Degradation of pectic compounds during pasteurised vegetable juice spoilage by Chryseomonas luteola: a predictive microbiology approach

International Journal of Food Microbiology 42 (1998) 159–166

Degradation of pectic compounds during pasteurised vegetable juice spoilage by Chryseomonas luteola: a predictive microbiology approach Jeanne-Marie Membre´ *, Martine Kubaczka ´ ´ ´ et Technologie Alimentaires, Institut National de la Recherche Agronomique, 369 rue Jules Laboratoire de Genie des Procedes Guesde, BP 39, 59651 Villeneuve d’ Ascq Cedex, France Received 17 October 1997; received in revised form 23 February 1998; accepted 21 April 1998

Abstract Predictive modelling consists in describing effects of environmental factors on microbial growth parameters. With food spoilage bacteria, this approach must be extended to both growth and food damage characterisation. In order to study the incidence of storage temperature on vegetable damage, using predictive microbiology tools, kinetics of pectic compound degradation were studied. Chryseomonas luteola has been chosen because of its ability to grow on post-harvested vegetables. Experiments were performed at refrigerated temperatures (0–108C) with low initial bacterial charges (10 1 –10 3.5 cfu / ml). Microbial specific growth rate ( m ), stability phase before pectic degradation (Sp ) and alteration percentage (A p ) were chosen as reference parameters. Then, sub-optimal temperature effects on these three parameters were estimated using modified Ratkowsky model. Results obtained in synthetic medium were compared with data observed in endive juice to appreciate the alteration of vegetable during post-harvest storage.  1998 Elsevier Science B.V. Keywords: Predictive modelling; Environmental factors; Microbial growth parameters; Food spoilage bacteria; Postharvested vegetables

1. Introduction For 15 years, modelling bacterial growth responses has been widely studied in food microbiology (Buchanan and Phillips, 1990; Zwietering et ´ 1995). Even if al., 1992; Dengremont and Membre, some food microbiologists were sceptical about models and asked for more accurate or reliable *Corresponding author. Tel.: 13 20 435424; fax: 13 20 435426; e-mail: [email protected]

predictions, modelling has played an important role in food microbiology research (McMeekin and Ross, 1996; Roberts, 1997; Schaffner and Labuza, 1997), particularly if predicted data have been validated in food products before being used as a reference in food spoilage. On the other hand, predictive microbiology has been mainly focused on pathogenic bacteria. However, bacteria or moulds food spoilage causes serious deterioration by producing off-odours, off-flavours or slime (Huis in’t Veld, 1996). Spoilage is more rapid

0168-1605 / 98 / $19.00  1998 Elsevier Science B.V. All rights reserved. PII: S0168-1605( 98 )00087-7

160

J. Membre´ , M. Kubaczka / International Journal of Food Microbiology 42 (1998) 159 – 166

in proteinaceous foods such as meat, poultry, fish (Dalgaard, 1995) or dairy products, and is nonnegligible in vegetable foods (Brocklehurst and Lund, 1981; Carlin et al., 1989). Psychrotrophic bacteria producing pectinolytic enzymes were reported to be responsible of vegetable spoilage before or after harvest. In fact, pectic compounds fill the rigid structure of the plant cell wall and so play an important role in many quality aspects of fruits and vegetable products (Thakur et al., 1997). Some papers have dealt with ready-to-use vegetable spoilage (Marchetti et al., 1992; Membre´ and Burlot, 1994; Membre´ et al., 1995; Garcia-Gimeno and Zurera-Cosano, 1997; Ng and Schaffner, 1997; Piagentini et al., 1997) in which microbial ecology and growth parameters were studied, but modelling of vegetable damage has not been mentioned yet. In this paper, degradation of pectic compounds at low temperatures is studied. Chryseomonas luteola, a psychrotrophic bacterium, isolated from harvested vegetables, has been chosen as reference because of its high production of pectinolytic enzymes. A predictive microbiology approach extended to microbial spoilage effects, and particularly to damage parameters, is proposed.

the same temperatures. Moreover, at the validation step, C. luteola broths were sub-cultured at 1, 4, 7 and 278C and cultured at 78C in order to study the sub-culture temperature incidence on growth and damage parameters.

2.2. Vegetable juice assays Packs of commercial ready-to-use endives were purchased 1 day after packaging from retail outlets. Juices were prepared by homogenising the fresh salads with a domestic blender. Then, they were centrifuged, pasteurised (Membre´ et al., 1995) and thus inoculated with C. luteola.

2.3. Pectic compound degradation During growth, the synthetic culture medium and vegetable juice were sampled and centrifuged to remove bacterial cells. Pectic content in the supernatant was analysed using the m-hydroxydiphenyl photocolorimetric assay as described by Blumenkrantz and Asboe-Hansen (1973) with the D-galacturonic acid as reference.

2.4. Statistical analysis 2. Materials and methods

2.1. Organism and culture conditions The organism used in this study was Chryseomonas luteola isolated from endives in the Laboratoire de Microbiologie du Froid of the Institut Universitaire et Technologique of Evreux (France). The synthetic growth medium contained NH 4 Cl 1 g / l, MgSO 4 0.5 g / l supplemented with piperazineN,N9-bis-[2-ethanesulfonic acid] 50 mmol / l, polygalacturonic acid (ICN Biomedicals Inc., OH, USA) 4 g / l and yeast extract (Biokar Diagnostics, France) 2 g / l, pH 7. Experiments were carried out in 250-ml flasks containing 50 ml medium. Initial biomass concentration was in a range of 10 1 –3.10 3 cfu / ml. Biomass concentration was measured by counting cell colonies (cfu / ml) on nutrient agar plates, incubated for 48 h at 308C. To establish temperature effect on growth and damage parameters, cultures were incubated at 0, 4, 7 and 108C, with sub-culture in synthetic medium at

The non-linear regression was computed with the Splus software (AT&T Bell Laboratories, Murray Hill, NJ, USA), the parameters of the regression were estimated by the maximum likelihood method. The Gauss–Marquardt algorithm was used in the numerical process. The linear regression was computed with the general linear model procedure of SAS software (SAS Institute Inc., Cary, NC, USA).

3. Results and discussion

3.1. Primary model 3.1.1. Microbial growth Predictive microbiology approach consists in developing models to describe bacterial growth kinetics with three parameters: the maximal growth rate m (h 21 ), the lag phase period L (h) and the maximal biomass quantity Nmax (cfu / ml). The objective is essentially to estimate m and L as accurately as possible in order to evaluate shelf-life of foods. The

J. Membre´ , M. Kubaczka / International Journal of Food Microbiology 42 (1998) 159 – 166

microbial curve, log N versus time, obtained for each combination of environmental factor levels, is fitted by a primary model. Recently, Baranyi et al. (1993) proposed a sigmoidal curve based on the logistic curve (Tomassone et al., 1993) with an additive term, a (t), which allowed lag phase period to be introduced as an adaptive period for micro-organisms before growing (Eq. (1)). dN ] 5 a (t) ? m (N) ? N dt 0 # a(t) # 1 N m (N) 5 m ? 1 2 ]] Nmax

5

S

(1)

D

t is the time (h) and N the cell density (cfu / ml). With this approach, m (N) depends only upon cell density as substrate concentration effect is not included. This model is now commonly employed in predictive modelling approaches (Garcia-Gimeno and ZureraCosano, 1997; McClure et al., 1997). After integration, microbial curves, log N versus time, were fitted by Eq. (2). ln N 5 ln No 1 m ? A(t) 2 ln

5

F

E

exp( m ? A(t)) 2 1 1 1 ]]]]] exp(ln Nmax 2 ln No

G

161

Astd 5 t 1 (1 /m ) ? ln[exp(2m ? t) 1 exp(2m ? L) 2 exp(2m ? t 2 m ? L)].

(3)

Kinetic profiles were established for up to 1100 h (46 days). No growth was observed at 08C. All our experimental growth curves, in synthetic medium as well as in endive juice, were fitted by the Baranyi model and modelling results seemed very satisfactory (Fig. 1). The maximal cell density, Nmax , was estimated up to 10 9 cfu / ml, whatever culture conditions. Specific growth rate and lag phase varied with incubation temperature: in synthetic medium, maximal density was achieved after 8 days at 78C and 33 days at 48C. In each case, growth took place before pectic compound degradation. Preliminary works had shown that initial substrate concentration ranging from 0.8 to 5 g / l did not change the growth rate value (personal data). Thus, the Baranyi model, based on the logistic curve (Eq. (1)), could be convenient for experimental data. However, with spoiling bacteria, microbial growth parameters were not sufficient to characterise the effective damage on foods. Therefore, the first primary model was completed by an other one concerning pectic compound concentration decrease kinetics.

(2)

A(t) 5 a(t) ? dt

No is the initial cell density (cfu / ml). Kalathenos et al. (1995) proposed to write A(t) as follows:

3.1.2. Pectic substance degradation Since it produces several pectinolytic enzymes, Chryseomonas luteola damages vegetable cells. The pectic substance degradation was followed by

Fig. 1. Growth (s) and pectic acid degradation (h) kinetics of C. luteola in synthetic medium at 48C. Observed values (symbols) and modelling results (dotted lines).

162

J. Membre´ , M. Kubaczka / International Journal of Food Microbiology 42 (1998) 159 – 166

measuring the decrease of polygalacturonic acid concentration during the microbial growth of C. luteola. This microorganism consumed pectic acid and vegetable pectinolytic compounds after a large stability phase corresponding to the end of exponential phase of bacterial curves (Fig. 1). The kinetic profile could be represented schematically in three parts: a stability phase before any consumption, a rather linear degradation phase and a stationary phase with a residual level of pectic compounds in culture medium. Therefore, to characterise the damage, three parameters could be introduced: stability phase period, Sp , degradation rate, Dp , and alteration percentage, A p (Eq. (4)). P 5 Pinit P 5 Pinit 2 Dp ? (t 2 Sp ) P 5 Pr 5 Pinit 2 Dp ? (Sr 2 Sp )

5

if t , Sp if Sp , t , Sr if t . Sr

(4)

P is the pectic acid concentration (g / l), Pinit the initial pectic acid concentration (g / l), Pr the residual pectic acid concentration (g / l), Sr the time from which Pr is observed (h). Thus, A p is written as A p 5(Pinit 2Pr ) /Pinit ?100 or as A p 5Dp ?(Sr 2Sp ) / Pinit ?100.

3.2. Secondary model: temperature effect In order to establish temperature effects on bacterial spoilage, microbial growth rate parameter, m,

was chosen as the microbial proliferation reference while pectic compound degradation stability time, Sp , was selected as reference for food product shelflife. The lag period before bacterial growth, L, was not considered as a representative food spoilage parameter. Likewise, alteration percentage, A p , was preferred to degradation rate Dp , to characterise damage in synthetic medium compounds as well as in foods.

3.2.1. Microbial proliferation parameter In Fig. 2, maximal specific growth rate, m, is plotted as a function of temperature. The relative importance of experimental error can be due to the low reproducibility of microbial kinetics at suboptimal temperatures, particularly with low initial bacterial charges. For instance, at 78C, with an inoculum of 10 1 cfu / ml, m was close to 0.11 h 21 , while with an inoculum size of 6.10 2 cfu / ml, m was close to 0.07 h 21 (Fig. 2). Ratkowsky’s model, written as œ] m 5 b ? (T 2 T min ) with T min corresponding to the minimal temperature of growth, was chosen to fit the experimental data. This model is also known as ‘the square root model’, or ‘the Belehradek type model’ (Ross, 1993) and it generally employed to characterise low temperature effect on growth rate (Stannard et al., 1985; GarciaGimeno and Zurera-Cosano, 1997). Because of its biological interpretation, it has been preferred to the quadratic model even if this latter model has often

Fig. 2. Low-temperature effect on specific growth rate, m, during C. luteola proliferation in synthetic medium. Observed data (symbols) and Ratkowsky model (dotted line).

J. Membre´ , M. Kubaczka / International Journal of Food Microbiology 42 (1998) 159 – 166

163

Fig. 3. Low-temperature effect on stability period before pectic compound degradation, Sp , during C. luteola proliferation in synthetic medium. Observed data (symbols) and Ratkowsky model (dotted line).

been used in response surface methodology (Buchanan and Phillips, 1990). The minimal temperature parameter, T min , was estimated to 228C. This value was in agreement with results generally obtained with psychrotrophic bacteria (Stannard et al., 1985) and, in our study, no growth of C. luteola was observed at 08C for 2 months.

3.2.2. Vegetable damage parameters In Figs. 3 and 4, lag phase period, Sp , and alteration percentage, A p , are plotted as a function of temperature. Incubation of bacterial cultures to an abusive temperature such as 108C had dramatic consequences on vegetable damage. In fact, lag period fell from 25 days at 48C to 3 days at 108C. Moreover, alteration percentage increased with temperature: 70–87% with temperature of 4 and 108C, respectively. To predict the effect of temperature on degradation of pectic compounds, Ratkowsky’s

] equation was adapted and written as œSp 5 b ? (T 2 T min ) with T min value estimated to 1.38C. To fit degradation percentage, the linear equation A p 5 b(T2Cte) seemed sufficient. As suggested by Ratkowsky (1993) or by Baranyi et al. (1996), in a predictive microbiology approach, a model should be as parsimonious as possible. The square root model was employed in this study to determine the temperature effects on the two different parameters, m and Sp . Moreover, in each case, T min values were similar and in agreement with experimental data (no growth at 08C). This homogeneity and also the rather accurate fitting results led us to use this model as reference in predictive model validation.

3.3. Application To establish the validity of a predictive model in food products, it is now generally admitted that

Fig. 4. Low-temperature effect on alteration percentage of pectic compounds, A p , during C. luteola proliferation in synthetic medium. Observed data (symbols) and linear model (dotted line).

164

J. Membre´ , M. Kubaczka / International Journal of Food Microbiology 42 (1998) 159 – 166

Fig. 5. Growth (s) and pectic acid degradation (h) kinetics of C. luteola in vegetable juice at 78C. Observed values (symbols) and modelling results (dotted lines).

results obtained in synthetic medium should be validated in various conditions. First of all, challenge tests in foods are commonly performed. Moreover, previous growth conditions seemed to play an important role in microbial behaviour after inoculation in culture. In our case, sub-culture temperature incidence was tested.

3.3.1. Microbial spoilage of pasteurised vegetable juice In this paper, additional experiments were carried out in endive juice to appreciate the variability due to experiments with biological raw materials, and thus to give the modelling limits. First, the two primary models were applied to growth and pectic compound degradation kinetics during proliferation of C. luteola on endive juice (Fig. 5). The Baranyi model for growth as well as the linear model for pectic degradation gave accurately fitted values. The ex-

perimental error observed in vegetable juice (Fig. 5) was equivalent to the one obtained in synthetic medium (Fig. 1). Then, m, Sp and A p parameters estimated in vegetable juices, were compared to predicted values obtained with the second modelling step (Table 1). Specific growth rate, m, was rather closed to predicted values whatever the temperature conditions. Likewise, the alteration percentage, A p , was not so far from the predicted ones at 7 and 108C, although the initial pectin concentration varied from 0.8 to 1.3 g / l. In contrast, the variability of the stability period data, Sp , was greater with vegetable juices than with synthetic medium. This latter could be due to the heterogeneity of the pectin constituent in endives and also to the great complexity of enzymatic mechanisms involved in vegetable spoilage. To validate the predictive model, it seems interesting to check if experimental conditions carried out in

Table 1 Comparison between predicted values (Pred.) and additional values observed with endive juice trials (Obs.) Culture conditions

78C

108C

m (h 21 )

Sp (h)

A p (%, w / w)

Obs.

Pred.

Obs.

Pred.

Obs.

Pred.

0.074 0.101 0.091 0.096 0.074 0.100 0.107 1.128

0.083

241 355 163 181 241 355 280 285

146

83 86 85 85 83 86 77 88

75

0.126

63

83

J. Membre´ , M. Kubaczka / International Journal of Food Microbiology 42 (1998) 159 – 166

165

Fig. 6. Sub-culture temperature effects on specific growth rate, m, and stability period before pectic compound degradation, Sp , in synthetic medium.

the laboratory correspond well to food industry processes. Post-harvested vegetables are stored at low temperatures for a few months. Several psychrotrophic bacterial strains are able to grow and damage vegetables by producing mainly pectinolytic enzymes (such as pectin lyase, pectate lyase, pectate hydrolase) but also cellulosic enzymes. C. luteola is a psychrotrophic microorganism being able to achieve 10 9 cfu / ml after 1 month and to damage 60% pectic acid after 2 months at 48C. Therefore, it seems to be representative of pectinolytic bacteria spoiling vegetables in cool rooms. Moreover, trials performed in endive juices showed that results obtained in synthetic medium could be extended, although the variability of experimental data was greater. Further studies will focus on damage characterisation of various post-harvested vegetables such as carrots and celery by analysing enzymatic activities at low temperatures.

3.3.2. Incidence of sub-culture temperatures To investigate an eventual adaptation to temperature of psychrotrophic bacteria, such as C. luteola, during vegetable damage, growth rate, m and pectic degradation lag period, Sp , were studied under various temperature conditions. C. luteola was subcultured at 1 or 48C and cultured at 78C to observe the effect of a positive shift in temperature. In contrast, a negative shift was carried out with a sub-culture incubated at 278C (optimal growth

value). These two shifts were compared to C. luteola cultured at 78C with a sub-culture at 78C. Specific growth rate was not affected by temperature changes between sub-cultures and cultures (Fig. 6). Dufrenne et al. (1997) arrived at the same conclusions with several pathogenic bacteria. Concerning lag period, for some pathogenic bacteria, lag time before growing seems affected by temperature variations (Gay et al., 1996; Dufrenne et al., 1997), but no study has reported significant changes in stability period before substrate consumption due to sub-culture condition modifications. With C. luteola, pectic acid kinetic profiles at cool temperatures, such as 78C, were not influenced by previous temperatures (Fig. 6). As spoilage takes place after log phase of microbial growth, it would be assumed that only a drastic change between sub-culture and culture conditions could lead to significant modifications in the stability period before food damaging.

4. Conclusion Works presented in this paper showed that predictive microbiology strategy could be extended to vegetable spoiling bacteria. In this latter case, damage parameters have to be estimated and analysed as functions of storage conditions. With Chyseomonas luteola, a vegetable spoiling microorganism, the stability period before pectic acid consumption, Sp ,

166

J. Membre´ , M. Kubaczka / International Journal of Food Microbiology 42 (1998) 159 – 166

seemed the best indicator. In fact, Sp represented the shelf-life of pectic compounds of vegetables during cool room storage. Temperature influenced greatly Sp value and, with Ratkowsky’s equation, this incidence could be estimated. Until now predictive modelling tools have been developed mainly for pathogenic bacteria. Recently, Dalgaard (1995) proposed models for fish spoilage, and Cuppers et al. (1997) described food spoilage moulds; the study of vegetable spoilage by pectinolytic bacteria allows food spoilage modelling to be extended towards other food microbiology fields.

References Baranyi, J., Roberts, T.A., McClure, P., 1993. A non-autonomous differential equation to model bacterial growth. Food Microbiol. 10, 43–59. Baranyi, J., Ross, T., McMeekin, T.A., Roberts, T.A., 1996. Effects of parametrization on the performance of empirical models used in predictive microbiology. Food Microbiol. 13, 83–91. Blumenkrantz, N., Asboe-Hansen, G., 1973. A new method for quantitative determination of uronic acids. Anal. Biochem. 54, 484–489. Brocklehurst, T.F., Lund, B.M., 1981. Properties of pseudomonads causing spoilage of vegetables stored at low temperature. J. Appl. Bacteriol. 50, 259–266. Buchanan, R.L., Phillips, J.G., 1990. Response surface model for predicting the effects of temperature pH, sodium chloride content, sodium nitrite concentration and atmosphere on the growth of Listeria monocytogenes. J. Food Protect. 53, 370– 376. Carlin, F., Nguyen-The, C., Cudennec, P., Reich, M., 1989. Microbiological spoilage of fresh, ready-to-use grated carrots. Sci. Aliments 9, 371–386. Cuppers, H.G.A.M., Oomes, S., Brul, S., 1997. A model for the combined effects of temperature and salt concentration on growth rate of food spoilage molds. Appl. Environ. Microbiol. 63, 3764–3769. Dalgaard, P., 1995. Modelling of microbial activity and prediction of shelf life of packed fish. Int. J. Food Microbiol. 26, 305–318. ´ J.M., 1995. Statistical approach for Dengremont, E., Membre, comparison of the growth rates of five strains of Staphylococcus aureus. Appl. Environ. Microbiol. 61, 4389–4395. Dufrenne, J., Delfgou, E., Ritmeester, W., Notermans, S., 1997. The effect of previous growth conditions on the lag phase time of some foodborne pathogenic micro-organisms. Int. J. Food Microbiol. 34, 89–94. Garcia-Gimeno, R.M., Zurera-Cosano, G., 1997. Determination of ready-to-eat vegetable salad shelf-life. Int. J. Food Microbiol. 36, 31–38.

Gay, M., Cerf, O., Davey, K.R., 1996. Significance of preincubation temperature and inoculum concentration on subsequent growth of Listeria monocytogenes at 148C. J. Appl. Bacteriol. 81, 433–438. Huis in’t Veld, J.H.J., 1996. Microbial and biochemical spoilage of foods: an overview. Int. J. Food Microbiol. 33, 1–18. Kalathenos, P., Baranyi, J., Sutherland, J.P., Roberts, T.A., 1995. A response surface study on the role of some environmental factors affecting the growth of Saccharomyces cerevisiae. Int. J. Food Microbiol. 25, 63–74. Marchetti, R., Casadei, M.A., Guerzoni, M.E., 1992. Microbial population dynamics in ready-to-use vegetable salads. Ital. J. Food Sci. 2, 97–108. McClure, P., Beaumont, A.L., Sutherland, J.P., Roberts, T.A., 1997. Predicitve modelling of growth of Listeria monocytogenes. The effects on growth of NaCl, pH, storage temperature and NaNO 2 . Int. J. Food Microbiol. 34, 221–232. McMeekin, T.A., Ross, T., 1996. Modeling applications. J. Food Protect. Suppl., 37–42. ´ J.M., Burlot, P.M., 1994. Effects of temperature, pH and Membre, NaCl on growth and pectinolytic activity of Pseudomonas marginalis. Appl. Environ. Microbiol. 60, 2017–2022. ´ J.M., Goubet, D., Kubaczka, M., 1995. Influence of Membre, salad constituents on growth of Pseudomonas marginalis: a predictive microbiology approach. J. Appl. Bacteriol. 79, 603– 608. Ng, T.M., Schaffner, D.W., 1997. Mathematical models for the effects of pH, temperature, and sodium chloride on the growth of Bacillus stearothermophilus in salty carrots. Appl. Environ. Microbiol. 63, 1237–1243. Piagentini, A.M., Pirovani, M.E., Guemes, D.R., Di Pentima, J.H., Tessi, M.A., 1997. Survival and growth of Salmonella hadar on minimally processed cabbage as influenced by storage abuse conditions. J. Food Sci. 62, 616–618, 631. Ratkowsky, D.A., 1993. Principles of nonlinear regression modeling. J. Ind. Microbiol. 12, 195–199. Roberts, T.A., 1997. Microbial growth and survival: developments in predicitve modelling. Food Technol. 51, 88–90. Ross, T., 1993. Belehradek-type models. J. Ind. Microbiol. 12, 180–189. Schaffner, D.W., Labuza, T.P., 1997. Predictive microbiology: where are we and where are we going?. Food Technol. 51, 95–99. Stannard, C.J., Williams, A.P., Gibbs, P.A., 1985. Temperature / growth relationships for psychrotrophic food-spoilage bacteria. Food Microbiol. 2, 115–122. Thakur, B.R., Singh, R.K., Handa, A.K., 1997. Chemistry and uses of pectin—a review. Crit. Rev. Food Sci. Nutr. 37, 47–73. ´ Tomassone, R., Dervin, C., Masson, J.P., 1993. Modelisation de ` ´ ´ ` systemes dynamiques. Modelisation de phenomenes biologiques, 19. Masson, Paris, 440 pp. Zwietering, M.H., Wijtzes, T., De Wit, J.C., Van’t Riet, K., 1992. A decision support system for prediction of the microbial spoilage in foods. J. Food Protect. 55, 973–979.