كتب في علم الميكروبيولوجي

~V

1237_Cll.fm Page 313 Wednesday, November 12, 2003 1:13 PM

"I "I Predictive Mycology

Philippe Dantigny

CONTENTS

11.1 Introduction

11.2 Concerns

11.2.1 Mycotoxins Production

11.2.2 Economic Losses

11.3 Mold Specificities

11.4 Models

1 1 .4. 1 Primary Models

11.4.2 Secondary Models

11.5 Perspectives
References

11.1 INTRODUCTION

For over 20 years, predictive microbiology has focused on bacterial food-borne
pathogens, and some spoilage bacteria. Few studies have been concerned with
modeling fungal development. Predictive modeling is a versatile tool that should
not be limited to bacteria, but should be extended to molds. Mathematical modeling
of fungal growth was reviewed earlier (Gibson and Hocking, 1997), but at that time
very few models were available. The concerns were growth and toxin production,
but germination was not examined. On one hand, most of food mycologists are not
familiar with modeling techniques, and they tend to use existing models that were
developed for describing bacterial growth in foods. On the other hand, people
involved in modeling may not be aware of mold specificities. Predictive mycology
aims at developing specific tools for describing fungal development.

11.2 CONCERNS

The occurrence of food-borne fungi was described extensively by Northolt et al.
(1995). Food raw material and products can be contaminated with spores or conidia
and mycelium fragments from the environment. Under favorable conditions, fungal
growth occurs. A large number of metabolites are formed during the breakdown of
carbohydrates, some of which can accumulate under certain conditions. The main
concern is production of mycotoxins, which cannot simply be destroyed by heat.

2004 by Robin C. McKellar and Xuewen Lu

1237_Cll.fm Page 314 Wednesday, November 12, 2003 1:13 PM

11.2.1 Mycotoxins Production

It has been reported that 25% of agriculture products are contaminated with myc-
otoxins (Mannon and Johnson, 1985). Mycotoxin ingestion by humans, which occurs
mainly through plant-based foods and the residues and metabolites present in animal-
derived foods, can lead to deterioration of liver or kidney function (Sweeney and
Dobson, 1998) and therefore constitutes a risk for human health. The main genera
responsible for toxins production are Fusarium, Aspergillus, and Penicillium. While
Fusarium species are destructive plant pathogens producing mycotoxins before or
immediately post harvesting, Penicillium and Aspergillus species are more com-
monly found as contaminants of commodities and foods during drying and subse-
quent storage (Sweeney and Dobson, 1998). Plant contamination by molds such as
Fusarium cannot be avoided at the field level because it depends largely on climatic
conditions. Predictive mycology would be useful for making predictions on the
extent of contamination, growth, and toxin production by these pathogens; however,
there are no models currently available. In contrast, controlling the environmental
factors during storage of raw materials can prevent the development of Penicillium
and Aspergillus. The prevalence of one species as compared to the other one is
related to temperature, Penicillium being capable of developing at lower temperatures
than is Aspergillus.

1 1 .2.2 Economic Losses

Because of the appearance of visible hyphae and production of unpleasant odors,
fungal spoilage of food causes economic losses. For example, in the baking industry,
these losses vary between 1 and 3% of products, depending on season, type of
product, and method of processing (Malkki and Rauha, 1978). The most widespread
and probably most important molds in terms of biodeterioration of bakery products
are species of Eurotium, Aspergillus, and Penicillium (Abellana et al., 1997). But
there are many other species responsible for food spoilage. The reason why a
particular species dominates in a product is certainly correlated with the species
characteristics and the properties of the product (Northolt et al., 1995). Therefore,
predictive mycology can well be applied to control fungal development through
product formulation, food processing, type of packaging, and conditions of storage.

11.3 MOLD SPECIFICITIES

Fungal growth involves germination and hyphal extension, eventually forming myce-
lium. Spores are widely disseminated in the environment, and they are principally
responsible for spoilage. Under favorable conditions, spores will swell. Thereafter,
when the length of the germ tube is between one half and twice the spore diameter
(depending on the source), the spore is considered to have germinated. Germination
can be considered as the main step to be focused on, because a product is spoiled
as soon as visible hyphae can be observed. However, few studies have concerned
germination kinetics. This limitation can be explained in part by the difficulties of
acquiring sufficient, reproducible data. In fact, this kind of study requires microscopic

2004 by Robin C. McKellar and Xuewen Lu

1237_Cll.fm Page 315 Wednesday, November 12, 2003 1:13 PM

~V

observation for evaluating the length of the germ tube. Moreover, observations and
measurements should be carried out without opening the dishes (Magan and Lacey,
1984) and experimental devices should be developed for this purpose (Sautour et al.,
2001a, 2001c). In contrast, more work was dedicated to the measurement of hyphal
extension rate, which is usually reported as radial growth rate (mm d -1 )-

Because of their ability of dividing, bacteria form single cells and they can be easily
enumerated, especially in liquid broth. In such a case, and at high cellular densities,
bacterial growth can be estimated automatically, for example, by using the Bioscreen®
device, which is based on turbidity measurements. At lower cellular densities and in
solid media, colony-forming units per gram (CFU/g) or CFU/ml can be determined.

In contrast, molds form mycelium, and the weight, except at the early stage of
growth, does not increase exponentially (Koch, 1975). It is therefore useless to deter-
mine the weight of the mycelium for estimating a growth rate parameter. In addition,
it is impossible to split the mycelium into individual cells. However, the CFU method
can be applied to the enumeration of spores (Vindel0v and Arneborg, 2002).

Temperature (T) is the main factor for controlling bacterial growth, but the effect
of water activity (a w ) on mold growth is more important than T (Holmquist et al.,
1983). Oxygen is necessary for the growth of food spoilage fungi. Therefore, the
use of modified atmospheres to prevent fungal growth and mycotoxin production
has been evaluated to extend the shelf life of some kinds of food (El Halouat and
Debevere, 1997; Taniwaki et al., 2001).

11.4 MODELS

1 1 .4.1 Primary Models

Two aspects of fungal growth can be modeled using primary models: spore germi-
nation and radial growth of colonies. The germination of spores of Fusarium monil-
iforme as a function of time was first studied at different a w (Marin et al., 1996).
The percentage of germination vs. time was modeled with the modified Gompertz
equation (see Chapter 2) at different water activities (Figure 11.1). In contrast to the
case with bacteria where the initial bacterial load (iV ) is a critical parameter to be
estimated, the initial percentage of germination was always equal to 0. The asymp-
totic value where the percentage of germination becomes constant was 100 in most
cases. But under harsh environmental conditions some spores are unable to initiate
a germ tube, thus leading to a maximum percentage of germination less than 100
(Figure 11.1).

There are two different ways of looking at spore germination: (1) the percentage
of germination at a certain time, and (2) the time to obtain a certain germination
percentage, or germination time. In the present example, the percent germination
after 24 h does not discriminate water activity levels very well (vertical dotted line
in Figure 11.1). The response was 100% for a w in the range 0.94 to 0.98, and 0%
in the range 0.88 and 0.92. In contrast, germination time (defined here as half the
maximum percent germination) is clearly dependent upon a w (horizontal dotted line
in Figure 11.1). It should be noted, however, that an accurate determination of the
germination time requires modeling of the whole germination curve.

2004 by Robin C. McKellar and Xuewen Lu

~V

1237_Cll.fm Page 316 Wednesday, November 12, 2003 1:13 PM

60 90

Time(h)

150

FIGURE 1 1 .1 Effect of water activity and time on germination (%) of spores of Fusarium
moniliforme (isolate 25N) on MMEA (maize meal extract agar) at 25°C. Water activity levels
were 0.98 (■), 0.96 (A), 0.94 (•), 0.92 (T), 0.90 (♦), and 0.88 (+). (Redrawn from Marin,
S., Sanchis, V., Teixido, A., Saenz, R., Ramos, A.J., Vinas, I., Magan, N. 1996. Can. J. Microbiol.
42, 1045-1050.) The vertical dotted line indicates the percent germination at 24 h, and the
horizontal dotted line shows the time required for 50% germination at each water activity.

The germination time can also be considered as the probability of a single spore
germinating. Accordingly, the logistic function that is usually dedicated to probabi-
listic models:

P =

P

max

(1-e

<*-[w])

)

(11.1)

was used for describing the germination kinetics of Mucor racemosus (Dantigny et
al., 2002). The parameter P max was substituted with 100% because all spores were
capable of germinating. With the objective of designing a secondary model, the
parameters of the logistic function were expressed as a function of environmental
factors. The rate factor k was constant whatever the temperature, whereas x (time
where P = P m „/2) was more discriminative.

ma A '

Shortly after the completion of germination, the mycelium is visible to the naked
eye (when the colony diameter reaches approximately 3 mm). Therefore fungal
growth can be easily estimated from macroscopic measurements of the radius of the
colony. The primary model developed by Baranyi (see Chapter 2) has been adapted
to fit colony diameter growth curves of Penicillium rocqueforti (Valik et al., 1999),
Aspergillus flavus (Gibson et al., 1994), and Penicillium brevicomp actum (Membre
and Kubaczka, 2000). In our laboratory we have had considerable success with a
simple linear model with breakpoint:

r = [[-(t-X)

(11.2)

2004 by Robin C. McKellar and Xuewen Lu

1 237_C 1 1 .fm Page 3 1 7 Wednesday, November 1 2, 2003 1 : 1 3 PM

where r is the colony radius (mm), |i is the radial growth rate (mm d _1 ), and X is
the lag time (d). The linear section of the graph (with growth rate of |i) is extrap-
olated to a zero increase in diameter, and the intercept on the time axis is defined
as the lag prior to growth (k). In most cases, the fits are excellent, with the regression
coefficients being greater than 0.995. Therefore, under these conditions, it is
unlikely that other models such as those of Baranyi or Gompertz would demonstrate
superiority over the linear approach. In addition, the parameters |l and X, can be
obtained even when the petri dish is not entirely covered with mycelium. It should
also be mentioned that early measurements of diameter of the colony improve the
accuracy of the lag period because this parameter is obtained by extrapolation of
the straight line.

1 1 .4.2 Secondary Models

pH, which is usually associated with other environmental factors to prevent bacterial
growth, has no marked influence on mold germination or growth. Water activity has
a greater effect on mold development than does temperature, whereas an interactive
effect between T and a w is noticed. The effects of temperature and water activity on
growth rate of food spoilage molds were compared using normalized variables |i dim ,
T dim , and a w dim within Belehradek-type equations : |i dim = |T dim ] a and |i dim = [a w dim f
(Sautour et al., 2002). It can be observed that for a = 2, the equation is equivalent
to the square-root model that was originally described by Ratkowsky et al. (1982).
It was reported that the molds studied were characterized by a-values ranging
from 0.81 to 1.54 and p-values from 1.50 to 2.44. Because of the lack of specific
models for molds there is a tendency to apply models that have been developed for
bacteria. For example, the square-root model was used to describe the effect of T
on the growth of Rhizopus microsporus (Han and Nout, 2000). It is clear that the
effect of temperature on molds cannot be modeled by the square-root model because
of a-values close to 1. It has been demonstrated that the use of the square-root model
when a is less than 2 leads to underestimation of r min (Dantigny and Molin, 2000).
Similarly, some doubt can be raised with the use of the cardinal model with inflexion
(CMI) described by Rosso et al. (1993) for describing the effect of T on fungal
growth rate. For example, a T min value as low as -12°C has been reported for P.
rocqueforti using the CMI model (Cuppers et al., 1997). In contrast, the CMI model
was used satisfactorily for describing the effect of a w on fungal growth rate (Rosso
and Robinson, 2001; Sautour et al., 2001b) as suggested by the P-values close to 2.
It should also be noted that, in contrast to the square-root model proposed by Gibson
et al. (1994) to describe fungal growth, the CMI model allows an estimation of a w
min , which is not easily determinable because fungal growth can well occur after
several months of incubation.

11.5 PERSPECTIVES

For the objective of modeling fungal kinetics, the tools that were developed for
bacteria can be used, but mold specificities should be taken into account. As a primary
step in modeling fungal development, attention should be focused on spore

2004 by Robin C. McKellar and Xuewen Lu

1 237_C 1 1 .fm Page 3 1 8 Wednesday, November 1 2, 2003 1 : 1 3 PM

germination. Although there is no widely accepted definition of germination time,
this variable provides a pertinent insight into how fast spores are germinating. To
determine this variable accurately, spore germination kinetics should be monitored
using regular microscopic observations and by developing a specific experimental
setup. A significant breakthrough in predictive mycology would be the automation
of spore observations. Some attempts were made using image analysis (Paul et al.,
1993), but clumping of spores should be avoided. In addition, culture media should
be clear enough to allow microscopic observations. Therefore, any model describing
germination kinetics could be hardly validated on food products by this technique.

Fungal growth, which is usually reported as radial growth rate, can be easily
determined by macroscopic observations. In order to substitute a microscopic obser-
vation for a macroscopic one, a relationship between the lag for growth and the
germination time was established (Dantigny et al., 2002). However, it was shown
that the lag time is very much dependent on the number of spores inoculated at the
same spot (Sautour et al., in press). This could be explained, because a large inoculum
will form a visible colony more rapidly than a small one, thus decreasing the lag.
Some more studies should be conducted to determine the relationship between the
lag and the number of spores inoculated. It should also be verified that such a
relationship is independent from the environmental factors.

Secondary models concern mainly the influence of environmental factors on
fungal growth rate. At present, very few models aimed at assessing the influence of
these factors on spore germination have been elucidated. Unfortunately, existing
models are polynomial, and cannot be extrapolated to other molds. Secondary models
based on parameters with biological significance (e.g., cardinal values) to determine
the influence of environmental factors on spore germination and mycotoxins pro-
duction should be developed. Eventually, other parameters such as preservatives
should be included in the list of environmental factors.

REFERENCES

Abellana, M., Torres, L., Sanchis, V., Ramos, A.J. 1997. Caracterizacion de differentes pro-
ductos de bolleria industrial. II. Estudio de la microflora. Alimentaria 287, 51-56.

Cuppers, H.G.A.M., Oomes, S., Brul, S. 1997. A model combined effects of temperature and
salt concentration on growth rate of food spoilage molds. Appl. Environ. Microbiol.
63, 3764-3769.

Dantigny, P., Molin, P. 2000. Influence of the modelling approach on the estimation of the
minimum temperature for growth in Belehradek-type models. Food Microbiol. 17,
597-604.

Dantigny, P., Soares Mansur, C, Sautour, M., Tchobanov, I., Bensoussan, M. 2002. Relation-
ship between spore germination kinetics and lag time during growth of Mucor race-
moms. Lett. Appl. Microbiol. 35, 395-398.

El Halouat, A., Debevere, J.M. 1997. Effect of water activity, modified atmosphere packaging
and storage temperature on spore germination of moulds isolated from prunes. Int.
J. Food Microbiol. 35, 41-48.

Gibson, A.M., Baranyi, J., Pitt, J.I., Eyles, M.J., Roberts, T.A. 1994. Predicting fungal growth:
the effect of water activity on Aspergillus flavus and related species. Int. J. Food
Microbiol. 23,419-431.

2004 by Robin C. McKellar and Xuewen Lu

1 237_C 1 1 .fm Page 3 1 9 Wednesday, November 1 2, 2003 1 : 1 3 PM

Gibson, A.M., Hocking, A.D. 1997. Advances in the predictive modelling of fungal growth

in food. Trends Food Sci. Technol. 8, 353-358.
Han, B.-Z., Nout, M.J. 2000. Effects of temperature, water activity and gas atmosphere on

mycelial growth of tempe fungi Rhizopus microsporias var microsporias and

R. microsporias var oligosporus. World J. Microbiol. Biotechnol. 16, 853-858.
Holmquist, G.U., Walker, H.W., Stahr, H.M. 1983. Influence of temperature, pH, water activity

and antifungal agents on growth of Aspergillus flavus and A. parasiticus. J. Food Sci.

48, 778-782.
Koch A.L. 1975. The kinetics of mycelial growth. /. Gen. Microbiol. 89, 209-216.
Magan, N., Lacey, J. 1984. Effect of temperature and pH on water relations of field and

storage fungi. Trans. Br. My col. Soc. 82, 71-81.
Malkki, Y., Rauha, O. 1978. Mold inhibition by aerosols. Baker's Digest. 52, 47-50.
Mannon, J., Johnson, E. 1985. Fungi down on the farm. New Scientist. 195, 12-16.
Mann, S., Sanchis, V., Teixido, A., Saenz, R., Ramos, A. J., Vinas, I., Magan, N. 1996. Water

and temperature relations and microconidial germination of Fusarium moniliforme

and Fusarium pro life ratum from maize. Can. J. Microbiol. 42, 1045-1050.
Membre, J.-M., Kubaczka M., 2000. Predictive modelling approach applied to spoilage fungi:

growth of Penicillium brevicomp actum on solid media. Lett. Appl. Microbiol. 31,

247-250.
Northolt, M.D., Frisvad, J.C., Samson, R.A. 1995. Occurrence of food-borne fungi and factors

for growth. In Introduction to Food-Borne Fungi, 4th ed. Samson, R.A., Hoekstra,

E.S., Frisvad, J.C., Filtenborg, O. (Eds.), Centraal Bureau voor Schimmelcultures,

Delft, The Netherlands, pp. 243-250.
Paul, G.C., Kent, C.A., Thomas, C.R. 1993. Viability testing and characterization of germi-
nation of fungal spores by automatic image analysis. Biotechnol. Bioeng. 42, 1 1-23.
Ratkowsky, D.A., Olley, J., McMeekin, T.A., Ball, A. 1982. Relationship between temperature

and growth rate of bacterial cultures. J. Bacteriol. 149, 1-5.
Rosso, L., Lobry, J.R., Bajard, S., Flandrois, J.P. 1993. An unexpected correlation between

cardinal temperatures of microbial growth highlighted by a new model. J. Theor. Biol.

162, 447-463.
Rosso, L., Robinson, T.P. 2001. A cardinal model to describe the effect of water activity on

the growth of moulds. Int. J. Food Microbiol. 63, 265-273.
Sautour, M., Dantigny, P., Divies, C, Bensoussan, M. 2001a. Application of Doehlert design

to determine the combined effects of temperature, water activity and pH on conidial

germination of Penicillium chrysogenum. J. Appl. Microbiol. 91, 900-906.
Sautour, M., Dantigny, P., Divies, C., Bensoussan, M. 2001b. A temperature- type model for

describing the relationship between fungal growth and water activity. Int. J. Food

Microbiol. 67, 63-69.
Sautour, M., Dantigny, P., Guilhem, M.-C. Bensoussan, M. In press. Influence of inoculum

preparation on the growth of Penicillium chrysogenum. J. Appl. Microbiol.
Sautour, M., Rouget, A., Dantigny, P., Divies, C., Bensoussan, M. 2001c. Prediction of conidial

germination of Penicillium chrysogenum as influenced by temperature, water activity

and pH. Lett. Appl. Microbiol. 32, 131-134.
Sautour, M., Soares Mansur, C, Divies, C, Bensoussan, M., Dantigny, P. 2002. Comparison

between the effects of temperature and water activity on growth rate of food spoilage

molds. /. Ind. Microbiol. Biotechnol. 28, 311-315.
Sweeney, M.J., Dobson, A.D.W. 1998. Mycotoxin production by Aspergillus, Fusarium and

Penicillium species. Int. J. Food Microbiol. 43, 141-158.
Taniwaki, M.H., Hocking, A.D., Pitt, J.I., Fleet, G.H. 2001. Growth of fungi and mycotoxin

production on cheese under modified atmospheres. Int. J. Food. Microbiol. 68, 125-133.

2004 by Robin C. McKellar and Xuewen Lu

1237_Cll.fm Page 320 Wednesday, November 12, 2003 1:13 PM

Valik, L., Baranyi, J., Gorner, F. 1999. Predicting fungal growth: the effect of water activity on
Penicillium rocqueforti. Int. J. Food. Microbiol. 47, 141-146.

Vindel0v, J., Arneborg, N. 2002. Effects of temperature, water activity, and syrup film com-
position on the growth of Wallemia sebi: development and assessment of model
predicting growth lags in syrup agar and crystalline sugar. Appl. Environ. Microbiol.
68, 1652-1657.

2004 by Robin C. McKellar and Xuewen Lu