New examinations of mycotoxin carryover to cocoa beans

Mycotoxins are secondary metabolites which can form on various foodstuffs through the growth of certain fungi. Ochratoxin A (OTA) and the aflatoxins B₁ B₂, G₁ and G₂ have been detected in low concentrations in cocoa and cocoa products. As regards the question of in what stages of the cocoa production process a contamination with the mycotoxin-producing moulds and the formation of mycotoxins takes place, it is assumed that in the case of cocoa the contamination is not concerning the individual beans but the fermentation units. A model test was carried out to provide information on the process by which a possible carryover of the above-mentioned mycotoxins to cocoa beans occurs during the fermentation process. For this purpose fresh cocoa beans were left to soak in an artificial mycotoxin-containing fermentation solution. The mycotoxin levels in the cocoa beans were regularly determined over a period of 12 days. New findings were made as regards the migration of mycotoxins during the fermentation process. We interpret the divergent uptake behaviour of the mycotoxins to indicate that the transport of OTA and that of aflatoxins does not take place in the same manner. This is possibly caused by chemico-physical effects, such as the different polarities of the mycotoxins. Presented at the 28^th Mykotoxin-Workshop, Bydgoszcz, Poland, May 29–31, 2006     

Sensitive method for determination of DON in cocoa by means of HPLC-techniques

The mycotoxin deoxynivalenol (DON) is one of a group of mycotoxins known as type B trichothecenes and is particularly formed by the mould speciesFusarium graminearum andFusarium culmorum. The frequency of the occurrence of DON in certain raw materials and the concentrations found make it one of the world’s most significant mycotoxin contaminants. Positive findings of the toxin especially have been established in cereal-based foods, as well as in oilseeds. The main objective of this study was to set up a current situation assessment of the possible occurrence of deoxynivalenol in cocoa and cocoa products. As there was no analytical method for determining DON in cocoa and cocoa products, a special method was developed. The applicability and consistency of the method was confirmed by performing recovery assays on various cocoa products. A special post-column derivatisation procedure was developed to increase selectivity and raise sensitivity by a factor of 80. The method was used to test 230 samples for possible DON content, ranging from cocoa beans to cocoa bean shells, nibs, cocoa liquor and cocoa powders through to finished cocoa-based products. The results suggest that DON may occasionally occur in cocoa beans in very low concentrations. Presented at the 29^th Mykotoxin-Workshop, Fellbach, Germany, May 14–16, 2007     

OTA-producing fungi isolated from stored cocoa beans

Aims:  The aim of this study was to identify fungal populations in unroasted cocoa beans stored in Spain in order to evaluate the ochratoxin A (OTA)-production ability of certain Aspergillus isolates.
Methods and Results:  Twenty batches of cocoa beans from different origins and with different OTA content were selected for this study. Three Aspergillus carbonarius and 13 Aspergillus niger aggregate strains isolated from these cocoa bean samples were selected to evaluate their OTA synthesis ability, being the only A. carbonarius isolates which are OTA producers [−1 culture medium; LOD = 6  μ g kg⁻¹ culture medium].
Conclusions:  No correspondence was found between the OTA levels in cocoa beans and the presence of OTA-producing fungi. Nonetheless, some samples contained A. carbonarius with a high OTA-producing ability and, consequently, specific fungal controls should be set up during storage to avoid this toxin.
Significance and Impact of the Study:  Toxigenic fungi in cocoa beans are not well understood. This study attempted to identify these fungi and evaluate their OTA-producing ability.     

Antifungal properties of selected lactic acid bacteria and application in the biological control of ochratoxin A producing fungi during cocoa fermentation

Thirteen Lactic acid bacteria strains isolated from fermenting cocoa and seven reference strains were used in order to assess their antifungal properties towards three ochratoxin A (OTA) producing fungi (Aspergillus carbonarius, Aspergillus niger and Aspergillus ochraceus). Furthermore, two of the isolates strains (A19 and A21) identified as belonging to the genus of Pediococcus as well as Lactobacillus plantarum B4496, Lactobacillus brevis 207 and Lactobacillus sanfranciscensis BB12 showed interesting in vitro broad antifungal activities towards the three ochratoxin-producing fungi with inhibition percentages ranging from 15% to 66.7%. Treatment of cell-free supernatant at 100°C affected antifungal activity suggesting that the main compounds responsible for this activity were of proteic nature, and hence could be bacteriocins. Application of isolate A19 in cocoa fermentation as starter inhibited the growth of each of the OTA-producing species. At the end of fermentation in boxes inoculated with A19, A. niger was not detectable while A. carbonarius concentration was found to be 2 Log CFU/g of wet beans. The assessment of the ochratoxin produced during fermentation of cocoa inoculated with A. carbonarius indicated that the use of isolate A19 as starter could reduce their level of growth so as to have only a toxin production of 0.0012 ± 0.0005 μg/kg after 40 days of storage, while this was 2.45 ± 0.35 μg/kg of fermented and dried cocoa beans in the absence of A19. This work is a contribution for the application of biological control of OTA-producing fungi during cocoa production.     

Study on distribution of mycotoxins in cocoa beans

Mycotoxins are not homogeneously distributed in foods which come in naturally small units, such as pistachios and peanuts, and may instead be extremely inhomogeneously distributed due to the occurrence of so-called hot spots. Tests conducted on pistachios, for example, show that a mouldy kernel can be so strongly contaminated with mycotoxins that it has a significant impact on the contamination profile of several thousand kernels. This makes a representative sampling of such foodstuffs very important but also a very difficult task. Whether cocoa beans also have a tendency to form so-called mycotoxin hot spots is hitherto unknown. A miniaturised analysis method was used in tests made on several independent batches of cocoa beans and although these tests showed that the mycotoxins ochratoxin A and the aflatoxins are not homogeneously distributed in cocoa, the tested batches revealed no real hot spots. Presented at the 27^th Mykotoxin-Workshop, Dortmund. Germany, June 13–15, 2005     

Genistein as an endogenous natural substrate of acidic peroxidases in lupin hypocotyls

Crude steam distillate from Ocimum gratissimum sprayed onto infection courts on detached cocoa pods moments after inoculation with Phytophthora palmivora completely inhibited the pathogen and blackpod lesion development on 75% of the infection courts. Disease suppression obtained with the extract was comparable to that obtained with a 2% Kocide 101 suspension. In the field, the O. gratissimum extract also suppressed lesion development although to a significantly lower (P = 0.05) extent in comparison with Kocide 101. Blackpod lesion expansion rates of 3.80, 3.56, 2.71 and 0.78 cm/day, respectively, were associated with pods treated in the field with C. citratus extract, tap water, O. gratissimum extract and 2% Kocide 101. The extract from Cymbopogon citratus was also ineffective on detached pods. Sporangia of P. palmivora from sporu-lating blackpod lesions on both detached and non-detached pods lost their infectivity within 1 h of treatment with the O. gratissimum extract. This effect was superior to that obtained with Kocide 101. Fungitoxicity of the extract on pods, however, was lost within 3 h of application. Thus, despite its in vivo effectiveness as an eradicant, the O. gratissimum extract, in its present form, has limited utility as a protectant fungicide.     

Rearing method and life-table data for the cocoa mirid bug Sahlbergella singularis Haglund (Hemiptera: Miridae)

The mirid bug Sahlbergella singularis feeds on cocoa pods and shoots, causing considerable crop losses. As laboratory experimentation requires numerous insects, this study aimed at improving available rearing method of S. singularis for Cameroon. Fifty second to fifth nymphal stages were collected at a cocoa farm and reared to the adult stage on cocoa pods in an insectary (T = 24.7 ± 0.9°C, RH = 84.5 ± 6.8%, photoperiod: 12 : 12 L : D). Newly emerged females were confined for 5–6 days on cocoa twigs for sexual maturation and each female paired with a male for 24 h. The pairs were returned to the field and enclosed in mousseline sleeves on attached cocoa pods, for egg laying. After 16 days (expected egg lifetime), the sleeve cages were checked daily to detect newly hatched nymphs. Then, the pods were collected and brought to the insectary, where nymphs continued to emerge and develop into adults. Our method allowed the production of 14.6 ± 6.7 nymphs per female per generation, for 15 consecutive generations. Nymphal survival was calculated to be 68.2% and the mean duration of the nymphal development was measured at 22.7 ± 3.1 days. The rearing performance was evaluated using life‐table calculations. The net reproductive rate (R₀) was 6.59; the intrinsic rate of increase (r) was 0.037 per female per day with a population doubling time (T_d) of 18.9 days. On average each female contributed 9.70 individuals to the population given a mean generation time (T_c) of 52.1 days. The percentage of reproductive females and the mean number of nymphs per female were significantly different between generations, with 86.8% and 18.1% in generation G7 as compared to 45.8% and 8.4% for generation G5, respectively. As rainfall showed concordant variations during the period of investigation, we discuss the impact of this factor on mirid fecundity.     

Cocoa agroforestry is less resilient to sub‐optimal and extreme climate than cocoa in full sun

Cocoa agroforestry is perceived as potential adaptation strategy to sub‐optimal or adverse environmental conditions such as drought. We tested this strategy over wet, dry and extremely dry periods comparing cocoa in full sun with agroforestry systems: shaded by (i) a leguminous tree species, Albizia ferruginea and (ii) Antiaris toxicaria, the most common shade tree species in the region. We monitored micro‐climate, sap flux density, throughfall, and soil water content from November 2014 to March 2016 at the forest‐savannah transition zone of Ghana with climate and drought events during the study period serving as proxy for projected future climatic conditions in marginal cocoa cultivation areas of West Africa. Combined transpiration of cocoa and shade trees was significantly higher than cocoa in full sun during wet and dry periods. During wet period, transpiration rate of cocoa plants shaded by A. ferruginea was significantly lower than cocoa under A. toxicaria and full sun. During the extreme drought of 2015/16, all cocoa plants under A. ferruginea died. Cocoa plants under A. toxicaria suffered 77% mortality and massive stress with significantly reduced sap flux density of 115 g cm^?2 day^?1, whereas cocoa in full sun maintained higher sap flux density of 170 g cm^?2 day^?1. Moreover, cocoa sap flux recovery after the extreme drought was significantly higher in full sun (163 g cm^?2 day^?1) than under A. toxicaria (37 g cm^?2 day^?1). Soil water content in full sun was higher than in shaded systems suggesting that cocoa mortality in the shaded systems was linked to strong competition for soil water. The present results have major implications for cocoa cultivation under climate change. Promoting shade cocoa agroforestry as drought resilient system especially under climate change needs to be carefully reconsidered as shade tree species such as the recommended leguminous A. ferruginea constitute major risk to cocoa functioning under extended severe drought.     

Allium cepa as a biomonitor of ochratoxin A toxicity and genotoxicity

Ochratoxin A (OTA) is a toxin produced by Aspergillus and Penicillum moulds. Since OTA has not yet been evaluated in plant systems, this paper focused on describing the controversial effect OTA in an Allium root test model, which has known sensitivity to genotoxins and could be useful in toxin screening. Analyses of root growth and the root meristematic zone in response to OTA treatment were undertaken. The results show OTA toxicity to root growth at a concentration of 10 ug·ml^?1 associated with inhibition of proliferation activity. Cytological changes observed in the Allium chromosome aberrations assay, at a concentration of 5.0 ug·ml^?1, showed that OTA was able to induce genotoxicity at the chromosome level. These results indicate that plants cells (Allium cepa) are very sensitive to the mycotoxin OTA, as observed at the highest concentration. Under these conditions, OTA produced toxicity and cytogenetic injury. Evidence in vitro and in vivo indicates that OTA can induce damage at the DNA level.     

Host preferences of the cocoa pod borer,Conopomorpha cramerella,the main threat to cocoa production in Southeast Asia

Insects ensure the survival of their offspring by depositing their eggs in suitable environments. Even generalist egg-laying insects often show preferences for specific host plants. The cocoa pod borer (CPB), Conopomorpha cramerella (Snellen) (Lepidoptera: Gracillariidae), successfully infests and reproduces on relatively few host plants, but has a major economic impact only on cocoa, Theobroma cacao L. (Malvaceae). Choice tests were performed in the laboratory to compare the frequency of insect visits, the duration of the visits, and the number of eggs laid on the fruits for each combination of host plants tested – that is, cocoa clones or fruits of rambutan (Nephelium lappaceum L.), Fijian longan (Pometia pinnata JR Forst & G Forst, both Sapindaceae), and langsat (Lansium domesticum Corrêa, Malvaceae). Our laboratory study showed that, when given the choice, CPB significantly favored cocoa pods over other host fruits (rambutan, langsat, or Fijian longan). Females also deposited more eggs on unripe fruits than on ripe fruits of similar size. The preference to lay eggs on specific cocoa clones in the bioassays did not reflect the level of the clone resistance to CPB damage reported from the field. Consequently, oviposition preference of female CPB does not seem to be the main factor explaining field resistance of some cocoa clones to CPB infestation.     

Adsorption of ochratoxin A and zearalenone by potential probiotic Saccharomyces cerevisiae strains and its relation with cell wall thickness

Aims: To examine Saccharomyces cerevisae strains with previously reported beneficial properties and aflatoxin B₁ binding capacity, for their ability to remove ochratoxin A (OTA) and zearalenone (ZEA) and to study the relation between cell wall thickness and detoxificant ability of yeast strains. Methods and Results: A mycotoxin binding assay at different toxin concentrations and the effect of gastrointestinal conditions on mycotoxin binding were evaluated. Ultrastructural studies of yeast cells were carried out with transmission electronic microscopy. All tested strains were capable of removing OTA and ZEA. Saccharomyces cerevisiae RC012 and RC016 showed the highest OTA removal percentage, whereas RC009 and RC012 strains showed the highest ZEA removal percentages. The cell diameter/cell wall thickness relation showed a correlation between cell wall amount and mycotoxin removal ability. After exposure to gastrointestinal conditions, a significant increase in mycotoxin binding was observed. Conclusions: All tested Saccharomyces cerevisiae strains were able to remove OTA and ZEA, and physical adsorption would be the main mechanism involved in ochratoxin A and ZEA removal. Gastrointestinal conditions would enhance adsorption and not decrease mycotoxin–adsorbent interactions. Significance and Impact of the Study: Live strains with mycotoxin binding ability and beneficial properties are potential probiotics that could be included in animal feed. Previous and present results suggest that the RC008 and RC016 strains are very promising candidates for functional feed product development.     

Mycotoxin Ochratoxin A-induced cell death and changes in oxidative metabolism of Arabidopsis thaliana

We evaluated the phytotoxicity of mycotoxin ochratoxin A (OTA) from Aspergillus and Penicillium strains on Arabidopsis thaliana. The results demonstrate that the growth of Arabidopsis thaliana on media containing OTA was inhibited significantly. Moreover, OTA induced necrotic lesions in detached leaves, which are reminiscent of hypersensitive response lesions that are activated during plant–pathogen interactions and other abiotic stress factors. From our study, we can see that OTA exposure stimulated a biphasic oxidative burst in the leaves, resulting in the generation of hydrogen peroxide (H₂O₂) and superoxide anion radicals (O₂^·−) and in the concomitant down-regulation of antioxidant enzyme defense responses and up-regulation of lipid peroxidation. These results suggested that OTA damage might result from reactive oxygen species pathways. Our experiments provide a useful model plant system for research on OTA-induced plant cell death.     

Effect of cocoa pod age on egg distribution and egg parasitism of the cocoa pod borer in Malaysia

The effects of pod age on oviposition, and on egg and larval survival of the cocoa pod borer (CPB), Conopomorpha cramerella (Snellen), were investigated. CPB were allowed to oviposit on a batch of 50 pods for a period of 7 days over several exposure dates. Numbers of CPB eggs/pod and survival to larval emergence from the pods were recorded. Pods were harvested when ripe and analyzed for larval survival within the pod. The oviposition pattern varied with pod age. Older pods less than 7 weeks before ripening (WBR) were preferred, suggesting greater nutritional value or chemical attraction of the older pods. The proportion of eggs parasitized by Trichogrammtoidea bactrae fumata Nagaraja (TBF), the number of eggs lost and the proportion surviving to larval emergence from the pod were independent of pod age. Egg parasitism was low, implying that TBF was not an effective natural enemy. Mass-releases of TBF should be intensified during cropping periods with higher proportions of susceptible pods (<10 WBR). Survival of larvae in the presclerotic layer (preSCL) was high and independent of pod age. Larval mortality at the SCL (sclerotic layer) was dependent on pod age and was high in older pods. Larval density influenced the survival of larvae inside the pod and was dependent on pod age. Larval mortality associated with SCL hardness and thickness is one of the criteria for the selection of CPB resistant cocoa clones.     

Application of chemical and biological agents for the management of frosty pod rot (Moniliophthora roreri) in Costa Rican cocoa (Theobroma cacao)

This article describes two field trials carried out at La Lola, Costa Rica, to assess control measures against frosty pod rot of cocoa (Theobroma cacao) caused by Moniliophthora (Crinipellis) roreri. In the first, factorial, trial the control agents were applied using motorised mistblowers (MMs) and hydraulic sprayers fitted with a narrow angle cone nozzle. There was an interaction between agents and application methods; together with previous application data for the most active fungicide (copper hydroxide), these trials indicate that best yields are achieved with sprays that maximise deposits on pods. We describe the droplet size spectra produced by a Stihl SR400 MM under a range of conditions because this has become the standard method of fungicide application in this series of trials at La Lola. The factor that had the largest effect on droplet size spectrum was the presence or the absence of a detachable baffle plate in front of the air‐shear nozzle. In both trials described here, MMs were fitted with baffle plates, a formulation pump and restrictor transmitting 550 mL min^?1 to deliver an estimated equivalent of 190 L ha^?1. Copper hydroxide as prophylactic applications at 1500 g a.i. ha^?1 have, to date, shown the most consistent (but incomplete) improvement in healthy pod yield. Use of copper fungicides may be cost effective when farm‐gate cocoa prices exceed approximately $1.25 kg^?1. In these trials, isolates of the hyperparasitic fungi Clonostachys byssicola and Trichoderma asperellum and two off‐patent triazole fungicides (bitertanol and triadimenol) made no significant improvement to healthy yields. The systemic oxathiin fungicide flutolanil, at a dosage of 300 g a.i. ha^?1, appears to protect pods substantially at early stages but gives proportionately less control of M. roreri than copper at later stages of pod development.     

Toxicity of ochratoxin A in aBrevibacillus brevis — Growth inhibition assay

Ochratoxin A (OTA) is a nephrotoxic, carcinogenic and immunosuppressive mycotoxin. It can be detoxified by various microorganisms, e.g. different yeast strains, via metabolisation into ochratoxin α (OTα). Within this study a growth inhibition assay was developed to compare the toxicity of OTA and its degradation product OTα. As an indicator organismBrevibacillus brevis was used. The assay was performed in microtiterplates. Growth inhibition was determined by comparing the optical density values ofBrevibacillus brevis cultures grown in medium supplemented with OTA/OTα and OTA/OTα-free medium, respectively. It could be shown thatB. brevis is sensitive to OTA (EC₁₀₀=0.5 mg/L±0.03 mg/L), which is not the case for its metabolite OTα. Therefore this bioassay is a useful tool to show the detoxification of OTA to OTα by microbial degradation. Presented at the 25^th Mykotoxin Workshop in Giessen, Germany, May 19–21, 2003     

Ochratoxin A in airborne dust and fungal conidia

Farm workers are often exposed to high concentrations of airborne organic dust and fungal conidia, especially when working with plant materials. The purpose of this investigation was to study the possibility of exposure to the mycotoxin ochratoxin A (OTA) through inhalation of organic dust and conidia. Dust and aerosol samples were collected from three local cowsheds. Aerosol samples for determination of total conidia and dust concentrations were collected by stationary sampling on polycarbonate filters. Total dust was analysed by gravimetry, and conidia were counted using scanning electron microscopy. A method was developed for extraction and determination of OTA in small samples of settled dust. OTA was extracted with a mixture of methanol, chloroform, HCI, and water, purified on immunoaffinity column, and analysed by ion-pair HPLC with fluorescence detection. Recovery of OTA from spiked dust samples (0.9–1.0 μg/kg) was 74% (quantitation limit 0.150 μg/kg). OTA was found in 6 out of 14 settled dust samples (0.2–70 μg/kg). The total concentration of airborne conidia ranged from < 1.1 × 10⁴ to 3.9 × 15⁵ per m³, and the airborne dust concentration ranged from 0.08 to 0.21 mg/m³. Conidia collected from cultures of Penicillium verrucosum and Aspergillus ochraceus contained 0.4–0.7 and 0.02–0.06 pg OTA per conidium, respectively. Testing of conidial extracts from these fungi in a Bacillus subtilis bioassay indicated the presence of toxic compounds in addition to OTA. The results show that airborne dust and fungal conidia can be sources of OTA. Peak exposures to airborne OTA may be significant, e.g., in agricultural environments. This revised version was published online in August 2006 with corrections to the Cover Date.     

Cocoa canker caused by Phytophthora palmivora

Canker disease of Theobroma cacao L. caused by Phytophthora palmivora (Butler) Butler gives rise to serious damage in Fiji and is common in the S.W. Pacific area. Fusarium decemcellulareBiick (imperfect stage of Calonectria rigidiuscula (Berk. & Br.) Sacc.) was constantly associated with P. palmivora in disorders of cocoa but was not found to be a primary pathogen. Most Amelonado cocoa trees were resistant to canker but there were differences in the reactions of other Forastero and of Trinitario types. The F₁ generation of a local Trinitario selection, itself resistant, showed segregation resulting in a proportion of very susceptible plants. Isolates of P. palmivora from black pod, canker and green twigs were equally pathogenic, both to pods and seedling stems. Sporangial dimensions, particularly the length/breadth ratio, differed from those reported elsewhere. It is uncertain whether the reason that cocoa canker is more serious in some countries than others is because of differences in the pathogenicity of P. palmivora strains to bark or the genetic differences of the cocoa grown.     

Geometric isomers of sex pheromone components do not affect attractancy of Conopomorpha cramerella in cocoa plantations

The cocoa pod borer (CPB), Conopomorpha cramerella (Snellen), sex pheromone was previously identified as a blend of (E,Z,Z)‐ and (E,E,Z)‐4,6,10‐hexadecatrienyl acetates and corresponding alcohols. These pheromone components were synthesized by modification of an existing method and the relative attractiveness of synthetic blends that included different levels of non‐target pheromone components and chemical purities was tested in a cocoa field using Delta traps. Male captures were not significantly different among traps baited with pheromone blends containing 5% to 47% (based on four identified pheromone components) of other geometric acetates [(E,Z,E)‐, (Z,Z,Z)‐, (Z,E,Z)‐ and (Z,E,E)‐4,6,10‐hexadecatrienyl acetates], indicating that C. cramerella males did not discriminate among the pheromone components and other geometric isomers in the blends. Therefore, neither antagonistic nor synergistic effects from other pheromone geometric isomers were observed. The modified synthetic pathway offers the prospect of more economical production of CPB sex pheromone. During 17 weeks when C. cramerella monitoring coincided with the main cocoa pod harvest period in 2013–2014, CPB trap catch data from some blends showed a good correlation with the number of pods with C. cramerella infestation symptoms.     

Occurrence of ochratoxin A in spices commercialized in Abidjan (Côte d’Ivoire)

Ochratoxin A (OTA) is a mycotoxin produced mostly by several species of Aspergillus and Penicillium. OTA is nephrotoxic in all animal species in which it has been tested and is cancerogenic in rodents. It is associated with Balkan endemic nephropathy. It is naturally present in many crop products such as cereals (barley, wheat, maize) and dried fruits, spices, coffee, wine, olives, and cocoa. The aim of this study was to assess the contamination of three Ivoirian spices with OTA (ginger, chili, and pepper) widely consumed by the population. A total of 90 spice samples (ginger: n?=?30; chili: n?=?30; pepper n?=?30) was taken from various sales outlets of Abidjan. OTA was quantified using an HPLC apparatus coupled with a fluorimetric detector. The chili and ginger samples were contaminated with OTA at a mean concentration of 57.48?±?174 and 0.12?±?0.15 μg/kg, respectively. No contamination of the pepper samples was detected. Eight (26.67 %) of the chili samples exceeded the maximum limit of 15 μg/kg established by European regulation. These results should serve as an alert on the risk to the consumer population of these products that are highly contaminated with OTA.     

The mechanisms involved in ochratoxin A elimination by Yarrowia lipolytica Y‐2

Biological control of mycotoxin in cereals, fruits and vegetables have emerged as a promising method. In a previous study, Yarrowia lipolytica Y‐2 isolated by our research team showed biocontrol effect on the post‐harvest decay of grapes and ochratoxin A (OTA) elimination in polytoma medium. The aim of this study was to elucidate the possible mechanisms of OTA elimination by Y. lipolytica Y‐2. The results indicated that OTA elimination by Y. lipolytica Y‐2 was attributed to the degradation action of intracellular enzymes but not extracellular enzymes. A degradation product was identified as ochratoxin alpha (OTα) by liquid chromatography‐tandem mass spectrometry. The intracellular enzymes precipitated with 65% saturation of ammonium sulphate degrade OTA the most quickly and 97.2% OTA was degraded within 4 h. Analysis of this fraction showed that two proteins of carboxypeptidase were expressed in Y. lipolytica Y‐2 but not in Y. lipolytica Polh without the ability to degrade OTA. The results of the protein identification combined with product identification indicated that OTA was degraded to OTα by Y. lipolytica Y‐2 through the hydrolysis activity of carboxypeptidases. Additionally, many proteins of Y. lipolytica Y‐2 involved in stress response and reactive O₂ species elimination also played essential role in OTA degradation.