Water and temperature relations of growth and ochratoxin A production by Aspergillus carbonarius strains from grapes in Europe and Israel


Prof. Naresh Magan, Applied Mycology Group, Biotechnology Centre, Cranfield University, Silsoe, Bedford MK45 4DT, UK
(e-mail: n.magan@cranfield.ac.uk).


Aims:  This study investigated the in vitro effects of water activity (aw; 0·85–0·987) and temperature (10–40°C) on growth and ochratoxin A (OTA) production by two strains of Aspergillus carbonarius isolated from wine grapes from three different European countries and Israel on a synthetic grape juice medium representative of mid-veraison (total of eight strains).

Methods and Results:  The synthetic grape juice medium was modified with glycerol or glucose and experiments carried out for up to 56 days for growth and 25 days for OTA production. The lag phase prior to growth, growth rates and ochratoxin production were quantified. Statistical comparisons were made of all factors and multiple regression analysis used to obtain surface response curves of aw × temperature for the eight strains and optimum growth and OTA production by A. carbonarius. The lag phase increased from <1 day at 25–35°C and 0·98 aw to >20 days at marginal temperatures and water availabilities. Generally, most A. carbonarius strains grew optimally at 30–35°C, regardless of solute used to modify aw, with no growth at <15°C. The optimum aw for growth varied from 0·93 to 0·987 depending on the strain, with the widest aw tolerance at 25–30°C. There was no direct relationship among growth, environmental factors and country of origin of individual strains. Optimum conditions for OTA production varied with strain. Some strains produced optimal OTA at 15–20°C and 0·95–98 aw. The maximum OTA produced after 10 days was about 0·6–0·7 μg g−1, with a mean production over all eight strains of 0·2 μg g−1 at optimum environmental conditions.

Conclusions:  This work demonstrates that optimum conditions for OTA production are very different from those for growth. While growth rates differed significantly between strains, integration of the OTA production data suggests possible benefits for use of the information on a regional basis.

Significance and Impact of the Study:  Very little detailed information has previously been available on the ecology of A. carbonarius. This knowledge is critical in the development and prediction of the risk models of contamination of grapes and grape products by this species under fluctuating and interacting environmental parameters.


Ochratoxin A (OTA) is a mycotoxin which has received increasing attention because of its potent nephrotoxic, immunosuppressive effects. It has been classified as a possible group 2b human carcinogen (IARC 1993). The occurrence of OTA in cereals has been well documented and for some time it was thought to be the major source of OTA in humans and animals. Over the last few years, OTA has been found in other food commodities including coffee beans, cocoa, beans and pulses (Kuiper-Goodman and Scott 1989; Pohland et al. 1992; Jörgensen and Rasmundsen 1996; Jörgensen 1998; Joosten et al. 2001). OTA has also been detected in wine, grape juice, beer and dried fruits (Zimmerli and Dick 1996).

Traditionally it was thought that OTA was produced by isolates of Aspergillus ochraceus and Penicillium verrucosum (Pitt and Hocking 1997). However, in the mid-1990s Aspergillus section Nigri were identified as being able to produce ochratoxin in grapes and wine for the first time (Belli et al. 2004). It was subsequently shown that within this group A. carbonarious was predominantly responsible for the production of OTA in grapes and wine (Bragulat et al. 2001; Battilani and Pietri 2002).

Fungal growth and OTA production are markedly affected by different environmental factors, the two most important being water availability (water activity, aw) and temperature influencing germination, growth and sporulation (Magan and Lacey 1984). Changes in aw and temperature have been shown to affect the growth of A. ochraceus and A. niger isolates (Ayerst 1969; Marin et al. 1998) and ability to compete with other spoilage fungi and influence OTA production (Ramakrishna et al. 1996; Magan et al. 2003).

Aspergillus carbonarius has been isolated in a range of southern European regions. At the present time no studies have compared the effect of water availability and temperature on the growth of such isolates and identified the conditions conducive to growth and OTA production. Such information is important in developing realistic forecasting systems for predicting risk of colonization of grapes by A. carbonarius and OTA production. One recent study has compared OTA production by A. carbonarius and other section Nigri strains from Italy and Spain in relation to water activity and time, but only at 25°C (Belli et al. 2004). They suggested that 5–10 days was optimum for OTA production. The objectives of the present work were to compare two strains of A. carbonarius from four different trans-European regions (Portugal, Italy, Greece, Israel) with regard to the effect of aw × temperature on lag phases prior to growth, growth rates and on temporal production of OTA. Surface response curves for optimum and marginal conditions for growth and OTA production were identified.

Materials and methods

Fungal isolates

In this study A. carbonarius strains isolated from grapes in Italy (MPVA A 1102, Ita1; MPVP A 933, Ita2); Portugal (01UAs219, Port1; 01UAs263, Port2); Greece (G 458,Gre1; G 444, Gre2) and Israel (1-4-1-9-10·8, Isr1; 1-4-1-9-7·7, Isr2). Identification of isolates was confirmed by Dr Z. Lawrence (CABI BioSciences, Egham, Surrey, UK) and were shown to produce OTA on Czapek yeast autolysate agar using the method developed by Bragulat et al. (2001).

Medium, water activity modification and inoculation

Studies were conducted in vitro on a synthetic grape juice medium (SGM) representative of grape composition at early veraison. This consisted of d(+) glucose 70 g, d(−) fructose 30 g, l(−) tartaric acid 7 g, l(−) malic acid 10 g, (NH4)2HPO4 0·67 g, KH2PO4 0·67 g, MgSO4·7H2O 1·5 g, NaCl 0·15 g, CaCl2 0·15 g, CuCl2 0·0015 g, FeSO4·7H2O 0·021 g, ZnSO4·7H2O 0·0075 g, (+) Catechin hydrate 0·05 g, agar 25 g in a litre of medium. This was adjusted with 2 m KOH to pH 4·0–4·2.

All experiments were carried out over the range of 0·987–0·85 aw by addition of either glucose or glycerol. The aw of the unmodified medium was 0·987 aw and this was used as the control treatment. The temperatures used were 10, 15, 20, 25, 30, 35 and 40°C. The aw of all cooled treatment media were checked using a Aqualab® WP4 (Decagon Devices, Inc., Pullman, WA, USA) connected to a PC using Hyperterminal software (Higraeve Inc., Monroe, MI, USA) and found to be within aw of treatment values.

All treatments and replicates were inoculated centrally with a small loop of spore suspension (5 μl) from a 105 spores ml−1 stock solution obtained from 10-day-old SGM media of the same aw. All studies were carried out with a minimum of three replicates per treatment and repeated at least twice. Replicates of the same treatment were stored in controlled environment chambers of the same aw/temperature to maintain treatment conditions for a maximum of 56 days.

Two types of measurement were made: (a) the lag phase (in days) prior to mycelial growth (>0·5 mm diameter) and (b) measurement of mycelial extension. Temporal mycelial extension rates were measured daily in two directions at right angles to each other until the medium was fully colonized. The radial extension rates were plotted against time and the growth rates calculated using linear regression (mm day−1) at each different aw and temperature, for each replicate and treatment.

Quantification of ochratoxin A production

A series of replicates (at least three per experiment) were destructively sampled for OTA quantification for each treatment condition. Six agar plugs were removed from each replicate plate using a cork borer (4·5 mm diameter) across the radius of each colony every 5 days for 20 days. Samples were frozen and subsequently analysed for OTA content. The method used was adapted from Bragulat et al. (2001). The plugs were placed into a 3 ml microtube (Eppendorf, Histon, UK) and weighed. One millilitre of HPLC grade methanol (Fisher, Loughborough, UK) was added and the samples shaken and incubated at room temperature for 90 min. The extracts were filtered (Millex® HV 13 mm; Millipore Corporation, Bellerica, MA, USA) directly into amber HPLC vials (Jaytee Biosciences Ltd, Whitstable, UK) and stored at 4°C until HPLC analysis was performed. This method gave an extraction efficiency of about 90%.

The HPLC system used consisted of a Millipore Water 600E system controller, a Millipore 712 WISP autosampler and a Millipore Waters 470 scanning fluorescence detector (Millipore Corporation) (excitation 330 nm, emission 460 nm). The samples were separated using a C18 Luna Spherisorb ODS2 column (150 × 4·6 mm, 5 μm) (Phenomenex, Macclesfield, UK), with a guard column of the same material. Run time for samples was 12 min with OTA detection after 5·75 min. The flow rate of the mobile phase (acetonitrile : water : acetic acid; 57 : 41 : 2) was 1 ml min−1. The detection limit was <0·01 μg OTA g−1 SGM, based on a signal to noise ratio of 3 : 1. Analysis of the results was carried out on a computer with Kroma systems 2000 software (Bio-tek Instruments, Milan, Italy). In this paper, data on OTA production after 10 days is presented although information on other times was also gathered.

Statistical analyses of the data

The data was statistically analysed in two ways. The effect of aw × temperature was examined by anova using Statistica (Statsoft, Tulsa, OK, USA). Differences in growth rates for isolates from each European region were also compared in relation to the medium modification with glucose/glycerol, and differences between strains. Using a polynomial multiple linear regression the surface response curves were obtained using the above programme.


Effect of aw × temperature on lag phase and growth of isolates of A. carbonarius

Figure 1 shows an example of the effect of different temperatures on the lag times prior to growth initiation at different steady-state aw levels on a strain from Italy (MPVA 1102). Very short lag phases occurred at 25–35°C (<1 day) and 0·985 to 0·95 aw. At lower temperatures the lag times increased to >20 days, especially at 0·93–0·88 aw. Similar trends were found for all strains studied (data not shown).

Figure 1.

Relationship between temperature and water activity (aw) level on the lag phase prior to growth of a strain of A. carbonarius on a synthetic grape juice medium. Dotted lines indicate extrapolation

Figure 2 compares growth rates in relation to water availability at three different temperatures using media modified with either glucose or glycerol for a strain of A. carbonarius from Portugal (01AUs263). For this isolate maximum growth occurred at 35°C and 0·98 aw with a decrease under wetter conditions (0·987 aw). Between 0·95 and 0·98, the growth rate of most strains at 25°C was 25% less than at 35°C. Solute type had a small effect on growth. Similar patterns in relation to aw were observed for all the strains examined although growth rates varied. Statistically there was no significant difference between growth rates on glucose/glycerol modified media at 15 or 35°C but for a few strains there was a significant difference (P = 0·05) at 25°C between 0·90 and 0·987 aw (data not shown). A large proportion of the strains did not show any growth at 0·88 aw. In an analysis of growth at 15°C there was a significant difference at all aw treatments. No strains grew at 10°C.

Figure 2.

Effect of water activity (aw) and solute used (glucose, glycerol) on growth rates of a strain of A. carbonarius (Port2) at three different steady-state temperature conditions on a synthetic grape juice medium

Comparisons were made of growth of strains from each of the four countries examined at different aw levels at 25°C (Fig. 3). There was no consistent pattern of growth for strains from different countries or between strains from the same country. Statistical comparison showed that for only a few strains within a country were there were statistically significant differences in growth rates (Table 1). Figure 4 shows three examples of the surface response profiles for growth of strains of A. carbonarius from Italy, Israel and Greece respectively. This confirms that optimum growth was at higher temperatures (25–35°C) and intermediate aw levels (0·97–0·985).

Figure 3.

Comparison of growth rates of two strains each of A. carbonarius from Italy, Portugal, Israel and Greece in relation to four steady-state water activity (aw) conditions at 25°C on a synthetic grape juice medium. Bars indicate standard errors. For key to strains see Materials and Methods

Table 1.  Statistical comparisons of mean growth rates of strains of Aspergillus carbonarius in relation to temperature × water availability for all strains using growth rate as the variable for analysis
Strains{Ita1} (M = 2·3000){Ita2} (M = 3·3597){Por1} (M = 1·9100){Por2} (M = 3·8155){Isr1} (M = 1·8671){Isr2} (M = 3·8172){Gre1} (M = 3·8172){Gre2} (M = 3·3552)
  1. P = 0·05.

  2. Figures in bold show statistically significant differences between strains.

  3. M, overall mean of a single strain over all conditions. For key to strains see Materials and methods.

{Ita1} 0·2860·6690·0980·6480·1200·1100·288
{Ita2}0·286 0·1320·6350·1340·6450·6450·996
{Por1}0·6690·132 0·0310·9630·0370·0370·133
{Por2}0·0970·6350·031 0·0340·9980·9980·631
{Isr1}0·6470·1340·9620·034 0·0400·0400·134
{Isr2}0·1100·6450·0370·9980·040 10·642
{Gre1}0·1100·6450·0370·9980·0401 0·642
Figure 4.

Surface response curves for three strains of A. carbonarius from Italy (Ita1), Portugal (Port) and Israel (Isr1) in relation to water activity and temperature. The numbers on the contour lines refer to growth rates (mm day−1). For key to strains see Materials and Methods

Effects of aw and temperature on ochratoxin production

Comparison of OTA production by all strains in relation to three aw levels at a steady-state temperature (20°C) after 10 days incubation is shown in Fig. 5. Significant differences between amounts of OTA produced by an individual strain, and between strains from the same country or different countries were observed. Statistical analysis showed that the overall, the mean OTA production by seven of the eight strains over all treatment conditions were relatively similar. Table 2 shows the statistical comparison for all eight strains and the statistically significant effect of single factors of aw, temperature, and two-way interactions of aw × temperature.

Figure 5.

Comparison of ochratoxin production (μg g−1) of two strains each of A. carbonarius from Italy, Portugal, Israel and Greece in relation to three steady-state water activity (aw) conditions at 20°C on a synthetic grape juice medium. For key to strains see Materials and Methods

Table 2.  Statistical comparison of strains, water activity, temperature and water activity × temperature effects on ochratoxin A production by eight strains of Aspergillus carbonarius
TreatmentsStandard errorSignificance
Water activity0·0340·000
Aw × temperature0·0170·000

The data for all eight strains were used to develop a surface response curve to identify the optimum conditions of aw and temperature and the range of conditions for production of different quantitates of OTA (Fig. 6). In contrast to results for growth this clearly identified 15–20°C and 0·95–0·98 aw as optimum conditions for OTA production.

Figure 6.

Surface response curve for OTA production by Aspergillus carbonarius based on data for all eight strains from the four countries in relation to water activity and temperature. The numbers on the contour lines refer to mean mycotoxin production (μg g−1)


This study compares the impact of aw × temperature regimes on growth of strains of A. carbonarius from different countries in Europe and from Israel for the first time. Both factors affected the lag phase prior to growth and mycelial extension of the strains. While the pattern of effects of environmental factors was similar, the growth rates of strains varied significantly. Generally, growth was reduced by up to 50% at ≤0·93 aw. Growth was also to a large extent inhibited at 0·88 aw for most strains even after 56 days incubation. There was some variation in the growth rates between countries, and also within countries. This shows the importance of using a wide range of strains from different climatic regions for building a database to determining environmental conditions where risk from OTA contamination may occur. These results are also very different from those obtained previously for isolates of the A. niger group prior to the detailed taxonomic knowledge of species in the section Nigri and the production of OTA by species in this section. Studies of A. niger by Ayerst (1969) and Marin et al. (1998) isolated from spoiled grain showed wider aw ranges of 0·83–0·995 with optimum growth at 0·995 aw and 35°C. Temperature ranges were found to be 10–40°C at optimum water availability. In comparison, OTA producing strains of A. ochraceus were found to grow optimally at 0·995 and 30°C with similar aw limits of about 0·83–0·85 aw and a temperature range of 10–35°C. These limits and ranges for growth are very different from those for A. carbonarius.

Temporal studies have shown that OTA was initially produced after 5–10 days incubation on the SGM medium. Previous studies by Cabanes et al. (2002) found that A. carbonarius isolates produced maximum OTA on unmodified yeast extract sucrose agar and Czapek yeast agar after about 14 days although some was detected after 7 days. In contrast, Belli et al. (2004) suggested that 5 days may be optimum for OTA production by A. carbonarius strains from Spain and Italy. They found OTA production of 4–5 μg g−1 of medium for a strain from Spain after 5 days with a subsequent decrease over time. However, the temperature (25°C) used now appears to be suboptimal for OTA production by many A. carbonarius strains. In the present study, the production of OTA over a range of environmental conditions provides important information as it could assist in predicting the possible risk of OTA production and in choosing appropriate harvesting dates to minimize entrance into wine and vine fruit production. The present study has also demonstrated the contrasting impact of both aw and temperature on growth and OTA production by the eight strains examined. The most important findings are that the optimum conditions for OTA production are generally very different from those for growth. Studies of A. cabonarius on irradiated coffee berries by Joosten et al. (2001) found that most OTA production at 25, 27 and 30°C and 0·99 aw, with production strongly reduced at 0·94 aw after 14 days incubation. However, only a few environmental conditions were tested, and the nutritional status of coffee berries is very different from grape-based media or grapes. The surface response curves presented in the present study on growth and OTA production may provide very useful guidelines for facilitating effective management of predicting risk from growth and OTA production during ripening and harvesting of grapes and during transport for wine production.

Over long incubation periods it has been observed that some decrease in OTA production can occur under some treatment conditions (Belli et al. 2004). Previous studies have suggested that A. carbonarius itself, or other black Aspergilli could degrade OTA in media resulting in the formation of breakdown products such as ochratoxin α (Abrunhosa et al. 2002; Varga et al. 2002). Other inputs into the system such as fungicides and pesticides may further stress these fungi influencing both growth and more importantly, OTA production.


We are grateful to Prof. P. Battilani, Dr A. Venancio, Prof. E. Tjamos and Dr A. Lichter for the supply of fungal strains and to the EU for financial support (Quality of Life Programme, Key Action 1 in Food Nutrition and Health; QLK1-CT-2001-01761; Wine Ochra Risk).