Ž .International Journal of Food Microbiology 69 2001 79–89www.elsevier.comrlocaterijfoodmicro

Variability of the lipolytic activity in Yarrowia lipolytica and itsdependence on environmental conditions

M.E. Guerzoni a,), R. Lanciotti a, L. Vannini a, F. Galgano b, F. Favati b, F. Gardini a,G. Suzzi b

a Dipartimento di Protezione e Valorizzazione Agroalimentare, UniÕersita degli Studi di Bologna, Õia S. Giacomo 7, 40126 Bologna, Italy`b Dipartimento di Biologia, Difesa e Biotecnologie Agroindustriali, UniÕersita degli Studi della Basilicata, Via Nazzario Sauro 85,`

85100 Potenza, Italy

Abstract

This work was aimed to the evaluation of the variability of lipolytic activity in Yarrowia lipolytica strains, as well as toŽ .asses for a selected strain, the response to the changes of physico-chemical variables such as pH, NaCl and lipid content , in

order to obtain predictive models describing their effects on the lipolysis pattern. The strains tested, having differentŽ .environmental origin, showed different patterns of the free fatty acids FFA released. The clustering of the free fatty acids

profiles evidenced that the unweighted average distance within the strains of the same species did not exceeded 30%.However, the lipolytic activity of some strains generated FFA profiles that differentiated from the majority of the strainsconsidered. Also, when a single strain was inoculated in model systems in which pH, NaCl and milk fat were modulated

Ž .according to a Central Composite Design CCD , chemico-physical characteristics of the system led to marked variations inthe lipolytic activity with consequent changes in individual fatty acids released.

In most cases, when the same Y. lipolytica strain was used, under the experimental conditions adopted, the modulation ofthe lactic acid, NaCl and lipid content did not generate differences in the fatty acid release exceeding 20–21%. However,some combinations of factors remarkably affected lipase expression or activity, and generated differences in the fatty acidreleased higher than those observed among different strains of the same species. q 2001 Elsevier Science B.V. All rightsreserved.

Keywords: Yarrowia lipolytica; Lipolytic activity; Free fatty acids

1. Introduction

The assessment of the role of a microbial group orspecies in the development of both desirable andundesirable flavours and textural features duringcheese ripening is generally obscured by the com-plex interactions of inoculated or naturally occurring

) Corresponding author. Tel.: q39-51-2099783; fax: q39-51-2099782.

Ž .E-mail address: guerzoni@foodsci.unibo.it M.E. Guerzoni .

species. In fact, all the components of the system,rennet, starters, milk bacteria, yeasts and moisturecontent are conditioned by the technological processand in turns affect the chemico-physical changesassociated with the fermentation and ripening pro-cess. The specific contribution of non-starter mi-croorganisms is, however, underestimated or not eas-ily distinguished. In particular, the yeast flora thatproliferates during maturation is an uncontrolledmixture of wild species whose favourable or un-favourable activity on cheese quality is unpre-

0168-1605r01r$ - see front matter q 2001 Elsevier Science B.V. All rights reserved.Ž .PII: S0168-1605 01 00575-X

( )M.E. Guerzoni et al.r International Journal of Food Microbiology 69 2001 79–8980

dictable. Therefore, any effects attributed to yeastcontamination, at levels often exceeding 107–108

cellrg, are not easily measurable because of the lackof an established quantitative relationship betweencell number and volatile or non-volatile metabolitesor lipid and protein breakdown products.

The mechanisms by which yeast growth is thoughtto influence the final quality of cheeses are thoseassociated to the fermentation of residual lactose,utilisation of lactic acid and their proteolytic and

Žlipolytic activities Fleet, 1992; Jakobsen and.Narvhus, 1996 . However, objective and quantitative

criteria to distinguish between desirable and undesir-able yeast species and activities for cheese ripeningare still lacking. Moreover, it should be possible toidentify, for the different metabolites or flavours,thresholds beyond which the desirable changes alsobecome detrimental to cheese quality or stability.

When non-conventional species are used asŽ .costarters, it is necessary to: a identify the factors

which control or limit the growth of the chosenŽ .species or strains under defined conditions; b iden-

tify the enzymatic activities which mostly contributeto the cheese attributes and explore in vitro anyopportunities provided by intraspecific variability;

Ž .and c evaluate the quantitative effects of environ-mental factors on the expression of selected enzy-matic activities.

The ability of Yarrowia lipolytica to predominatein real system on the naturally occurring yeasts andits compatibility with starter cultures have been evi-

Ž .denced by Guerzoni et al. 1998 and van den Tem-Ž .pel and Jakobsen 2000 . Moreover, this species has

been suggested for a deliberate use in cheese ripen-ing due to its extracellular enzymatic activities, par-

Žticularly its lipase Guerzoni et al., 1998; van den.Tempel and Jakobsen, 2000 . However, the released

fatty acids can further be transformed into desirableor undesirable volatile or non-volatile compoundswith characteristic aroma. Therefore, the selection ofstrains of Y. lipolytica to be used as costarter shouldbe based on a deeper knowledge of their milk fathydrolysis patterns.

Ž .The specific aims of this work were: 1 to inves-tigate the variability in the lipolysis patterns of strainsof Y. lipolytica of different environmental origin;

Ž .and 2 to assess a selected strain for its response tochanges in pH, NaCl and lipid content, in order to

obtain a predictive model describing their effects onthe lipolysis pattern.

2. Materials and methods

2.1. Microorganisms

Twelve strains of Y. lipolytica, belonging to theculture collection of the Dipartimento di Protezione eValorizzazione Agroalimentare, were employed. Thestrains used in this work originated from previous

Ž .examinations of chilled foods Y3, Y5, Y14 , lightŽ .butter RO13, RO21 , superficial water of lagoon of

Ž .the Po river delta PO1, PO10, PO17 and irradiatedŽ . Žpoultry meat 5D, 5B, 27D, 1A Guerzoni et al.,

.1993; Sinigaglia et al., 1994

2.2. Lipolytic actiÕity of 12 strains

The 12 strains of Y. lipolytica were inoculated atlevels of 103 cfurg on plates containing 10 g ofcommercial butter and incubated at 25 8C. After 7days, the samples were subjected to the analysis of

Ž .the free fatty acids FFA .

2.3. Effect of compositional factors on lipase actiÕityexpression

The individual and interactive effects of environ-mental factors on lipase activity expression of a

Ž .single Y. lipolytica strain Y3 were evaluated byŽ .varying milk fat Parmalat, Parma, Italy , NaCl and

Ž .lactic acid Carlo Erba, Milan, Italy concentrationsin pasteurised skimmed milk according to a three-

Ž .factor five-level Central Composite Design CCDŽ .Box et al., 1978 . The 17 combinations obtained arereported in Table 1. After preparation, the 17 combi-nations were inoculated with about 104 cfurg of Y.lipolytica Y3 and incubated at 25 8C for 22 days.The determination of free fatty acids was performedafter 3, 7 and 22 days of incubation at 25 8C.

2.4. Free fatty acids determination

Ž .The determination of free fatty acids FFA bothof butter samples and of the 17 different milk sam-ples was carried out using a slight modification of

( )M.E. Guerzoni et al.r International Journal of Food Microbiology 69 2001 79–89 81

Table 1Different combinations obtained by modulating NaCl, lactic acidand milk fat content according to a Central Composite Design

Run NaCl Lactic MilkŽ . Ž . Ž .% acid % fat %

1 1.00 0.25 2.002 2.00 0.25 2.003 1.00 0.75 2.004 2.00 0.75 2.005 1.00 0.25 4.006 2.00 0.25 4.007 1.00 0.75 4.008 2.00 0.75 4.009 1.50 0.50 3.00

10 1.50 0.50 3.0011 0.50 0.50 3.0012 2.50 0.50 3.0013 1.50 0.00 3.0014 1.50 1.00 3.0015 1.50 0.50 1.0016 1.50 0.50 5.0017 1.50 0.50 3.00

Ž .the procedure proposed by Lencioni et al. 1998 . Inparticular, the extraction of lipids and measurementsof FFA were performed after 7 days for buttersamples and after 3, 7 and 22 days for the differentmilk samples. All the chemicals used for the lipidextraction and isolation of FFA were from CarloErba.

2.4.1. Lipid extraction procedureTwo grams of culture medium was put in a

screw-capped glass centrifuge tube, and 2 ml ofethanol and 0.3 ml of 0.5 M H SO was added. The2 4

mixture was extracted three times with 3 ml of aŽ .diethyl etherrhexane 1:1 . Each time, the slurry was

heated to 40 8C, shaken for 5 min using a VortexŽmixer and clarified by a short centrifugation 2 min,

.2000=g, 4 8C . After centrifugation, the upper sol-vent layer was transferred to a screw-capped glasscentrifuge tube containing 1 g of anhydrous Na SO ,2 4

in order to adsorb the residual water. The lipidŽ .extract ca. 9 ml was transferred to a 50-ml volu-

metric flask and brought up to volume with hexane.

2.4.2. Isolation of the FFASeparation of the FFA fraction from other lipid

classes in the extract was conducted utilising SPE

Ž . Žaminopropyl columns 500 mg Supelco, Milan,.Italy . After conditioning the column with 10 ml of

hexane, the whole lipid extract was loaded onto thecolumn and eluted drop by drop. The neutral lipidswere washed out by using 4 ml of a chloroformr

Ž .hexane 1:1 mixture, and the FFA fraction was theneluted using 4 ml of diethyl ether containing 2%formic acid. After the recovery of the FFA fraction,the column was washed with 10 ml of methanol andreconditioned with 10 ml of hexane before beingreused.

The sample was then spiked with 250 ml of a C7:0Ž .solution 1 mgrml as internal standard.

2.4.3. Gas chromatographic analysisFFA analysis was conducted using a Carlo Erba

Ž .Mega 5160 gas chromatograph Carlo Erba equippedwith a flame ionisation detector and a cold on col-

Ž .umn injector. The sample 1 ml was injected onto aŽ .Nukol wide-bore column 15 m=0.53 mm ID hav-

Ž .ing a film thickness of 0.5 mm Supelco . The carrierŽ .gas helium flow rate was set at 20 mlrmin, the

detector temperature at 220 8C, and detector attenua-tion at 10. The oven temperature was programmedfrom 90 to 195 8C at 10 8Crmin.

2.5. Statistical analysis of data

ŽA software package Statistica for Windows, Stat-.soft, Tusla, USA was used for data analysis. Poly-

nomial equations describing the effects of the inde-Žpendent variables lactic acid, NaCl and milk fat

.content as individual or quadratic terms, and of theirinteractive effects on the dependent variables, wereobtained. The goodness of fit of the models obtained

Ž .was evaluated using R regression coefficient , theŽ .Fisher F-test and derived p-values and the stan-

Ž .dard error of estimate SE . The plots of the responsesurface were obtained as a function of two indepen-

Ždent variable at a time e.g. lactic acid and NaCl. Žcontent holding the other independent variable e.g.

. Žmilk fat content constant at a fixed level for in-.stance, the intermediate value . These plots can be

easily derived from the polynomial models obtainedby calculating the values assumed by one indepen-dent variable when the second varies from the mini-mum to the maximum values considered and the

( )M.E. Guerzoni et al.r International Journal of Food Microbiology 69 2001 79–8982

Ždependent variable has given values Cheynier et al.,.1983 .

The cluster analysis was performed by means ofthe unweighted pair group using Manahattan dis-tance algorithm.

3. Results

Twelve strains of Y. lipolytica were inoculated onbutter and the free fatty acids in the lipid extractswere examined after 7 days of incubation at 25 8C.Generally, lipase activity is tested on tryacylglyc-

Žerols composed of long chain fatty acids Jensen,.1983; Peled and Krenz, 1981 . In this trial, butter has

been regarded as a suitable substrate due to its stateof uniform fine emulsion because the lipase reactionvaries directly with the surface area of substrate

Ž .available to the enzyme Hadeball, 1991 . The analy-sis of the total fatty acids indicated that some fattyacids had been assimilated and reconverted into othercellular fatty acids. In particular, an increase of

Ž .pentadecanoic acid C and anteiso heptadecanoic15

acid was observed in butter inoculated with somestrains. In fact, branched fatty acids have been found

Žin Y. lipolytica cellular membranes Sinigaglia et al.,.1994 . Table 2 shows the free fatty acids released

Ž .after 7 days as percentage in comparison with theuninoculated control. The changes of the latter can

be attributed to naturally occurring butter lipases.Ž .Oleic acid C is the principal fatty acid re-18:1 D9

leased in butter inoculated with the majority ofthe strains. It accounted for 76.7% of the free fattyacids in the samples inoculated with Y5, but itspresence did not exceed 15% in the samples inocu-lated with PO10 and RO13. On the contrary, thebutter samples inoculated with these strains werecharacterised by a high proportion of free linolenic

Ž . Ž .acid C and stearic acid C . Moreover, RO1318:2 18:0

induced a selective release or production of margaricŽ .acid C , C anteiso, C iso and heptade-17:0 17:0 17:0

Ž .cenoic acid C .17:1

The large variability observed within this group ofstrains can have an indirect effect on the subsequenttransformation of the fatty acids into aroma com-pounds.

In particular, the release of polyunsaturated fattyacids can have a relevant role in the aroma ofsurface-ripened cheeses. Linoleic acid is regarded asa precursor of eight carbon components containingtwo double bonds which can further be degraded to

Žless volatile 10-carbon oxo and hydroxy acids Jelen.and Wasowicz, 1998 . On the other hand, the mono-

saturated C and C carbon chain fatty acids are16 18

precursors of hydroxy acids which, in turn, are con-verted in g and d lactones.

The practical significance attributed to such invitro strain differences needs to be evaluated. Also,

Table 2Free fatty acids released by different Y. lipolytica strains inoculated in butter samples

Strain Free fatty acid percentages

C C C C C C C C C C C C C C C C12:0 14:0 14:1 14:2 15:0 16:0 16:1 16:2 17:0 17:1 17:0 17:0 18:0 18:1D9 18:1D11 18:2

anteiso iso

PO1 2.1 10.3 2.9 0.0 3.1 3.9 0.4 0.0 0.5 0.3 0.3 0.0 16.0 54.0 0.0 6.2PO10 1.4 9.2 1.1 0.0 3.4 18.4 2.6 0.0 0.3 0.0 0.0 0.0 29.4 14.9 0.0 19.3PO17 0.2 5.7 0.2 0.0 0.2 3.5 0.6 0.0 0.2 0.1 0.1 0.0 14.8 64.2 0.2 10.05D 2.9 8.7 0.6 0.0 0.5 7.3 1.4 0.0 0.6 0.4 0.0 0.3 11.4 52.4 0.3 13.25B 0.2 8.1 1.1 0.0 0.3 3.6 0.7 0.0 0.1 0.0 0.0 0.0 12.4 64.4 0.1 9.01A 0.6 9.1 1.8 0.1 0.7 3.3 0.5 0.0 0.7 0.8 0.2 0.0 11.1 63.8 0.3 7.027D 0.4 1.0 0.0 0.0 0.5 34.3 0.6 0.0 0.0 0.0 0.0 0.0 8.7 48.6 0.0 5.9Y3 0.8 11.6 2.8 0.1 1.4 32.5 0.3 0.2 1.0 0.5 0.4 0.0 9.0 35.0 0.4 4.0Y5 0.2 2.3 0.4 0.0 0.2 5.9 0.7 0.0 0.1 0.0 0.0 0.0 6.9 76.7 0.0 6.6Y14 0.2 10.5 1.8 0.0 3.6 30.9 2.5 0.0 0.2 0.0 0.0 0.0 10.1 36.0 0.0 4.2RO13 4.4 2.3 5.4 0.0 0.8 9.5 1.3 0.8 2.8 2.0 0.6 0.1 34.5 15.0 1.0 19.5RO21 0.3 1.0 0.2 0.0 0.4 4.2 0.6 0.0 0.0 0.0 0.0 0.0 13.5 70.1 0.2 9.5

aControl 2.4 0.7 1.8 0.0 4.2 20.7 3.6 0.0 0.5 0.4 0.1 0.0 47.3 1.9 0.1 16.3

a Not-inoculated sample.

( )M.E. Guerzoni et al.r International Journal of Food Microbiology 69 2001 79–89 83

Table 3Total free fatty acids released from skimmed milk based modelsystems after 3, 7 and 22 days of incubation with Y. lipolytica Y3under different conditions of fat, salt and lactic acid

a Ž .Condition Total FFA mgrl

3 days 7 days 22 days

1 88.34 95.09 397.332 86.60 182.49 502.113 89.83 99.20 189.084 31.58 59.75 247.035 84.95 159.66 211.526 139.73 369.16 222.637 97.25 224.66 372.448 44.60 118.65 137.739 73.98 129.70 141.45

10 49.16 115.83 149.5311 91.60 323.32 452.4012 77.80 108.78 304.1113 222.05 456.02 592.3314 56.44 162.90 343.2415 53.46 95.88 71.0416 53.39 99.14 93.8717 83.44 145.77 93.75

aSee Table 1 for conditions.

the expression and activity of the lipases of Y.lipolytica, as affected by environmental conditions,need to be assessed. The availability of lipid source,

oxygen pressure, pH and water activity as well ascell morphology in Y. lipolytica are regarded as

Ž .determinants of lipase biosynthesis Ota et al., 1987 .Moreover, the selective enzymatic release and accu-mulation of individual free fatty acids as well assubsequent transformation of the released FFA, by b

or a oxidation, to other molecules such as lactonesdepend on fatty acid composition and position speci-ficity in the tryglycerides. Therefore, the in vitrocharacterisation of the specificity of the lipases andof the other relevant enzymes of the different strainsare insufficient to predict the lipid transformationpatterns in the real cheese system.

In order to assess the relative importance of theenvironmental conditions with respect to strain char-acteristics, the effect of factors such as milk fat,NaCl, and lactic acid content on lipase activity ex-pression was evaluated in model systems.

These variables were modulated according to aŽ .Central Composite Design CCD .

Ž .Table 3 shows the amounts as mgrl of the totalfree fatty acids released after 3, 7 and 22 days ofincubation in the various runs of the CCD. In themajority of the runs, the total FFA increased up to 22days. However, broad differences were observedwithin the runs, in fact: the FFA ranged between 71

Table 4Free fatty acids released after 7 days by Y. lipolytica Y3 in the different runs of the Central Composite Design

a Ž .Condition Free fatty acids mgrl

C C C C C C C C C C C6:0 8:0 10:0 12:0 14:0 15:0 16:0 16:1 18:0 18:1 18:2

b1 20.22 15.47 7.22 3.09 3.09 – 17.39 – 9.86 18.75 –2 26.85 20.12 10.98 6.38 6.38 13.09 43.00 12.89 19.23 23.57 –3 18.74 14.80 7.22 2.70 2.70 – 21.15 – 9.67 22.22 –4 18.74 14.58 6.82 2.31 2.31 – 13.99 – – – –5 18.50 19.89 9.20 5.99 41.73 14.74 11.30 14.89 7.91 14.51 –6 37.42 26.09 18.32 11.02 36.27 13.68 79.90 16.88 31.52 89.47 8.597 30.54 17.02 9.59 5.99 18.69 12.35 43.54 12.67 18.45 40.91 13.918 23.11 15.03 7.81 3.09 11.14 – 23.84 – 11.03 22.60 –9 25.38 15.91 7.42 2.70 11.69 – 26.16 – 12.21 27.23 –

10 23.66 15.91 7.02 2.12 9.48 – 22.22 – 12.01 22.41 –11 29.06 19.45 14.75 10.63 30.30 13.46 77.03 15.33 33.47 79.84 –12 11.52 16.57 8.80 5.79 10.77 – 22.05 – 10.25 22.03 –13 58.55 38.04 25.65 18.17 56.66 15.85 130.77 14.22 35.62 62.49 –14 19.97 16.13 7.81 6.18 21.83 13.09 25.99 12.22 13.38 25.30 –15 17.76 14.80 7.02 1.93 12.24 – 15.42 – 8.69 17.02 –16 23.16 17.02 7.61 2.89 12.43 – 13.99 – 8.11 13.93 –17 23.16 16.35 9.20 4.63 12.43 – 26.52 12.67 11.62 28.19 –

aSee Table 1 for conditions.b Not detectable.

( )M.E. Guerzoni et al.r International Journal of Food Microbiology 69 2001 79–8984

Table 5Best fit equations relative to free fatty acids released in the different runs of CCD by Y. lipolytica Y3 after 7 days

a b cFatty acid Equations F R SE2Ž . Ž . w x w x w x w xEq. 1 C s32.28 NaCl q73.01 lactic acid y61.90 NaCl lactic acid 67.65 0.967 7.576:0

2Ž . Ž . w x w x w xEq. 2 C s11.65 NaCl q5.40 milk fat q40.37 lactic acid 168.61 0.992 2.728:0w xw x w xw xq21.23 NaCl lactic acid y7.94 lactic acid milk fat

2 2Ž . Ž . w x w x w xEq. 3 C s5.01 milk fat q3.15 NaCl q36.55 lactic acid 46.62 0.975 2.9610:0w x w x w xw xy17.77 NaCl lactic acid y7.22 lactic acid milk fat

2 2Ž . Ž . w x w x w xEq. 4 C s1.49 milk fat q4.08 NaCl q28.73 lactic acid 29.14 0.948 2.5212:0w x w xy24.07 NaCl lactic acid

2Ž . Ž . w x w x w xEq. 5 C sy60.78 lactic acid q17.21 milk fat q98.16 lactic acid 25.156 0.941 8.9014:0w xw xy22.61 lactic acid milk fat

2Ž . Ž . w x w x w xEq. 6 C sy40.56 lactic acid q8.97 milk fat q7.10 NaCl 14.74 0.943 3.6615:02w x w xw x w xw xq61.28 lactic acid y20.02 NaCl lactic acid y4.99 NaCl milk fat

2 2Ž . Ž . w x w x w xEq. 7 C s9.45 milk fat q28.87 NaCl q204.96 lactic acid 16.87 0.915 21.2816:0w xw xy172.07 lactic acid NaCl

2 2Ž . Ž . w x w x w xEq. 8 C s2.42 milk fat q4.45 NaCl q40.91 lactic acid 9.80 0.866 5.5016:1w xw xy32.70 lactic acid NaCl

2 2Ž . Ž . w x w x w xEq. 9 C s10.60 milk fat q6.67 NaCl q51.05 lactic acid 26.40 0.957 6.0718:02w x w xw xy1.44 milk fat y NaCl milk fat

2Ž . Ž . w x w xEq. 10 C s166.17 lactic acid y3.90 milk fat 18.29 0.921 17.2118:1w xw x w xw xy133.31 NaCl lactic acid q19.31 milk fat NaCl

aF-value.bRegression coefficient.cStandard error.

and 592 mgrl. Some FFA, such as C or palmi-18:2Ž . Ž .toleic acid C and pentadecanoic acid C ,16:1 15:0

were specifically present only in certain runs, asshown in Table 4, relative to FFA released after 7days of incubation. Moreover, while the release orproduction of short chain fatty acids, such as butyric

Ž . Ž . Ž .acid C , caproic acid C , caprilic acid C ,4:0 6:0 8:0Ž . Ž .capric acid C or palmitic acid C , was10:0 16:0

precocious, the release of long chain fatty acids, suchŽ .as oleic acid C , continued to increase over18:1 D9

Ž .the 22 days data not shown .

The final growth of Y. lipolytica in the variousruns did not present significant differences.

The data relative to the individual fatty acids wereanalysed in order to obtain equations describing theeffects of the three variables considered on theirrelease.

In Table 5, the most significant equations relativeto the free fatty acids released after 7 days of incuba-tion are shown. The comparison of the equationsevidenced the general importance of the interactive

w xw xterm lactic acid NaCl , indicating that these inde-

Table 6Best fit equations relative to total free fatty acids released in the different runs of CCD by Y. lipolytica Y3 after 3, 7 and 22 days

a b cŽ .Incubation time days Equations F R SE2Ž . Ž . w x w x w xw xEq. 1 3 days s125.92 NaCl q274.42 lactic acid y255.17 NaCl lactic acid 95.07 0.976 22.33

2 2Ž . Ž . w x w x w xEq. 2 7 days s112.06 milk fat q77.52 NaCl q588.55 lactic acid 70.66 0.983 79.852w x w xw xy13.84 milk fat y534.307 NaCl lactic acid

2 2Ž . Ž . w x w x w xEq. 3 22 days s240.05 NaCl q166.50 milk fat q1001.55 lactic acid 30.91 0.963 97.69w xw x w xw xy122.72 milk fat NaCl y784.67 NaCl lactic acid

aF-values.bRegression coefficient.cStandard error.

( )M.E. Guerzoni et al.r International Journal of Food Microbiology 69 2001 79–89 85

pendent variables showed combined effects on therelease of all the free fatty acids detected. Thenegative sign of the coefficient of the interactive

w xw xterm lactic acid NaCl indicated a synergistic in-hibiting effect of low pH and high NaCl content onlipase activity. On the other hand, the lipid content,both as linear or quadratic term, seemed to have ageneral stimulating effect on the lipase activity. Thehighest level of NaCl stimulated the total fatty acidsrelease, as indicated by the positive sign of the

Ž .quadratic term in Eq. 3 of Table 6, relative to thebest fit equations of total FFA released after 3, 7 and22 days of incubation.

The analysis of the data after 3 days led tosignificant models just for the short chain fatty acids,such as caproic acid, caprylic and capric acid, aswell as for oleic and palmitic acids, whose release or

Ž .production were more precocious data not shown .Fig. 1a shows the effect of the interaction of NaCl

and lactic acid content on the total free fatty acidsreleased after 7 days. Both NaCl and lactic acid

Ž .concentrations had, in agreement with Eq. 2 ofTable 6, an individual stimulating effect on lipolyticactivity and, particularly, on total FFA release. Onthe contrary, when both factors were at their highestvalues, the total FFA release was inhibited.

w xw xFig. 1b shows the effect of milk fat NaCl inter-action on oleic acid release. The results indicatedthat high salt concentration inhibited oleic acid re-lease only when the lipid content ranged between1.0% and 2.5%. On the other hand, oleic acid releaseincreased with lipid content. When both factors wereat their highest levels, the oleic acid released attainedrelevant proportions.

The release of linoleic acid was less precociousand occurred just in combinations 1, 2 and 13 of theCCD, characterised by a weak lactic acid concentra-tion. Therefore, the data could not be modelled.

The stimulating effect of lipid content on oleicacid concentration is better shown by Fig. 1c, whichalso emphasised the inhibiting effect of high doses oflactic acid in combination with a low lipid content.

The dependance of palmitic, oleic and total FFAw xw xon the interaction lactic acid NaCl was similar, as

indicated by the comparison of Figs. 2a,b and 1a.w xw xFig. 3a,b,c shows the effect of NaCl lactic acid

interaction on caproic acid, caprylic acid and capricacid accumulation. In agreement with the equations

w xw xFig. 1. Effects of the interaction NaCl lactic acid on the totalŽ . w xw x Ž .free fatty acids a , and of the interactions milk fat NaCl b and

w xw x Ž .lactic acid milk fat c on oleic acid released by Y. lipolytica inthe different runs of the Central Composite Design after 7 days.

( )M.E. Guerzoni et al.r International Journal of Food Microbiology 69 2001 79–8986

w xw xFig. 2. Effects of the interaction NaCl lactic acid on palmiticŽ . Ž .acid a , and on oleic acid b released by Y. lipolytica in the

different runs of the Central Composite Design after 7 days.

included in Table 4, the accumulation after 7 days ofthese short chain fatty acids was strongly stimulatedby NaCl and lactic acid as individual factors. How-ever, when both factors were at their maximumlevels, the accumulation of such fatty acids wasinhibited by the synergistic effect of the low pH andhigh NaCl concentration on Y. lipolytica growth orits lipolytic activity.

w xw xFig. 3. Effects of the interactions NaCl lactic acid on caproicŽ . Ž . Ž .a , caprylic b and capric c acids released by Y. lipolytica inthe different runs of the Central Composite Design after 7 days.

( )M.E. Guerzoni et al.r International Journal of Food Microbiology 69 2001 79–89 87

The synergistic effect of lactic acid and NaClplayed an inhibiting role on the release of both shortchain fatty acids such as esanoic, octanoic and de-

Ž .canoic acids and long chain fatty acids Table 4 .The model systems, which had been subjected to

Ž .a gentle homogenisation 50 bar , presented a texturesimilar to that of the fermented milk, in which theincorporation of NaCl presumably accelerated theexchange of Ca2q with Naq. Therefore, it can beassumed that the three factors and their interactionsaffected the activity of Y. lipolytica both directly,due to A and pH decrease, and indirectly, byw

means of the triglicerydes exposition due to mi-crostructure changes.

The results of the CCD showed that, also when asingle strain was used, changes of the chemico-physical characteristics of the system led to markedvariation in the lipolytic activity, with consequentchanges in individual and in the relative proportionof fatty acids released.

In order to assess the relative importance of strainbiodiversity and environmental conditions on lipaseactivity and, particularly, FFA release, the overalldifferences concerning free fatty acids due to straincharacteristics or to chemico-physical conditionswere compared. Two distances matrices, concerning

Ž .the released individual fatty acids after 7 days ,

were calculated analysing the quantitative differ-ences both among the 12 strains inoculated in butterand among the 17 runs of the CCD inoculated with 1of the 12 strains.

Fig. 4 shows the dendrogram relative to the meanabsolute distances obtained analysing the proportionsof the individual fatty acids released from 12 strainsin butter, and indicates that three groups of strainscan be obtained. These groups did not correspond tothe environmental origin of the strains. A group ofseven strains showed similar lipolysis profiles and

Ž w x .distances not exceeding 30% as DrD =100 .max

However, this group showed, respectively, a distanceover 60% and near 100%, with respect to the othertwo groups. Fig. 5 represents the dendrogram rela-

Ž w x .tive to mean absolute distance as DrD =100max

obtained analysing the extent of the individual fattyacids released by the strain Y3 in the different runsof the CCD. The mean absolute distance among 11combinations of the CCD did not exceed 21%. Onthe other hand, 3 of the 17 free fatty acid profilesappeared very different, although the values of NaCl,lactic acid and fat content ranged between narrowintervals of values.

This preliminary approach suggested that in mostcases, under the experimental conditions considered,the modulation of the lactic acid, NaCl and lipid

ŽFig. 4. Dendrogram obtained comparing the fatty acids released after 7 days by 12 Y. lipolytica strains in butter for information on the.origin of strains see Materials and methods .

( )M.E. Guerzoni et al.r International Journal of Food Microbiology 69 2001 79–8988

Fig. 5. Dendrogram obtained comparing the fatty acids released by Y. lipolytica Y3 in the different runs of the Central Composite DesignŽ . Ž .CCD for information on conditions, see Table 1 .

content does not generate differences in the fattyacid release exceeding 20–21% when the same Y.lipolytica strain was used as starter. However, somecombinations of factors remarkably affected the li-pase expression or activity of one single strain, andgenerated differences in the fatty acid released higherthan those observed when different strains of thesame species are used.

4. Discussion

The lipolytic activity of Y. lipolytica has beenŽwidely studied Ota et al., 1987; Hadeball, 1991;

.Roostita and Fleet, 1996; Freitas et al., 1999 . Zviag-Ž .intzeva et al. 1980 reported a single cell wall

Ž .anchored lipase, whereas, Kalle et al. 1972 de-scribed two cell-bounded activities, a constitutiveglucose insensitive lipase and a second activity in-

Ž .duced by sorbitan monooleate. Ota et al. 1987described both an extracellular activity and a cellbound lipase. The extracellular lipase required oleicacid as stabiliserractivator, whereas the cell boundlipase did not and differed by several properties fromthe extracellular enzyme. According to HadeballŽ .1991 , strains of Y. lipolytica display a lipolyticactivity which acts prevalently on oleyl residues atpositions 1 and 3 of the glyceride. However, the invitro characterisation of extracellular lipase is not

sufficient to predict the lipolysis pattern under realconditions. In fact, in addition to their effects onyeasts growth, the chemico-physical and composi-tional factors have also an indirect effect on lipaseactivity due to their influence on the microstructureof the system. It is, in fact, well established that

Žlipase is active at the aqueous–lipid interface Dere-.wenda and Sharp, 1993; Verger, 1997 . The lipase

reaction rate varies directly with the surface area ofthe substrate available and the exposition of tryglyc-

Ž .erides to the enzyme activity Hadeball, 1991 .Moreover, the lipase access to substrate molecule isaffected by the tryglyceride tendency to form mi-celles which in turn depends also on chain length

Žand saturation level of fatty acids Kordel et al.,.1991; Kojima et al., 1994 .

The strains used in this work showed differentlipolytic patterns. The clustering of the free fattyacids profiles evidenced that the mean absolute dis-tances within majority of the strains of the samespecies did not exceeded 40%. Therefore, theorganoleptic properties associated to the metabolisa-tion or chemical and enzymatic conversion of suchFFA are in general expected to be similar. However,the lipolytic activity of some strains such as RO13and PO10 generates FFA profiles that differed fromthe majority of the strains considered. The impor-tance of the environmental conditions, and particu-

( )M.E. Guerzoni et al.r International Journal of Food Microbiology 69 2001 79–89 89

larly lipid, salt and lactic acid content, on the expres-sion of the lipase was emphasised by the comparisonof the data regarding the FFA released by an individ-ual strain under different conditions. In fact, specificcombinations of factors generated higher differencesin the FFA than those deriving from the natural bio-diversity within the same species.

It should be emphasised that yeasts, including Y.lipolytica, have a long history of widespread humanconsumption and there is currently no evidence tosuggest that their use in food transformation poseany danger. Y. lipolytica fulfills a number of specificcriteria to be regarded as a costarter for cheesemak-ing. However, in the strain selection phase, it isimportant to check whether some individual peculiar-ities can modify the ripening profile and generateunpleasant flavour.

References

Box, G.E.P., Hunter, W.G., Hunter, J.S., 1978. Statistics forExperimenters: An Introduction to Design Data Analysis andModel Buildings. Wiley, New York.

Cheynier, V., Feinberg, M., Chararas, C., Ducauze, C., 1983.Application of response surface methodology to the evaluationof bioconversion experimental conditions. Appl. Environ. Mi-crobiol. 45, 634–639.

Derewenda, Z.S., Sharp, A.M., 1993. News from the interface: themolecular structures of triacylglyceride lipases. TrendsBiochem. Sci. 18, 20–25.

Fleet, G., 1992. Spoilage yeasts. Crit. Rev. Biotechnol. 12, 1–44.Freitas, A.C., Pintado, A.E., Pintado, M.E., Malcata, F.X., 1999.

Role of dominant microflora Picante cheese on proteolysis andlipolysis. Int. Dairy J. 9, 593–603.

Guerzoni, M.E., Lanciotti, R., Marchetti, R., 1993. Survey of thephysiological properties of the most frequent yeasts associatedwith commercial chilled foods. Int. J. Food Microbiol. 17,329–341.

Guerzoni, M.E., Gobbetti, M., Lanciotti, R., Vannini, L., ChavesLopez, C., 1998. Yarrowia lipolytica as potential ripeningagent in milk products. In: Jakobsen, M., Narvhus, J., Viljoen,

Ž .B.C. Eds. , IDF Symp. Yeasts in Dairy Industry: Positive and

Negative Aspects, Chopenagen, Denmark, 1996. InternationalDairy Federation, Brussels, Belgium, pp. 23–33.

Hadeball, W., 1991. Production of lipase by Yarrowia lipolytica:Ž .I. Lipases from yeasts Review . Acta Biotechnol. 11, 159–

167.Jakobsen, M., Narvhus, J., 1996. Yeasts and their benefical and

negative effects on the quality of dairy products. Int. Dairy J.6, 755–768.

Jelen, H., Wasowicz, E., 1998. Volatile fungal metabolites andtheir relation to the spoilage and agricultural commodities.Food Rev. Int. 14, 391–426.

ŽJensen, R.B., 1983. Detection and determination of lipase acyl-.glycerol hydrolase activity from various sources. Lipids 18,

650–657.Kalle, G.P., Gadkari, S.V., Deshpande, Y., 1972. Inducibility of

lipase in Candida lipolytica. Indian J. Biochem. Biophys. 9,171–175.

Kojima, Y., Yokoe, M., Mase, T., 1994. Purification and charac-terization of an alkaline lipase from Pseudomonas fluorescensAK 102. Biosci., Biotechnol., Biochem. 58, 1564–1568.

Kordel, M., Hofmann, B., Schomburg, B., Schmid, R.D., 1991.Extracellular lipase of Pseudomonas sp. strain ATCC 21808:purifation, characterization, crystallization, and preliminaryX-ray diffraction data. J. Bacteriol. 173, 4836–4841.

Lencioni, L., Bertuccioli, M., Galgano, F., 1998. Metodo diestrazione per la determinazione degli acidi grassi liberi nelformaggio. Ind. Aliment. XXXVII, 171–175.

Ota, Y., Morimoto, Y., Sugiura, T., Minoda, Y., 1987. Soybeanfraction increasing the extracellular lipase production by Sac-charomycopsis lipolytica. Agric. Biol. Chem. 42, 1937–1938.

Peled, N., Krenz, M.C., 1981. A new assay of microbial lipaseswith emulsified trioleoyl glycerol. Ann. Biochem. 112, 219–222.

Roostita, R., Fleet, G.H., 1996. The occurrence and growth ofyeasts in Camembert and Blue-veined cheeses. Int. J. FoodMicrbiol. 28, 393–404.

Sinigaglia, M., Lanciotti, R., Guerzoni, M.E., 1994. Biochemicaland physiological chracteristics of Yarrowia lipolytica strainsin relation to isolation source. Can. J. Microbiol. 40, 54–59.

van den Tempel, T., Jakobsen, M., 2000. The technologicalcharacteristics of Debaryomyces hansenii and Yarrowialipolytica and their potential as starter cultures for productionof Danablu. Int. Dairy J. 10, 263–270.

Verger, R., 1997. AInterfacial activationB of lipases: facts andartifacts. TIBTECH 15, 32–38.

Zviagintzeva, I.S., Dmitriev, V.V., Ripan, E.L., Fichte, B.A.,1980. Localization of an extracellular lipase of the yeastCandida lipolytica. Microbiology 49, 417–420.