1864

Journal of Food Protection Vol 66 No 10 2003 Pages 1864ndash1872Copyright q International Association for Food Protection

Modi ed Atmosphere Packaging for Prevention of MoldSpoilage of Bakery Products with Different pH and Water

Activity Levels

M E GUYNOT S MARIN V SANCHIS AND A J RAMOS

Food Technology Department Lleida University Unitat de Tecnologia de Productes Vegetals-Centre de Referencia en Tecnologia drsquoAliments RoviraRoure 191 25198 Lleida Spain

MS 02-440 Received 26 November 2002Accepted 6 April 2003

ABSTRACT

A sponge cake analog was used to study the in uence of pH water activity (aw) and carbon dioxide (CO2) levels onthe growth of seven fungal species commonly causing bakery product spoilage (Eurotium amstelodami Eurotium herbariorumEurotium repens Eurotium rubrum Aspergillus niger Aspergillus avus and Penicillium corylophilum) A full factorial designwas used Water activity CO2 and their interaction were the main factors signi cantly affecting fungal growth Water activityat levels of 080 to 090 had a signi cant in uence on fungal growth and determined the concentration of CO2 needed toprevent cake analog spoilage At an aw level of 085 lag phases increased twofold when the level of CO2 in the headspaceincreased from 0 to 70 In general no fungal growth was observed for up to 28 days of incubation at 258C when sampleswere packaged with 100 CO2 regardless of the aw level Partial least squares projection to latent structures regression wasused to build a polynomial model to predict sponge cake shelf life on the basis of the lag phases of all seven species testedThe model developed explained quite well (R2 5 79) the growth of almost all species which responded similarly to changesin tested factors The results of this study emphasize the importance of combining several hurdles such as modi ed atmospherepackaging aw and pH that have synergistic or additive effects on the inhibition of mold growth

Bakery products are important staple foods in mostcountries and cultures However mold growth and stalingare two problems that limit the shelf lives of both high-and intermediate-moisture bakery products (35) Fungalcontaminants have been isolated from bakery products andidenti ed as species of Eurotium Aspergillus and Penicil-lium Other species such as those of Cladosporium Mucorand Rhizopus have been found less frequently (3)

Through the use of good sanitary conditions it is pos-sible to reduce the number of unwanted fungal spores onprocessed bakery products However the presence of suchspores is expected and special attention must be paid tothem for food preservation Many traditional preservationmethods are based on the application of different lsquolsquohur-dlesrsquorsquo or lsquolsquobarriersrsquorsquo that act synergistically to inhibit orretard microbial growth (21) pH and water activity (aw) areamong the most common variables used to ensure the mi-crobial stability of a particular foodstuff Also the additionof weak organic acids such as sorbic propionic and ben-zoic acids is widely used in the preservation of food prod-ucts (5) However it is well known that the antimicrobialactivity of these acids depends on their undissociated mol-ecule Several researchers have reported the inef cacy ofthis type of preservative in preventing the spoilage of prod-ucts with a near-neutral pH (18 23 24) A more recent andincreasingly popular way of preserving foods is controlledatmosphere storage or modi ed atmosphere packaging

Author for correspondence Tel 34 973-702811 Fax 34 973-702596E-mail ajramostecaludles

(MAP) These methods take advantage of a combination ofthe inhibitory effects of low oxygen levels and elevatedcarbon dioxide levels on many deterioration processes infoods and are effective in preventing microbial spoilage(15 19)

The three major gases used in commercial MAP areoxygen (O2) nitrogen (N2) and carbon dioxide (CO2) (8)Nitrogen is an inert tasteless gas that displays little or noantimicrobial activity on its own Because of its low solu-bility in water the presence of N2 in MAP food can preventpackages from collapsing for products that can absorb CO2CO2 is the most important gas in the gas mixture it is bothbacteriostatic and fungistatic (7) Its preserving effect varieswith concentration incubation temperature organism andthe aw of the medium (15 33) El Halouat and Debevere(10) reported that a reduction in the aw level increased theinhibitory effects of high levels of CO2 on the germinationand mycelium growth of Aspergillus niger Eurotium am-stelodami Penicillium chrysogenum and Fusarium oxys-porum

Since molds are strictly aerobic for the attainment oflong shelf lives the levels of residual O2 must be kept below1 (27) Abellana et al (2) reported that the CO2 concen-tration in the headspace aw and the interaction betweenCO2 and O2 levels were the most signi cant factors affect-ing Eurotium spp growth on a sponge cake analog Toattain a residual O2 level below 1 the lowest vacuumlevel has to be reached before the shielding gas is reinjectedinto the package The lm permeability for CO2 and the

J Food Prot Vol 66 No 10 MAP FOR FUNGAL PREVENTION IN CAKES 1865

TABLE 1 Mean oxygen concentrations for cake analogs as af-fected by aw and modi ed atmosphere packaging measured after28 days of incubation at 258Ca

Atmosphere

Mean O2 concn () for aw

080 085 090

Air100 N2

70 N2 30 CO2

50 N2 50 CO2

30 N2 70 CO2

100 CO2

1765 A

060 B

052 B

077 B

075 B

105 B

114 AB

018 B

019 AB

067 AB

070 AB

135 A

107 A

068 A

046 A

003 A

053 A

120 A

a Means with different letters in the same column are signi cantlydifferent by the Duncan test

TA

BL

E2

Ana

lysi

sof

cova

rian

ce(w

ith

tim

eas

the

cova

riab

le)

ofth

eef

fect

sof

mod

ied

atm

osph

ere

pack

agin

g(M

AP

)pH

an

dw

ater

acti

vity

(aw)

and

thei

rin

tera

ctio

nson

Eur

otiu

msp

p

Asp

ergi

llus

spp

an

dP

cory

loph

ilum

colo

nyra

dius

eson

cake

anal

ogsa

Fact

or

E

amst

elod

ami

MS

F

E

herb

ario

rum

MS

F

E

repe

ns

MS

F

E

rubr

um

MS

F

A

nige

r

MS

F

A

avu

s

MS

F

P

cory

loph

ilum

MS

F

MA

PM

AP

3pH

MA

P3

a wpH pH

3a w

a w

126

495

251

113

980

314

125

711

29

159

75

317

17

66

0

400

1689

83

195

057

167

425

637

004

243

31

296

29

182

96

157

240

5

000

228

121

89

159

553

226

620

277

057

161

41

230

78

141

74

201

180

1

005

143

109

33

227

042

630

319

057

021

427

102

854

173

54

482

14

57

0

020

3378

62

477

9913

10

201

2211

07

187

885

67

729

3

200

307

0

169

028

135

13

458

2815

07

243

1133

23

108

42

114

69

605

2

199

321

1

439

1

4327

927

108

12

216

731

510

5347

77

173

6

355

10

80

2

420

8476

71

aM

S

mea

nsq

uare

P

0

05

P

0

01

characteristics of the product such as its ability to trap O2

molecules inside its structure have to be considered (2736)

The MAP technique involving various mixtures ofCO2 O2 and N2 has been used to extend the chemical andmicrobiological shelf lives of several food products suchas meat and sh (9) peanuts and pecans sh rice andbakery products (12 15 34) Much literature has describedthe effects of gaseous environments on fungal germinationand growth on food and synthetic media However reportson the use of MAP to control fungal spoilage of bakeryproducts are scarce

Despite the inherent complexities associated with thequanti cation of fungal growth (17) predictive modelingof lamentous fungal growth has been carried out by sev-eral researchers (4 29 31) Many of these researchers havefocused on the development of models for the growth rateor lag phase as a function of temperature pH and aw andonly a few studies have included carbon dioxide or oxygenconcentration as a factor (16)

The objectives of the present work were (i) to deter-mine the antifungal effects of different CO2 concentrationsin the MAP of a sponge cake analog representative of aSpanish bakery product with a near-neutral pH at differentaw and pH levels and (ii) to develop models for the pre-diction of spoilage in bakery products

MATERIALS AND METHODS

Fungal isolates Seven isolates from different bakery prod-ucts were used Five of them Eurotium amstelodami (3205) Eu-rotium herbariorum (3209) Eurotium rubrum (3228) Aspergil-lus avus (3226) and Aspergillus niger (3227) were isolatedfrom Spanish bakery products by Abellana et al (3) These iso-lates belong to the Food Technology Department microorganismcollection of Lleida University The other two isolates Eurotiumrepens (IBT18000) and Penicillium corylophilum (IBT6978)were kindly provided by the Department of Biotechnology of theTechnical University of Denmark and had been isolated from Dan-ish bakery products

Experimental design The factors used in this study includedaw (at levels of 080 085 and 090) pH (at 6 and 75) anddifferent gas compositions (air [21 O2 70 N2] 100 N2 70N2 and 30 CO2 50 N2 and 50 CO2 30 N2 and 70 CO2and 100 CO2) The aw and pH levels were selected on the basis

J Food Prot Vol 66 No 101866 GUYNOT ET AL

FIGURE 1 Growth of E amstelodami (a) and E herbariorum (b)in MAP combined effects of CO2 aw and pH level after 28 daysof incubation at 258C

FIGURE 2 Growth of E repens (a) and E rubrum (b) in MAPcombined effects of CO2 aw and pH level after 28 days of in-cubation at 258C

of the range of levels that might occur in bakery products Colonydiameter was the response recorded A full factorial design wasused and all treatments were carried out three times separately

Preparation of the analog Sponge cake analogs were pre-pared as previously reported by Abellana et al (1) Each analogwas composed of 273 g of wheat our 211 g of vegetable oil258 g of eggs and 4 g of baking powder After baking the cakehad a pH of about 75 and its initial aw value was ca 075 Toattain the required pH (60) citric acid was added to the mix ofsolid ingredients before baking whereas aw was adjusted by plac-ing the cake analogs in petri plates containing water-glycerolagar(1058 690 and 391 g of glycerol in 100 ml of distilled waterplus 15 of agar to adjust the aw values of cake analogs to 080085 and 090 respectively) Previously two calibration curveswere constructed one to determine the amount of citric acid nec-essary to reach the desired pH (24) and another to determine theconcentration of glycerol in the agar needed to increase and main-tain the aw of the dough during the incubation period (data notshown) After baking the cakes were aseptically cut into squarepieces (5 by 5 cm) and placed in the sterile 9-cm petri dishescontaining water-glycerol agar at the desired aw and the cakeswere sealed with Para lm for 48 h for the equilibration of mois-ture between the agar and the analogs

Inoculation conditions Fungi were grown on DG18 (10 gof glucose 5 g of peptone 1 g of KH2PO4 05 g of MgSO4middot7H2O220 g of glycerol 15 g of agar 02 mg of dichloran and 100 mgof chloramphenicol in 1000 ml of distilled water) for 14 days at258C to obtain sporulating cultures Spores were harvested with aloop (3 mm) and transferred to sterile distilled water (0005Tween 80) The concentration of spores was determined with aThoma chamber and adjusted to 1 3 106 to 5 3 106 spores perml Cake analogs were needle-inoculated in the center Analogswith the same aw and pH levels were inoculated with the sevendifferent isolates and packed in the same bag (eight plates [sevenisolates and one blank] per bag) The uninoculated plate (blank)was included in each bag to check aw and pH levels at the end ofthe incubation period The plastic bags used were composed of apolyethylene and polyamide coextrusion mix lm (FontPackComercial del Paper Font SA Spain) with a total thickness of90 mm The transmission rates for the lm were as follows ox-ygen 19913 cm3 m22 24 h21 atm21 [1 atm 5 10129 kPa] carbondioxide 164903 cm3 m22 24 h21 atm21 water vapor 260 cm3

m22 24 h21 atm21 The bags were lled with air or the desiredmixture of N2 and CO2 (Carburos Metalicos SE de CarburosMetalicos SA Barcelona Spain) at a productgas ratio of 13(volvol) The packaging was carried out with an EGAR VACBasic-9 Digital compensated vacuum machine (Egarvac SCPTerrassa Barcelona Spain) by creating a vacuum in the bag andthen ushing the gas mixture at around 150 kPa prior to heatsealing The bags were then stored at 258C for 28 days

J Food Prot Vol 66 No 10 MAP FOR FUNGAL PREVENTION IN CAKES 1867

FIGURE 3 Growth of A niger (a) A a-vus (b) and P corylophilum (c) in MAPcombined effects of CO2 aw and pH levelafter 28 days of incubation at 258C

Growth measurements Colony diameters were measureddaily or as required with the aid of a binocular magni er orientedin two directions at right angles to each other At the end of theexperiment the gas composition in the headspace of each packagewas sampled and analyzed by gas chromatography A micro-gaschromatograph CP2002 (Chrompack International MidderlburgThe Netherlands) tted with a thermal conductivity detector wasused A packed column COP-Molsieve (5 Aƒ 4 m by 032 cmthickness of stationary phase [df] 5 10 mm) was kept at 90 kPafor the quanti cation of O2 and N2 (558C) The CO2 concentrationwas analyzed with a Pora-PLOT Q column (10 m by 032 mmdf 5 10 mm) at 758C and 140 kPa

Data analysis An analysis of covariance of colony radiusesmeasured during the storage period with time as a covariable wascarried out for each fungal species separately in order to determinesigni cant differences between the levels of the factors assayedand their interactions For this purpose the Statistical AnalysisSystem package (SAS Version 802 SAS Institute Inc CaryNC) was used The Gompertz model was used as the tting equa-tion (25) to estimate time before visible growth (lag phase) andthis parameter was analyzed by partial least squares projection to

latent structures (PLS) in order to establish suitable secondarypredictive models PLS is a regression extension of principal com-ponents analysis which is used when it is of interest to connectthe information in two blocks of variables to each other (14) Theanalysis gives the percentage of variance of the response ex-plained by the model (R2) and the predictive power of the modelaccording to cross validation (Q2) The latter analysis was per-formed with MODDE software (Version 40 Umetrics UmeaSweden)

RESULTS

Oxygen levels in packages Table 1 presents the meanconcentrations of O2 measured in inoculated bags at the endof the incubation period At aw levels of 085 and 090 allgas atmospheres contained 135 O2 and almost no sig-ni cant differences were observed among them Howeverat an aw level of 080 the level of O2 detected was nearly17 for air-packaged samples suggesting that the limitingfactor for fungal growth was water availability The ex-haustion of O2 in air-packaged samples at high aw levels

J Food Prot Vol 66 No 101868 GUYNOT ET AL

TABLE 3 Lag phases (days to visible growth estimated by the Gompertz equation) at pH 6 under modi ed atmosphere packaging(MAP) and at different aw levelsa

aw CO2

Lag phase (days) for organism

E amstelodami E herbariorum E repens E rubrum A niger A avus P corylophilum

080 0305070

100

26 6 2028282828

214 6 3628282828

187 6 3328282828

19 6 3027 6 15

282828

2828282828

2828282828

2828282828

085 0305070

100

15 6 50121 6 22172 6 35

2828

101 6 42143 6 21118 6 30

2828

81 6 20124 6 24172 6 20

2828

98 6 17179 6 26

18 6 202828

155 6 50121 6 38

27 6 402828

126 6 1928282828

28212 6 50

282828

090 0305070

100

66 6 4445 6 2138 6 3021 6 40

28

48 6 1654 6 0839 6 4593 6 23

28

64 6 0649 6 1364 6 1683 6 1126 6 40

44 6 0771 6 2155 6 3981 6 47

28

41 6 0660 6 37

130 6 55256 6 03

28

39 6 0547 6 1457 6 2096 6 44

28

2863 6 3520 6 20

2828

a Values presented are means 6 asymptotic standard errors Twenty-eight days was the maximal incubation period

(085 to 090) was probably the reason fungal growthstopped after some days of incubation

Impact of CO2 headspace concentration on colonyradius The analysis of covariance revealed that MAP (airand 0 to 100 of CO2 balanced with N2) aw (at levels of080 085 and 090) and their interaction had a statisticallysigni cant effect (P 001) on fungal growth (Table 2)Despite the signi cant effect of pH or the interaction of pHand atmosphere on the growth of some species (E amste-lodami E rubrum A avus and P corylophilum) no clearlink was found between CO2 antifungal activity and pHlevel In general all species were affected in the same waythey grew faster as the aw level increased and the level ofCO2 in the headspace decreased Some slight differencesamong fungal speciesrsquo levels of sensitivity to gas compo-sition were found with P corylophilum being the speciesthat was most affected by the limiting factors applied fol-lowed by Aspergillus spp and Eurotium spp Under theconditions that were most favorable for fungal growth (pre-sented by cakes with high aw levels packaged in air) themaximum colony radii were ca 25 mm for Eurotium sppand 5 mm for Aspergillus spp and P corylophilum

Water activity had a signi cant in uence on fungalgrowth and determined the level of CO2 needed to preventcake analog spoilage (Figs 1 through 3) It should be notedthat the maximal colony radii for cake analogs packagedwith 30 to 70 CO2 were ca 5 mm for Eurotium spp and2 mm for Aspergillus spp and P corylophilum In generalfungal growth was prevented at all aw levels in packageswith high CO2 levels (100) However E repens and Eherbariorum were able to grow under this condition (withthe maximum colony radius observed being 25 mm) at anaw level of 090 (Figs 1 and 2)

Impact of CO2 headspace concentration on lagphases In general longer lag phases were estimated forthe conditions that were the most critical for fungal growth

(Table 3) A linear correlation between the inverse of thelag phase and the growth rate was found for all isolateswith R2 values ranging from 084 to 096 depending on theisolate Longer lag phases corresponded to lower growthrates and smaller maximal colony radii It is important tomake use of these kinds of experimental data by construct-ing models with which given a combination of factors anda storage period one can predict the approximate degree ofspoilage Although approximate these models can helpfood technologists to better predict shelf life and safety

Lag phases for the seven isolates were modeled si-multaneously with the use of all factors (pH aw and CO2

concentration) and their quadratic terms and cross-term in-teractions However since pH was nonsigni cant (data notshown) a new regression model including only signi cantterms (aw and CO2 concentration their interaction and qua-dratic terms) was calculated This new PLS regression ledto two principal components models (PC1 and PC2) (Fig4) PC1 explained most of the variance in the lag phasedate (R2 5 759) PC2 is represented by a line orthogonalto PC1 and improves the approximation of the data set asmuch as possible carrying the maximum residual infor-mation (not taken into account in PC1) (18 variance ex-plained by PC2) The predictive power of the model (Q2)was near 720

The model explained the variation for almost all spe-cies quite well with fungal responses located close to oneanother in the loadings plot indicating that all species re-sponded to changes in tested factors in similar ways (Fig4) The greater the distance of a factor or response fromthe origin the stronger its effect on the model Thus re-sponses were well explained and aw CO2 concentrationand their interaction had a strong in uence of on lag phaseswhile quadratic terms had little effect on PC1 which meantthat they had little effect on lag phases If a factor is locatedopposite a fungus a negative correlation between the levelsof the factor and lag phase of that fungus exists

J Food Prot Vol 66 No 10 MAP FOR FUNGAL PREVENTION IN CAKES 1869

FIGURE 4 Plot of principal components models 1 and 2 (obtained by PLS regression) showing the relationships between aw CO2

level and their cross and quadratic terms and lag phase before growth (days)

TABLE 4 Polynomial model equation generated by PLS of lag phase (time to visible growth) against aw CO2 and quadratic andcross termsa

Isolate Q2 R2 Equation

E amstelodamiE herbariorumE repensE rubrumA nigerA avusP corylophilum

815649815809712823447

865678857843763848588

lag phase 5 239158 1 120093aw 2 156CO2 2 85129aw2 1 (496 3 1024)(CO2)2 1 196awCO2

lag phase 5 263002 1 176363aw 2 175CO2 2 118192aw2 1 (828 3 1024)(CO2)2 1 211awCO2

lag phase 5 211152 1 51190aw 2 121CO2 2 43105aw2 1 (111 3 1024)(CO2)2 1 160awCO2

lag phase 5 26409 1 38873aw 2 111CO2 2 35036aw2 1 (503 3 1025)(CO2)2 1 148awCO2

lag phase 5 257149 1 161086aw 2 157CO2 2 107631aw2 1 (761 3 1024)(CO2)2 1 189awCO2

lag phase 5 274534 1 206957aw 2 198CO2 2 137798aw2 1 (985 3 1024)(CO2)2 1 237awCO2

lag phase 5 249724 1 134113aw 2 096CO2 2 85260aw2 1 (698 3 1024)(CO2)2 1 108awCO2

a Q2 percentage of variation predicted by the model R2 percentage of variation explained by the model

Lag phases were described by a polynomial modelequation with six coef cients (b0 b1 b22) (Table 4)lag phase 5 b0 1 b1aw 1 b2CO2 1 b12awCO2 1 b11aw

2

1 b22CO22 It should be noted that these coef cients are

unscaled and uncentered in order to allow the direct cal-culation of lag phases so they cannot be used to assess thesigni cance of the effects of the factors on lag phases

Most of the models are highly signi cant with R2 val-ues of 70 The response of P corylophilum was slightlydifferent and the model explained only 588 of the lagphase variance for this organism (Table 4) This result canalso be observed in the loadings plot (Fig 4) in which Pcorylophilum appears quite separate from the other fungalspecies

The models were subsequently used to generate two-dimensional contour plots Plots of the combined effects ofaw and CO2 are shown in Figure 5 Since similar plots wereobtained for all Eurotium species and for all Aspergillusspecies one species of each genus is represented Lag phas-es for Eurotium spp at an aw level of 087 increased two-fold when the level of CO2 in the headspace increased from0 to 70 (eg E amstelodami growth started after 20days) Storage in 100 CO2 irrespective of the aw level

resulted in lag times of 20 days When Figure 5 is con-sidered it should be taken into account that the absence ofgrowth is equivalent to a lag phase of 28 days the durationof the experiment Thus when a shelf life of ca 28 daysis predicted from the models it is probable that no spoilagewill occur

DISCUSSION

In this study the use of MAP combined with differentpH and aw values to prevent the spoilage of bakery productsby Eurotium Aspergillus and Penicillium species was eval-uated Bakery products with near-neutral pHs were used forthe experiment The results obtained corroborate those fromother studies which show that CO2 has an important fun-gistatic effect (2 15 26) Spoilage was prevented with 70CO2 in the headspace (with a residual O2 level of 135)for products with an aw level of 080 while spoilage wassigni cantly delayed for products with aw levels of 085 to090 These results con rm those of previous studies inwhich reductions in the aw level were shown to increasethe effect of high levels of CO2 (10 19 20 22) It wasfound that with 30 to 50 CO2 in the headspace Eurotiumspp growth was totally prevented at an aw level of 080

J Food Prot Vol 66 No 101870 GUYNOT ET AL

FIGURE 5 Contour plots based on models obtained by PLS re-gression showing lag phases (days) as affected by CO2 level andaw for (a) E amstelodami (b) A niger and (c) P corylophilum

as water availability increased growth was only delayedOnly a 100 CO2 atmosphere prevented spoilage by almostall isolates regardless of the aw level

The inhibitory effect of CO2 on microorganisms in aculture medium or in a food system depends on many fac-tors These factors include partial CO2 pressure O2 con-centration headspace gas volume temperature acidity andaw (15) The effect of temperature was not assayed in thisstudy and the temperature level chosen was the one thatwas most likely to be encountered during distribution andstorage in the retail market The preserving effect of CO2

strongly depended on aw while interaction with pH was

not important probably because of the similar pH valuesinvolved Similarly Ellis et al (11 12) demonstrated thatthe growth of A avus under different atmospheric CO2

and O2 concentrations was highly dependent on aw andtemperature However these authors did not nd importantdifferences between the effects of gas composition on fun-gal growth at pH 6 and those at pH 8 Haasum and Nielsen(20) in a study of the growth of fungi in a cheese envi-ronment also found no signi cant differences in the inhib-itory effects of CO2 at different pH levels (4ndash8) The ef-fectiveness of CO2 also varies with the type of organismunder consideration with molds being more sensitive thanyeasts (19 22 30)

In this study all O2 concentrations measured at the endof the incubation period except those for air-packaged bagswere between 02 and 135 Since molds by de nition arestrict aerobes suf cient residual oxygen must be present inthe package headspace to allow mold growth (10 34) Ithas been demonstrated that molds can tolerate and evengrow in air with headspace oxygen concentrations as lowas 1 to 2 (13 32) Moreover several studies have shownthat molds can grow in the presence of elevated CO2 levelsif O2 is present (11 12) Smith et al (34) reported that aminimum of 04 O2 was needed for the growth of A nigerand Penicillium spp in a CO2N2 (6040) atmosphereAbellana et al (2) also found that low levels (002 to 05)of O2 did not modify fungal growth when levels of CO2 inthe bags were high However these authors found that asthe CO2 concentration decreased O2 levels had more in u-ence on growth kinetics (small differences in O2 levelscould make fungi grow) Similarly Nielsen and Rios (26)found that the growth of Aspergillus and Penicillium spe-cies was delayed only at reduced O2 levels (002 to 003)in either pure nitrogen or 50 N2 and 50 CO2 Howeverin pure CO2 the molds were strongly inhibited despite theresidual O2 in the package The total elimination of air dur-ing MAP represents a serious hurdle owing to both the lmrsquos permeability and the productrsquos capacity to trap O2

molecules (28) When a vacuum evacuation of the packageis carried out prior to gas injection the level of residual O2

in the headspace can be 1 (30) However the smoothand fragile texture of sponge cakes must be taken into ac-count the cakes could be deformed (27) An innovativemethod for controlling oxygen without physically deform-ing products involves the incorporation of O2 absorbers intothe package (32)

CO2 is both water and lipid soluble and when gaseousCO2 is applied to a product part of the CO2 goes to theliquid phase (water or oil) as carbonic acid (about 2) (7)This development can produce a small decrease in pH re-sulting in changes in the organoleptic characteristics of theproduct (15 30) The solubility of this gas in a culturemedium or in a food system increases as pH increases asaw increases as the volume of headspace gas increases oras temperature decreases The pHs of cake analogs pack-aged aerobically and with the different gas compositionstested were measured and no appreciable differences werefound (data not shown) This result may suggest that thesolubility of CO2 in the cake analog used was not signi -

J Food Prot Vol 66 No 10 MAP FOR FUNGAL PREVENTION IN CAKES 1871

cant or was not suf cient to produce a decrease in the an-alog pH

In the last few years predictive microbiology has be-come an important area of research The advantage of theresponse surface modeling approach is that it allows thesimultaneous examination of several variables Also math-ematical models can be generated to predict the packagingand environmental storage conditions that would controlfungal growth (11) Abellana et al (2) predicted lag phasesradial growth rates and maximum growth in relation to awO2 and CO2 concentrations for three Eurotium species ona sponge cake analog A model such as that described heremay be an important tool to aid food technologists in pre-dicting shelf lives of products or may give an idea of theeffects of certain changes in a productrsquos formulation Alsothis study emphasizes the importance of combining severalbarriers such as MAP aw and pH whose effects on moldgrowth are synergistic or additive

ACKNOWLEDGMENTS

The authors acknowledge the EC Quality of Life Programme (QoL)Key Action 1 (KA1) on Food Nutrition and Health (PL98-4075) theSpanish government (CICYT ALI99-0831) and the Catalonian govern-ment (CIRIT Comissio Interdepartamental de Recerca i Innovacio Tec-nologica)

REFERENCES

1 Abellana M J BenedDagger V Sanchis and A J Ramos 1999 Wateractivity and temperature effects on germination and growth of Eu-rotium amstelodami E chevalieri and E herbariorum isolates frombakery products J Appl Microbiol 87371ndash380

2 Abellana M V Sanchis A J Ramos and P V Nielsen 2000Effect of modi ed atmosphere packaging and water activity ongrowth of Eurotium amstelodami E chevalieri and E herbariorumon a sponge cake analogue J Appl Microbiol 88606ndash616

3 Abellana M L Torres V Sanchis and A J Ramos 1997 Car-acterizacion de diferentes productos de bollerDaggera industrial II Estudiode la mico ora Alimentaria 28751ndash56

4 Baranyi J A M Gibson J I Pitt M J Eyles and T A Roberts1996 Predictive models as means of measuring the relatedness ofsome Aspergillus species Food Microbiol 14347ndash351

5 Chirife J and G J Favetto 1992 Some physico-chemical basisof food preservation by combined methods Food Res Int 25389ndash396

6 Daifas D P J P Smith B Blanch eld and J W Austin 1999Growth and toxin production by Clostridium botulinum in English-style crumpets packaged under modi ed atmosphere J Food Prot62349ndash355

7 Daniels J A R Krishnamurthi and S S H Rizvi 1985 A reviewof effects of carbon dioxide on microbial growth and food qualityJ Food Prot 48532ndash537

8 Davies A R 1995 Advances in modi ed-atmosphere packaging p304ndash320 In G W Gould (ed) New methods of food preservationBlackie Academic and Professional London

9 Dixon N M and D B Kell 1989 The inhibition by CO2 of thegrowth and metabolism of micro-organisms J Appl Bacteriol 67109ndash136

10 El Halouat A and J M Debevere 1997 Effect of water activitymodi ed atmosphere packing and storage temperature on spore ger-mination of moulds isolated from prunes Int J Food Microbiol 3541ndash48

11 Ellis W O J P Smith B K Simpson S Khanizadeh and J HOldhem 1993 Control of growth and a atoxin production of As-pergillus avus under modi ed atmosphere packaging (MAP) con-ditions Food Microbiol 109ndash21

12 Ellis W O J P Smith B K Simpson and H Ramaswamy 1993

Effect of inoculum level on a atoxin production by Aspergillus a-vus under modi ed atmosphere packaging (MAP) conditions FoodMicrobiol 10525ndash535

13 Ellis W O J P Smith B K Simpson H Ramaswamy and GDoyon 1994 Growth of and a atoxin production by Aspergillus avus in peanuts stored under modi ed atmosphere packaging(MAP) conditions Int J Food Microbiol 22173ndash187

14 Eriksson L E Johansson N Kettaneh-Wold and S Wold 1999Introduction to multi- and megavariate data analysis using projectionmethods (PCA amp PLS) Umetrics AB Sweden

15 Faber J M 1991 Microbial aspects of modi ed-atmosphere pack-aging technologymdasha review J Food Prot 5458ndash70

16 GarcDaggera-Gimeno R M C Sanz-MartDaggernez J M GarcDaggera-Martos andG Zurera-Cosano 2002 Modeling Botrytis cinerea spores growthin carbon dioxide enriched atmospheres Food Microbiol Saf 671904ndash1907

17 Gibson A M and A D Hocking 1997 Advances in the predictivemodelling of fungal growth in food Trends Food Sci Technol 8353ndash358

18 Guynot M E A J Ramos D Sala V Sanchis and S MarDaggern2002 Combined effects of weak acid preservatives pH and wateractivity on growth of Eurotium species on a sponge cake Int JFood Microbiol 7639ndash46

19 Haasum I and P V Nielsen 1998 Ecophysiological characteriza-tion of common food-borne fungi in relation to pH and water activityunder various atmospheric compositions J Appl Microbiol 84451ndash460

20 Haasum I and P V Nielsen 1998 Physiological characterizationof common fungi associated with cheese J Food Sci 63157ndash161

21 Leistner L 1992 Food preservation by combined methods FoodRes Int 25151ndash158

22 Magan N and J Lacey 1984 Effects of gas composition and wateractivity on growth of eld and storage fungi and their interactionsTrans Br Mycol Soc 82305ndash314

23 MarDaggern S M E Guynot P Neira M Bernado V Sanchis and AJ Ramos 2002 Risk assessment of the use of sub-optimal levels ofweak-acid preservatives in the control of mould growth on bakeryproducts Int J Food Microbiol 79203ndash211

24 MarDaggern S M E Guynot V Sanchis J Arbones and A J Ramos2002 Aspergillus avus Aspergillus niger and Penicillium corylo-philum spoilage prevention of bakery products by means of weak-acid preservatives J Food Sci 672271ndash2277

25 MarDaggern S V Sanchis A Teixido et al 1996 Water and tem-perature relations and microconidial germination of Fusariummoniliforme and F proliferatum from maize Can J Microbiol421045ndash1054

26 Nielsen P V and R Rios 2000 Inhibition of fungal growth onbread by volatile components from species and herbs and the pos-sible application in active packaging with special emphasis on mus-tard essential oil Int J Food Microbiol 60219ndash229

27 Ortola C and C Santacreu 1998 Principios de aplicacion del en-vasado en atmosfera modi cada a los productos de pani cacion ybollerDaggera Aliment Equip Tecnol 17111ndash117

28 Piergiovanni L and P Fava 1997 Minimizing the residual oxygenin modi ed atmosphere packaging of bakery products Food AdditContam 14765ndash773

29 Pitt R E 1993 A descriptive model of mold growth and a atoxinformation as affected by environmental conditions J Food Prot 56139ndash146

30 RodrDaggerguez Castilla M V and R Jordano 1997 Envasado de prod-uctos de panaderDaggera y bollerDaggera en atmosferas modi cadas composi-cion de la atmosfera de envasado lms equipos de envasado y efec-tos sobre los productos de panaderDaggera y bollerDaggera Alimentaria 28679ndash99

31 Sautour M P Dantigny C Divies and M Bensoussan 2001A temperature-type model for describing the relationship be-tween fungal growth and water activity Int J Food Microbiol6763ndash69

32 Smith J P J Hoshino and Y Abe 1995 Interactive packaginginvolving sachet technology p 143ndash176 In M L Rooney (ed)

J Food Prot Vol 66 No 10 MAP FOR FUNGAL PREVENTION IN CAKES 1865

TABLE 1 Mean oxygen concentrations for cake analogs as af-fected by aw and modi ed atmosphere packaging measured after28 days of incubation at 258Ca

Atmosphere

Mean O2 concn () for aw

080 085 090

Air100 N2

70 N2 30 CO2

50 N2 50 CO2

30 N2 70 CO2

100 CO2

1765 A

060 B

052 B

077 B

075 B

105 B

114 AB

018 B

019 AB

067 AB

070 AB

135 A

107 A

068 A

046 A

003 A

053 A

120 A

a Means with different letters in the same column are signi cantlydifferent by the Duncan test

TA

BL

E2

Ana

lysi

sof

cova

rian

ce(w

ith

tim

eas

the

cova

riab

le)

ofth

eef

fect

sof

mod

ied

atm

osph

ere

pack

agin

g(M

AP

)pH

an

dw

ater

acti

vity

(aw)

and

thei

rin

tera

ctio

nson

Eur

otiu

msp

p

Asp

ergi

llus

spp

an

dP

cory

loph

ilum

colo

nyra

dius

eson

cake

anal

ogsa

Fact

or

E

amst

elod

ami

MS

F

E

herb

ario

rum

MS

F

E

repe

ns

MS

F

E

rubr

um

MS

F

A

nige

r

MS

F

A

avu

s

MS

F

P

cory

loph

ilum

MS

F

MA

PM

AP

3pH

MA

P3

a wpH pH

3a w

a w

126

495

251

113

980

314

125

711

29

159

75

317

17

66

0

400

1689

83

195

057

167

425

637

004

243

31

296

29

182

96

157

240

5

000

228

121

89

159

553

226

620

277

057

161

41

230

78

141

74

201

180

1

005

143

109

33

227

042

630

319

057

021

427

102

854

173

54

482

14

57

0

020

3378

62

477

9913

10

201

2211

07

187

885

67

729

3

200

307

0

169

028

135

13

458

2815

07

243

1133

23

108

42

114

69

605

2

199

321

1

439

1

4327

927

108

12

216

731

510

5347

77

173

6

355

10

80

2

420

8476

71

aM

S

mea

nsq

uare

P

0

05

P

0

01

characteristics of the product such as its ability to trap O2

molecules inside its structure have to be considered (2736)

The MAP technique involving various mixtures ofCO2 O2 and N2 has been used to extend the chemical andmicrobiological shelf lives of several food products suchas meat and sh (9) peanuts and pecans sh rice andbakery products (12 15 34) Much literature has describedthe effects of gaseous environments on fungal germinationand growth on food and synthetic media However reportson the use of MAP to control fungal spoilage of bakeryproducts are scarce

Despite the inherent complexities associated with thequanti cation of fungal growth (17) predictive modelingof lamentous fungal growth has been carried out by sev-eral researchers (4 29 31) Many of these researchers havefocused on the development of models for the growth rateor lag phase as a function of temperature pH and aw andonly a few studies have included carbon dioxide or oxygenconcentration as a factor (16)

The objectives of the present work were (i) to deter-mine the antifungal effects of different CO2 concentrationsin the MAP of a sponge cake analog representative of aSpanish bakery product with a near-neutral pH at differentaw and pH levels and (ii) to develop models for the pre-diction of spoilage in bakery products

MATERIALS AND METHODS

Fungal isolates Seven isolates from different bakery prod-ucts were used Five of them Eurotium amstelodami (3205) Eu-rotium herbariorum (3209) Eurotium rubrum (3228) Aspergil-lus avus (3226) and Aspergillus niger (3227) were isolatedfrom Spanish bakery products by Abellana et al (3) These iso-lates belong to the Food Technology Department microorganismcollection of Lleida University The other two isolates Eurotiumrepens (IBT18000) and Penicillium corylophilum (IBT6978)were kindly provided by the Department of Biotechnology of theTechnical University of Denmark and had been isolated from Dan-ish bakery products

Experimental design The factors used in this study includedaw (at levels of 080 085 and 090) pH (at 6 and 75) anddifferent gas compositions (air [21 O2 70 N2] 100 N2 70N2 and 30 CO2 50 N2 and 50 CO2 30 N2 and 70 CO2and 100 CO2) The aw and pH levels were selected on the basis

J Food Prot Vol 66 No 101866 GUYNOT ET AL

FIGURE 1 Growth of E amstelodami (a) and E herbariorum (b)in MAP combined effects of CO2 aw and pH level after 28 daysof incubation at 258C

FIGURE 2 Growth of E repens (a) and E rubrum (b) in MAPcombined effects of CO2 aw and pH level after 28 days of in-cubation at 258C

of the range of levels that might occur in bakery products Colonydiameter was the response recorded A full factorial design wasused and all treatments were carried out three times separately

Preparation of the analog Sponge cake analogs were pre-pared as previously reported by Abellana et al (1) Each analogwas composed of 273 g of wheat our 211 g of vegetable oil258 g of eggs and 4 g of baking powder After baking the cakehad a pH of about 75 and its initial aw value was ca 075 Toattain the required pH (60) citric acid was added to the mix ofsolid ingredients before baking whereas aw was adjusted by plac-ing the cake analogs in petri plates containing water-glycerolagar(1058 690 and 391 g of glycerol in 100 ml of distilled waterplus 15 of agar to adjust the aw values of cake analogs to 080085 and 090 respectively) Previously two calibration curveswere constructed one to determine the amount of citric acid nec-essary to reach the desired pH (24) and another to determine theconcentration of glycerol in the agar needed to increase and main-tain the aw of the dough during the incubation period (data notshown) After baking the cakes were aseptically cut into squarepieces (5 by 5 cm) and placed in the sterile 9-cm petri dishescontaining water-glycerol agar at the desired aw and the cakeswere sealed with Para lm for 48 h for the equilibration of mois-ture between the agar and the analogs

Inoculation conditions Fungi were grown on DG18 (10 gof glucose 5 g of peptone 1 g of KH2PO4 05 g of MgSO4middot7H2O220 g of glycerol 15 g of agar 02 mg of dichloran and 100 mgof chloramphenicol in 1000 ml of distilled water) for 14 days at258C to obtain sporulating cultures Spores were harvested with aloop (3 mm) and transferred to sterile distilled water (0005Tween 80) The concentration of spores was determined with aThoma chamber and adjusted to 1 3 106 to 5 3 106 spores perml Cake analogs were needle-inoculated in the center Analogswith the same aw and pH levels were inoculated with the sevendifferent isolates and packed in the same bag (eight plates [sevenisolates and one blank] per bag) The uninoculated plate (blank)was included in each bag to check aw and pH levels at the end ofthe incubation period The plastic bags used were composed of apolyethylene and polyamide coextrusion mix lm (FontPackComercial del Paper Font SA Spain) with a total thickness of90 mm The transmission rates for the lm were as follows ox-ygen 19913 cm3 m22 24 h21 atm21 [1 atm 5 10129 kPa] carbondioxide 164903 cm3 m22 24 h21 atm21 water vapor 260 cm3

m22 24 h21 atm21 The bags were lled with air or the desiredmixture of N2 and CO2 (Carburos Metalicos SE de CarburosMetalicos SA Barcelona Spain) at a productgas ratio of 13(volvol) The packaging was carried out with an EGAR VACBasic-9 Digital compensated vacuum machine (Egarvac SCPTerrassa Barcelona Spain) by creating a vacuum in the bag andthen ushing the gas mixture at around 150 kPa prior to heatsealing The bags were then stored at 258C for 28 days

J Food Prot Vol 66 No 10 MAP FOR FUNGAL PREVENTION IN CAKES 1867

FIGURE 3 Growth of A niger (a) A a-vus (b) and P corylophilum (c) in MAPcombined effects of CO2 aw and pH levelafter 28 days of incubation at 258C

Growth measurements Colony diameters were measureddaily or as required with the aid of a binocular magni er orientedin two directions at right angles to each other At the end of theexperiment the gas composition in the headspace of each packagewas sampled and analyzed by gas chromatography A micro-gaschromatograph CP2002 (Chrompack International MidderlburgThe Netherlands) tted with a thermal conductivity detector wasused A packed column COP-Molsieve (5 Aƒ 4 m by 032 cmthickness of stationary phase [df] 5 10 mm) was kept at 90 kPafor the quanti cation of O2 and N2 (558C) The CO2 concentrationwas analyzed with a Pora-PLOT Q column (10 m by 032 mmdf 5 10 mm) at 758C and 140 kPa

Data analysis An analysis of covariance of colony radiusesmeasured during the storage period with time as a covariable wascarried out for each fungal species separately in order to determinesigni cant differences between the levels of the factors assayedand their interactions For this purpose the Statistical AnalysisSystem package (SAS Version 802 SAS Institute Inc CaryNC) was used The Gompertz model was used as the tting equa-tion (25) to estimate time before visible growth (lag phase) andthis parameter was analyzed by partial least squares projection to

latent structures (PLS) in order to establish suitable secondarypredictive models PLS is a regression extension of principal com-ponents analysis which is used when it is of interest to connectthe information in two blocks of variables to each other (14) Theanalysis gives the percentage of variance of the response ex-plained by the model (R2) and the predictive power of the modelaccording to cross validation (Q2) The latter analysis was per-formed with MODDE software (Version 40 Umetrics UmeaSweden)

RESULTS

Oxygen levels in packages Table 1 presents the meanconcentrations of O2 measured in inoculated bags at the endof the incubation period At aw levels of 085 and 090 allgas atmospheres contained 135 O2 and almost no sig-ni cant differences were observed among them Howeverat an aw level of 080 the level of O2 detected was nearly17 for air-packaged samples suggesting that the limitingfactor for fungal growth was water availability The ex-haustion of O2 in air-packaged samples at high aw levels

J Food Prot Vol 66 No 101868 GUYNOT ET AL

TABLE 3 Lag phases (days to visible growth estimated by the Gompertz equation) at pH 6 under modi ed atmosphere packaging(MAP) and at different aw levelsa

aw CO2

Lag phase (days) for organism

E amstelodami E herbariorum E repens E rubrum A niger A avus P corylophilum

080 0305070

100

26 6 2028282828

214 6 3628282828

187 6 3328282828

19 6 3027 6 15

282828

2828282828

2828282828

2828282828

085 0305070

100

15 6 50121 6 22172 6 35

2828

101 6 42143 6 21118 6 30

2828

81 6 20124 6 24172 6 20

2828

98 6 17179 6 26

18 6 202828

155 6 50121 6 38

27 6 402828

126 6 1928282828

28212 6 50

282828

090 0305070

100

66 6 4445 6 2138 6 3021 6 40

28

48 6 1654 6 0839 6 4593 6 23

28

64 6 0649 6 1364 6 1683 6 1126 6 40

44 6 0771 6 2155 6 3981 6 47

28

41 6 0660 6 37

130 6 55256 6 03

28

39 6 0547 6 1457 6 2096 6 44

28

2863 6 3520 6 20

2828

a Values presented are means 6 asymptotic standard errors Twenty-eight days was the maximal incubation period

(085 to 090) was probably the reason fungal growthstopped after some days of incubation

Impact of CO2 headspace concentration on colonyradius The analysis of covariance revealed that MAP (airand 0 to 100 of CO2 balanced with N2) aw (at levels of080 085 and 090) and their interaction had a statisticallysigni cant effect (P 001) on fungal growth (Table 2)Despite the signi cant effect of pH or the interaction of pHand atmosphere on the growth of some species (E amste-lodami E rubrum A avus and P corylophilum) no clearlink was found between CO2 antifungal activity and pHlevel In general all species were affected in the same waythey grew faster as the aw level increased and the level ofCO2 in the headspace decreased Some slight differencesamong fungal speciesrsquo levels of sensitivity to gas compo-sition were found with P corylophilum being the speciesthat was most affected by the limiting factors applied fol-lowed by Aspergillus spp and Eurotium spp Under theconditions that were most favorable for fungal growth (pre-sented by cakes with high aw levels packaged in air) themaximum colony radii were ca 25 mm for Eurotium sppand 5 mm for Aspergillus spp and P corylophilum

Water activity had a signi cant in uence on fungalgrowth and determined the level of CO2 needed to preventcake analog spoilage (Figs 1 through 3) It should be notedthat the maximal colony radii for cake analogs packagedwith 30 to 70 CO2 were ca 5 mm for Eurotium spp and2 mm for Aspergillus spp and P corylophilum In generalfungal growth was prevented at all aw levels in packageswith high CO2 levels (100) However E repens and Eherbariorum were able to grow under this condition (withthe maximum colony radius observed being 25 mm) at anaw level of 090 (Figs 1 and 2)

Impact of CO2 headspace concentration on lagphases In general longer lag phases were estimated forthe conditions that were the most critical for fungal growth

(Table 3) A linear correlation between the inverse of thelag phase and the growth rate was found for all isolateswith R2 values ranging from 084 to 096 depending on theisolate Longer lag phases corresponded to lower growthrates and smaller maximal colony radii It is important tomake use of these kinds of experimental data by construct-ing models with which given a combination of factors anda storage period one can predict the approximate degree ofspoilage Although approximate these models can helpfood technologists to better predict shelf life and safety

Lag phases for the seven isolates were modeled si-multaneously with the use of all factors (pH aw and CO2

concentration) and their quadratic terms and cross-term in-teractions However since pH was nonsigni cant (data notshown) a new regression model including only signi cantterms (aw and CO2 concentration their interaction and qua-dratic terms) was calculated This new PLS regression ledto two principal components models (PC1 and PC2) (Fig4) PC1 explained most of the variance in the lag phasedate (R2 5 759) PC2 is represented by a line orthogonalto PC1 and improves the approximation of the data set asmuch as possible carrying the maximum residual infor-mation (not taken into account in PC1) (18 variance ex-plained by PC2) The predictive power of the model (Q2)was near 720

The model explained the variation for almost all spe-cies quite well with fungal responses located close to oneanother in the loadings plot indicating that all species re-sponded to changes in tested factors in similar ways (Fig4) The greater the distance of a factor or response fromthe origin the stronger its effect on the model Thus re-sponses were well explained and aw CO2 concentrationand their interaction had a strong in uence of on lag phaseswhile quadratic terms had little effect on PC1 which meantthat they had little effect on lag phases If a factor is locatedopposite a fungus a negative correlation between the levelsof the factor and lag phase of that fungus exists

J Food Prot Vol 66 No 10 MAP FOR FUNGAL PREVENTION IN CAKES 1869

FIGURE 4 Plot of principal components models 1 and 2 (obtained by PLS regression) showing the relationships between aw CO2

level and their cross and quadratic terms and lag phase before growth (days)

TABLE 4 Polynomial model equation generated by PLS of lag phase (time to visible growth) against aw CO2 and quadratic andcross termsa

Isolate Q2 R2 Equation

E amstelodamiE herbariorumE repensE rubrumA nigerA avusP corylophilum

815649815809712823447

865678857843763848588

lag phase 5 239158 1 120093aw 2 156CO2 2 85129aw2 1 (496 3 1024)(CO2)2 1 196awCO2

lag phase 5 263002 1 176363aw 2 175CO2 2 118192aw2 1 (828 3 1024)(CO2)2 1 211awCO2

lag phase 5 211152 1 51190aw 2 121CO2 2 43105aw2 1 (111 3 1024)(CO2)2 1 160awCO2

lag phase 5 26409 1 38873aw 2 111CO2 2 35036aw2 1 (503 3 1025)(CO2)2 1 148awCO2

lag phase 5 257149 1 161086aw 2 157CO2 2 107631aw2 1 (761 3 1024)(CO2)2 1 189awCO2

lag phase 5 274534 1 206957aw 2 198CO2 2 137798aw2 1 (985 3 1024)(CO2)2 1 237awCO2

lag phase 5 249724 1 134113aw 2 096CO2 2 85260aw2 1 (698 3 1024)(CO2)2 1 108awCO2

a Q2 percentage of variation predicted by the model R2 percentage of variation explained by the model

Lag phases were described by a polynomial modelequation with six coef cients (b0 b1 b22) (Table 4)lag phase 5 b0 1 b1aw 1 b2CO2 1 b12awCO2 1 b11aw

2

1 b22CO22 It should be noted that these coef cients are

unscaled and uncentered in order to allow the direct cal-culation of lag phases so they cannot be used to assess thesigni cance of the effects of the factors on lag phases

Most of the models are highly signi cant with R2 val-ues of 70 The response of P corylophilum was slightlydifferent and the model explained only 588 of the lagphase variance for this organism (Table 4) This result canalso be observed in the loadings plot (Fig 4) in which Pcorylophilum appears quite separate from the other fungalspecies

The models were subsequently used to generate two-dimensional contour plots Plots of the combined effects ofaw and CO2 are shown in Figure 5 Since similar plots wereobtained for all Eurotium species and for all Aspergillusspecies one species of each genus is represented Lag phas-es for Eurotium spp at an aw level of 087 increased two-fold when the level of CO2 in the headspace increased from0 to 70 (eg E amstelodami growth started after 20days) Storage in 100 CO2 irrespective of the aw level

resulted in lag times of 20 days When Figure 5 is con-sidered it should be taken into account that the absence ofgrowth is equivalent to a lag phase of 28 days the durationof the experiment Thus when a shelf life of ca 28 daysis predicted from the models it is probable that no spoilagewill occur

DISCUSSION

In this study the use of MAP combined with differentpH and aw values to prevent the spoilage of bakery productsby Eurotium Aspergillus and Penicillium species was eval-uated Bakery products with near-neutral pHs were used forthe experiment The results obtained corroborate those fromother studies which show that CO2 has an important fun-gistatic effect (2 15 26) Spoilage was prevented with 70CO2 in the headspace (with a residual O2 level of 135)for products with an aw level of 080 while spoilage wassigni cantly delayed for products with aw levels of 085 to090 These results con rm those of previous studies inwhich reductions in the aw level were shown to increasethe effect of high levels of CO2 (10 19 20 22) It wasfound that with 30 to 50 CO2 in the headspace Eurotiumspp growth was totally prevented at an aw level of 080

J Food Prot Vol 66 No 101870 GUYNOT ET AL

FIGURE 5 Contour plots based on models obtained by PLS re-gression showing lag phases (days) as affected by CO2 level andaw for (a) E amstelodami (b) A niger and (c) P corylophilum

as water availability increased growth was only delayedOnly a 100 CO2 atmosphere prevented spoilage by almostall isolates regardless of the aw level

The inhibitory effect of CO2 on microorganisms in aculture medium or in a food system depends on many fac-tors These factors include partial CO2 pressure O2 con-centration headspace gas volume temperature acidity andaw (15) The effect of temperature was not assayed in thisstudy and the temperature level chosen was the one thatwas most likely to be encountered during distribution andstorage in the retail market The preserving effect of CO2

strongly depended on aw while interaction with pH was

not important probably because of the similar pH valuesinvolved Similarly Ellis et al (11 12) demonstrated thatthe growth of A avus under different atmospheric CO2

and O2 concentrations was highly dependent on aw andtemperature However these authors did not nd importantdifferences between the effects of gas composition on fun-gal growth at pH 6 and those at pH 8 Haasum and Nielsen(20) in a study of the growth of fungi in a cheese envi-ronment also found no signi cant differences in the inhib-itory effects of CO2 at different pH levels (4ndash8) The ef-fectiveness of CO2 also varies with the type of organismunder consideration with molds being more sensitive thanyeasts (19 22 30)

In this study all O2 concentrations measured at the endof the incubation period except those for air-packaged bagswere between 02 and 135 Since molds by de nition arestrict aerobes suf cient residual oxygen must be present inthe package headspace to allow mold growth (10 34) Ithas been demonstrated that molds can tolerate and evengrow in air with headspace oxygen concentrations as lowas 1 to 2 (13 32) Moreover several studies have shownthat molds can grow in the presence of elevated CO2 levelsif O2 is present (11 12) Smith et al (34) reported that aminimum of 04 O2 was needed for the growth of A nigerand Penicillium spp in a CO2N2 (6040) atmosphereAbellana et al (2) also found that low levels (002 to 05)of O2 did not modify fungal growth when levels of CO2 inthe bags were high However these authors found that asthe CO2 concentration decreased O2 levels had more in u-ence on growth kinetics (small differences in O2 levelscould make fungi grow) Similarly Nielsen and Rios (26)found that the growth of Aspergillus and Penicillium spe-cies was delayed only at reduced O2 levels (002 to 003)in either pure nitrogen or 50 N2 and 50 CO2 Howeverin pure CO2 the molds were strongly inhibited despite theresidual O2 in the package The total elimination of air dur-ing MAP represents a serious hurdle owing to both the lmrsquos permeability and the productrsquos capacity to trap O2

molecules (28) When a vacuum evacuation of the packageis carried out prior to gas injection the level of residual O2

in the headspace can be 1 (30) However the smoothand fragile texture of sponge cakes must be taken into ac-count the cakes could be deformed (27) An innovativemethod for controlling oxygen without physically deform-ing products involves the incorporation of O2 absorbers intothe package (32)

CO2 is both water and lipid soluble and when gaseousCO2 is applied to a product part of the CO2 goes to theliquid phase (water or oil) as carbonic acid (about 2) (7)This development can produce a small decrease in pH re-sulting in changes in the organoleptic characteristics of theproduct (15 30) The solubility of this gas in a culturemedium or in a food system increases as pH increases asaw increases as the volume of headspace gas increases oras temperature decreases The pHs of cake analogs pack-aged aerobically and with the different gas compositionstested were measured and no appreciable differences werefound (data not shown) This result may suggest that thesolubility of CO2 in the cake analog used was not signi -

J Food Prot Vol 66 No 10 MAP FOR FUNGAL PREVENTION IN CAKES 1871

cant or was not suf cient to produce a decrease in the an-alog pH

In the last few years predictive microbiology has be-come an important area of research The advantage of theresponse surface modeling approach is that it allows thesimultaneous examination of several variables Also math-ematical models can be generated to predict the packagingand environmental storage conditions that would controlfungal growth (11) Abellana et al (2) predicted lag phasesradial growth rates and maximum growth in relation to awO2 and CO2 concentrations for three Eurotium species ona sponge cake analog A model such as that described heremay be an important tool to aid food technologists in pre-dicting shelf lives of products or may give an idea of theeffects of certain changes in a productrsquos formulation Alsothis study emphasizes the importance of combining severalbarriers such as MAP aw and pH whose effects on moldgrowth are synergistic or additive

ACKNOWLEDGMENTS

The authors acknowledge the EC Quality of Life Programme (QoL)Key Action 1 (KA1) on Food Nutrition and Health (PL98-4075) theSpanish government (CICYT ALI99-0831) and the Catalonian govern-ment (CIRIT Comissio Interdepartamental de Recerca i Innovacio Tec-nologica)

REFERENCES

1 Abellana M J BenedDagger V Sanchis and A J Ramos 1999 Wateractivity and temperature effects on germination and growth of Eu-rotium amstelodami E chevalieri and E herbariorum isolates frombakery products J Appl Microbiol 87371ndash380

2 Abellana M V Sanchis A J Ramos and P V Nielsen 2000Effect of modi ed atmosphere packaging and water activity ongrowth of Eurotium amstelodami E chevalieri and E herbariorumon a sponge cake analogue J Appl Microbiol 88606ndash616

3 Abellana M L Torres V Sanchis and A J Ramos 1997 Car-acterizacion de diferentes productos de bollerDaggera industrial II Estudiode la mico ora Alimentaria 28751ndash56

4 Baranyi J A M Gibson J I Pitt M J Eyles and T A Roberts1996 Predictive models as means of measuring the relatedness ofsome Aspergillus species Food Microbiol 14347ndash351

5 Chirife J and G J Favetto 1992 Some physico-chemical basisof food preservation by combined methods Food Res Int 25389ndash396

6 Daifas D P J P Smith B Blanch eld and J W Austin 1999Growth and toxin production by Clostridium botulinum in English-style crumpets packaged under modi ed atmosphere J Food Prot62349ndash355

7 Daniels J A R Krishnamurthi and S S H Rizvi 1985 A reviewof effects of carbon dioxide on microbial growth and food qualityJ Food Prot 48532ndash537

8 Davies A R 1995 Advances in modi ed-atmosphere packaging p304ndash320 In G W Gould (ed) New methods of food preservationBlackie Academic and Professional London

9 Dixon N M and D B Kell 1989 The inhibition by CO2 of thegrowth and metabolism of micro-organisms J Appl Bacteriol 67109ndash136

10 El Halouat A and J M Debevere 1997 Effect of water activitymodi ed atmosphere packing and storage temperature on spore ger-mination of moulds isolated from prunes Int J Food Microbiol 3541ndash48

11 Ellis W O J P Smith B K Simpson S Khanizadeh and J HOldhem 1993 Control of growth and a atoxin production of As-pergillus avus under modi ed atmosphere packaging (MAP) con-ditions Food Microbiol 109ndash21

12 Ellis W O J P Smith B K Simpson and H Ramaswamy 1993

Effect of inoculum level on a atoxin production by Aspergillus a-vus under modi ed atmosphere packaging (MAP) conditions FoodMicrobiol 10525ndash535

13 Ellis W O J P Smith B K Simpson H Ramaswamy and GDoyon 1994 Growth of and a atoxin production by Aspergillus avus in peanuts stored under modi ed atmosphere packaging(MAP) conditions Int J Food Microbiol 22173ndash187

14 Eriksson L E Johansson N Kettaneh-Wold and S Wold 1999Introduction to multi- and megavariate data analysis using projectionmethods (PCA amp PLS) Umetrics AB Sweden

15 Faber J M 1991 Microbial aspects of modi ed-atmosphere pack-aging technologymdasha review J Food Prot 5458ndash70

16 GarcDaggera-Gimeno R M C Sanz-MartDaggernez J M GarcDaggera-Martos andG Zurera-Cosano 2002 Modeling Botrytis cinerea spores growthin carbon dioxide enriched atmospheres Food Microbiol Saf 671904ndash1907

17 Gibson A M and A D Hocking 1997 Advances in the predictivemodelling of fungal growth in food Trends Food Sci Technol 8353ndash358

18 Guynot M E A J Ramos D Sala V Sanchis and S MarDaggern2002 Combined effects of weak acid preservatives pH and wateractivity on growth of Eurotium species on a sponge cake Int JFood Microbiol 7639ndash46

19 Haasum I and P V Nielsen 1998 Ecophysiological characteriza-tion of common food-borne fungi in relation to pH and water activityunder various atmospheric compositions J Appl Microbiol 84451ndash460

20 Haasum I and P V Nielsen 1998 Physiological characterizationof common fungi associated with cheese J Food Sci 63157ndash161

21 Leistner L 1992 Food preservation by combined methods FoodRes Int 25151ndash158

22 Magan N and J Lacey 1984 Effects of gas composition and wateractivity on growth of eld and storage fungi and their interactionsTrans Br Mycol Soc 82305ndash314

23 MarDaggern S M E Guynot P Neira M Bernado V Sanchis and AJ Ramos 2002 Risk assessment of the use of sub-optimal levels ofweak-acid preservatives in the control of mould growth on bakeryproducts Int J Food Microbiol 79203ndash211

24 MarDaggern S M E Guynot V Sanchis J Arbones and A J Ramos2002 Aspergillus avus Aspergillus niger and Penicillium corylo-philum spoilage prevention of bakery products by means of weak-acid preservatives J Food Sci 672271ndash2277

25 MarDaggern S V Sanchis A Teixido et al 1996 Water and tem-perature relations and microconidial germination of Fusariummoniliforme and F proliferatum from maize Can J Microbiol421045ndash1054

26 Nielsen P V and R Rios 2000 Inhibition of fungal growth onbread by volatile components from species and herbs and the pos-sible application in active packaging with special emphasis on mus-tard essential oil Int J Food Microbiol 60219ndash229

27 Ortola C and C Santacreu 1998 Principios de aplicacion del en-vasado en atmosfera modi cada a los productos de pani cacion ybollerDaggera Aliment Equip Tecnol 17111ndash117

28 Piergiovanni L and P Fava 1997 Minimizing the residual oxygenin modi ed atmosphere packaging of bakery products Food AdditContam 14765ndash773

29 Pitt R E 1993 A descriptive model of mold growth and a atoxinformation as affected by environmental conditions J Food Prot 56139ndash146

30 RodrDaggerguez Castilla M V and R Jordano 1997 Envasado de prod-uctos de panaderDaggera y bollerDaggera en atmosferas modi cadas composi-cion de la atmosfera de envasado lms equipos de envasado y efec-tos sobre los productos de panaderDaggera y bollerDaggera Alimentaria 28679ndash99

31 Sautour M P Dantigny C Divies and M Bensoussan 2001A temperature-type model for describing the relationship be-tween fungal growth and water activity Int J Food Microbiol6763ndash69

32 Smith J P J Hoshino and Y Abe 1995 Interactive packaginginvolving sachet technology p 143ndash176 In M L Rooney (ed)

J Food Prot Vol 66 No 101866 GUYNOT ET AL

FIGURE 1 Growth of E amstelodami (a) and E herbariorum (b)in MAP combined effects of CO2 aw and pH level after 28 daysof incubation at 258C

FIGURE 2 Growth of E repens (a) and E rubrum (b) in MAPcombined effects of CO2 aw and pH level after 28 days of in-cubation at 258C

of the range of levels that might occur in bakery products Colonydiameter was the response recorded A full factorial design wasused and all treatments were carried out three times separately

Preparation of the analog Sponge cake analogs were pre-pared as previously reported by Abellana et al (1) Each analogwas composed of 273 g of wheat our 211 g of vegetable oil258 g of eggs and 4 g of baking powder After baking the cakehad a pH of about 75 and its initial aw value was ca 075 Toattain the required pH (60) citric acid was added to the mix ofsolid ingredients before baking whereas aw was adjusted by plac-ing the cake analogs in petri plates containing water-glycerolagar(1058 690 and 391 g of glycerol in 100 ml of distilled waterplus 15 of agar to adjust the aw values of cake analogs to 080085 and 090 respectively) Previously two calibration curveswere constructed one to determine the amount of citric acid nec-essary to reach the desired pH (24) and another to determine theconcentration of glycerol in the agar needed to increase and main-tain the aw of the dough during the incubation period (data notshown) After baking the cakes were aseptically cut into squarepieces (5 by 5 cm) and placed in the sterile 9-cm petri dishescontaining water-glycerol agar at the desired aw and the cakeswere sealed with Para lm for 48 h for the equilibration of mois-ture between the agar and the analogs

Inoculation conditions Fungi were grown on DG18 (10 gof glucose 5 g of peptone 1 g of KH2PO4 05 g of MgSO4middot7H2O220 g of glycerol 15 g of agar 02 mg of dichloran and 100 mgof chloramphenicol in 1000 ml of distilled water) for 14 days at258C to obtain sporulating cultures Spores were harvested with aloop (3 mm) and transferred to sterile distilled water (0005Tween 80) The concentration of spores was determined with aThoma chamber and adjusted to 1 3 106 to 5 3 106 spores perml Cake analogs were needle-inoculated in the center Analogswith the same aw and pH levels were inoculated with the sevendifferent isolates and packed in the same bag (eight plates [sevenisolates and one blank] per bag) The uninoculated plate (blank)was included in each bag to check aw and pH levels at the end ofthe incubation period The plastic bags used were composed of apolyethylene and polyamide coextrusion mix lm (FontPackComercial del Paper Font SA Spain) with a total thickness of90 mm The transmission rates for the lm were as follows ox-ygen 19913 cm3 m22 24 h21 atm21 [1 atm 5 10129 kPa] carbondioxide 164903 cm3 m22 24 h21 atm21 water vapor 260 cm3

m22 24 h21 atm21 The bags were lled with air or the desiredmixture of N2 and CO2 (Carburos Metalicos SE de CarburosMetalicos SA Barcelona Spain) at a productgas ratio of 13(volvol) The packaging was carried out with an EGAR VACBasic-9 Digital compensated vacuum machine (Egarvac SCPTerrassa Barcelona Spain) by creating a vacuum in the bag andthen ushing the gas mixture at around 150 kPa prior to heatsealing The bags were then stored at 258C for 28 days

J Food Prot Vol 66 No 10 MAP FOR FUNGAL PREVENTION IN CAKES 1867

FIGURE 3 Growth of A niger (a) A a-vus (b) and P corylophilum (c) in MAPcombined effects of CO2 aw and pH levelafter 28 days of incubation at 258C

Growth measurements Colony diameters were measureddaily or as required with the aid of a binocular magni er orientedin two directions at right angles to each other At the end of theexperiment the gas composition in the headspace of each packagewas sampled and analyzed by gas chromatography A micro-gaschromatograph CP2002 (Chrompack International MidderlburgThe Netherlands) tted with a thermal conductivity detector wasused A packed column COP-Molsieve (5 Aƒ 4 m by 032 cmthickness of stationary phase [df] 5 10 mm) was kept at 90 kPafor the quanti cation of O2 and N2 (558C) The CO2 concentrationwas analyzed with a Pora-PLOT Q column (10 m by 032 mmdf 5 10 mm) at 758C and 140 kPa

Data analysis An analysis of covariance of colony radiusesmeasured during the storage period with time as a covariable wascarried out for each fungal species separately in order to determinesigni cant differences between the levels of the factors assayedand their interactions For this purpose the Statistical AnalysisSystem package (SAS Version 802 SAS Institute Inc CaryNC) was used The Gompertz model was used as the tting equa-tion (25) to estimate time before visible growth (lag phase) andthis parameter was analyzed by partial least squares projection to

latent structures (PLS) in order to establish suitable secondarypredictive models PLS is a regression extension of principal com-ponents analysis which is used when it is of interest to connectthe information in two blocks of variables to each other (14) Theanalysis gives the percentage of variance of the response ex-plained by the model (R2) and the predictive power of the modelaccording to cross validation (Q2) The latter analysis was per-formed with MODDE software (Version 40 Umetrics UmeaSweden)

RESULTS

Oxygen levels in packages Table 1 presents the meanconcentrations of O2 measured in inoculated bags at the endof the incubation period At aw levels of 085 and 090 allgas atmospheres contained 135 O2 and almost no sig-ni cant differences were observed among them Howeverat an aw level of 080 the level of O2 detected was nearly17 for air-packaged samples suggesting that the limitingfactor for fungal growth was water availability The ex-haustion of O2 in air-packaged samples at high aw levels

J Food Prot Vol 66 No 101868 GUYNOT ET AL

TABLE 3 Lag phases (days to visible growth estimated by the Gompertz equation) at pH 6 under modi ed atmosphere packaging(MAP) and at different aw levelsa

aw CO2

Lag phase (days) for organism

E amstelodami E herbariorum E repens E rubrum A niger A avus P corylophilum

080 0305070

100

26 6 2028282828

214 6 3628282828

187 6 3328282828

19 6 3027 6 15

282828

2828282828

2828282828

2828282828

085 0305070

100

15 6 50121 6 22172 6 35

2828

101 6 42143 6 21118 6 30

2828

81 6 20124 6 24172 6 20

2828

98 6 17179 6 26

18 6 202828

155 6 50121 6 38

27 6 402828

126 6 1928282828

28212 6 50

282828

090 0305070

100

66 6 4445 6 2138 6 3021 6 40

28

48 6 1654 6 0839 6 4593 6 23

28

64 6 0649 6 1364 6 1683 6 1126 6 40

44 6 0771 6 2155 6 3981 6 47

28

41 6 0660 6 37

130 6 55256 6 03

28

39 6 0547 6 1457 6 2096 6 44

28

2863 6 3520 6 20

2828

a Values presented are means 6 asymptotic standard errors Twenty-eight days was the maximal incubation period

(085 to 090) was probably the reason fungal growthstopped after some days of incubation

Impact of CO2 headspace concentration on colonyradius The analysis of covariance revealed that MAP (airand 0 to 100 of CO2 balanced with N2) aw (at levels of080 085 and 090) and their interaction had a statisticallysigni cant effect (P 001) on fungal growth (Table 2)Despite the signi cant effect of pH or the interaction of pHand atmosphere on the growth of some species (E amste-lodami E rubrum A avus and P corylophilum) no clearlink was found between CO2 antifungal activity and pHlevel In general all species were affected in the same waythey grew faster as the aw level increased and the level ofCO2 in the headspace decreased Some slight differencesamong fungal speciesrsquo levels of sensitivity to gas compo-sition were found with P corylophilum being the speciesthat was most affected by the limiting factors applied fol-lowed by Aspergillus spp and Eurotium spp Under theconditions that were most favorable for fungal growth (pre-sented by cakes with high aw levels packaged in air) themaximum colony radii were ca 25 mm for Eurotium sppand 5 mm for Aspergillus spp and P corylophilum

Water activity had a signi cant in uence on fungalgrowth and determined the level of CO2 needed to preventcake analog spoilage (Figs 1 through 3) It should be notedthat the maximal colony radii for cake analogs packagedwith 30 to 70 CO2 were ca 5 mm for Eurotium spp and2 mm for Aspergillus spp and P corylophilum In generalfungal growth was prevented at all aw levels in packageswith high CO2 levels (100) However E repens and Eherbariorum were able to grow under this condition (withthe maximum colony radius observed being 25 mm) at anaw level of 090 (Figs 1 and 2)

Impact of CO2 headspace concentration on lagphases In general longer lag phases were estimated forthe conditions that were the most critical for fungal growth

(Table 3) A linear correlation between the inverse of thelag phase and the growth rate was found for all isolateswith R2 values ranging from 084 to 096 depending on theisolate Longer lag phases corresponded to lower growthrates and smaller maximal colony radii It is important tomake use of these kinds of experimental data by construct-ing models with which given a combination of factors anda storage period one can predict the approximate degree ofspoilage Although approximate these models can helpfood technologists to better predict shelf life and safety

Lag phases for the seven isolates were modeled si-multaneously with the use of all factors (pH aw and CO2

concentration) and their quadratic terms and cross-term in-teractions However since pH was nonsigni cant (data notshown) a new regression model including only signi cantterms (aw and CO2 concentration their interaction and qua-dratic terms) was calculated This new PLS regression ledto two principal components models (PC1 and PC2) (Fig4) PC1 explained most of the variance in the lag phasedate (R2 5 759) PC2 is represented by a line orthogonalto PC1 and improves the approximation of the data set asmuch as possible carrying the maximum residual infor-mation (not taken into account in PC1) (18 variance ex-plained by PC2) The predictive power of the model (Q2)was near 720

The model explained the variation for almost all spe-cies quite well with fungal responses located close to oneanother in the loadings plot indicating that all species re-sponded to changes in tested factors in similar ways (Fig4) The greater the distance of a factor or response fromthe origin the stronger its effect on the model Thus re-sponses were well explained and aw CO2 concentrationand their interaction had a strong in uence of on lag phaseswhile quadratic terms had little effect on PC1 which meantthat they had little effect on lag phases If a factor is locatedopposite a fungus a negative correlation between the levelsof the factor and lag phase of that fungus exists

J Food Prot Vol 66 No 10 MAP FOR FUNGAL PREVENTION IN CAKES 1869

FIGURE 4 Plot of principal components models 1 and 2 (obtained by PLS regression) showing the relationships between aw CO2

level and their cross and quadratic terms and lag phase before growth (days)

TABLE 4 Polynomial model equation generated by PLS of lag phase (time to visible growth) against aw CO2 and quadratic andcross termsa

Isolate Q2 R2 Equation

E amstelodamiE herbariorumE repensE rubrumA nigerA avusP corylophilum

815649815809712823447

865678857843763848588

lag phase 5 239158 1 120093aw 2 156CO2 2 85129aw2 1 (496 3 1024)(CO2)2 1 196awCO2

lag phase 5 263002 1 176363aw 2 175CO2 2 118192aw2 1 (828 3 1024)(CO2)2 1 211awCO2

lag phase 5 211152 1 51190aw 2 121CO2 2 43105aw2 1 (111 3 1024)(CO2)2 1 160awCO2

lag phase 5 26409 1 38873aw 2 111CO2 2 35036aw2 1 (503 3 1025)(CO2)2 1 148awCO2

lag phase 5 257149 1 161086aw 2 157CO2 2 107631aw2 1 (761 3 1024)(CO2)2 1 189awCO2

lag phase 5 274534 1 206957aw 2 198CO2 2 137798aw2 1 (985 3 1024)(CO2)2 1 237awCO2

lag phase 5 249724 1 134113aw 2 096CO2 2 85260aw2 1 (698 3 1024)(CO2)2 1 108awCO2

a Q2 percentage of variation predicted by the model R2 percentage of variation explained by the model

Lag phases were described by a polynomial modelequation with six coef cients (b0 b1 b22) (Table 4)lag phase 5 b0 1 b1aw 1 b2CO2 1 b12awCO2 1 b11aw

2

1 b22CO22 It should be noted that these coef cients are

unscaled and uncentered in order to allow the direct cal-culation of lag phases so they cannot be used to assess thesigni cance of the effects of the factors on lag phases

Most of the models are highly signi cant with R2 val-ues of 70 The response of P corylophilum was slightlydifferent and the model explained only 588 of the lagphase variance for this organism (Table 4) This result canalso be observed in the loadings plot (Fig 4) in which Pcorylophilum appears quite separate from the other fungalspecies

The models were subsequently used to generate two-dimensional contour plots Plots of the combined effects ofaw and CO2 are shown in Figure 5 Since similar plots wereobtained for all Eurotium species and for all Aspergillusspecies one species of each genus is represented Lag phas-es for Eurotium spp at an aw level of 087 increased two-fold when the level of CO2 in the headspace increased from0 to 70 (eg E amstelodami growth started after 20days) Storage in 100 CO2 irrespective of the aw level

resulted in lag times of 20 days When Figure 5 is con-sidered it should be taken into account that the absence ofgrowth is equivalent to a lag phase of 28 days the durationof the experiment Thus when a shelf life of ca 28 daysis predicted from the models it is probable that no spoilagewill occur

DISCUSSION

In this study the use of MAP combined with differentpH and aw values to prevent the spoilage of bakery productsby Eurotium Aspergillus and Penicillium species was eval-uated Bakery products with near-neutral pHs were used forthe experiment The results obtained corroborate those fromother studies which show that CO2 has an important fun-gistatic effect (2 15 26) Spoilage was prevented with 70CO2 in the headspace (with a residual O2 level of 135)for products with an aw level of 080 while spoilage wassigni cantly delayed for products with aw levels of 085 to090 These results con rm those of previous studies inwhich reductions in the aw level were shown to increasethe effect of high levels of CO2 (10 19 20 22) It wasfound that with 30 to 50 CO2 in the headspace Eurotiumspp growth was totally prevented at an aw level of 080

J Food Prot Vol 66 No 101870 GUYNOT ET AL

FIGURE 5 Contour plots based on models obtained by PLS re-gression showing lag phases (days) as affected by CO2 level andaw for (a) E amstelodami (b) A niger and (c) P corylophilum

as water availability increased growth was only delayedOnly a 100 CO2 atmosphere prevented spoilage by almostall isolates regardless of the aw level

The inhibitory effect of CO2 on microorganisms in aculture medium or in a food system depends on many fac-tors These factors include partial CO2 pressure O2 con-centration headspace gas volume temperature acidity andaw (15) The effect of temperature was not assayed in thisstudy and the temperature level chosen was the one thatwas most likely to be encountered during distribution andstorage in the retail market The preserving effect of CO2

strongly depended on aw while interaction with pH was

not important probably because of the similar pH valuesinvolved Similarly Ellis et al (11 12) demonstrated thatthe growth of A avus under different atmospheric CO2

and O2 concentrations was highly dependent on aw andtemperature However these authors did not nd importantdifferences between the effects of gas composition on fun-gal growth at pH 6 and those at pH 8 Haasum and Nielsen(20) in a study of the growth of fungi in a cheese envi-ronment also found no signi cant differences in the inhib-itory effects of CO2 at different pH levels (4ndash8) The ef-fectiveness of CO2 also varies with the type of organismunder consideration with molds being more sensitive thanyeasts (19 22 30)

In this study all O2 concentrations measured at the endof the incubation period except those for air-packaged bagswere between 02 and 135 Since molds by de nition arestrict aerobes suf cient residual oxygen must be present inthe package headspace to allow mold growth (10 34) Ithas been demonstrated that molds can tolerate and evengrow in air with headspace oxygen concentrations as lowas 1 to 2 (13 32) Moreover several studies have shownthat molds can grow in the presence of elevated CO2 levelsif O2 is present (11 12) Smith et al (34) reported that aminimum of 04 O2 was needed for the growth of A nigerand Penicillium spp in a CO2N2 (6040) atmosphereAbellana et al (2) also found that low levels (002 to 05)of O2 did not modify fungal growth when levels of CO2 inthe bags were high However these authors found that asthe CO2 concentration decreased O2 levels had more in u-ence on growth kinetics (small differences in O2 levelscould make fungi grow) Similarly Nielsen and Rios (26)found that the growth of Aspergillus and Penicillium spe-cies was delayed only at reduced O2 levels (002 to 003)in either pure nitrogen or 50 N2 and 50 CO2 Howeverin pure CO2 the molds were strongly inhibited despite theresidual O2 in the package The total elimination of air dur-ing MAP represents a serious hurdle owing to both the lmrsquos permeability and the productrsquos capacity to trap O2

molecules (28) When a vacuum evacuation of the packageis carried out prior to gas injection the level of residual O2

in the headspace can be 1 (30) However the smoothand fragile texture of sponge cakes must be taken into ac-count the cakes could be deformed (27) An innovativemethod for controlling oxygen without physically deform-ing products involves the incorporation of O2 absorbers intothe package (32)

CO2 is both water and lipid soluble and when gaseousCO2 is applied to a product part of the CO2 goes to theliquid phase (water or oil) as carbonic acid (about 2) (7)This development can produce a small decrease in pH re-sulting in changes in the organoleptic characteristics of theproduct (15 30) The solubility of this gas in a culturemedium or in a food system increases as pH increases asaw increases as the volume of headspace gas increases oras temperature decreases The pHs of cake analogs pack-aged aerobically and with the different gas compositionstested were measured and no appreciable differences werefound (data not shown) This result may suggest that thesolubility of CO2 in the cake analog used was not signi -

J Food Prot Vol 66 No 10 MAP FOR FUNGAL PREVENTION IN CAKES 1871

cant or was not suf cient to produce a decrease in the an-alog pH

In the last few years predictive microbiology has be-come an important area of research The advantage of theresponse surface modeling approach is that it allows thesimultaneous examination of several variables Also math-ematical models can be generated to predict the packagingand environmental storage conditions that would controlfungal growth (11) Abellana et al (2) predicted lag phasesradial growth rates and maximum growth in relation to awO2 and CO2 concentrations for three Eurotium species ona sponge cake analog A model such as that described heremay be an important tool to aid food technologists in pre-dicting shelf lives of products or may give an idea of theeffects of certain changes in a productrsquos formulation Alsothis study emphasizes the importance of combining severalbarriers such as MAP aw and pH whose effects on moldgrowth are synergistic or additive

ACKNOWLEDGMENTS

The authors acknowledge the EC Quality of Life Programme (QoL)Key Action 1 (KA1) on Food Nutrition and Health (PL98-4075) theSpanish government (CICYT ALI99-0831) and the Catalonian govern-ment (CIRIT Comissio Interdepartamental de Recerca i Innovacio Tec-nologica)

REFERENCES

1 Abellana M J BenedDagger V Sanchis and A J Ramos 1999 Wateractivity and temperature effects on germination and growth of Eu-rotium amstelodami E chevalieri and E herbariorum isolates frombakery products J Appl Microbiol 87371ndash380

2 Abellana M V Sanchis A J Ramos and P V Nielsen 2000Effect of modi ed atmosphere packaging and water activity ongrowth of Eurotium amstelodami E chevalieri and E herbariorumon a sponge cake analogue J Appl Microbiol 88606ndash616

3 Abellana M L Torres V Sanchis and A J Ramos 1997 Car-acterizacion de diferentes productos de bollerDaggera industrial II Estudiode la mico ora Alimentaria 28751ndash56

4 Baranyi J A M Gibson J I Pitt M J Eyles and T A Roberts1996 Predictive models as means of measuring the relatedness ofsome Aspergillus species Food Microbiol 14347ndash351

5 Chirife J and G J Favetto 1992 Some physico-chemical basisof food preservation by combined methods Food Res Int 25389ndash396

6 Daifas D P J P Smith B Blanch eld and J W Austin 1999Growth and toxin production by Clostridium botulinum in English-style crumpets packaged under modi ed atmosphere J Food Prot62349ndash355

7 Daniels J A R Krishnamurthi and S S H Rizvi 1985 A reviewof effects of carbon dioxide on microbial growth and food qualityJ Food Prot 48532ndash537

8 Davies A R 1995 Advances in modi ed-atmosphere packaging p304ndash320 In G W Gould (ed) New methods of food preservationBlackie Academic and Professional London

9 Dixon N M and D B Kell 1989 The inhibition by CO2 of thegrowth and metabolism of micro-organisms J Appl Bacteriol 67109ndash136

10 El Halouat A and J M Debevere 1997 Effect of water activitymodi ed atmosphere packing and storage temperature on spore ger-mination of moulds isolated from prunes Int J Food Microbiol 3541ndash48

11 Ellis W O J P Smith B K Simpson S Khanizadeh and J HOldhem 1993 Control of growth and a atoxin production of As-pergillus avus under modi ed atmosphere packaging (MAP) con-ditions Food Microbiol 109ndash21

12 Ellis W O J P Smith B K Simpson and H Ramaswamy 1993

Effect of inoculum level on a atoxin production by Aspergillus a-vus under modi ed atmosphere packaging (MAP) conditions FoodMicrobiol 10525ndash535

13 Ellis W O J P Smith B K Simpson H Ramaswamy and GDoyon 1994 Growth of and a atoxin production by Aspergillus avus in peanuts stored under modi ed atmosphere packaging(MAP) conditions Int J Food Microbiol 22173ndash187

14 Eriksson L E Johansson N Kettaneh-Wold and S Wold 1999Introduction to multi- and megavariate data analysis using projectionmethods (PCA amp PLS) Umetrics AB Sweden

15 Faber J M 1991 Microbial aspects of modi ed-atmosphere pack-aging technologymdasha review J Food Prot 5458ndash70

16 GarcDaggera-Gimeno R M C Sanz-MartDaggernez J M GarcDaggera-Martos andG Zurera-Cosano 2002 Modeling Botrytis cinerea spores growthin carbon dioxide enriched atmospheres Food Microbiol Saf 671904ndash1907

17 Gibson A M and A D Hocking 1997 Advances in the predictivemodelling of fungal growth in food Trends Food Sci Technol 8353ndash358

18 Guynot M E A J Ramos D Sala V Sanchis and S MarDaggern2002 Combined effects of weak acid preservatives pH and wateractivity on growth of Eurotium species on a sponge cake Int JFood Microbiol 7639ndash46

19 Haasum I and P V Nielsen 1998 Ecophysiological characteriza-tion of common food-borne fungi in relation to pH and water activityunder various atmospheric compositions J Appl Microbiol 84451ndash460

20 Haasum I and P V Nielsen 1998 Physiological characterizationof common fungi associated with cheese J Food Sci 63157ndash161

21 Leistner L 1992 Food preservation by combined methods FoodRes Int 25151ndash158

22 Magan N and J Lacey 1984 Effects of gas composition and wateractivity on growth of eld and storage fungi and their interactionsTrans Br Mycol Soc 82305ndash314

23 MarDaggern S M E Guynot P Neira M Bernado V Sanchis and AJ Ramos 2002 Risk assessment of the use of sub-optimal levels ofweak-acid preservatives in the control of mould growth on bakeryproducts Int J Food Microbiol 79203ndash211

24 MarDaggern S M E Guynot V Sanchis J Arbones and A J Ramos2002 Aspergillus avus Aspergillus niger and Penicillium corylo-philum spoilage prevention of bakery products by means of weak-acid preservatives J Food Sci 672271ndash2277

25 MarDaggern S V Sanchis A Teixido et al 1996 Water and tem-perature relations and microconidial germination of Fusariummoniliforme and F proliferatum from maize Can J Microbiol421045ndash1054

26 Nielsen P V and R Rios 2000 Inhibition of fungal growth onbread by volatile components from species and herbs and the pos-sible application in active packaging with special emphasis on mus-tard essential oil Int J Food Microbiol 60219ndash229

27 Ortola C and C Santacreu 1998 Principios de aplicacion del en-vasado en atmosfera modi cada a los productos de pani cacion ybollerDaggera Aliment Equip Tecnol 17111ndash117

28 Piergiovanni L and P Fava 1997 Minimizing the residual oxygenin modi ed atmosphere packaging of bakery products Food AdditContam 14765ndash773

29 Pitt R E 1993 A descriptive model of mold growth and a atoxinformation as affected by environmental conditions J Food Prot 56139ndash146

30 RodrDaggerguez Castilla M V and R Jordano 1997 Envasado de prod-uctos de panaderDaggera y bollerDaggera en atmosferas modi cadas composi-cion de la atmosfera de envasado lms equipos de envasado y efec-tos sobre los productos de panaderDaggera y bollerDaggera Alimentaria 28679ndash99

31 Sautour M P Dantigny C Divies and M Bensoussan 2001A temperature-type model for describing the relationship be-tween fungal growth and water activity Int J Food Microbiol6763ndash69

32 Smith J P J Hoshino and Y Abe 1995 Interactive packaginginvolving sachet technology p 143ndash176 In M L Rooney (ed)

J Food Prot Vol 66 No 10 MAP FOR FUNGAL PREVENTION IN CAKES 1867

FIGURE 3 Growth of A niger (a) A a-vus (b) and P corylophilum (c) in MAPcombined effects of CO2 aw and pH levelafter 28 days of incubation at 258C

Growth measurements Colony diameters were measureddaily or as required with the aid of a binocular magni er orientedin two directions at right angles to each other At the end of theexperiment the gas composition in the headspace of each packagewas sampled and analyzed by gas chromatography A micro-gaschromatograph CP2002 (Chrompack International MidderlburgThe Netherlands) tted with a thermal conductivity detector wasused A packed column COP-Molsieve (5 Aƒ 4 m by 032 cmthickness of stationary phase [df] 5 10 mm) was kept at 90 kPafor the quanti cation of O2 and N2 (558C) The CO2 concentrationwas analyzed with a Pora-PLOT Q column (10 m by 032 mmdf 5 10 mm) at 758C and 140 kPa

Data analysis An analysis of covariance of colony radiusesmeasured during the storage period with time as a covariable wascarried out for each fungal species separately in order to determinesigni cant differences between the levels of the factors assayedand their interactions For this purpose the Statistical AnalysisSystem package (SAS Version 802 SAS Institute Inc CaryNC) was used The Gompertz model was used as the tting equa-tion (25) to estimate time before visible growth (lag phase) andthis parameter was analyzed by partial least squares projection to

latent structures (PLS) in order to establish suitable secondarypredictive models PLS is a regression extension of principal com-ponents analysis which is used when it is of interest to connectthe information in two blocks of variables to each other (14) Theanalysis gives the percentage of variance of the response ex-plained by the model (R2) and the predictive power of the modelaccording to cross validation (Q2) The latter analysis was per-formed with MODDE software (Version 40 Umetrics UmeaSweden)

RESULTS

Oxygen levels in packages Table 1 presents the meanconcentrations of O2 measured in inoculated bags at the endof the incubation period At aw levels of 085 and 090 allgas atmospheres contained 135 O2 and almost no sig-ni cant differences were observed among them Howeverat an aw level of 080 the level of O2 detected was nearly17 for air-packaged samples suggesting that the limitingfactor for fungal growth was water availability The ex-haustion of O2 in air-packaged samples at high aw levels

J Food Prot Vol 66 No 101868 GUYNOT ET AL

TABLE 3 Lag phases (days to visible growth estimated by the Gompertz equation) at pH 6 under modi ed atmosphere packaging(MAP) and at different aw levelsa

aw CO2

Lag phase (days) for organism

E amstelodami E herbariorum E repens E rubrum A niger A avus P corylophilum

080 0305070

100

26 6 2028282828

214 6 3628282828

187 6 3328282828

19 6 3027 6 15

282828

2828282828

2828282828

2828282828

085 0305070

100

15 6 50121 6 22172 6 35

2828

101 6 42143 6 21118 6 30

2828

81 6 20124 6 24172 6 20

2828

98 6 17179 6 26

18 6 202828

155 6 50121 6 38

27 6 402828

126 6 1928282828

28212 6 50

282828

090 0305070

100

66 6 4445 6 2138 6 3021 6 40

28

48 6 1654 6 0839 6 4593 6 23

28

64 6 0649 6 1364 6 1683 6 1126 6 40

44 6 0771 6 2155 6 3981 6 47

28

41 6 0660 6 37

130 6 55256 6 03

28

39 6 0547 6 1457 6 2096 6 44

28

2863 6 3520 6 20

2828

a Values presented are means 6 asymptotic standard errors Twenty-eight days was the maximal incubation period

(085 to 090) was probably the reason fungal growthstopped after some days of incubation

Impact of CO2 headspace concentration on colonyradius The analysis of covariance revealed that MAP (airand 0 to 100 of CO2 balanced with N2) aw (at levels of080 085 and 090) and their interaction had a statisticallysigni cant effect (P 001) on fungal growth (Table 2)Despite the signi cant effect of pH or the interaction of pHand atmosphere on the growth of some species (E amste-lodami E rubrum A avus and P corylophilum) no clearlink was found between CO2 antifungal activity and pHlevel In general all species were affected in the same waythey grew faster as the aw level increased and the level ofCO2 in the headspace decreased Some slight differencesamong fungal speciesrsquo levels of sensitivity to gas compo-sition were found with P corylophilum being the speciesthat was most affected by the limiting factors applied fol-lowed by Aspergillus spp and Eurotium spp Under theconditions that were most favorable for fungal growth (pre-sented by cakes with high aw levels packaged in air) themaximum colony radii were ca 25 mm for Eurotium sppand 5 mm for Aspergillus spp and P corylophilum

Water activity had a signi cant in uence on fungalgrowth and determined the level of CO2 needed to preventcake analog spoilage (Figs 1 through 3) It should be notedthat the maximal colony radii for cake analogs packagedwith 30 to 70 CO2 were ca 5 mm for Eurotium spp and2 mm for Aspergillus spp and P corylophilum In generalfungal growth was prevented at all aw levels in packageswith high CO2 levels (100) However E repens and Eherbariorum were able to grow under this condition (withthe maximum colony radius observed being 25 mm) at anaw level of 090 (Figs 1 and 2)

Impact of CO2 headspace concentration on lagphases In general longer lag phases were estimated forthe conditions that were the most critical for fungal growth

(Table 3) A linear correlation between the inverse of thelag phase and the growth rate was found for all isolateswith R2 values ranging from 084 to 096 depending on theisolate Longer lag phases corresponded to lower growthrates and smaller maximal colony radii It is important tomake use of these kinds of experimental data by construct-ing models with which given a combination of factors anda storage period one can predict the approximate degree ofspoilage Although approximate these models can helpfood technologists to better predict shelf life and safety

Lag phases for the seven isolates were modeled si-multaneously with the use of all factors (pH aw and CO2

concentration) and their quadratic terms and cross-term in-teractions However since pH was nonsigni cant (data notshown) a new regression model including only signi cantterms (aw and CO2 concentration their interaction and qua-dratic terms) was calculated This new PLS regression ledto two principal components models (PC1 and PC2) (Fig4) PC1 explained most of the variance in the lag phasedate (R2 5 759) PC2 is represented by a line orthogonalto PC1 and improves the approximation of the data set asmuch as possible carrying the maximum residual infor-mation (not taken into account in PC1) (18 variance ex-plained by PC2) The predictive power of the model (Q2)was near 720

The model explained the variation for almost all spe-cies quite well with fungal responses located close to oneanother in the loadings plot indicating that all species re-sponded to changes in tested factors in similar ways (Fig4) The greater the distance of a factor or response fromthe origin the stronger its effect on the model Thus re-sponses were well explained and aw CO2 concentrationand their interaction had a strong in uence of on lag phaseswhile quadratic terms had little effect on PC1 which meantthat they had little effect on lag phases If a factor is locatedopposite a fungus a negative correlation between the levelsof the factor and lag phase of that fungus exists

J Food Prot Vol 66 No 10 MAP FOR FUNGAL PREVENTION IN CAKES 1869

FIGURE 4 Plot of principal components models 1 and 2 (obtained by PLS regression) showing the relationships between aw CO2

level and their cross and quadratic terms and lag phase before growth (days)

TABLE 4 Polynomial model equation generated by PLS of lag phase (time to visible growth) against aw CO2 and quadratic andcross termsa

Isolate Q2 R2 Equation

E amstelodamiE herbariorumE repensE rubrumA nigerA avusP corylophilum

815649815809712823447

865678857843763848588

lag phase 5 239158 1 120093aw 2 156CO2 2 85129aw2 1 (496 3 1024)(CO2)2 1 196awCO2

lag phase 5 263002 1 176363aw 2 175CO2 2 118192aw2 1 (828 3 1024)(CO2)2 1 211awCO2

lag phase 5 211152 1 51190aw 2 121CO2 2 43105aw2 1 (111 3 1024)(CO2)2 1 160awCO2

lag phase 5 26409 1 38873aw 2 111CO2 2 35036aw2 1 (503 3 1025)(CO2)2 1 148awCO2

lag phase 5 257149 1 161086aw 2 157CO2 2 107631aw2 1 (761 3 1024)(CO2)2 1 189awCO2

lag phase 5 274534 1 206957aw 2 198CO2 2 137798aw2 1 (985 3 1024)(CO2)2 1 237awCO2

lag phase 5 249724 1 134113aw 2 096CO2 2 85260aw2 1 (698 3 1024)(CO2)2 1 108awCO2

a Q2 percentage of variation predicted by the model R2 percentage of variation explained by the model

Lag phases were described by a polynomial modelequation with six coef cients (b0 b1 b22) (Table 4)lag phase 5 b0 1 b1aw 1 b2CO2 1 b12awCO2 1 b11aw

2

1 b22CO22 It should be noted that these coef cients are

unscaled and uncentered in order to allow the direct cal-culation of lag phases so they cannot be used to assess thesigni cance of the effects of the factors on lag phases

Most of the models are highly signi cant with R2 val-ues of 70 The response of P corylophilum was slightlydifferent and the model explained only 588 of the lagphase variance for this organism (Table 4) This result canalso be observed in the loadings plot (Fig 4) in which Pcorylophilum appears quite separate from the other fungalspecies

The models were subsequently used to generate two-dimensional contour plots Plots of the combined effects ofaw and CO2 are shown in Figure 5 Since similar plots wereobtained for all Eurotium species and for all Aspergillusspecies one species of each genus is represented Lag phas-es for Eurotium spp at an aw level of 087 increased two-fold when the level of CO2 in the headspace increased from0 to 70 (eg E amstelodami growth started after 20days) Storage in 100 CO2 irrespective of the aw level

resulted in lag times of 20 days When Figure 5 is con-sidered it should be taken into account that the absence ofgrowth is equivalent to a lag phase of 28 days the durationof the experiment Thus when a shelf life of ca 28 daysis predicted from the models it is probable that no spoilagewill occur

DISCUSSION

In this study the use of MAP combined with differentpH and aw values to prevent the spoilage of bakery productsby Eurotium Aspergillus and Penicillium species was eval-uated Bakery products with near-neutral pHs were used forthe experiment The results obtained corroborate those fromother studies which show that CO2 has an important fun-gistatic effect (2 15 26) Spoilage was prevented with 70CO2 in the headspace (with a residual O2 level of 135)for products with an aw level of 080 while spoilage wassigni cantly delayed for products with aw levels of 085 to090 These results con rm those of previous studies inwhich reductions in the aw level were shown to increasethe effect of high levels of CO2 (10 19 20 22) It wasfound that with 30 to 50 CO2 in the headspace Eurotiumspp growth was totally prevented at an aw level of 080

J Food Prot Vol 66 No 101870 GUYNOT ET AL

FIGURE 5 Contour plots based on models obtained by PLS re-gression showing lag phases (days) as affected by CO2 level andaw for (a) E amstelodami (b) A niger and (c) P corylophilum

as water availability increased growth was only delayedOnly a 100 CO2 atmosphere prevented spoilage by almostall isolates regardless of the aw level

The inhibitory effect of CO2 on microorganisms in aculture medium or in a food system depends on many fac-tors These factors include partial CO2 pressure O2 con-centration headspace gas volume temperature acidity andaw (15) The effect of temperature was not assayed in thisstudy and the temperature level chosen was the one thatwas most likely to be encountered during distribution andstorage in the retail market The preserving effect of CO2

strongly depended on aw while interaction with pH was

not important probably because of the similar pH valuesinvolved Similarly Ellis et al (11 12) demonstrated thatthe growth of A avus under different atmospheric CO2

and O2 concentrations was highly dependent on aw andtemperature However these authors did not nd importantdifferences between the effects of gas composition on fun-gal growth at pH 6 and those at pH 8 Haasum and Nielsen(20) in a study of the growth of fungi in a cheese envi-ronment also found no signi cant differences in the inhib-itory effects of CO2 at different pH levels (4ndash8) The ef-fectiveness of CO2 also varies with the type of organismunder consideration with molds being more sensitive thanyeasts (19 22 30)

In this study all O2 concentrations measured at the endof the incubation period except those for air-packaged bagswere between 02 and 135 Since molds by de nition arestrict aerobes suf cient residual oxygen must be present inthe package headspace to allow mold growth (10 34) Ithas been demonstrated that molds can tolerate and evengrow in air with headspace oxygen concentrations as lowas 1 to 2 (13 32) Moreover several studies have shownthat molds can grow in the presence of elevated CO2 levelsif O2 is present (11 12) Smith et al (34) reported that aminimum of 04 O2 was needed for the growth of A nigerand Penicillium spp in a CO2N2 (6040) atmosphereAbellana et al (2) also found that low levels (002 to 05)of O2 did not modify fungal growth when levels of CO2 inthe bags were high However these authors found that asthe CO2 concentration decreased O2 levels had more in u-ence on growth kinetics (small differences in O2 levelscould make fungi grow) Similarly Nielsen and Rios (26)found that the growth of Aspergillus and Penicillium spe-cies was delayed only at reduced O2 levels (002 to 003)in either pure nitrogen or 50 N2 and 50 CO2 Howeverin pure CO2 the molds were strongly inhibited despite theresidual O2 in the package The total elimination of air dur-ing MAP represents a serious hurdle owing to both the lmrsquos permeability and the productrsquos capacity to trap O2

molecules (28) When a vacuum evacuation of the packageis carried out prior to gas injection the level of residual O2

in the headspace can be 1 (30) However the smoothand fragile texture of sponge cakes must be taken into ac-count the cakes could be deformed (27) An innovativemethod for controlling oxygen without physically deform-ing products involves the incorporation of O2 absorbers intothe package (32)

CO2 is both water and lipid soluble and when gaseousCO2 is applied to a product part of the CO2 goes to theliquid phase (water or oil) as carbonic acid (about 2) (7)This development can produce a small decrease in pH re-sulting in changes in the organoleptic characteristics of theproduct (15 30) The solubility of this gas in a culturemedium or in a food system increases as pH increases asaw increases as the volume of headspace gas increases oras temperature decreases The pHs of cake analogs pack-aged aerobically and with the different gas compositionstested were measured and no appreciable differences werefound (data not shown) This result may suggest that thesolubility of CO2 in the cake analog used was not signi -

J Food Prot Vol 66 No 10 MAP FOR FUNGAL PREVENTION IN CAKES 1871

cant or was not suf cient to produce a decrease in the an-alog pH

In the last few years predictive microbiology has be-come an important area of research The advantage of theresponse surface modeling approach is that it allows thesimultaneous examination of several variables Also math-ematical models can be generated to predict the packagingand environmental storage conditions that would controlfungal growth (11) Abellana et al (2) predicted lag phasesradial growth rates and maximum growth in relation to awO2 and CO2 concentrations for three Eurotium species ona sponge cake analog A model such as that described heremay be an important tool to aid food technologists in pre-dicting shelf lives of products or may give an idea of theeffects of certain changes in a productrsquos formulation Alsothis study emphasizes the importance of combining severalbarriers such as MAP aw and pH whose effects on moldgrowth are synergistic or additive

ACKNOWLEDGMENTS

The authors acknowledge the EC Quality of Life Programme (QoL)Key Action 1 (KA1) on Food Nutrition and Health (PL98-4075) theSpanish government (CICYT ALI99-0831) and the Catalonian govern-ment (CIRIT Comissio Interdepartamental de Recerca i Innovacio Tec-nologica)

REFERENCES

1 Abellana M J BenedDagger V Sanchis and A J Ramos 1999 Wateractivity and temperature effects on germination and growth of Eu-rotium amstelodami E chevalieri and E herbariorum isolates frombakery products J Appl Microbiol 87371ndash380

2 Abellana M V Sanchis A J Ramos and P V Nielsen 2000Effect of modi ed atmosphere packaging and water activity ongrowth of Eurotium amstelodami E chevalieri and E herbariorumon a sponge cake analogue J Appl Microbiol 88606ndash616

3 Abellana M L Torres V Sanchis and A J Ramos 1997 Car-acterizacion de diferentes productos de bollerDaggera industrial II Estudiode la mico ora Alimentaria 28751ndash56

4 Baranyi J A M Gibson J I Pitt M J Eyles and T A Roberts1996 Predictive models as means of measuring the relatedness ofsome Aspergillus species Food Microbiol 14347ndash351

5 Chirife J and G J Favetto 1992 Some physico-chemical basisof food preservation by combined methods Food Res Int 25389ndash396

6 Daifas D P J P Smith B Blanch eld and J W Austin 1999Growth and toxin production by Clostridium botulinum in English-style crumpets packaged under modi ed atmosphere J Food Prot62349ndash355

7 Daniels J A R Krishnamurthi and S S H Rizvi 1985 A reviewof effects of carbon dioxide on microbial growth and food qualityJ Food Prot 48532ndash537

8 Davies A R 1995 Advances in modi ed-atmosphere packaging p304ndash320 In G W Gould (ed) New methods of food preservationBlackie Academic and Professional London

9 Dixon N M and D B Kell 1989 The inhibition by CO2 of thegrowth and metabolism of micro-organisms J Appl Bacteriol 67109ndash136

10 El Halouat A and J M Debevere 1997 Effect of water activitymodi ed atmosphere packing and storage temperature on spore ger-mination of moulds isolated from prunes Int J Food Microbiol 3541ndash48

11 Ellis W O J P Smith B K Simpson S Khanizadeh and J HOldhem 1993 Control of growth and a atoxin production of As-pergillus avus under modi ed atmosphere packaging (MAP) con-ditions Food Microbiol 109ndash21

12 Ellis W O J P Smith B K Simpson and H Ramaswamy 1993

Effect of inoculum level on a atoxin production by Aspergillus a-vus under modi ed atmosphere packaging (MAP) conditions FoodMicrobiol 10525ndash535

13 Ellis W O J P Smith B K Simpson H Ramaswamy and GDoyon 1994 Growth of and a atoxin production by Aspergillus avus in peanuts stored under modi ed atmosphere packaging(MAP) conditions Int J Food Microbiol 22173ndash187

14 Eriksson L E Johansson N Kettaneh-Wold and S Wold 1999Introduction to multi- and megavariate data analysis using projectionmethods (PCA amp PLS) Umetrics AB Sweden

15 Faber J M 1991 Microbial aspects of modi ed-atmosphere pack-aging technologymdasha review J Food Prot 5458ndash70

16 GarcDaggera-Gimeno R M C Sanz-MartDaggernez J M GarcDaggera-Martos andG Zurera-Cosano 2002 Modeling Botrytis cinerea spores growthin carbon dioxide enriched atmospheres Food Microbiol Saf 671904ndash1907

17 Gibson A M and A D Hocking 1997 Advances in the predictivemodelling of fungal growth in food Trends Food Sci Technol 8353ndash358

18 Guynot M E A J Ramos D Sala V Sanchis and S MarDaggern2002 Combined effects of weak acid preservatives pH and wateractivity on growth of Eurotium species on a sponge cake Int JFood Microbiol 7639ndash46

19 Haasum I and P V Nielsen 1998 Ecophysiological characteriza-tion of common food-borne fungi in relation to pH and water activityunder various atmospheric compositions J Appl Microbiol 84451ndash460

20 Haasum I and P V Nielsen 1998 Physiological characterizationof common fungi associated with cheese J Food Sci 63157ndash161

21 Leistner L 1992 Food preservation by combined methods FoodRes Int 25151ndash158

22 Magan N and J Lacey 1984 Effects of gas composition and wateractivity on growth of eld and storage fungi and their interactionsTrans Br Mycol Soc 82305ndash314

23 MarDaggern S M E Guynot P Neira M Bernado V Sanchis and AJ Ramos 2002 Risk assessment of the use of sub-optimal levels ofweak-acid preservatives in the control of mould growth on bakeryproducts Int J Food Microbiol 79203ndash211

24 MarDaggern S M E Guynot V Sanchis J Arbones and A J Ramos2002 Aspergillus avus Aspergillus niger and Penicillium corylo-philum spoilage prevention of bakery products by means of weak-acid preservatives J Food Sci 672271ndash2277

25 MarDaggern S V Sanchis A Teixido et al 1996 Water and tem-perature relations and microconidial germination of Fusariummoniliforme and F proliferatum from maize Can J Microbiol421045ndash1054

26 Nielsen P V and R Rios 2000 Inhibition of fungal growth onbread by volatile components from species and herbs and the pos-sible application in active packaging with special emphasis on mus-tard essential oil Int J Food Microbiol 60219ndash229

27 Ortola C and C Santacreu 1998 Principios de aplicacion del en-vasado en atmosfera modi cada a los productos de pani cacion ybollerDaggera Aliment Equip Tecnol 17111ndash117

28 Piergiovanni L and P Fava 1997 Minimizing the residual oxygenin modi ed atmosphere packaging of bakery products Food AdditContam 14765ndash773

29 Pitt R E 1993 A descriptive model of mold growth and a atoxinformation as affected by environmental conditions J Food Prot 56139ndash146

30 RodrDaggerguez Castilla M V and R Jordano 1997 Envasado de prod-uctos de panaderDaggera y bollerDaggera en atmosferas modi cadas composi-cion de la atmosfera de envasado lms equipos de envasado y efec-tos sobre los productos de panaderDaggera y bollerDaggera Alimentaria 28679ndash99

31 Sautour M P Dantigny C Divies and M Bensoussan 2001A temperature-type model for describing the relationship be-tween fungal growth and water activity Int J Food Microbiol6763ndash69

32 Smith J P J Hoshino and Y Abe 1995 Interactive packaginginvolving sachet technology p 143ndash176 In M L Rooney (ed)

J Food Prot Vol 66 No 101868 GUYNOT ET AL

TABLE 3 Lag phases (days to visible growth estimated by the Gompertz equation) at pH 6 under modi ed atmosphere packaging(MAP) and at different aw levelsa

aw CO2

Lag phase (days) for organism

E amstelodami E herbariorum E repens E rubrum A niger A avus P corylophilum

080 0305070

100

26 6 2028282828

214 6 3628282828

187 6 3328282828

19 6 3027 6 15

282828

2828282828

2828282828

2828282828

085 0305070

100

15 6 50121 6 22172 6 35

2828

101 6 42143 6 21118 6 30

2828

81 6 20124 6 24172 6 20

2828

98 6 17179 6 26

18 6 202828

155 6 50121 6 38

27 6 402828

126 6 1928282828

28212 6 50

282828

090 0305070

100

66 6 4445 6 2138 6 3021 6 40

28

48 6 1654 6 0839 6 4593 6 23

28

64 6 0649 6 1364 6 1683 6 1126 6 40

44 6 0771 6 2155 6 3981 6 47

28

41 6 0660 6 37

130 6 55256 6 03

28

39 6 0547 6 1457 6 2096 6 44

28

2863 6 3520 6 20

2828

a Values presented are means 6 asymptotic standard errors Twenty-eight days was the maximal incubation period

(085 to 090) was probably the reason fungal growthstopped after some days of incubation

Impact of CO2 headspace concentration on colonyradius The analysis of covariance revealed that MAP (airand 0 to 100 of CO2 balanced with N2) aw (at levels of080 085 and 090) and their interaction had a statisticallysigni cant effect (P 001) on fungal growth (Table 2)Despite the signi cant effect of pH or the interaction of pHand atmosphere on the growth of some species (E amste-lodami E rubrum A avus and P corylophilum) no clearlink was found between CO2 antifungal activity and pHlevel In general all species were affected in the same waythey grew faster as the aw level increased and the level ofCO2 in the headspace decreased Some slight differencesamong fungal speciesrsquo levels of sensitivity to gas compo-sition were found with P corylophilum being the speciesthat was most affected by the limiting factors applied fol-lowed by Aspergillus spp and Eurotium spp Under theconditions that were most favorable for fungal growth (pre-sented by cakes with high aw levels packaged in air) themaximum colony radii were ca 25 mm for Eurotium sppand 5 mm for Aspergillus spp and P corylophilum

Water activity had a signi cant in uence on fungalgrowth and determined the level of CO2 needed to preventcake analog spoilage (Figs 1 through 3) It should be notedthat the maximal colony radii for cake analogs packagedwith 30 to 70 CO2 were ca 5 mm for Eurotium spp and2 mm for Aspergillus spp and P corylophilum In generalfungal growth was prevented at all aw levels in packageswith high CO2 levels (100) However E repens and Eherbariorum were able to grow under this condition (withthe maximum colony radius observed being 25 mm) at anaw level of 090 (Figs 1 and 2)

Impact of CO2 headspace concentration on lagphases In general longer lag phases were estimated forthe conditions that were the most critical for fungal growth

(Table 3) A linear correlation between the inverse of thelag phase and the growth rate was found for all isolateswith R2 values ranging from 084 to 096 depending on theisolate Longer lag phases corresponded to lower growthrates and smaller maximal colony radii It is important tomake use of these kinds of experimental data by construct-ing models with which given a combination of factors anda storage period one can predict the approximate degree ofspoilage Although approximate these models can helpfood technologists to better predict shelf life and safety

Lag phases for the seven isolates were modeled si-multaneously with the use of all factors (pH aw and CO2

concentration) and their quadratic terms and cross-term in-teractions However since pH was nonsigni cant (data notshown) a new regression model including only signi cantterms (aw and CO2 concentration their interaction and qua-dratic terms) was calculated This new PLS regression ledto two principal components models (PC1 and PC2) (Fig4) PC1 explained most of the variance in the lag phasedate (R2 5 759) PC2 is represented by a line orthogonalto PC1 and improves the approximation of the data set asmuch as possible carrying the maximum residual infor-mation (not taken into account in PC1) (18 variance ex-plained by PC2) The predictive power of the model (Q2)was near 720

The model explained the variation for almost all spe-cies quite well with fungal responses located close to oneanother in the loadings plot indicating that all species re-sponded to changes in tested factors in similar ways (Fig4) The greater the distance of a factor or response fromthe origin the stronger its effect on the model Thus re-sponses were well explained and aw CO2 concentrationand their interaction had a strong in uence of on lag phaseswhile quadratic terms had little effect on PC1 which meantthat they had little effect on lag phases If a factor is locatedopposite a fungus a negative correlation between the levelsof the factor and lag phase of that fungus exists

J Food Prot Vol 66 No 10 MAP FOR FUNGAL PREVENTION IN CAKES 1869

FIGURE 4 Plot of principal components models 1 and 2 (obtained by PLS regression) showing the relationships between aw CO2

level and their cross and quadratic terms and lag phase before growth (days)

TABLE 4 Polynomial model equation generated by PLS of lag phase (time to visible growth) against aw CO2 and quadratic andcross termsa

Isolate Q2 R2 Equation

E amstelodamiE herbariorumE repensE rubrumA nigerA avusP corylophilum

815649815809712823447

865678857843763848588

lag phase 5 239158 1 120093aw 2 156CO2 2 85129aw2 1 (496 3 1024)(CO2)2 1 196awCO2

lag phase 5 263002 1 176363aw 2 175CO2 2 118192aw2 1 (828 3 1024)(CO2)2 1 211awCO2

lag phase 5 211152 1 51190aw 2 121CO2 2 43105aw2 1 (111 3 1024)(CO2)2 1 160awCO2

lag phase 5 26409 1 38873aw 2 111CO2 2 35036aw2 1 (503 3 1025)(CO2)2 1 148awCO2

lag phase 5 257149 1 161086aw 2 157CO2 2 107631aw2 1 (761 3 1024)(CO2)2 1 189awCO2

lag phase 5 274534 1 206957aw 2 198CO2 2 137798aw2 1 (985 3 1024)(CO2)2 1 237awCO2

lag phase 5 249724 1 134113aw 2 096CO2 2 85260aw2 1 (698 3 1024)(CO2)2 1 108awCO2

a Q2 percentage of variation predicted by the model R2 percentage of variation explained by the model

Lag phases were described by a polynomial modelequation with six coef cients (b0 b1 b22) (Table 4)lag phase 5 b0 1 b1aw 1 b2CO2 1 b12awCO2 1 b11aw

2

1 b22CO22 It should be noted that these coef cients are

unscaled and uncentered in order to allow the direct cal-culation of lag phases so they cannot be used to assess thesigni cance of the effects of the factors on lag phases

Most of the models are highly signi cant with R2 val-ues of 70 The response of P corylophilum was slightlydifferent and the model explained only 588 of the lagphase variance for this organism (Table 4) This result canalso be observed in the loadings plot (Fig 4) in which Pcorylophilum appears quite separate from the other fungalspecies

The models were subsequently used to generate two-dimensional contour plots Plots of the combined effects ofaw and CO2 are shown in Figure 5 Since similar plots wereobtained for all Eurotium species and for all Aspergillusspecies one species of each genus is represented Lag phas-es for Eurotium spp at an aw level of 087 increased two-fold when the level of CO2 in the headspace increased from0 to 70 (eg E amstelodami growth started after 20days) Storage in 100 CO2 irrespective of the aw level

resulted in lag times of 20 days When Figure 5 is con-sidered it should be taken into account that the absence ofgrowth is equivalent to a lag phase of 28 days the durationof the experiment Thus when a shelf life of ca 28 daysis predicted from the models it is probable that no spoilagewill occur

DISCUSSION

In this study the use of MAP combined with differentpH and aw values to prevent the spoilage of bakery productsby Eurotium Aspergillus and Penicillium species was eval-uated Bakery products with near-neutral pHs were used forthe experiment The results obtained corroborate those fromother studies which show that CO2 has an important fun-gistatic effect (2 15 26) Spoilage was prevented with 70CO2 in the headspace (with a residual O2 level of 135)for products with an aw level of 080 while spoilage wassigni cantly delayed for products with aw levels of 085 to090 These results con rm those of previous studies inwhich reductions in the aw level were shown to increasethe effect of high levels of CO2 (10 19 20 22) It wasfound that with 30 to 50 CO2 in the headspace Eurotiumspp growth was totally prevented at an aw level of 080

J Food Prot Vol 66 No 101870 GUYNOT ET AL

FIGURE 5 Contour plots based on models obtained by PLS re-gression showing lag phases (days) as affected by CO2 level andaw for (a) E amstelodami (b) A niger and (c) P corylophilum

as water availability increased growth was only delayedOnly a 100 CO2 atmosphere prevented spoilage by almostall isolates regardless of the aw level

The inhibitory effect of CO2 on microorganisms in aculture medium or in a food system depends on many fac-tors These factors include partial CO2 pressure O2 con-centration headspace gas volume temperature acidity andaw (15) The effect of temperature was not assayed in thisstudy and the temperature level chosen was the one thatwas most likely to be encountered during distribution andstorage in the retail market The preserving effect of CO2

strongly depended on aw while interaction with pH was

not important probably because of the similar pH valuesinvolved Similarly Ellis et al (11 12) demonstrated thatthe growth of A avus under different atmospheric CO2

and O2 concentrations was highly dependent on aw andtemperature However these authors did not nd importantdifferences between the effects of gas composition on fun-gal growth at pH 6 and those at pH 8 Haasum and Nielsen(20) in a study of the growth of fungi in a cheese envi-ronment also found no signi cant differences in the inhib-itory effects of CO2 at different pH levels (4ndash8) The ef-fectiveness of CO2 also varies with the type of organismunder consideration with molds being more sensitive thanyeasts (19 22 30)

In this study all O2 concentrations measured at the endof the incubation period except those for air-packaged bagswere between 02 and 135 Since molds by de nition arestrict aerobes suf cient residual oxygen must be present inthe package headspace to allow mold growth (10 34) Ithas been demonstrated that molds can tolerate and evengrow in air with headspace oxygen concentrations as lowas 1 to 2 (13 32) Moreover several studies have shownthat molds can grow in the presence of elevated CO2 levelsif O2 is present (11 12) Smith et al (34) reported that aminimum of 04 O2 was needed for the growth of A nigerand Penicillium spp in a CO2N2 (6040) atmosphereAbellana et al (2) also found that low levels (002 to 05)of O2 did not modify fungal growth when levels of CO2 inthe bags were high However these authors found that asthe CO2 concentration decreased O2 levels had more in u-ence on growth kinetics (small differences in O2 levelscould make fungi grow) Similarly Nielsen and Rios (26)found that the growth of Aspergillus and Penicillium spe-cies was delayed only at reduced O2 levels (002 to 003)in either pure nitrogen or 50 N2 and 50 CO2 Howeverin pure CO2 the molds were strongly inhibited despite theresidual O2 in the package The total elimination of air dur-ing MAP represents a serious hurdle owing to both the lmrsquos permeability and the productrsquos capacity to trap O2

molecules (28) When a vacuum evacuation of the packageis carried out prior to gas injection the level of residual O2

in the headspace can be 1 (30) However the smoothand fragile texture of sponge cakes must be taken into ac-count the cakes could be deformed (27) An innovativemethod for controlling oxygen without physically deform-ing products involves the incorporation of O2 absorbers intothe package (32)

CO2 is both water and lipid soluble and when gaseousCO2 is applied to a product part of the CO2 goes to theliquid phase (water or oil) as carbonic acid (about 2) (7)This development can produce a small decrease in pH re-sulting in changes in the organoleptic characteristics of theproduct (15 30) The solubility of this gas in a culturemedium or in a food system increases as pH increases asaw increases as the volume of headspace gas increases oras temperature decreases The pHs of cake analogs pack-aged aerobically and with the different gas compositionstested were measured and no appreciable differences werefound (data not shown) This result may suggest that thesolubility of CO2 in the cake analog used was not signi -

J Food Prot Vol 66 No 10 MAP FOR FUNGAL PREVENTION IN CAKES 1871

cant or was not suf cient to produce a decrease in the an-alog pH

In the last few years predictive microbiology has be-come an important area of research The advantage of theresponse surface modeling approach is that it allows thesimultaneous examination of several variables Also math-ematical models can be generated to predict the packagingand environmental storage conditions that would controlfungal growth (11) Abellana et al (2) predicted lag phasesradial growth rates and maximum growth in relation to awO2 and CO2 concentrations for three Eurotium species ona sponge cake analog A model such as that described heremay be an important tool to aid food technologists in pre-dicting shelf lives of products or may give an idea of theeffects of certain changes in a productrsquos formulation Alsothis study emphasizes the importance of combining severalbarriers such as MAP aw and pH whose effects on moldgrowth are synergistic or additive

ACKNOWLEDGMENTS

The authors acknowledge the EC Quality of Life Programme (QoL)Key Action 1 (KA1) on Food Nutrition and Health (PL98-4075) theSpanish government (CICYT ALI99-0831) and the Catalonian govern-ment (CIRIT Comissio Interdepartamental de Recerca i Innovacio Tec-nologica)

REFERENCES

1 Abellana M J BenedDagger V Sanchis and A J Ramos 1999 Wateractivity and temperature effects on germination and growth of Eu-rotium amstelodami E chevalieri and E herbariorum isolates frombakery products J Appl Microbiol 87371ndash380

2 Abellana M V Sanchis A J Ramos and P V Nielsen 2000Effect of modi ed atmosphere packaging and water activity ongrowth of Eurotium amstelodami E chevalieri and E herbariorumon a sponge cake analogue J Appl Microbiol 88606ndash616

3 Abellana M L Torres V Sanchis and A J Ramos 1997 Car-acterizacion de diferentes productos de bollerDaggera industrial II Estudiode la mico ora Alimentaria 28751ndash56

4 Baranyi J A M Gibson J I Pitt M J Eyles and T A Roberts1996 Predictive models as means of measuring the relatedness ofsome Aspergillus species Food Microbiol 14347ndash351

5 Chirife J and G J Favetto 1992 Some physico-chemical basisof food preservation by combined methods Food Res Int 25389ndash396

6 Daifas D P J P Smith B Blanch eld and J W Austin 1999Growth and toxin production by Clostridium botulinum in English-style crumpets packaged under modi ed atmosphere J Food Prot62349ndash355

7 Daniels J A R Krishnamurthi and S S H Rizvi 1985 A reviewof effects of carbon dioxide on microbial growth and food qualityJ Food Prot 48532ndash537

8 Davies A R 1995 Advances in modi ed-atmosphere packaging p304ndash320 In G W Gould (ed) New methods of food preservationBlackie Academic and Professional London

9 Dixon N M and D B Kell 1989 The inhibition by CO2 of thegrowth and metabolism of micro-organisms J Appl Bacteriol 67109ndash136

10 El Halouat A and J M Debevere 1997 Effect of water activitymodi ed atmosphere packing and storage temperature on spore ger-mination of moulds isolated from prunes Int J Food Microbiol 3541ndash48

11 Ellis W O J P Smith B K Simpson S Khanizadeh and J HOldhem 1993 Control of growth and a atoxin production of As-pergillus avus under modi ed atmosphere packaging (MAP) con-ditions Food Microbiol 109ndash21

12 Ellis W O J P Smith B K Simpson and H Ramaswamy 1993

Effect of inoculum level on a atoxin production by Aspergillus a-vus under modi ed atmosphere packaging (MAP) conditions FoodMicrobiol 10525ndash535

13 Ellis W O J P Smith B K Simpson H Ramaswamy and GDoyon 1994 Growth of and a atoxin production by Aspergillus avus in peanuts stored under modi ed atmosphere packaging(MAP) conditions Int J Food Microbiol 22173ndash187

14 Eriksson L E Johansson N Kettaneh-Wold and S Wold 1999Introduction to multi- and megavariate data analysis using projectionmethods (PCA amp PLS) Umetrics AB Sweden

15 Faber J M 1991 Microbial aspects of modi ed-atmosphere pack-aging technologymdasha review J Food Prot 5458ndash70

16 GarcDaggera-Gimeno R M C Sanz-MartDaggernez J M GarcDaggera-Martos andG Zurera-Cosano 2002 Modeling Botrytis cinerea spores growthin carbon dioxide enriched atmospheres Food Microbiol Saf 671904ndash1907

17 Gibson A M and A D Hocking 1997 Advances in the predictivemodelling of fungal growth in food Trends Food Sci Technol 8353ndash358

18 Guynot M E A J Ramos D Sala V Sanchis and S MarDaggern2002 Combined effects of weak acid preservatives pH and wateractivity on growth of Eurotium species on a sponge cake Int JFood Microbiol 7639ndash46

19 Haasum I and P V Nielsen 1998 Ecophysiological characteriza-tion of common food-borne fungi in relation to pH and water activityunder various atmospheric compositions J Appl Microbiol 84451ndash460

20 Haasum I and P V Nielsen 1998 Physiological characterizationof common fungi associated with cheese J Food Sci 63157ndash161

21 Leistner L 1992 Food preservation by combined methods FoodRes Int 25151ndash158

22 Magan N and J Lacey 1984 Effects of gas composition and wateractivity on growth of eld and storage fungi and their interactionsTrans Br Mycol Soc 82305ndash314

23 MarDaggern S M E Guynot P Neira M Bernado V Sanchis and AJ Ramos 2002 Risk assessment of the use of sub-optimal levels ofweak-acid preservatives in the control of mould growth on bakeryproducts Int J Food Microbiol 79203ndash211

24 MarDaggern S M E Guynot V Sanchis J Arbones and A J Ramos2002 Aspergillus avus Aspergillus niger and Penicillium corylo-philum spoilage prevention of bakery products by means of weak-acid preservatives J Food Sci 672271ndash2277

25 MarDaggern S V Sanchis A Teixido et al 1996 Water and tem-perature relations and microconidial germination of Fusariummoniliforme and F proliferatum from maize Can J Microbiol421045ndash1054

26 Nielsen P V and R Rios 2000 Inhibition of fungal growth onbread by volatile components from species and herbs and the pos-sible application in active packaging with special emphasis on mus-tard essential oil Int J Food Microbiol 60219ndash229

27 Ortola C and C Santacreu 1998 Principios de aplicacion del en-vasado en atmosfera modi cada a los productos de pani cacion ybollerDaggera Aliment Equip Tecnol 17111ndash117

28 Piergiovanni L and P Fava 1997 Minimizing the residual oxygenin modi ed atmosphere packaging of bakery products Food AdditContam 14765ndash773

29 Pitt R E 1993 A descriptive model of mold growth and a atoxinformation as affected by environmental conditions J Food Prot 56139ndash146

30 RodrDaggerguez Castilla M V and R Jordano 1997 Envasado de prod-uctos de panaderDaggera y bollerDaggera en atmosferas modi cadas composi-cion de la atmosfera de envasado lms equipos de envasado y efec-tos sobre los productos de panaderDaggera y bollerDaggera Alimentaria 28679ndash99

31 Sautour M P Dantigny C Divies and M Bensoussan 2001A temperature-type model for describing the relationship be-tween fungal growth and water activity Int J Food Microbiol6763ndash69

32 Smith J P J Hoshino and Y Abe 1995 Interactive packaginginvolving sachet technology p 143ndash176 In M L Rooney (ed)

J Food Prot Vol 66 No 10 MAP FOR FUNGAL PREVENTION IN CAKES 1869

FIGURE 4 Plot of principal components models 1 and 2 (obtained by PLS regression) showing the relationships between aw CO2

level and their cross and quadratic terms and lag phase before growth (days)

TABLE 4 Polynomial model equation generated by PLS of lag phase (time to visible growth) against aw CO2 and quadratic andcross termsa

Isolate Q2 R2 Equation

E amstelodamiE herbariorumE repensE rubrumA nigerA avusP corylophilum

815649815809712823447

865678857843763848588

lag phase 5 239158 1 120093aw 2 156CO2 2 85129aw2 1 (496 3 1024)(CO2)2 1 196awCO2

lag phase 5 263002 1 176363aw 2 175CO2 2 118192aw2 1 (828 3 1024)(CO2)2 1 211awCO2

lag phase 5 211152 1 51190aw 2 121CO2 2 43105aw2 1 (111 3 1024)(CO2)2 1 160awCO2

lag phase 5 26409 1 38873aw 2 111CO2 2 35036aw2 1 (503 3 1025)(CO2)2 1 148awCO2

lag phase 5 257149 1 161086aw 2 157CO2 2 107631aw2 1 (761 3 1024)(CO2)2 1 189awCO2

lag phase 5 274534 1 206957aw 2 198CO2 2 137798aw2 1 (985 3 1024)(CO2)2 1 237awCO2

lag phase 5 249724 1 134113aw 2 096CO2 2 85260aw2 1 (698 3 1024)(CO2)2 1 108awCO2

a Q2 percentage of variation predicted by the model R2 percentage of variation explained by the model

Lag phases were described by a polynomial modelequation with six coef cients (b0 b1 b22) (Table 4)lag phase 5 b0 1 b1aw 1 b2CO2 1 b12awCO2 1 b11aw

2

1 b22CO22 It should be noted that these coef cients are

unscaled and uncentered in order to allow the direct cal-culation of lag phases so they cannot be used to assess thesigni cance of the effects of the factors on lag phases

Most of the models are highly signi cant with R2 val-ues of 70 The response of P corylophilum was slightlydifferent and the model explained only 588 of the lagphase variance for this organism (Table 4) This result canalso be observed in the loadings plot (Fig 4) in which Pcorylophilum appears quite separate from the other fungalspecies

The models were subsequently used to generate two-dimensional contour plots Plots of the combined effects ofaw and CO2 are shown in Figure 5 Since similar plots wereobtained for all Eurotium species and for all Aspergillusspecies one species of each genus is represented Lag phas-es for Eurotium spp at an aw level of 087 increased two-fold when the level of CO2 in the headspace increased from0 to 70 (eg E amstelodami growth started after 20days) Storage in 100 CO2 irrespective of the aw level

resulted in lag times of 20 days When Figure 5 is con-sidered it should be taken into account that the absence ofgrowth is equivalent to a lag phase of 28 days the durationof the experiment Thus when a shelf life of ca 28 daysis predicted from the models it is probable that no spoilagewill occur

DISCUSSION

In this study the use of MAP combined with differentpH and aw values to prevent the spoilage of bakery productsby Eurotium Aspergillus and Penicillium species was eval-uated Bakery products with near-neutral pHs were used forthe experiment The results obtained corroborate those fromother studies which show that CO2 has an important fun-gistatic effect (2 15 26) Spoilage was prevented with 70CO2 in the headspace (with a residual O2 level of 135)for products with an aw level of 080 while spoilage wassigni cantly delayed for products with aw levels of 085 to090 These results con rm those of previous studies inwhich reductions in the aw level were shown to increasethe effect of high levels of CO2 (10 19 20 22) It wasfound that with 30 to 50 CO2 in the headspace Eurotiumspp growth was totally prevented at an aw level of 080

J Food Prot Vol 66 No 101870 GUYNOT ET AL

FIGURE 5 Contour plots based on models obtained by PLS re-gression showing lag phases (days) as affected by CO2 level andaw for (a) E amstelodami (b) A niger and (c) P corylophilum

as water availability increased growth was only delayedOnly a 100 CO2 atmosphere prevented spoilage by almostall isolates regardless of the aw level

The inhibitory effect of CO2 on microorganisms in aculture medium or in a food system depends on many fac-tors These factors include partial CO2 pressure O2 con-centration headspace gas volume temperature acidity andaw (15) The effect of temperature was not assayed in thisstudy and the temperature level chosen was the one thatwas most likely to be encountered during distribution andstorage in the retail market The preserving effect of CO2

strongly depended on aw while interaction with pH was

not important probably because of the similar pH valuesinvolved Similarly Ellis et al (11 12) demonstrated thatthe growth of A avus under different atmospheric CO2

and O2 concentrations was highly dependent on aw andtemperature However these authors did not nd importantdifferences between the effects of gas composition on fun-gal growth at pH 6 and those at pH 8 Haasum and Nielsen(20) in a study of the growth of fungi in a cheese envi-ronment also found no signi cant differences in the inhib-itory effects of CO2 at different pH levels (4ndash8) The ef-fectiveness of CO2 also varies with the type of organismunder consideration with molds being more sensitive thanyeasts (19 22 30)

In this study all O2 concentrations measured at the endof the incubation period except those for air-packaged bagswere between 02 and 135 Since molds by de nition arestrict aerobes suf cient residual oxygen must be present inthe package headspace to allow mold growth (10 34) Ithas been demonstrated that molds can tolerate and evengrow in air with headspace oxygen concentrations as lowas 1 to 2 (13 32) Moreover several studies have shownthat molds can grow in the presence of elevated CO2 levelsif O2 is present (11 12) Smith et al (34) reported that aminimum of 04 O2 was needed for the growth of A nigerand Penicillium spp in a CO2N2 (6040) atmosphereAbellana et al (2) also found that low levels (002 to 05)of O2 did not modify fungal growth when levels of CO2 inthe bags were high However these authors found that asthe CO2 concentration decreased O2 levels had more in u-ence on growth kinetics (small differences in O2 levelscould make fungi grow) Similarly Nielsen and Rios (26)found that the growth of Aspergillus and Penicillium spe-cies was delayed only at reduced O2 levels (002 to 003)in either pure nitrogen or 50 N2 and 50 CO2 Howeverin pure CO2 the molds were strongly inhibited despite theresidual O2 in the package The total elimination of air dur-ing MAP represents a serious hurdle owing to both the lmrsquos permeability and the productrsquos capacity to trap O2

molecules (28) When a vacuum evacuation of the packageis carried out prior to gas injection the level of residual O2

in the headspace can be 1 (30) However the smoothand fragile texture of sponge cakes must be taken into ac-count the cakes could be deformed (27) An innovativemethod for controlling oxygen without physically deform-ing products involves the incorporation of O2 absorbers intothe package (32)

CO2 is both water and lipid soluble and when gaseousCO2 is applied to a product part of the CO2 goes to theliquid phase (water or oil) as carbonic acid (about 2) (7)This development can produce a small decrease in pH re-sulting in changes in the organoleptic characteristics of theproduct (15 30) The solubility of this gas in a culturemedium or in a food system increases as pH increases asaw increases as the volume of headspace gas increases oras temperature decreases The pHs of cake analogs pack-aged aerobically and with the different gas compositionstested were measured and no appreciable differences werefound (data not shown) This result may suggest that thesolubility of CO2 in the cake analog used was not signi -

J Food Prot Vol 66 No 10 MAP FOR FUNGAL PREVENTION IN CAKES 1871

cant or was not suf cient to produce a decrease in the an-alog pH

In the last few years predictive microbiology has be-come an important area of research The advantage of theresponse surface modeling approach is that it allows thesimultaneous examination of several variables Also math-ematical models can be generated to predict the packagingand environmental storage conditions that would controlfungal growth (11) Abellana et al (2) predicted lag phasesradial growth rates and maximum growth in relation to awO2 and CO2 concentrations for three Eurotium species ona sponge cake analog A model such as that described heremay be an important tool to aid food technologists in pre-dicting shelf lives of products or may give an idea of theeffects of certain changes in a productrsquos formulation Alsothis study emphasizes the importance of combining severalbarriers such as MAP aw and pH whose effects on moldgrowth are synergistic or additive

ACKNOWLEDGMENTS

The authors acknowledge the EC Quality of Life Programme (QoL)Key Action 1 (KA1) on Food Nutrition and Health (PL98-4075) theSpanish government (CICYT ALI99-0831) and the Catalonian govern-ment (CIRIT Comissio Interdepartamental de Recerca i Innovacio Tec-nologica)

REFERENCES

1 Abellana M J BenedDagger V Sanchis and A J Ramos 1999 Wateractivity and temperature effects on germination and growth of Eu-rotium amstelodami E chevalieri and E herbariorum isolates frombakery products J Appl Microbiol 87371ndash380

2 Abellana M V Sanchis A J Ramos and P V Nielsen 2000Effect of modi ed atmosphere packaging and water activity ongrowth of Eurotium amstelodami E chevalieri and E herbariorumon a sponge cake analogue J Appl Microbiol 88606ndash616

3 Abellana M L Torres V Sanchis and A J Ramos 1997 Car-acterizacion de diferentes productos de bollerDaggera industrial II Estudiode la mico ora Alimentaria 28751ndash56

4 Baranyi J A M Gibson J I Pitt M J Eyles and T A Roberts1996 Predictive models as means of measuring the relatedness ofsome Aspergillus species Food Microbiol 14347ndash351

5 Chirife J and G J Favetto 1992 Some physico-chemical basisof food preservation by combined methods Food Res Int 25389ndash396

6 Daifas D P J P Smith B Blanch eld and J W Austin 1999Growth and toxin production by Clostridium botulinum in English-style crumpets packaged under modi ed atmosphere J Food Prot62349ndash355

7 Daniels J A R Krishnamurthi and S S H Rizvi 1985 A reviewof effects of carbon dioxide on microbial growth and food qualityJ Food Prot 48532ndash537

8 Davies A R 1995 Advances in modi ed-atmosphere packaging p304ndash320 In G W Gould (ed) New methods of food preservationBlackie Academic and Professional London

9 Dixon N M and D B Kell 1989 The inhibition by CO2 of thegrowth and metabolism of micro-organisms J Appl Bacteriol 67109ndash136

10 El Halouat A and J M Debevere 1997 Effect of water activitymodi ed atmosphere packing and storage temperature on spore ger-mination of moulds isolated from prunes Int J Food Microbiol 3541ndash48

11 Ellis W O J P Smith B K Simpson S Khanizadeh and J HOldhem 1993 Control of growth and a atoxin production of As-pergillus avus under modi ed atmosphere packaging (MAP) con-ditions Food Microbiol 109ndash21

12 Ellis W O J P Smith B K Simpson and H Ramaswamy 1993

Effect of inoculum level on a atoxin production by Aspergillus a-vus under modi ed atmosphere packaging (MAP) conditions FoodMicrobiol 10525ndash535

13 Ellis W O J P Smith B K Simpson H Ramaswamy and GDoyon 1994 Growth of and a atoxin production by Aspergillus avus in peanuts stored under modi ed atmosphere packaging(MAP) conditions Int J Food Microbiol 22173ndash187

14 Eriksson L E Johansson N Kettaneh-Wold and S Wold 1999Introduction to multi- and megavariate data analysis using projectionmethods (PCA amp PLS) Umetrics AB Sweden

15 Faber J M 1991 Microbial aspects of modi ed-atmosphere pack-aging technologymdasha review J Food Prot 5458ndash70

16 GarcDaggera-Gimeno R M C Sanz-MartDaggernez J M GarcDaggera-Martos andG Zurera-Cosano 2002 Modeling Botrytis cinerea spores growthin carbon dioxide enriched atmospheres Food Microbiol Saf 671904ndash1907

17 Gibson A M and A D Hocking 1997 Advances in the predictivemodelling of fungal growth in food Trends Food Sci Technol 8353ndash358

18 Guynot M E A J Ramos D Sala V Sanchis and S MarDaggern2002 Combined effects of weak acid preservatives pH and wateractivity on growth of Eurotium species on a sponge cake Int JFood Microbiol 7639ndash46

19 Haasum I and P V Nielsen 1998 Ecophysiological characteriza-tion of common food-borne fungi in relation to pH and water activityunder various atmospheric compositions J Appl Microbiol 84451ndash460

20 Haasum I and P V Nielsen 1998 Physiological characterizationof common fungi associated with cheese J Food Sci 63157ndash161

21 Leistner L 1992 Food preservation by combined methods FoodRes Int 25151ndash158

22 Magan N and J Lacey 1984 Effects of gas composition and wateractivity on growth of eld and storage fungi and their interactionsTrans Br Mycol Soc 82305ndash314

23 MarDaggern S M E Guynot P Neira M Bernado V Sanchis and AJ Ramos 2002 Risk assessment of the use of sub-optimal levels ofweak-acid preservatives in the control of mould growth on bakeryproducts Int J Food Microbiol 79203ndash211

24 MarDaggern S M E Guynot V Sanchis J Arbones and A J Ramos2002 Aspergillus avus Aspergillus niger and Penicillium corylo-philum spoilage prevention of bakery products by means of weak-acid preservatives J Food Sci 672271ndash2277

25 MarDaggern S V Sanchis A Teixido et al 1996 Water and tem-perature relations and microconidial germination of Fusariummoniliforme and F proliferatum from maize Can J Microbiol421045ndash1054

26 Nielsen P V and R Rios 2000 Inhibition of fungal growth onbread by volatile components from species and herbs and the pos-sible application in active packaging with special emphasis on mus-tard essential oil Int J Food Microbiol 60219ndash229

27 Ortola C and C Santacreu 1998 Principios de aplicacion del en-vasado en atmosfera modi cada a los productos de pani cacion ybollerDaggera Aliment Equip Tecnol 17111ndash117

28 Piergiovanni L and P Fava 1997 Minimizing the residual oxygenin modi ed atmosphere packaging of bakery products Food AdditContam 14765ndash773

29 Pitt R E 1993 A descriptive model of mold growth and a atoxinformation as affected by environmental conditions J Food Prot 56139ndash146

30 RodrDaggerguez Castilla M V and R Jordano 1997 Envasado de prod-uctos de panaderDaggera y bollerDaggera en atmosferas modi cadas composi-cion de la atmosfera de envasado lms equipos de envasado y efec-tos sobre los productos de panaderDaggera y bollerDaggera Alimentaria 28679ndash99

31 Sautour M P Dantigny C Divies and M Bensoussan 2001A temperature-type model for describing the relationship be-tween fungal growth and water activity Int J Food Microbiol6763ndash69

32 Smith J P J Hoshino and Y Abe 1995 Interactive packaginginvolving sachet technology p 143ndash176 In M L Rooney (ed)

J Food Prot Vol 66 No 101870 GUYNOT ET AL

FIGURE 5 Contour plots based on models obtained by PLS re-gression showing lag phases (days) as affected by CO2 level andaw for (a) E amstelodami (b) A niger and (c) P corylophilum

as water availability increased growth was only delayedOnly a 100 CO2 atmosphere prevented spoilage by almostall isolates regardless of the aw level

The inhibitory effect of CO2 on microorganisms in aculture medium or in a food system depends on many fac-tors These factors include partial CO2 pressure O2 con-centration headspace gas volume temperature acidity andaw (15) The effect of temperature was not assayed in thisstudy and the temperature level chosen was the one thatwas most likely to be encountered during distribution andstorage in the retail market The preserving effect of CO2

strongly depended on aw while interaction with pH was

not important probably because of the similar pH valuesinvolved Similarly Ellis et al (11 12) demonstrated thatthe growth of A avus under different atmospheric CO2

and O2 concentrations was highly dependent on aw andtemperature However these authors did not nd importantdifferences between the effects of gas composition on fun-gal growth at pH 6 and those at pH 8 Haasum and Nielsen(20) in a study of the growth of fungi in a cheese envi-ronment also found no signi cant differences in the inhib-itory effects of CO2 at different pH levels (4ndash8) The ef-fectiveness of CO2 also varies with the type of organismunder consideration with molds being more sensitive thanyeasts (19 22 30)

In this study all O2 concentrations measured at the endof the incubation period except those for air-packaged bagswere between 02 and 135 Since molds by de nition arestrict aerobes suf cient residual oxygen must be present inthe package headspace to allow mold growth (10 34) Ithas been demonstrated that molds can tolerate and evengrow in air with headspace oxygen concentrations as lowas 1 to 2 (13 32) Moreover several studies have shownthat molds can grow in the presence of elevated CO2 levelsif O2 is present (11 12) Smith et al (34) reported that aminimum of 04 O2 was needed for the growth of A nigerand Penicillium spp in a CO2N2 (6040) atmosphereAbellana et al (2) also found that low levels (002 to 05)of O2 did not modify fungal growth when levels of CO2 inthe bags were high However these authors found that asthe CO2 concentration decreased O2 levels had more in u-ence on growth kinetics (small differences in O2 levelscould make fungi grow) Similarly Nielsen and Rios (26)found that the growth of Aspergillus and Penicillium spe-cies was delayed only at reduced O2 levels (002 to 003)in either pure nitrogen or 50 N2 and 50 CO2 Howeverin pure CO2 the molds were strongly inhibited despite theresidual O2 in the package The total elimination of air dur-ing MAP represents a serious hurdle owing to both the lmrsquos permeability and the productrsquos capacity to trap O2

molecules (28) When a vacuum evacuation of the packageis carried out prior to gas injection the level of residual O2

in the headspace can be 1 (30) However the smoothand fragile texture of sponge cakes must be taken into ac-count the cakes could be deformed (27) An innovativemethod for controlling oxygen without physically deform-ing products involves the incorporation of O2 absorbers intothe package (32)

CO2 is both water and lipid soluble and when gaseousCO2 is applied to a product part of the CO2 goes to theliquid phase (water or oil) as carbonic acid (about 2) (7)This development can produce a small decrease in pH re-sulting in changes in the organoleptic characteristics of theproduct (15 30) The solubility of this gas in a culturemedium or in a food system increases as pH increases asaw increases as the volume of headspace gas increases oras temperature decreases The pHs of cake analogs pack-aged aerobically and with the different gas compositionstested were measured and no appreciable differences werefound (data not shown) This result may suggest that thesolubility of CO2 in the cake analog used was not signi -

J Food Prot Vol 66 No 10 MAP FOR FUNGAL PREVENTION IN CAKES 1871

cant or was not suf cient to produce a decrease in the an-alog pH

In the last few years predictive microbiology has be-come an important area of research The advantage of theresponse surface modeling approach is that it allows thesimultaneous examination of several variables Also math-ematical models can be generated to predict the packagingand environmental storage conditions that would controlfungal growth (11) Abellana et al (2) predicted lag phasesradial growth rates and maximum growth in relation to awO2 and CO2 concentrations for three Eurotium species ona sponge cake analog A model such as that described heremay be an important tool to aid food technologists in pre-dicting shelf lives of products or may give an idea of theeffects of certain changes in a productrsquos formulation Alsothis study emphasizes the importance of combining severalbarriers such as MAP aw and pH whose effects on moldgrowth are synergistic or additive

ACKNOWLEDGMENTS

The authors acknowledge the EC Quality of Life Programme (QoL)Key Action 1 (KA1) on Food Nutrition and Health (PL98-4075) theSpanish government (CICYT ALI99-0831) and the Catalonian govern-ment (CIRIT Comissio Interdepartamental de Recerca i Innovacio Tec-nologica)

REFERENCES

1 Abellana M J BenedDagger V Sanchis and A J Ramos 1999 Wateractivity and temperature effects on germination and growth of Eu-rotium amstelodami E chevalieri and E herbariorum isolates frombakery products J Appl Microbiol 87371ndash380

2 Abellana M V Sanchis A J Ramos and P V Nielsen 2000Effect of modi ed atmosphere packaging and water activity ongrowth of Eurotium amstelodami E chevalieri and E herbariorumon a sponge cake analogue J Appl Microbiol 88606ndash616

3 Abellana M L Torres V Sanchis and A J Ramos 1997 Car-acterizacion de diferentes productos de bollerDaggera industrial II Estudiode la mico ora Alimentaria 28751ndash56

4 Baranyi J A M Gibson J I Pitt M J Eyles and T A Roberts1996 Predictive models as means of measuring the relatedness ofsome Aspergillus species Food Microbiol 14347ndash351

5 Chirife J and G J Favetto 1992 Some physico-chemical basisof food preservation by combined methods Food Res Int 25389ndash396

6 Daifas D P J P Smith B Blanch eld and J W Austin 1999Growth and toxin production by Clostridium botulinum in English-style crumpets packaged under modi ed atmosphere J Food Prot62349ndash355

7 Daniels J A R Krishnamurthi and S S H Rizvi 1985 A reviewof effects of carbon dioxide on microbial growth and food qualityJ Food Prot 48532ndash537

8 Davies A R 1995 Advances in modi ed-atmosphere packaging p304ndash320 In G W Gould (ed) New methods of food preservationBlackie Academic and Professional London

9 Dixon N M and D B Kell 1989 The inhibition by CO2 of thegrowth and metabolism of micro-organisms J Appl Bacteriol 67109ndash136

10 El Halouat A and J M Debevere 1997 Effect of water activitymodi ed atmosphere packing and storage temperature on spore ger-mination of moulds isolated from prunes Int J Food Microbiol 3541ndash48

11 Ellis W O J P Smith B K Simpson S Khanizadeh and J HOldhem 1993 Control of growth and a atoxin production of As-pergillus avus under modi ed atmosphere packaging (MAP) con-ditions Food Microbiol 109ndash21

12 Ellis W O J P Smith B K Simpson and H Ramaswamy 1993

Effect of inoculum level on a atoxin production by Aspergillus a-vus under modi ed atmosphere packaging (MAP) conditions FoodMicrobiol 10525ndash535

13 Ellis W O J P Smith B K Simpson H Ramaswamy and GDoyon 1994 Growth of and a atoxin production by Aspergillus avus in peanuts stored under modi ed atmosphere packaging(MAP) conditions Int J Food Microbiol 22173ndash187

14 Eriksson L E Johansson N Kettaneh-Wold and S Wold 1999Introduction to multi- and megavariate data analysis using projectionmethods (PCA amp PLS) Umetrics AB Sweden

15 Faber J M 1991 Microbial aspects of modi ed-atmosphere pack-aging technologymdasha review J Food Prot 5458ndash70

16 GarcDaggera-Gimeno R M C Sanz-MartDaggernez J M GarcDaggera-Martos andG Zurera-Cosano 2002 Modeling Botrytis cinerea spores growthin carbon dioxide enriched atmospheres Food Microbiol Saf 671904ndash1907

17 Gibson A M and A D Hocking 1997 Advances in the predictivemodelling of fungal growth in food Trends Food Sci Technol 8353ndash358

18 Guynot M E A J Ramos D Sala V Sanchis and S MarDaggern2002 Combined effects of weak acid preservatives pH and wateractivity on growth of Eurotium species on a sponge cake Int JFood Microbiol 7639ndash46

19 Haasum I and P V Nielsen 1998 Ecophysiological characteriza-tion of common food-borne fungi in relation to pH and water activityunder various atmospheric compositions J Appl Microbiol 84451ndash460

20 Haasum I and P V Nielsen 1998 Physiological characterizationof common fungi associated with cheese J Food Sci 63157ndash161

21 Leistner L 1992 Food preservation by combined methods FoodRes Int 25151ndash158

22 Magan N and J Lacey 1984 Effects of gas composition and wateractivity on growth of eld and storage fungi and their interactionsTrans Br Mycol Soc 82305ndash314

23 MarDaggern S M E Guynot P Neira M Bernado V Sanchis and AJ Ramos 2002 Risk assessment of the use of sub-optimal levels ofweak-acid preservatives in the control of mould growth on bakeryproducts Int J Food Microbiol 79203ndash211

24 MarDaggern S M E Guynot V Sanchis J Arbones and A J Ramos2002 Aspergillus avus Aspergillus niger and Penicillium corylo-philum spoilage prevention of bakery products by means of weak-acid preservatives J Food Sci 672271ndash2277

25 MarDaggern S V Sanchis A Teixido et al 1996 Water and tem-perature relations and microconidial germination of Fusariummoniliforme and F proliferatum from maize Can J Microbiol421045ndash1054

26 Nielsen P V and R Rios 2000 Inhibition of fungal growth onbread by volatile components from species and herbs and the pos-sible application in active packaging with special emphasis on mus-tard essential oil Int J Food Microbiol 60219ndash229

27 Ortola C and C Santacreu 1998 Principios de aplicacion del en-vasado en atmosfera modi cada a los productos de pani cacion ybollerDaggera Aliment Equip Tecnol 17111ndash117

28 Piergiovanni L and P Fava 1997 Minimizing the residual oxygenin modi ed atmosphere packaging of bakery products Food AdditContam 14765ndash773

29 Pitt R E 1993 A descriptive model of mold growth and a atoxinformation as affected by environmental conditions J Food Prot 56139ndash146

30 RodrDaggerguez Castilla M V and R Jordano 1997 Envasado de prod-uctos de panaderDaggera y bollerDaggera en atmosferas modi cadas composi-cion de la atmosfera de envasado lms equipos de envasado y efec-tos sobre los productos de panaderDaggera y bollerDaggera Alimentaria 28679ndash99

31 Sautour M P Dantigny C Divies and M Bensoussan 2001A temperature-type model for describing the relationship be-tween fungal growth and water activity Int J Food Microbiol6763ndash69

32 Smith J P J Hoshino and Y Abe 1995 Interactive packaginginvolving sachet technology p 143ndash176 In M L Rooney (ed)

J Food Prot Vol 66 No 10 MAP FOR FUNGAL PREVENTION IN CAKES 1871

cant or was not suf cient to produce a decrease in the an-alog pH

In the last few years predictive microbiology has be-come an important area of research The advantage of theresponse surface modeling approach is that it allows thesimultaneous examination of several variables Also math-ematical models can be generated to predict the packagingand environmental storage conditions that would controlfungal growth (11) Abellana et al (2) predicted lag phasesradial growth rates and maximum growth in relation to awO2 and CO2 concentrations for three Eurotium species ona sponge cake analog A model such as that described heremay be an important tool to aid food technologists in pre-dicting shelf lives of products or may give an idea of theeffects of certain changes in a productrsquos formulation Alsothis study emphasizes the importance of combining severalbarriers such as MAP aw and pH whose effects on moldgrowth are synergistic or additive

ACKNOWLEDGMENTS

The authors acknowledge the EC Quality of Life Programme (QoL)Key Action 1 (KA1) on Food Nutrition and Health (PL98-4075) theSpanish government (CICYT ALI99-0831) and the Catalonian govern-ment (CIRIT Comissio Interdepartamental de Recerca i Innovacio Tec-nologica)

REFERENCES

1 Abellana M J BenedDagger V Sanchis and A J Ramos 1999 Wateractivity and temperature effects on germination and growth of Eu-rotium amstelodami E chevalieri and E herbariorum isolates frombakery products J Appl Microbiol 87371ndash380

2 Abellana M V Sanchis A J Ramos and P V Nielsen 2000Effect of modi ed atmosphere packaging and water activity ongrowth of Eurotium amstelodami E chevalieri and E herbariorumon a sponge cake analogue J Appl Microbiol 88606ndash616

3 Abellana M L Torres V Sanchis and A J Ramos 1997 Car-acterizacion de diferentes productos de bollerDaggera industrial II Estudiode la mico ora Alimentaria 28751ndash56

4 Baranyi J A M Gibson J I Pitt M J Eyles and T A Roberts1996 Predictive models as means of measuring the relatedness ofsome Aspergillus species Food Microbiol 14347ndash351

5 Chirife J and G J Favetto 1992 Some physico-chemical basisof food preservation by combined methods Food Res Int 25389ndash396

6 Daifas D P J P Smith B Blanch eld and J W Austin 1999Growth and toxin production by Clostridium botulinum in English-style crumpets packaged under modi ed atmosphere J Food Prot62349ndash355

7 Daniels J A R Krishnamurthi and S S H Rizvi 1985 A reviewof effects of carbon dioxide on microbial growth and food qualityJ Food Prot 48532ndash537

8 Davies A R 1995 Advances in modi ed-atmosphere packaging p304ndash320 In G W Gould (ed) New methods of food preservationBlackie Academic and Professional London

9 Dixon N M and D B Kell 1989 The inhibition by CO2 of thegrowth and metabolism of micro-organisms J Appl Bacteriol 67109ndash136

10 El Halouat A and J M Debevere 1997 Effect of water activitymodi ed atmosphere packing and storage temperature on spore ger-mination of moulds isolated from prunes Int J Food Microbiol 3541ndash48

11 Ellis W O J P Smith B K Simpson S Khanizadeh and J HOldhem 1993 Control of growth and a atoxin production of As-pergillus avus under modi ed atmosphere packaging (MAP) con-ditions Food Microbiol 109ndash21

12 Ellis W O J P Smith B K Simpson and H Ramaswamy 1993

Effect of inoculum level on a atoxin production by Aspergillus a-vus under modi ed atmosphere packaging (MAP) conditions FoodMicrobiol 10525ndash535

13 Ellis W O J P Smith B K Simpson H Ramaswamy and GDoyon 1994 Growth of and a atoxin production by Aspergillus avus in peanuts stored under modi ed atmosphere packaging(MAP) conditions Int J Food Microbiol 22173ndash187

14 Eriksson L E Johansson N Kettaneh-Wold and S Wold 1999Introduction to multi- and megavariate data analysis using projectionmethods (PCA amp PLS) Umetrics AB Sweden

15 Faber J M 1991 Microbial aspects of modi ed-atmosphere pack-aging technologymdasha review J Food Prot 5458ndash70

16 GarcDaggera-Gimeno R M C Sanz-MartDaggernez J M GarcDaggera-Martos andG Zurera-Cosano 2002 Modeling Botrytis cinerea spores growthin carbon dioxide enriched atmospheres Food Microbiol Saf 671904ndash1907

17 Gibson A M and A D Hocking 1997 Advances in the predictivemodelling of fungal growth in food Trends Food Sci Technol 8353ndash358

18 Guynot M E A J Ramos D Sala V Sanchis and S MarDaggern2002 Combined effects of weak acid preservatives pH and wateractivity on growth of Eurotium species on a sponge cake Int JFood Microbiol 7639ndash46

19 Haasum I and P V Nielsen 1998 Ecophysiological characteriza-tion of common food-borne fungi in relation to pH and water activityunder various atmospheric compositions J Appl Microbiol 84451ndash460

20 Haasum I and P V Nielsen 1998 Physiological characterizationof common fungi associated with cheese J Food Sci 63157ndash161

21 Leistner L 1992 Food preservation by combined methods FoodRes Int 25151ndash158

22 Magan N and J Lacey 1984 Effects of gas composition and wateractivity on growth of eld and storage fungi and their interactionsTrans Br Mycol Soc 82305ndash314

23 MarDaggern S M E Guynot P Neira M Bernado V Sanchis and AJ Ramos 2002 Risk assessment of the use of sub-optimal levels ofweak-acid preservatives in the control of mould growth on bakeryproducts Int J Food Microbiol 79203ndash211

24 MarDaggern S M E Guynot V Sanchis J Arbones and A J Ramos2002 Aspergillus avus Aspergillus niger and Penicillium corylo-philum spoilage prevention of bakery products by means of weak-acid preservatives J Food Sci 672271ndash2277

25 MarDaggern S V Sanchis A Teixido et al 1996 Water and tem-perature relations and microconidial germination of Fusariummoniliforme and F proliferatum from maize Can J Microbiol421045ndash1054

26 Nielsen P V and R Rios 2000 Inhibition of fungal growth onbread by volatile components from species and herbs and the pos-sible application in active packaging with special emphasis on mus-tard essential oil Int J Food Microbiol 60219ndash229

27 Ortola C and C Santacreu 1998 Principios de aplicacion del en-vasado en atmosfera modi cada a los productos de pani cacion ybollerDaggera Aliment Equip Tecnol 17111ndash117

28 Piergiovanni L and P Fava 1997 Minimizing the residual oxygenin modi ed atmosphere packaging of bakery products Food AdditContam 14765ndash773

29 Pitt R E 1993 A descriptive model of mold growth and a atoxinformation as affected by environmental conditions J Food Prot 56139ndash146

30 RodrDaggerguez Castilla M V and R Jordano 1997 Envasado de prod-uctos de panaderDaggera y bollerDaggera en atmosferas modi cadas composi-cion de la atmosfera de envasado lms equipos de envasado y efec-tos sobre los productos de panaderDaggera y bollerDaggera Alimentaria 28679ndash99

31 Sautour M P Dantigny C Divies and M Bensoussan 2001A temperature-type model for describing the relationship be-tween fungal growth and water activity Int J Food Microbiol6763ndash69

32 Smith J P J Hoshino and Y Abe 1995 Interactive packaginginvolving sachet technology p 143ndash176 In M L Rooney (ed)

J Food Prot Vol 66 No 101872 GUYNOT ET AL

Active food packaging Blackie Academic amp Professional Lon-don

33 Smith J P S Khanizadeh F R van de Voort R Hardin B Oor-aikul and E D Jackson 1988 Use of response surface methodologyin shelf life extension studies of a bakery product Food Microbiol5163ndash176

34 Smith J P B Ooraikul W J Koersen E D Jackson and R ALawrence 1986 Novel approach to oxygen control in modi ed at-

mosphere packaging of bakery products Food Microbiol 3315ndash320

35 Smith J P and B K Simpson 1995 Modi ed atmosphere pack-aging of bakery and pasta products p 207ndash242 In K L Dodds andJ Faber (ed) Principles of modi ed-atmosphere and sous vide prod-uct packaging Technomic Publishing Company Lancaster Pa

36 Sparakawski W 1993 Longer shelf life and freshness without ar-ti cial preservatives Food Mark Technol 744ndash48