors. Geneva. Available at: http://iso.org.

Jay, J.M., 1986. Microbial spoilage indicators and metabolites. In: Pierson, M.D., Sterm, N.J.(Eds.), Food-Borne Microorganisms and Their Toxins: Developing Methodology.Marcel Dekker, Inc., New York, N.Y., pp. 219–240.

Joffraud, J.J., Leroi, F., Roy, C., Berdague, J.L., 2001. Characterisation of volatile compoundsproduced by bacteria isolated from the spoilage flora of cold-smoked salmon. Int. J.Food Microbiol. 66, 175–181.

Jonsdottir, R.,Olafsdottir, G.,Chanie, E.,Haugen, J.E., 2008. Volatile compounds suitable forrapid detection as quality indicators of cold smoked salmon (Salmo salar). FoodChem. 109, 184–195.

Jorgensen, B.R.,Huss, H.H., 1989. Growth and activity of Shewanella putrefaciens isolatedfrom spoiling fish. Int. J. Food Microbiol. 9, 51–62.

Jorgensen, L.V.,Huss, H.H.,Dalgaard, P., 2001. Significance of volatile compounds producedby spoilage bacteria in vacuum-packed cold-smoked salmon (Salmo salar) analysedby GC–MS and multivariate regression. J. Agric. Food Chem. 49, 2376–2381.

Kawai, T., 1996. Fish flavor. Crit. Rev. Food Sci. Nutr. 36, 257–298.Koutsoumanis, K.,Nychas, G.-J.E., 1999. Chemical and sensory changes associated with

microbial flora of Mediterranean boque (Boops boops) stored aerobically at 0, 3, 7,and 10 °C. Appl. Environ. Microbiol. 65, 698–706.

Koutsoumanis, K.,Nychas, G.-J.E., 2000. Application of a systematic experimental proce-dure to develop a microbial model for rapid fish shelf-life predictions. Int. J. FoodMicrobiol. 60, 171–184.

Koutsoumanis, K., Lampropoulou, K.,Nychas, G.-J.E., 1999. Biogenic amines and sensorychanges associated with the microbial flora of Mediterranean gilt-head sea breamstored aerobically at 0, 8 and 15 °C. J. Food Prot. 62, 398–402.

Koutsoumanis, K.P., Taoukis, P.S.,Drosinos, E.H.,Nychas, G.-J.E., 2000. Applicability of anArrhenius model for the combined effect of temperature and CO2 packaging on thespoilage microflora of fish. Appl. Environ. Microbiol. 66 (8), 3528–3534.

Kyrana, V.R., Lougovois, V.P., 2002. Sensory, chemical and microbiological assessment offarm-raised European sea bass (Dicentrarchus labrax) stored in melting ice. Int. J.Food Sci. Technol. 37 (3), 319–328.

Kyrana, V.R.,Lougovois, V.P.,Valsamis, D.S., 1997. Assessment of shelf-life of mariculturedgilthead sea bream (Sparus aurata) stored in ice. Int. J. Food Sci. Technol. 32,339–347.

Laursen, B.G., Leisner, J.J.,Dalgaard, P., 2006. Carnobacterium species: effect of metabolicactivity and interaction with Brochothrix thermosphacta on sensory characteristicsof modified atmosphere packed shrimp. J. Agric. Food Chem. 54, 3604–3611.

Leduc, F.,Tournayre, P.,Kondjoyan, N.,Mercier, F.,Malle, P.,Kol, O.,Berdagué, J.L.,Duflos, G.,2012. Evolution of volatile odorous compounds during the storage of Europeanseabass (Dicentrarchus labrax). Food Chem. 131, 1304–1311.

Moreira, N.,Valente, L.M.P., Castro-Cunha, M., Cunha, L.M.,de Pinho, P.G., 2013. Effect ofstorage time and heat processing on the volatile profile of Senegalese sole (Soleasenegalensis Kaup, 1858) muscle. Food Chem. 138, 2365–2373.

Noseda, B., Tariqul Islam, Md, Eriksson, M., Heyndrickx, M., De Reu, K., Van Langenhove,H., Devlieghere, F., 2012. Microbiological spoilage of vacuum and modified atmo-sphere packaged Vietnamese Pangasius hypophthalmus fillets. Food Microbiol. 30,408–419.

Noseda, B.,Vermeulen, A.,Ragaert, P.,Devlieghere, F., 2014. Packaging of fish and fisheryproducts. In: Boziaris, I.S. (Ed.), Seafood Processing. Technology, Quality & Safety.IFST Advances in Food Science Series, Wiley-Blackwell, pp. 237–261.

Nychas, G.-J.E., Arkoudelos, J.S., 1991. The influence of Brochothrix thermosphacta on thequality of minced meat. J. Agric. Res. 15, 103–115.

Nychas, G.J.E.,Drosinos, E., Board, R.G., 1998. The microbiology of meat and poultry. In:Board, R.G., Davies, A.R. (Eds.), Chemical Changes in Stored Meat. Blackie Academicand Professional, London, pp. 288–326.

Nychas, G.J.E.,Marshall, D., Sofos, J., 2007. Food microbiology fundamentals and frontiers,3rd edition. In: Doyle, M.P., Beuchat, L.R., Montville, T.J. (Eds.), Meat Poultry andSeafood. ASM: Press, pp. 105–140.

Nychas, G.E.,Skandamis, P.N.,Tassou, C.C.,Koutsoumanis, K.P., 2008. Meat spoilage duringdistribution. Meat Sci. 78, 77–89.

Oehlenschläger, J., 2014. Seafood quality assessment. In: Boziaris, I.S. (Ed.), Seafood Pro-cessing. Technology, Quality & Safety. IFST Advances in Food Science Series, Wiley-Blackwell, pp. 361–386.

Olafsdottir, G.,Martinsdbttir, E., Oehlenschbger, J., Dalgaard, P., Jensen, B., Undeland, I.,Mackie, I.M.,Henehan, G.,Nielsen, J.,Nilsen, H., 1997. Methods to evaluate fish fresh-ness in research and industry. A review. Trends Food Sci. Technol. 8, 258–265.

Olafsdottir, G.,Jonsdottir, R.,Lauzon, H.L.,Luten, J.,Kristbergsson, K., 2005. Characterizationof volatile compounds in chilled cod (Gadus morhua) fillets by gas chromatographyand detection of quality indicators by an electronic nose. J. Agric. Food Chem. 53,10140–10147.

Özogul, F., Kuley, E., Özogul, Y., 2007. Sensory, chemical and microbiological qualityparameters in sea bream (Sparus aurata) stored in ice or wrapped in cling film orin aluminium foil at 2 ± 1 °C. Int. J. Food Sci. Technol. 42, 903–909.

Parlapani, F.F., 2013. Specific Spoilage Organisms and Their Effect on Quality and Fate ofPathogenic Bacteria in Fish Products. PhD Thesis (In Greek). School of AgriculturalSciences, University of Thessaly, Greece.

Parlapani, F.F.,Meziti, A.,Kormas, K.Ar.,Boziaris, I.S., 2013. Indigenous and spoilage micro-biota of farmed sea bream stored in ice identified by phenotypic and 16S rRNA geneanalysis. Food Microbiol. 33, 85–89.

Pournis, N., Papavergou, A., Badeka, A.,Kontominas, M.G., Savvaidis, I.N., 2005. Shelf-lifeextension of refrigerated Mediterranean mullet (Mullus surmuletus) using modifiedatmosphere packaging. J. Food Prot. 68, 2201–2207.

Rodriguez, O., Barros-Velazquez, J.,Ojea, A., Pineiro, C.,Aubourg, S.P., 2003. Evaluation ofsensory and microbiological changes and identification of proteolytic bacteria duringthe iced storage of farmed turbot (Psetta maxima). J. Food Sci. 68 (5), 2764–2771.

Selli, S., Cayhan, G.G., 2009. Analysis of volatile compounds of wild gilthead sea bream(Sparus aurata) by simultaneousdistillation–extraction (SDE) andGC–MS.Microchem.J. 93, 232–235.

Soncin, S., Chiesa, M.L., Panseri, S., Biondi, P., Cantoni, C., 2008. Determination of volatilecompounds of precooked prawn (Penaeus vannamei) and cultured gilthead seabream (Sparus aurata) stored in ice as possible spoilage markers using solid phasemicroextraction and gas chromatography/mass spectrometry. J. Sci. Food Agric. 89,436–442.

163F.F. Parlapani et al. / International Journal of Food Microbiology 189 (2014) 153–163

Tryfinopoulou, P., Tsakalidou, E., Nychas, G.-J.E., 2002. Characterization of Pseudomonasspp. associated with spoilage of gilt-head sea bream stored under various conditions.Appl. Environ. Microbiol. 68, 65–72.

Tryfinopoulou, P.,Tsakalidou, E.,Vancanneyt, M.,Hoste, B.,Swings, J.,Nychas, G.-J. E., 2007.Diversity of Shewanella population in fish Sparus aurata harvested in the Aegean Sea.J. Appl. Microbiol. 103, 711–721.

Tsigarida, E.,Boziaris, I.S.,Nychas, G.-J.E., 2003. Bacterial synergism or antagonism in a GelCassette System. Appl. Environ. Microbiol. 69, 7204–7209.

Tsironi, T.N.,Taoukis, P.S., 2011. Shelf-life extension of gilthead seabream fillets by osmotictreatment and antimicrobial agents. J. Appl. Microbiol. 112, 316–328.

Tsironi, T.,Salapa, I., Taoukis, P., 2009. Shelf-life modelling of osmotically treated chilledgilthead seabream fillets. Innovative Food Sci. Emerg. Technol. 10, 23–31.

Tuckey, N.P.L.,Day, J.R.,Miller, M.R., 2013. Determination of volatile compounds in NewZealand Greenshell™ mussels (Perna canaliculus) during chilled storage using solidphase microextraction gas chromatography–mass spectrometry. Food Chem. 136,218–223.

Vyncke, W.,Luten, J.,Brünner, K.,Moermans, R., 1987. Determination of total volatile basesin fish: a collaborative study by the West European Fish Technologists' Association(WEFTA). Z. Lebensm. Unters. Forsch. 184, 110–114.

Wierda, R.L., Fletcher, G., Xu, L., Dufour, J.-P., 2006. Analysis of volatile compounds asspoilage indicators in fresh king salmon (Oncorhynchus tshawytscha) during storageusing SPME–GC–MS. J. Agric. Food Chem. 54, 8480–8490.

Yasuhura, A.,Shibamoto, T., 1995. Quantitative analysis of volatile aldehydes formed fromvarious kinds of fish flesh during heat treatment. J. Agric. Food Chem. 43, 94–97.