1090 Journal of Food Protection, Vol. 80, No. 7, 2017, Pages 1090–1098 doi:10.4315/0362-028X.JFP-16-408 Copyright Ó, International Association for Food Protection
Research Paper
Growth Modeling of Aspergillus niger Strains Isolated from Citrus Fruit as a Function of Temperature on a Synthetic Medium from Lime (Citrus latifolia T.) Pericarp 3 A. GSCHAEDLER,4 L. GARRIDO-SANCHEZ, 5 ´ ´ ´ 2 A. VILLARRUEL-LOPEZ, T. SANDOVAL-CONTRERAS,1 S. MARIN, AND F. ASCENCIO1* 1Centro
de Investigaciones Biol´ogicas del Noroeste, A.C. Av. Instituto Polit´ecnico Nacional 195, 23097 La Paz, Baja California Sur, M´exico; 2Ci´encia i Tecnologia Agr`aria i Aliment`aria, Departament de Tecnologia d’Aliments, Universitat de Lleida. Av. Rovira Roure 191, 25198 Lleida, Spain; 3Centro Universitario de Ciencias Exactas e Ingenier´ıas, Universidad de Guadalajara, Marcelino Garc´ıa Barraga´n 145, 44430, Guadalajara, Jalisco, M´exico; 4Centro de Investigaci´ on y Asistencia en Tecnolog´ıa y Dise˜no del Estado de Jalisco, A.C. Camino Arenero 1227, 45019 Zapopan, Jalisco, M´exico; and 5Instituto Tecnol´ ogico de Estudios Superiores de Occidente, A.C. Perif´erico Sur Manuel G´omez Mor´ın 8585, 45604 Tlaquepaque, Jalisco, M´exico MS 16-408: Received 26 September 2016/Accepted 11 February 2017/Published Online 2 June 2017
ABSTRACT Molds are responsible for postharvest spoilage of citrus fruits. The objective of this study was to evaluate the effect of temperature on growth rate and the time to visible growth of Aspergillus niger strains isolated from citrus fruits. The growth of these strains was studied on agar lime medium (AL) at different temperatures, and growth rate was estimated using the Baranyi and Roberts model (Int. J. Food Microbiol. 23:277–294, 1994). The Rosso et al. cardinal model with inflexion (L. Rosso, J. R. Lobry, S. Bajard, and J. P. Flandrois, J. Theor. Biol. 162:447–463, 1993) was used as a secondary model to describe the effect of temperature on growth rate and the lag phase. We hypothesized that the same model could be used to calculate the time for the mycelium to become visible (tv) by substituting the lag phase (1/k and 1/kopt) with the time to visible colony (1/tv-opt and 1/tv), respectively, in the Rosso et al. model. High variability was observed at suboptimal conditions. Extremes of temperature of growth for A. niger seem to have a normal variability. For the growth rate and time tv, the model was satisfactorily compared with results of previous studies. An external validation was performed in lime fruits; the bias and accuracy factors were 1.3 and 1.5, respectively, for growth rate and 0.24 and 3.72, respectively, for the appearance time. The discrepancy may be due to the influence of external factors. A. niger grows significantly more slowly on lime fruit than in culture medium, probably because the nutrients are more easily available in medium than in fruits, where the peel consistency may be a physical barrier. These findings will help researchers understand the postharvest behavior of mold on lime fruits, host-pathogen interactions, and environmental conditions infecting fruit and also help them develop guidelines for future work in the field of predictive mycology to improve models for control of postharvest fungi. Key words: Environmental conditions; Growth simulation; Phytopathogenic molds
Pathogenic fungi can attack citrus fruits at any postharvest stage. The majority of the microorganisms that have caused postharvest citrus fruit rot are opportunists invading tissues directly or through wounds either inflicted naturally or as the result of inadequate postharvest handling and processing for sale. After rot has developed, decay spreads by contact from one fruit to another (41). The subsequent production of spores on decaying fruits and their deposition on adjacent healthy fruits results in a latent infection (51). Latent contamination involves fungal spores on the host surface that germinate only when the host reaches maturity or senescence or is wounded by insects or other means (37). The most common causes of citrus fruit decay are blue and green molds (Penicillium italicum and Penicillium digitatum), which cause some of the most economically * Author for correspondence. Tel: (52) 612 123 8484; Fax: (52) 612 123 8525; E-mail:
[email protected].
important postharvest diseases of citrus worldwide (32). Several factors related to the pathogen, host fruit, weather, and postharvest conditions determine the incidence and severity of disease (18). Less common but serious spoilage of citrus can be produced by a variety of other fungi such as Aspergillus. Even though Aspergillus species are not considered a major cause of citrus disease, these fungi are responsible for several disorders in various fruits, plants, and plant products (19). Aspergillus niger is prevalent in warm climates, both in the field and in stored foods, and is by far the most common Aspergillus species responsible for postharvest decay of fresh fruit including apples, pears, peaches, citrus, grapes, figs, strawberries, mangoes, tomatoes, and melons and some vegetables, especially onions, garlic, and yams (30, 40, 41). A. niger and related species have been isolated from citrus rot in various studies (5, 31, 36, 40, 47, 52). The most common rot species are A. niger and Aspergillus flavus, followed by Aspergillus parasiticus,
J. Food Prot., Vol. 80, No. 7
Aspergillus ochraceus, Aspergillus carbonarius, and Aspergillus alliaceus. A. niger is capable of dominating other Aspergillus species in coculture (35), which is probably why A. niger is the most commonly isolated Aspergillus species in fresh fruits and vegetables. This fungus can contaminate agricultural products at different stages, including preharvest, harvest, processing, and handling. Like P. italicum and P. digitatum, Aspergillus is an opportunistic pathogen that penetrates fruits through wounds inflicted naturally or from inadequate postharvest handling (1). These molds have high enzymatic activity that produces severe damage to the pericarp and reduces shelf life (5, 19). In warm grapevine production areas (e.g., Spain, France, and Italy), rot caused by Aspergillus section Nigri is particularly severe. The fungus survives mainly on soil, and the conidia are disseminated by warm air currents (7). However, the most notable consequence of Aspergillus infection is mycotoxin contamination of foods and feeds (40). To improve the quality and safety of fresh food, techniques for predicting fungal growth are needed. Prediction of mold growth is necessary for quantitative risk assessments from the orchard to the table. For years predictive models have focused on bacteria (16, 23). Although predictive mycology techniques have been developed recently, few approaches for describing mold behavior in real products have been validated (3), and A. niger models are scarce. Some models have been developed for A. niger growth in vitro as a function of temperature and water activity (aw) to predict food shelf life (8, 11, 13, 25, 28, 39). The development of visible mycelium is one of the most significant quality problems in many products, with important economic implications for industry because a product will be rejected for sale when mycelium becomes visible (15, 25). Gougouli et al. (25) developed a predictive model for the effect of storage temperature and inoculum size on A. niger mycelial growth in the dairy industry, and Dagnas and Membr´e (12) developed a modelling approach to predict the effect of aw and temperature on Aspergillus candidus growth; both studies had the final aim of predicting mold appearance during product storage. The objective of the present study was to model the effect of temperature on growth rate (lmax) and time of mycelium appearance for five strains of A. niger isolated from Persian limes (Citrus latifolia T.) after culture on lime agar (LA).
MATERIALS AND METHODS Fungal strains. Five isolates of A. niger were included in this research: A. niger-02, A. niger-24, A. niger-33, A. niger-36, and A. niger-37. All are native molds previously isolated from Persian limes (Citrus latifolia T.) in a citrus orchard in Baja California Sur, Mexico. Characterization and identification of all A. niger isolates was carried out at the Laboratory of Microbial Pathogenesis (Northwest Biological Research Center, La Paz, Baja California Sur, Mexico). The mold DNA was extracted by alkaline lysis and enzymatic digestion according to the protocol of Sambrook and Russell (45) with a modified buffer and enzymes (crushing buffer, chitinase, and proteinase K). PCRs were performed to amplify the rDNA internal transcribed spacer (ITS) regions (ITS1 to 5.8S to ITS2) using the universal primers ITS1 and ITS4 (53). The ITS of
MODELING PHYTOPATHOGENIC MOLDS
1091
nuclear rDNA is accepted as the official DNA barcode for fungi (48). The PCR products were sent to Genewiz, Inc. (Cambridge, MA) for sequencing, and the results were analyzed and the species identified using the Basic Local Alignment Search Tool from the National Center of Biotechnological Information (Bethesda, MD). The A. niger aggregate was the mold with the highest relative frequency of isolation (47). A. niger-24 was pathogenic for Persian lime in previous studies (2). Preparation of inoculum. The isolates were subcultured on potato dextrose agar and incubated at 258C for 7 days to obtain heavily sporulated cultures. Spores were collected by flooding the surface with an aqueous solution of 0.05% Tween 80. Mycelium filaments were removed from the suspension by filtering with a sterile medical tissue. Spores were counted using a Neubauer chamber and adjusted to 107 spores per mL in a 0.05% Tween solution. The spore suspension was further diluted to reach a final level of 5 3 103 spores per mL. Medium. The medium used in this study was LA containing 57.5 g of finely ground Persian lime pericarp and 15 g of agar per 1,000 mL at pH 5.5 6 0.2. The medium was autoclaved and poured into 50-mm sterile petri dishes. The aw (AquaLab Pre Water Activity Meter, Decagon Devices, Pullman, WA) did not differ significantly from that at the beginning of the experiment (aw ¼ 0.968; SD ¼ 0.03). Inoculation and incubation. The LA plates were inoculated centrally with 10 lL of suspension (approximately 50 spores) in three replicates per temperature. This quantity was used for the trial inoculum based on results of previous studies in which the mycobiota of the surface of citrus fruits from different orchards and packing facilities was quantified. In those studies, molds were reported as 2.59 6 0.46 log CFU/cm2, or approximately 40 spores per 0.1 cm2 (possible size of a wound on the surface of a fruit) (47). This initial level is similar to that reported by Huchet et al. (29), who used 5 3 103 spores per mL (approximately 50 spores per inoculum) as their initial inoculum, representative of the amount of mold that can be found naturally in fruits. Other researchers have used inocula of 105 to 108 spores per mL to model the rate or time in which the colony is visible as a function of temperature and/or aw (3, 8, 12, 21, 33, 38). After inoculation, petri dishes were sealed in polyethylene bags to prevent water loss. Bags were incubated inside plastic boxes containing a glycerol solution (10%) to minimize water transfer to or from the medium and to maintain a constant aw (24, 38). The temperatures of incubation were 8, 15, 20, 25, 30, 35, 37, and 408C. The diameter of the inoculated spore solution drop was about 0.6 mm. Growth assessment. After inoculation, the plates were examined every day for visible growth. As soon as growth started, colony diameters were measured with a ruler and a binocular magnifier. Fungal growth was observed daily for 25 days. Model development. A typical two-step modeling approach, including primary and secondary modeling, was used to quantify the effect of temperature on the kinetic parameters of all A. niger strains. Primary model. Initially, growth rate estimates of fungi were obtained by plotting colony radii against time. For each treatment, a nonlinear regression was applied to estimate the
1092
SANDOVAL-CONTRERAS ET AL.
J. Food Prot., Vol. 80, No. 7
maximum growth rate (lmax, mm/day), latency before growth (k), and maximum colony radius (R, mm) by fitting the experimental data to the primary model of Baranyi and Roberts (6) (equations 1 and 2): expðlmax AÞ 1 R ¼ R0 þ lmax A ln 1 þ ð1Þ expðRmax R0 Þ A¼ tþ
1
lmax
3 ln½expðlmax tÞ þ expðlmax kÞ expðlmax t lmax kÞ
ð2Þ
where R0 is the colony radius at time 0, Rmax is the maximum colony radius in petri dishes, A is an integral variable running from 0 to t as a function of the curvature of the plot, k (days) is the lag time, and t (days) is the time. The curve fitting procedure used Marquardt’s algorithm in Centurion XV.II (Statgraphics, Warrenton, VA) with 95% confidence. Secondary model. To describe the effect of temperature on fungal growth, cardinal temperature (Tmin, Tmax, and Topt) values were estimated using the cardinal model with inflexion (CMI) developed by Rosso et al. (43) (equation 3): h i lmax ¼ lopt ðT Tmax ÞðT Tmin Þ2 n 4 Topt Tmin 3 Topt Tmin T Topt Topt Tmax o ð3Þ 3 Topt þ Tmin 2T where lopt is the growth rate at optimal temperature, and T is temperature (8C). Marquardt’s algorithm was used for model fitting with Centurion XV.II with 95% confidence. The goodness of fit was evaluated graphically and by alculating the coefficient of determination R2 and the root mean square errors (RMSE). No correlation was found between lmax mean and variance, indicating homoscedasticity of data; thus, no transformation of data was required (12). The minimal lag time (1/kopt) may be modeled using the Rosso et al. CMI (equation 4) (27): h i 1=k ¼ 1=kopt ðT Tmax ÞðT Tmin Þ2 n 4 Topt Tmin 3 Topt Tmin T Topt o ð4Þ Topt Tmax Topt þ Tmin 2T The same model could be used for tv (the time for the mycelium to be unambiguously visible on petri dishes corresponding to a diameter of 2 mm (26, 29) by substituting 1/k from equation 4 with 1/tv and 1/kopt with 1/tvopt (equation 5): h i 1=tv ¼ 1=tvopt ðT Tmax ÞðT Tmin Þ2 n 4 Topt Tmin 3 Topt Tmin T Topt o ð5Þ Topt Tmax Topt þ Tmin 2T
Time tv was calculated using the Solver function of Excel (Microsoft, Redmond, WA) on the Baranyi and Roberts model. In this case, correlations were found between the 1/tv mean and variance, so a root-square transformation was used for 1/tv to homogenize its variance (14) (equation 6):
h i ð1=tv Þ0:5 ¼ 1=tvopt ðT Tmax ÞðT Tmin Þ2 n 4 Topt Tmin 3 Topt Tmin T Topt Topt Tmax o 0:5 3 Topt þ Tmin 2T
ð6Þ
For bacteria, the square-root or logarithm transformation is used for modelling the effect of environmental factors and is always related to a variance analysis. For molds, Dantigny and Bensoussan (14) suggested the use of the square-root transformation to stabilize the variance of the data because the resulting model will imply lower accuracy. Time to obtain visible mycelium on limes. To validate the predictions of the models, an experiment was performed on Persian limes to evaluate the behavior of mold on the surface of actual fruit. Fruits were disinfected with 1% (v/v) sodium hypochlorite for 2 min and then rinsed with sterile water. The fruits were inoculated through a wound, because the main pathogens for Persian lime penetrate through wounds. A wound (3 mm wide by 1 mm deep) was made on each fruit by puncturing with a sterile scalpel. Wounds were inoculated with 10 lL of an A. niger-24 spore suspension at 5 3 103 spores per mL. This native mold was found as a pathogen on Persian lime (2). Once inoculated, 10 limes per trial were incubated in a chamber with a saturated atmosphere (relative humidity, 85 to 90%) at 25, 30, 35, and 378C (3). When mycelium was visible, the time was recorded to compare it with the tv obtained from the model, and the diameter of the colony was measured daily for 10 days. The radius of the colony was plotted against time, and the same models were fitted to calculate the growth rate.
RESULTS The growth data modeled in our study consisted of the growth curves in triplicate of five strains of A. niger aggregate at eight temperatures: 8, 15, 20, 25, 30, 35, 37, and 408C. No growth in LA was observed at 8, 15, and 408C for any strain. Growth of A. niger isolates followed in general a biphasic Baranyi and Roberts model. Growth rate and lag phase were estimated through this model, resulting in 15 graphs for each temperature. The curves obtained were similar in shape. The goodness of fit was satisfactory for all of them (R2 . 95; data not show). All sets of data were examined for variance homogeneity. Significant differences were observed among growth rates (mean 6 standard error) of all strains at the extreme temperatures: 1.7 6 0.3 mm/day at 208C, 4.6 6 1.3 mm/day at 358C, and 2.5 6 1.2 mm/day at 378C. At temperatures of 25 and 308C (near the optimal conditions of growth), there were no significant differences for all strains: 3.4 6 0.9 mm/day at 258C and 4.4 6 1.2 mm/ day at 308C. These results indicate that temperatures near to the optimum, between 25 to 308C, did not result in variable
J. Food Prot., Vol. 80, No. 7
MODELING PHYTOPATHOGENIC MOLDS
1093
FIGURE 1. Radial growth rate (lmax, mm/day) versus temperature (T, 8 C) for five strains of Aspergillus niger isolated from citrus fruits. Points are observed data, and lines indicate the fit of the data to the cardinal model with inflexion. (A) A. niger-24; (B) A. niger-02; (C) A. niger-33; (D) A. niger-36; (E) A. niger-37.
growth, whereas extreme temperatures (20, 35, and 378C) did. Secondary models were then fitted using data from each strain. Using the secondary cardinal model (equation 3), the effect of temperature on lmax was determined, and the cardinal values were estimated. The fitted curves showing the influence of temperature on lmax are presented in Figure 1. Coefficient estimations of equation 3 are presented in Table 1 for all strains. The lopt calculated for the each strain was 4.5 to 5.5 mm/day with a coefficient of variation (CV) of 6.7%. The theoretical Tmax was about 408C and the theoretical Topt was 30 to 338C; both had a CV of ,5%. The highest level of variation was for the theoretical Tmin, with a CV of 16.85% and values of 10 to 158C. All parameters obtained are in the 95% confidence interval. The inverse of the time for each strain to develop visible colonies (1/tv) was also modeled. Direct fitting of function was unlikely to be satisfactory, so we transformed the 1/tv
measure by introducing a square-root transformation (21) (equation 6), allowing stabilization of the variance (13, 22). The fitted models are presented in Figure 2. The estimated coefficients are shown in Table 2. Tmin, Tmax, and Topt had a CV of ,5%, and their estimated values paralleled those from the previous model. The values estimated for tv at optimal conditions was 0.56 to 0.93 days, with a CV of 22.54%, the highest variation. The majority of strains (except A. niger-24 and A. niger-36) were in the 95% confidence interval. The time tv obtained at extreme temperatures was higher (1.26 6 0.2 days at 208C and 1.7 6 0.4 days at 378C) than the tv obtained near the optimal conditions (0.93 6 0.3 days at 308C). The shortest tv in LA was for cultures incubated at 358C (0.66 6 0.3 days). pffiffiffiffiffiffiffiffi Correlations were analyzed between lopt and 1=tv for all strains; Pearson correlation coefficients were negative for all strains but only significant (P , 0.05) for A. niger-24 (0.63), A. niger-33 (0.64), and A. niger-37 (0.52).
TABLE 1. Estimated coefficients for the Rosso et al. model (43) fitted to growth rates for five isolates of Aspergillus niger on lime agara A. niger isolate
24 02 33 36 37 a
Tmin (8C)
10.14 14.26 14.46 14.53 11.19
6 6 6 6 6
2.0 1.26 1.55 0.78 1.92
Tmax (8C)
40.00 40.00 39.87 39.79 40.05
6 6 6 6 6
0.18 0.15 0.24 0.39 0.26
Topt (8C)
32.68 31.12 30.36 30.42 31.66
6 6 6 6 6
0.55 0.30 0.46 0.78 0.55
Values are means 6 standard errors. Initial inoculum was approximately 50 spores.
lopt (mm/day)
R2
RMSE
6 6 6 6 6
94.82 97.76 95.18 87.35 92.17
0.30 0.11 0.19 0.60 0.42
5.07 5.43 4.73 4.88 5.57
0.21 0.15 0.17 0.31 0.25
1094
SANDOVAL-CONTRERAS ET AL.
J. Food Prot., Vol. 80, No. 7
FIGURE 2. Root square of 1/tv versus temperature (T), indicating the time for each strain to become visible (colony diameter of 2 mm is considered visible to the naked eye) for five strains of Aspergillus niger isolated from citrus fruits. Points are observed data, and lines indicate the fit of the data to the cardinal model with inflexion. (A) A. niger-24; (B) A. niger-02; (C) A. niger-33; (D) A. niger-36; (E) A. niger-37.
The time needed to obtain visible mycelium on limes was determined by infecting lime fruits with A. niger-24 after confirming that the pH of the pericarp was 5.5 6 0.5 and aw was 0.974 6 0.01. With a low inoculum level (5 3 103 spores per mL), the percentage of lime fruits infected was low with high variation (28.6% 6 14.3%) at all temperatures tested. Growth rate was calculated with the same models (Table 3). No significant differences were observed at 30 and 358C with high variability (Fig. 3A). To evaluate the performance of the model, predicted and observed maximum growth rates on limes were compared. The bias factor was 1.3, and the accuracy factor was 1.5, indicating conservative predictions. The tv was obtained directly on limes. High variation was observed at the extreme temperatures 25 and 378C (Fig. 3B), with no significant differences. The performance of the model was calculated. The resulting bias factor was 0.24, and the accuracy factor was 3.7, indicating discrepancy, which
means that the model is not fail-safe or that the predicted colony sizes are larger than those actually observed on limes. For a perfect fit, both the bias and accuracy factors would be 1.0. A. niger grew significantly more slowly on lime fruit than in LA, especially at low temperatures. A positive correlation was observed between log(tv predicted/ tv observed) and temperature; at high temperatures, the values tended to zero. For a perfect fit, log(tv predicted/tv observed) ¼ 0.
DISCUSSION Lime fruits contain mycoflora, which are normally present on the fruit surface during harvest and postharvest processing (41). A. niger, despite not being one of the specific fungi of citrus fruit, is one of the most commonly isolated fungi in studies of surface microorganisms because of its wide distribution in nature (30, 40, 41, 52). In a previous study in our laboratory, A. niger aggregate was the mold with the highest relative frequency of isolation from
TABLE 2. Estimated coefficients for the Rosso et al. model (43) fitted to A. niger isolate
24 02 33 36 37 a
Tmin (8C)
14.74 14.85 14.86 11.48 14.80
6 6 6 6 6
1.05 0.58 0.80 0.55 0.55
Tmax (8C)
40.08 40.15 40.15 40.24 40.12
6 6 6 6 6
0.24 0.23 0.32 0.39 0.19
pffiffiffiffiffiffiffiffiffiffiffiffi ð1=tv Þ for five isolates of Aspergillus niger on lime agara
Topt (8C)
31.45 29.88 29.38 28.33 30.10
6 6 6 6 6
1.4 0.86 1.16 0.78 0.80
tvopt (days)
R2
RMSE
6 6 6 6 6
84.24 88.50 80.4 84.66 90.1
0.07 0.03 0.06 0.02 0.03
0.56 0.70 0.62 0.93 0.57
0.08 0.07 0.08 0.07 0.05
Values are means 6 standard errors. The diameter for the colonies to become visible was fixed at 2 mm.
J. Food Prot., Vol. 80, No. 7
MODELING PHYTOPATHOGENIC MOLDS
TABLE 3. Estimated coefficients for Baranyi and Roberts model (6) fitted to lmax for A. niger-24 on lime fruit a Temp (8C)
25 30 35 37 a
b
lmax (mm/day)b
R2
6 6 6 6
97.8 95.1 92.4 95.3
3.3 3.6 4.1 1.9
0.5 1.5 1.7 1.8
tv (days)
4.4 2.9 3.0 4.3
6 6 6 6
1.4 0.9 1.5 2.1
Values are means 6 standard errors. Time (tv) was measured directly on Persian limes. For all lmax, P , 0.05.
the surface of citrus fruits in some regions of Mexico. In published reports, some A. niger isolates have been found to be pathogens for lime fruit and produce aflatoxins under a wide range of pH and temperature conditions. A. niger-24 on the Persian lime can also synthesize aflatoxins, which diffuse into the fruit (unpublished data). Most pathogens causing postharvest citrus fruit rot are opportunists invading tissues normally through wounds made during processing (41). Two of the most important environmental parameters that determine the ability of molds to grow on foods are aw and temperature. Fruits in general have an aw close to 0.99. From a food quality and safety point of view, most of the work has been directed toward the development of models for fungal growth as a function of temperature and aw. Because lime aw is near the optimal level for fungal growth, the most important parameter to determine the ability of mold to grow on lime fruit is temperature. In our study, the aw effect was not taken into account because the high quality lime fruit used had a stable aw during the storage period of 6 to 8 weeks (49). In our research, data obtained may represent most of the situations that can be encountered when dealing with fungal growth during handling of citrus. No growth observed at 15 and 408C does not mean that A. niger cannot develop at these temperatures, but it may grow after a longer time, beyond the period covered by the experiment. Higher and lower growth temperatures (Tmax and Tmin) have been reported (41). In this study, linear growth occurred after a lag time and a short period of nonlinear growth; thus, the sigmoidal Baranyi and Roberts model had the best fit (34). The effect of temperature on A. niger growth was compared with that reported previously. For many years, a group of fungi has been morphologically identified as Aspergillus section Nigri. Thus, older studies referring to A. niger may involve species other than A. niger in section Nigri. More recently, a number of species in this section have been
1095
identified through molecular techniques. The strains used in the present study have been identified based on the ITS region of the nuclear rDNA, now accepted as the official DNA barcode for fungi (46). The ITS region is considered superior to others markers for fungus species discrimination (48). Pitt and Hocking (41) reported very similar values for Tmin of 8 to 128C, Tmax of 45 to 478C, and Topt of 35 to 378C. Gougouli and Koutsoumanis (26) modeled the growth of A. niger in malt extract agar at aw 0.997 using the Rosso et al. cardinal model with similar results for Tmin (10.138C), Imax (43.138C), and Topt (31.448C), but the lopt value was 0.840 mm/h (radial growth rate of 0.42 mm/h), whereas in the present study the lopt was 4.5 to 6.0 mm/day (radial growth rate of 0.19 to 0.25 mm/h). The differences were most likely due to the difference in culture medium because molds behave differently depending on matrix or environmental conditions (11, 44), in particular pH level, which in our study was 5.5 to simulate the pH of lime pericarp. Gougouli et al. (25), working with yoghurt (pH 4.2), found that A. niger had faster mycelium growth and a shorter lag time with Tmin ¼ 9.68C, Tmax ¼ 46.98C, and Topt ¼ 34.38C, higher temperatures than those in our study except for Tmin, which was lower. Gougouli et al. found a higher lopt (1.32 mm/h, as a function of diameter) than that in our study. Growth differences may be owing to where the isolates were originally found. Garc´ıa-Cela et al. (22) studied the ecophysiological characteristics of A. niger isolates from Spanish vineyards and compared these characteristics with those from isolates from other countries. Strains of A. niger from Morocco grew faster at aw 0.95 and 258C, whereas isolates from Europe grew faster at aw 0.95 and 25 to 308C. The maximum growth rates in our research were similar to those observed by Bell´ı et al. (8), with lmax ¼ 6.14 to 6.29 mm/day for 30 to 378C with an aw similar to that of lime fruits but pH (4.5) different from that of LA. Parra and Magan (39) reported a radial growth rate of 4.30 to 5.04 mm/day at 308C and an aw of 0.99, parameters near optimum values obtained in our research. The tendency of the growth rate to increase as a function of temperature close to 308C and decrease under marginal conditions is similar to that found in previously published investigations. The lowest temperature reported for A. niger growth was 6 to 88C, and the maximum was 45 to 478C (41). Most published studies on predictive mycology generally have focused on growth rate, a very important parameter for predicting the extent of fungal contamination; however, less attention has been paid to modeling of the lag phase. In the particular case of molds, the prediction of the time to FIGURE 3. Plots of the time to visible mycelium (tv) observed directly on Persian limes versus temperature for A. niger-24. Circles are observed data, and dotted lines are the tendency of the model. (A) Plot of radial growth rate (l) versus temperature. (B) Plot of tv versus temperature.
1096
SANDOVAL-CONTRERAS ET AL.
visible molding is of particular interest (10, 20) because spore germination and microscopic growth occur during this time (10). For most fresh fruit, the product will be rejected as soon as the mycelium becomes visible, which occurs in a very short time after the lag phase. Thus, it is more useful to model time to visible growth (tv), i.e., when the colony becomes visible to the naked eye, instead of the lag phase (3, 10, 29, 50). In this context, parameters such as time for a visible colony of 2 to 3 mm may be more interesting (34). Time to visible growth is easier to estimate than germination time because microscopic observations are not required. Nevertheless, when the inoculum level is low (e.g., one spore), the ratio of time to visible growth over germination time is greater and depends on the aw (10). In our study, significant p negative ffiffiffiffiffiffiffiffi correlation was found between lopt and estimated 1=tv in the majority of the molds analyzed, indicating that when tv gets shorter, the resulting colony grows faster. In our case, the fungal growth data had a higher variance at high growth rates under optimal temperature conditions and a lower variance at low growth rates; thus, transformation of the data was necessary. Similar findings were obtained by Dantigny and Bensoussan (14), who modeled various fungi under different environmental conditions to find the most appropriate statistical processing method for the data. They found that the square-root transformation was the only one of the tested transformations that did not exhibit any correlation between the mean and the variance. However, the choice of transformation was not related to the magnitude of the growth and depended on characteristics of the mold (14). Regarding the effect of temperature on tv, the same trend was observed as for growth rate, i.e., decreasing time to visible mycelium as the temperature rises to optimal. To our knowledge, few previous studies have been done on modeling of tv on fruits. The growth of Aspergillus section Nigri in mango nectar has been modeled by Silva et al. (50) using a polynomial model with different pH values and temperatures (aw was constant at 0.98). At pH 4.5 and 228C, lopt was 4.4 mm/day and lag time was 2.1 days, both similar to those found in our study, indicating similarity of strains. Dagnas et al. (13), modelling A. niger as a bakery product spoilage mold, found that the effect of aw and temperature was equivalent for growth rate and lag time, and pffiffiffiffiffiffiffiffi cardinal parameters for 1=k were Tmin ¼ 9.08C, Topt ¼ pffiffiffiffiffiffiffiffiffiffiffiffi 37.08C, and Tmax ¼ 44.58C, with a 1=kopt ¼ 2.01/days. The parameter values obtained were slightly higher than those in our study, perhaps because of the difference in matrix (11, 44). To understand mold behavior on fruit surfaces, experiments on Persian limes were performed with one representative strain. Near optimal conditions of growth, temperature did not influence the growth rate (lmax) on the limes. As the temperature increased, the time decreased, for a shorter tv. No relevant differences in tv were observed near to optimal conditions, 25 to 308C, where tv was shortest. In our research, growth of mold on lime fruit was significantly slower at the lower and upper temperatures, with high variability. Garcia et al. (21) studied the effect of suboptimal environmental conditions on the intraspecific variability of A. carbonarius growth and ochratoxin A production using a
J. Food Prot., Vol. 80, No. 7
large number of isolates. These researchers found significant differences among growth rate, lag phase, and ochratoxin A concentrations within the isolates with wide variation. The behavior of all strains was consistent throughout their study. In our study, the five strains of A. niger were variable in growth rate at suboptimal conditions, which suggests that the results will be affected by this variability. Extreme temperatures of growth for A. niger seem to have a normal and natural variability. In this case, one of the strains was used to represent the others to validate the experiment without undermining the model. Baert et al. (4) modeled the effect of temperature on growth rate and lag phase of Penicillium expansum in apples and found differences in the lag phase between mold grown on apple-based medium and mold grown on apple fruits. The effect of the inoculum size on the growth of P. expansum also was modeled (3). A low inoculum level resulted in longer lag phases and higher variability of the estimated lag phase, i.e., P. expansum grew more slowly. In our study, A. niger infection occurred in approximately 28% of lime fruit inoculated with a low inoculum level (5 3 103 spores per mL), which indicates that not all spores present germinate, influencing the parameter estimates for A. niger and resulting in higher values of tv. The differences in growth rate and tv could be due to the effects of external factors. When A. niger infects lime fruit, environmental conditions affect the development of rot. Spoilage depends on factors favoring germination, such as inoculum size (10), temperature, and water distribution and availability (38). However, in our research, the inoculum size and aw did not change; thus, temperature was the main factor that influenced mold growth. Accurate information is needed on the relationship between these key factors and their impact on the marginal and optimal levels for A. niger germination, growth, and toxin production on Persian limes. Germination is the most important step to consider because as soon mold is visible the fruit is considered spoiled (17). Variability in germination time has been observed when mold spores are stressed. Dagnas et al. (11) studied the biological variability in spore germination of Penicillium corylophilum in various concentrations of red cabbage seed extract and found that when extract concentration increased, the germination ratio was smaller. However, for some molds an active pathogenic process starts immediately after spores land on the wounded tissue, whereas other pathogens can breach the unripe fruit cuticle and then remain inactive for months until the harvested fruit ripens (42). For limes, chemical and mechanical barriers in the peel protect the fruit, and processes such as lignification of wounds induce resistance to fungal infection. Citral, as essentail oil present in the fruit flavedo layer, has activity against P. digitatum (9). This kind of protection could slow development of A. niger on limes by reducing the availability of nutrients, as in culture medium, and thus reducing growth. Germination and growth of A, niger on limes is unlikely to depend on the presence of other fungal spores, unless these spores also germinate and grow. In such a case, the limes would also become spoiled. This work was a first approximation for modeling growth rate and time to visible mycelium for A. niger in lime
J. Food Prot., Vol. 80, No. 7
fruits. For the isolates tested, predicted growth rate and time to visible mycelium were not significantly different than those values reported by others, but in actual lime fruits, the molds had a different behavior. Extreme temperatures of growth for A. niger seem to have a normal and natural variability, and differences due to intraspecific variability and/or the lime itself most probably resulted in the differences between the predicted and observed growth rate and time to visible mycelium. Chemical and physical barriers in the peel protect citrus fruits; therefore, differences in predictions developed for agar media versus those for the actual food matrix should be taken into account when choosing the best model to be used to develop control strategies for fungal contamination.
ACKNOWLEDGMENTS We thank Diana Dorantes and Acoyani Garrido for editing the English in this article, the Center for Biological Research of the Northwest and the Mexican National Council of Science and Technology for a doctoral scholarship to T. Sandoval-Contreras, and the University of Guadalajara for use of their facilities for the laboratory work.
REFERENCES 1. Agrios, G. N. 2005. Plant pathology, 5th ed. Elsevier, Amsterdam. 2. Ascencio, F. 2012. Validaci´on de un paquete tecnol´ogico para la protecci´on de c´ıtricos contra hongos pat´ogenos en poscosecha mediante la aplicaci´on de levaduras. Secretar´ıa de Agricultura, Ganader´ıa, Desarrollo Rural, Pesca y Alimentaci´on, Consejo Nacional de Ciencia y Tecnolog´ıa, La Paz, M´exico. 3. Baert, K., F. Devlieghere, L. Debevere, L. Bo, and B. De Meulenaer. 2008. The effect of inoculum size on the growth of Penicillium expansum in apples. Food Microbiol. 25:212–217. 4. Baert, K., A. Valero, B. De Meulenaer, S. Samapundo, M. M. Ahmed, L. Bo, L. Debevere, and F. Devlieghere. 2007. Modeling the effect of temperature on the growth rate and lag phase of Penicillium expansum in apples. Int. J. Food Microbiol. 118:139–150. 5. Bamba, R., and G. Sumbali. 2005. Co-occurrence of aflatoxin B1 and cyclopiazonic acid in sour lime (Citrus aurantifolia Swingle) during post-harvest pathogenesis by Aspergillus flavus. Mycopathologia 159:407–411. 6. Baranyi, J., and T. A. Roberts. 1994. A dynamic approach to predicting bacterial growth in food. Int. J. Food Microbiol. 23:277– 294. 7. Battilani, P., C. Barbano, and A. Logrieco. 2008. A risk assessment and safety evaluation of mycotoxins in fruits, p. 1–26. In R. BarkaiGolan and N. Paster (ed.), Mouldy fruits and vegetables as a source of mycotoxins. Elsevier, Amsterdam. 8. Bell´ı, N., S. Mar´ın, V. Sanchis, and A. J. Ramos. 2004. Influence of water activity and temperature on growth of isolates of Aspergillus section Nigri obtained from grapes. Int. J. Food Microbiol. 96:19–27. 9. Ben-Yehoshua, S., V. Rodov, D. Qiu Fang, and J. Jin Kim. 1995. Preformed antifungal compounds of citrus fruit: effect of postharvest treatments with heat and growth regulators. J. Agric. Food Chem. 43:1062–1066. 10. Burgain, A., M. Bensoussan, and P. Dantigny. 2013. Effect of inoculum size and water activity on the time to visible growth of Penicillium chrysogenum colony. Int. J. Food Microbiol. 163:180– 183. 11. Dagnas, S., M. Gougouli, B. Onno, K. P. Koutsoumanis, and J. M. Membr´e. 2015. Modeling red cabbage seed extract effect on Penicillium corylophilum: relationship between germination time, individual and population lag time. Int. J. Food Microbiol. 211:86– 94. 12. Dagnas, S., and J. M. Membr´e. 2013. Predicting and preventing mold spoilage of food products. J. Food Prot. 76:538–551.
MODELING PHYTOPATHOGENIC MOLDS
1097
13. Dagnas, S., B. Onno, and J. M. Membr´e. 2014. Modeling growth of three bakery product spoilage molds as a function of water activity, temperature and pH. Int. J. Food Microbiol. 186:95–104. 14. Dantigny, P., and M. Bensoussan. 2008. The logarithmic transformation should be avoided for stabilising the variance of mould growth rate. Int. J. Food Microbiol. 121:225–228. 15. Dantigny, P., and S. Nanguy. 2009. Significance of the physiological state of fungal spores. Int. J. Food Microbiol. 134:16–20. 16. Dantigny, P., S. Nanguy, D. Judet-Correia, and M. Bensoussan. 2011. A new germination model for fungi. Int. J. Food Microbiol. 146:176– 181. 17. Dantigny, P., C. Soares-Mansur, M. Sautour, I. Tchobanov, and M. Bensoussan. 2002. Relationship between spore germination kinetics and lag time during growth of Mucor racemosus. Lett. Appl. Microbiol. 35:395–398. 18. Eckert, J. W., and I. L. Eaks. 1989. Postharvest disorders and diseases of citrus fruits, p. 179–260. In W. Reuter, E. C. Calavan, and G. E. Carman (ed.), The citrus industry, vol. 5. Crop protection, postharvest technology, and early history of citrus research in California. Division of Agriculture and Natural Resources, University of California, Oakland. 19. Filtenborg, O., J. C. Frisvad, and U. Thrane. 1996. Moulds in food spoilage. Int. J. Food Microbiol. 33:85–102. 20. Garc´ıa, D., A. J. Ramos, V. Sanchis, and S. Mar´ın. 2009. Predicting mycotoxins in foods: a review. Food Microbiol. 26:757–769. 21. Garc´ıa, D., A. J. Ramos, V. Sanchis, and S. Mar´ın. 2011. Is intraspecific variability of growth and mycotoxin production dependent on environmental conditions? A study with Aspergillus carbonarius isolates. Int. J. Food Microbiol. 144:432–439. 22. Garc´ıa-Cela, E., A. Crespo-Sempere, A. J. Ramos, V. Sanchis, and S. Mar´ın. 2014. Ecophysiological characterization of Aspergillus carbonarius, Aspergillus tubingensis and Aspergillus niger isolated from grapes in Spanish vineyards. Int. J. Food Microbiol. 173:89–98. 23. Gibson, A. M., J. Baranyi, I. J. Pitt, M. J. Eyles, and Y. A. Roberts. 1994. Predicting fungal growth: the effect of water activity on Aspergillus flavus and related species. Int. J. Food Microbiol. 23:419– 431. 24. Gibson, A. M., and A. D. Hocking. 1997. Advances in the predictive modelling of fungal growth in food. Trends Food Sci. Technol. 8:353–358. 25. Gougouli, M., K. Kalantzi, E. Beletsiotis, and K. P. Koutsoumanis. 2011. Development and application of predictive models for fungal growth as tools to improve quality control in yogurt production. Food Microbiol. 28:1453–1462. 26. Gougouli, M., and K. P. Koutsoumanis. 2010. Modelling growth of Penicillium expansum and Aspergillus niger at constant and fluctuating temperature conditions. Int. J. Food Microbiol. 140:254– 262. 27. Gougouli, M., and K. P. Koutsoumanis. 2012. Modeling germination of fungal spores at constant and fluctuating temperature conditions. Int. J. Food Microbiol. 152:153–161. 28. Gougouli, M., and K. P. Koutsoumanis. 2013. Relation between germination and mycelium growth of individual fungal spores. Int. J. Food Microbiol. 161:231–239. 29. Huchet, V., S. Pavan, A. Lochardet, M. L. Divanac’h, F. Postollec, and D. Thuault. 2013. Development and application of a predictive model of Aspergillus candidus growth as a tool to improve shelf life of bakery products. Food Microbiol. 36:254–259. 30. Joint FAO/WHO Expert Committee on Food Additives. 2001. Ochratoxin A. In Evaluation of certain mycotoxins in food. Fiftysixth report of the Joint FAO/WHO Expert Committee on Food Additives. Available at: http://www.inchem.org/documents/jecfa/ jecmono/v47je01.htm. Accessed 30 November 2015. 31. Lugauskas, A., and J. Stak`enien`e. 2002. Toxin producing micromycetes on fruit, berries, and vegetables. Ann. Agric. Environ. Med. 9:183–197. 32. Magan, N., and D. Aldred. 2007. Mycotoxins from the field to the table. Post-harvest control strategies: minimizing mycotoxins in the food chain. Int. J. Food Microbiol. 119:131–139.
1098
SANDOVAL-CONTRERAS ET AL.
33. Mar´ın, S., C. Colom, V. Sanchis, and A. J. Ramos. 2009. Modelling of growth of aflatoxigenic A. flavus isolates from red chilli powder as a function of water availability. Int. J. Food Microbiol. 128:491–496. 34. Mar´ın, S., E. Companys, V. Sanchis, A. J. Ramos, and N. Magan. 1998. Effect of water activity and temperature on competing abilities of common maize fungi. Mycol. Res. 120:959–964. 35. Mar´ın, S., D. Cuevas, A. J. Ramos, and V. Sanchis. 2008. Fitting of colony diameter and ergosterol as indicators of food borne mould growth to known growth models in solid medium. Int. J. Food Microbiol. 121:139–149. 36. Marino, A., A. Nostro, and C. Fiorentino. 2009. Ochratoxin A production by Aspergillus westerdjkiae in orange fruit and juice. Int. J. Food Microbiol. 132:185–189. 37. Michailides, T. J., D. P. Morgan, and Y. Luo. 2010. Epidemiological assessments and postharvest disease incidence, p. 69–86. In D. Prusky and M. L. Gullino (ed.), Post-harvest pathology: plant pathology in the 21st century, vol. 2. Springer, New York. 38. Pardo, E., U. Lagunas, V. Sanchis, A. J. Ramos, and S. Mar´ın. 2005. Influence of water activity and temperature on conidial germination and mycelial growth of ochratoxigenic isolates of Aspergillus ochraceus on grape juice synthetic medium. Predictive models. J. Sci. Food Agric. 85:1681–1686. 39. Parra, R., and N. Magan. 2004. Modelling the effect of temperature and water activity on growth of Aspergillus niger strains and applications for food spoilage moulds. J. Appl. Microbiol. 97:429– 438. 40. Perrone, G. A., G. Susca, K. Cozzi, J. Ehrlich, J. C. Varga, M. Frisvad, P. Meijer, W. Noonim, L. Mahakarnchanakul, and R. A. Samson. 2007. Biodiversity of Aspergillus species in some important agricultural products. Stud. Mycol. 59:53–66. 41. Pitt, J. I., and A. D. Hocking. 2009. Fungi and food spoilage. Springer, New York. 42. Prusky, D., N. Alkan, I. Miyara, S. Barad, M. Davidzon, I. Kobiler, S. Brown-Horowitz, A. Lichter, A. Sherman, and R. Fluhr. 2010. Mechanisms modulating postharvest pathogen colonization of decaying fruits, p. 43–55. In D. Prusky and M. L. Gullino (ed.), Postharvest pathology: plant pathology in the 21st century, vol. 2. Springer, New York. 43. Rosso, L., J. R. Lobry, S. Bajard, and J. P. Flandrois. 1993. An unexpected correlation between cardinal temperatures of microbial growth highlighted by a new model. J. Theor. Biol. 162:447–463.
J. Food Prot., Vol. 80, No. 7
44. Rychlik, M., and P. Schieberle. 2001. Model studies on the diffusion behavior of the mycotoxin patulin in apples, tomatoes, and wheat bread. Eur. Food Res. Technol. 212:274–278. 45. Sambrook, J., and D. W. Russell. 2001. Preparation of plasmid DNA by lysis with SDS, p. 1.55–1.58. In J. Sambrook and D. W. Russell (ed.), Molecular cloning: a laboratory manual, 3rd ed., vol. 1. Cold Spring Harbor Laboratory Press, Cold Spring Harbor, NY. 46. Samson, R. A., C. M. Visagie, J. Houbraken, S.-B. Hong, V. Hubka, C. H. W. Klaassen, G. Perrone, K. A. Seifert, A. Susca, J. B. Tanney, J. Varga, S. Kocsub´e, G. Szigeti, T. Yaguchi, and J. C. Frisvad. 2014. Phylogeny, identification and nomenclature of the genus Aspergillus. Stud. Mycol. 78:141–173. 47. Sandoval-Contreras, T. 2011. Caracterizaci´on fenot´ıpica y genot´ıpica de mohos y levaduras en lim´on mexicano (Citrus aurantifolia S.) y lim´on persa (Citrus latifolia T.) procedentes de campos de cultivo y empacadoras de Jalisco y Colima. Master’s thesis. Universidad de Guadalajara, Guadalajara, M´exico. 48. Schoch, C. L., K. A. Seifert, S. Huhndorf, V. Robert, J. L. Spouge, C. A. Levesque, W. Chen, and Fungal Barcoding Consortium. 2012. Nuclear ribosomal internal transcribed spacer (ITS) region as a universal DNA barcode marker for Fungi. Proc. Natl. Acad. Sci. USA 109:6241–6246. 49. Secretar´ıa de Agricultura, Ganader´ıa, Desarrollo Rural, Pesca y Alimentaci´on. 2004. Pliego de Condiciones para el uso de la marca oficial M´exico Calidad Selecta en lim´on persa. PC-012-2004. Secretar´ıa de Agricultura, Ganader´ıa, Desarrollo Rural, Pesca y Alimentaci´on, M´exico City, M´exico. 50. Silva, A. R., A. S. Sant’Ana, and P. R. Massaguer. 2010. Modelling the lag time and growth rate of Aspergillus section Nigri IOC 4573 in mango nectar as a function of temperature and pH. J. Appl. Microbiol. 109:1105–1116. 51. Smilanick, J. L., D. A. Margosan, F. Mlikota, J. Usall, and I. J. Michael. 1999. Control of citrus green mold by carbonate and bicarbonate salts and the influence of commercial postharvest practices on their efficacy. Plant Dis. 83:139–145. 52. Tournas, V. H., and E. Katsoudas. 2005. Mould and yeast flora in fresh berries, grapes and citrus fruits. Int. J. Food Microbiol. 105:11– 17. 53. White, T. J., T. Bruns, S. Lee, and J. Taylor. 1990. Amplification and direct sequencing of ribosomal RNA genes for phylogenetics, p. 315– 322. In M. A. Innis, D. H. Gelford, J. J. Sninsky, and T. J. White (ed.), PCR protocols: a guide to methods and applications. Academic Press, San Diego, CA.
