Study of ochratoxin A-producing strains in coffee processing

International Journal of Food Science and Technology 2004, 39, 501 507 501 Study of ochratoxin A-producing strains in coffee processing Mirna Suárez-Quiroz, 1 Oscar González-Rios, 1 Michel Barel, 2 Bernard Guyot, 2 Sabine Schorr-Galindo 1 & Joseph-Pierre Guiraud 1 * 1 UMR-IR2B (ENSAM/INRA/UM2), Universite Montpellier II, Place Eugène Bataillon, 34095 Cedex 5, France 2 CIRAD-CP, TA80/16 34398 Montpellier Cedex 5, France (Received 15 July 2003; Accepted in revised form 21 November 2003) Summary Keywords Ochratoxin A (OTA) is the main mycotoxin that has been detected in coffee. The occurrence of OTA in coffee beans can be because of environmental conditions and/or processingconditions. Three coffee processes were evaluated (wet, mechanical and dry processes), at different stages from harvesting to storage, and fungi producing OTA were enumerated and identified. The frequency of potential OTA-producingfungi and their ability to produce OTA was also studied. By direct plating, the levels of contamination found in the coffee processes were 80, 72 and 92%, respectively, for parchment and dry cherry coffee and 20, 34 and 15% for green coffee. Aspergillus ochraceus isolated from the three processes accounted for 6.6, 8.3 and 3.3%, and Aspergillus niger for 15, 13 and 25% of the strains isolated, respectively. The toxigenic potential of five A. ochraceus and two A. niger strains was tested in FDA medium and coffee medium usingthe HPLC technique. There was no difference between the processes studied in terms of isolation and occurrence of ochratoxigenic fungi. Aspergillus niger, Aspergillus ochraceus, coffee mycoflora, fungi, post-harvest process. Introduction Ochratoxin A is a secondary metabolite produced by several toxigenic species of Aspergillus and Penicillium. OTA has been shown to exhibit nephrotoxic, immunosuppressive, teratogenic and carcinogenic properties (Ho hler, 1998). The International Agency for Research on Cancer (IARC) has classified OTA as a possible human carcinogen (group 2B), based on sufficient evidence for carcinogenicity in experimental animal studies and inadequate evidence in humans (IARC, 1993). OTA frequently occurs as a contaminant in cereals worldwide and it is the main mycotoxic contamination reported to be found in coffee (Le- Bars & Le-Bars, 2000). Aspergillus ochraceus is thought to be the most important OTA-producing mould in relation to coffee beans (Pitt, 2000). Recently, it was suggested that Aspergillus niger *Correspondent: Fax: +33 467 144292; e-mail: guiraud@arpb.univ-montp2.fr and closely related species such as Aspergillus carbonarius could be OTA producers in coffee (Bucheli et al., 2000; Frank, 2001; Joosten et al., 2001). The presence of OTA in coffee is undesirable because it may be used as a barrier to trade, affectingthe economies of producingcountries. The European Commission has not yet fixed maximum levels for OTA in coffee, and has only given one recommendation, in which a reference level of 3 lgkg )1 was suggested to EU member states (Romani et al., 2000). Nine countries have specific regulations for OTA. Legislative limits range from 5 to 50 lgkg )1 (Boutrif, 1999). If a maximum limit for OTA in coffee were to be established, it could affect international trade for producingcountries which do not control this parameter. Without such limits (uncontrolled) this could have an impact on human health in coffee importingcountries. The occurrence of OTA in coffee beans can be due to both environmental conditions (climate, doi:10.1111/j.1365-2621.2004.00810.x

502 Ochratoxigenic strains in coffee processing M. Suárez-Quiroz et al. length of storage and transportation) and processingconditions (wet, mechanical or dry processes) (Romani et al., 2000). The occurrence and the formation of OTA in the dry process has been studied by several authors (Bucheli et al., 1998, 2000; Urbano et al., 2001). OTA was present before storage, indicating the possibility that harvestingand post-harvest handlingof coffee cherries could be the critical steps leadingto contamination (Bucheli et al., 1998, 2000). There is currently little information available on the presence of OTA-producingmoulds in coffee beans used in the wet and mechanical processes and the impact of these processes on the production or presence of OTA. Published data indicate that the depulpingprocess would significantly reduce the risk of OTA contamination duringthe subsequent fermentation and dryingsteps, but mycological studies of these processes are necessary (Frank, 2001). In order to protect coffee from OTA formation, there is a real need to identify moulds able to produce this mycotoxin, and their relation with the processingmethod. This research was done so as to enumerate and identify fungi colonizing coffee beans and to determinate the frequency of potential OTAproducingfungi in wet, dry and mechanical processes at different stages from harvesting to storage. Their ability to produce OTA was also studied. Materials and methods Coffee samples Coffee cherries (Coffea arabica var. bourbon, typica and catimor) were harvested by manual pickingduringthe 2000 2001 harvest peaks in a plantation in the Coatepec area (Xalapa, Mexico). Coffee processing Three coffee processingmethods, wet, mechanical and dry, were tested on 40-kgbatches of coffee cherries. The trials were repeated four times. The wet process consisted of the removal of the outer skin by a mechanical disk pulper and mucilage by microbial action (fermentation), washingwas by hand. In the mechanical process, a conic vertical coffee pulper (Penagos Hnos & CIA LTDA, Colombia) was used and the mucilage was removed mechanically with an ascendant vertical mucilage remover (Penagos Hnos & CIA LTDA, Colombia). In the dry process, the entire cherry was directly dried, immediately after harvest. An air dryer at 45 C was used for the three methods (rather than the usual sun drying, to ensure more uniform drying). The first two processes gave parchment coffee and the other gave dry cherry coffee. Parchment and husk were removed by mechanical hulling. Sampling Samples (300 g) for mycological analysis were collected for each method of coffee processingat different stages of processing from fresh cherries to storage (Fig. 1). Samples (300 g) of parchment coffee stored for 1 year were collected usingthe same process. Mycological analysis Subsamples of cherries or coffee beans were stirred for 5 min at room temperature in 0.1% peptone water (1:10 w/v) and 0.1 ml was spread-plated onto Dichloran Rose Bengal Chloramphenicol agar (DRBC) (Samson, 1991). The results were expressed in CFU g )1. Subsamples (fifty beans) of parchment coffee, dry cherry or green coffee were also plated directly onto Petri dishes (ten beans per plate) containingdichloran 18% Glycerol agar (DG18) (Hocking& Pitt, 1980; Guiraud, 1998) without prior superficial disinfection. The plates were incubated at 25 C for 5 7 days, the results beingexpressed as a percentage of infected beans (% infection). The moulds detected were isolated and subcultured on Czapek Dox agar for identification purposes. Predominant moulds were identified accordingto the identification key for common food-borne fungi (Samson et al., 1995) and the strains recognized as A. ochraceus Wilhelm and A. niger van Tieghem were sent to the BCCM TM /MULC Culture Collection (Leuven Catholic University, B-1348 Louvain-la-Neuve, Belgium) for confirmation of the identity. Isolates identified as A. ochraceus and A. niger were counted for each sample, and percentage occurrences were calculated as a proportion of total fungi. International Journal of Food Science and Technology 2004, 39, 501 507

Ochratoxigenic strains in coffee processing M. Suárez-Quiroz et al. 503 Figure 1 Coffee processing tested for ochratoxigenic fungi from southeastern Mexico (*sampling). OTA production by potential producers Isolates of A. ochraceus and A. niger were grown on rice media accordingto the methodology recommended by FDA (Tournas et al., 2001). OTA production was then detected and quantified by HPLC (Nakajima et al., 1997). In order to study OTA production in coffee, A. ochraceus and A. niger were grown on PDA agar, ph 3.5, at 25 C for 5 days. Spores were suspended in 0.01% Tween 80 solution and 40 gof coffee beans, obtained by the wet and mechanical processes, were inoculated with a sample of 15 ml (43 10 6 conidia ml )1 ) and incubated at 25 C for 15 days (Mantle & Chow, 2000). OTA production was determined by HPLC (see below). OTA analysis in coffee beans Coffee samples were frozen at )80 C then ground to pass through a 0.5-mm sieve and analysed for OTA (Nakajima et al., 1997). The samples were extracted for 30 min with a solution of methanol/ 3% sodium bicarbonate (50:50), the extracts were filtered and diluted with phosphate-buffered saline and applied to an immunoaffinity column (OchraprepÒ,Rhoˆ ne Diagnostics, Scotland). OTA was eluted with 3-mL HPLC grade methanol. The eluate was evaporated to dryness under a stream of nitrogen at 70 C and the residue was redissolved in 1 ml of HPLC mobile phase and then quantified by HPLC (Shimadzu LC-10ADVP, Japan, with a fluorescence detector). The mobile phase consisted of distilled water/acetonitrile/ International Journal of Food Science and Technology 2004, 39, 501 507

504 Ochratoxigenic strains in coffee processing M. Suárez-Quiroz et al. glacial acetic acid (51:48:1). The flow rate was 1 ml min )1. OTA was detected by absorption at 333 nm excitation and 460 nm emission, and a retention time of 13.3 13.5 min. Standard OTA curves were established with an ochratoxin standard (1000 ngml )1 ; ref PD 226 R. Biopharm Rhoˆ ne Ltd, Scotland); the detection limit was 0.075 ngml )1. Results Evolution of mould flora during the processes Table 1 shows the mould counts in beans collected from the three processes at different stages. Mould counts reached maximum values of 4.3 ± 0.9 10 2 CFU g )1 in cherries. There was a qualitative uniform distribution of filamentous fungi in samples from the coffee processes tested. All fungi genera identified had already been recorded in coffee bean samples from Brazil (Silva et al., 2000; Urbano et al., 2001; Batista et al., 2003). No strains of OTA-producingfungi were isolated from any samples at this stage, but this may have been because of the detection limit of the technique used. Consequently, the use of direct plating of beans in Petri dishes on DG18 agar for samples giving a negative result (Table 2) was employed. A high level of infection by moulds was observed in all processes after drying(on parchment and dry cherries). It was found that 80% of the beans from the wet process, 72% from the mechanical process and 92% from the dry process, analysed as described earlier, were contaminated with fungi, Penicillium, Mucor, Cladosporium and Aspergillus spp. includingwell-known potential OTA-producingfungi (A. ochraceus and A. niger). However, the percentage of infection in green coffee was lower than that observed in parchment and in dry cherry coffee, and no OTA-producingfungi were found. Table 2 shows the percentage occurrence of A. ochraceus and A. niger isolated from parchment, dry and green coffee for each process; this corresponded to the analysis of 200 beans at each stage of processing. One strain of A. ochraceus was also isolated from parchment coffee produced by the wet process and stored for 1 year. OTA production The toxigenic potential of the isolated fungal strains was tested and compared to A. ochraceus Table 1 Total mould count (CFU g )1 ) a on DRBC agar and main isolates in coffee beans collected at different stages of processing Process Stage of processing Mould count b (CFU g )1 ) Main strains isolated Wet process Cherries 4.3 ± 0.9 10 2 Penicillium citrinum, Mucor hiemalis, P. glabrum, P. minioluteum Pulped beans 1.8 ± 1.1 10 2 Penicillium citrinum, P. minioluteum Fermented beans 1.1 ± 0.8 10 2 Geotrichum candidum Parchment coffee Mechanical process Cherries 4.3 ± 0.9 10 2 Penicillium citrinum, Mucor hiemalis, Penicillium glabrum, Penicillium minioluteum Pulped beans 1.3 ± 0.7 10 2 Penicillium citrinum Beans with mechanically 7.5 ± 8.6 10 Geotrichum candidum removed mucilage Parchment coffee Dry process Cherries 4.3 ± 0.9 10 2 Penicillium citrinum, Mucor hiemalis, P. glabrum, P. minioluteum Dry cherries <50 b <50 b a Spread-plated. b Mean of four samplings. c Limit of method. International Journal of Food Science and Technology 2004, 39, 501 507

Ochratoxigenic strains in coffee processing M. Suárez-Quiroz et al. 505 Table 2 Rate of coffee bean infection by moulds and percentage occurrence of Aspergillus species potentially capable of producingochratoxin A, from different coffee processes % occurrence Process Stage of processing % infection a A. ochraceus A. niger Other species isolated Wet process Parchment coffee 80 6.6 15 Penicillium glabrum, Mucor sp., Cladosporium sp., Aspergillus fumigatus, A. candidum, A. tamarii, Botrytis sp., P. italicum 20 0 0 Eurotium sp., Fusarium sp., Aspergillus fumigatus Mechanical process Parchment coffee 72 8.3 13 Mucor sp, Aspergillus fumigatus, Penicillium citrinum, P. glabrum, P. granulatum 34 0 0 Eurotium chevalieri, Aspergillus fumigatus, Cladosporium cladosporoides Dry process Dry cherry 92 3.3 25 Penicillium glabrum, Mucor sp., Cladosporium sp., Penicillium sp. 15 0 0 Penicillium minioluteum, Eurotium sp. a Direct plating of 200 beans on DG18 agar. MULC 39534 BCCM TM /MULC Culture Collection (Table 3). Of the twelve strains of A. ochraceus detected, five were chosen for their morphological differences and tested for their ability to produce OTA: three strains were positive after culture on rice and coffee media. Production ranged from 0.3 to 679 lgkg )1 of OTA on rice medium and from 1 to 697 lgkg )1 on coffee beans. From the thirtynine strains of A. niger isolated, two strains, chosen at random, were tested for the production of OTA in rice and coffee media and all gave positive results, ranging from 2.2 to 114 lgkg )1 of OTA. No strains of A. carbonarius and P. verrucosum were isolated in this study. Table 4 shows the levels of OTA (lgkg )1 ) from samples testingpositive for A. ochraceus and A. niger. For all samples, the analysis was done after hullingon green coffee, parchment or husk. The presence of OTA was correlated with the presence of OTA-producingstrains. Discussion The mould counts from coffee were generally low and no well-known potential OTA-producing fungi could be isolated before drying. Similar observations were made by several authors (Frank, 1999; Bucheli et al., 2000), who failed to isolate ochratoxigenic moulds from fresh cherries; this was attributed to the presence of species adapted to high moisture (yeast, bacteria or other moulds) that may inhibit their growth. However, a Table 3 Comparison of OTA production in different media by isolates of Aspergillus ochraceus and Aspergillus niger from different coffee processingmethods Isolate Origin OTA production (lg kg )1 ) FDA rice medium Coffee bean a medium Coffee bean b medium A. ochraceus Reference strain 19 1.8 1 A. ochraceus 232 Wet process nd 2.3 1.3 A. ochraceus 243 Wet process nd nd nd A. ochraceus 237 Mechanical process 0.3 2.6 1.3 A. ochraceus 246 Mechanical process nd nd nd A. ochraceus 259 Wet process 679 697 188 (1999 harvest) A. niger 250 Wet process 5.7 6.2 9.8 A. niger 296 Dry process 114 5.1 2.2 a Beans from mechanical process. b Beans from wet process. nd, not detected. International Journal of Food Science and Technology 2004, 39, 501 507

506 Ochratoxigenic strains in coffee processing M. Suárez-Quiroz et al. Table 4 Relationship between levels of OTA contamination in green coffee, parchment or husk and isolates of Aspergillus ochraceus and Aspergillus niger present in coffee beans Sample a Level of OTA contamination (lg kg )1 ) b Parchment or husk Species of potential OTA-producing fungi isolated 232 0.2 0.3 A. ochraceus 243 nd nd A. ochraceus 237 nd 0.1 A. ochraceus 246 nd nd A. ochraceus 259 0.2 Trace A. ochraceus 250 0.2 0.3 A. niger 296 0.2 0.3 A. niger a Sample number assigned for laboratory control. b HPLC. nd, not detected. high level of contamination by yeast and moulds, including A. ochraceus and A. niger, was observed in raw coffee beans collected in different producing regions of Brazil (Urbano et al., 2001). Independently of the method used, during processinga substantial decrease in the mould count was found, which was expressed in an apparent absence (limit of detection) of ochratoxigenic moulds in parchment and dry cherry coffee and later on in green coffee. Using the viable count methodology it was not possible to prove the existence of ochratoxigenic strains in parchment and dry cherry coffee. Nevertheless, by the direct platingtechnique, several strains of A. niger and A. ochraceus were isolated from beans. It can be seen that the average rate of parchment and dry cherry coffee infection by these fungi was low (3 25%). This method made it possible to determine the percentage of beans contaminated by moulds, which was high in parchment coffee (71 92%) decreasinglater on in green coffee (15 34%). As regards the isolation of A. ochraceus, no great difference was seen between the processes studied, but the dry method seemed to promote the presence of A. niger. A study of the toxigenic strength of the isolates demonstrated that not all of them had the same toxigenic potential. In the FDA media, large differences were observed in the ability of the isolates tested to produce OTA. Each time a strain produced OTA in FDA medium, it was found to produce OTA in the coffee medium, though for one strain of A. ochraceus, OTA production was observed in the coffee medium but not in the FDA medium: 75 and 3% of A. ochraceus and A. niger isolates, respectively, were OTA producers (Taniwaki et al., 2003). Studies of factors influencingtoxigenesis showed that toxin production can vary by as much as 1000-fold amongisolates from the same species (Le-Bars & Le-Bars, 2000). This study is in accordance with the observations made by other authors (Taniwaki et al., 2003), who reported that infection by A. ochraceus and A. niger did not necessarily indicate high OTA contamination. However, the results obtained in this study indicated that toxigenesis does not appear to be affected by the process to which the coffee is subjected. When a toxigenic strain was isolated from a coffee sample, the sample contained OTA. This shows that the presence of toxigenic strains involves a great risk of OTA presence. Contrary to what might be expected from the results in Table 2, it was shown that hullingcoffee containingota did not totally eliminate the toxin, which was subsequently found in green coffee. In this work, it was not possible to monitor changes in the toxigenic strains right from the first stage (coffee cherries). This work will be completed by a study based on pilot fermentations with artificially contaminated coffee cherries. Acknowledgments We thank Mr C.I. Herna ndez-palacios, Mexico, for providingthe coffee samples and processing facilities, UNIDA of the Veracruz Technology Institute and the LATEX Laboratory at Veracruz University for laboratory facilities. References Batista, L.R., Chalfoun, S.M., Prado, G., Schwan, R.F. & Wheals, A.E. (2003). Toxigenic fungi associated with processed (green) coffee beans (Coffea arabica L.). International Journal of Food Microbiology, 85, 293 300. Boutrif, E. (1999). Recent international developments in the field of mycotoxins. In: 18th ASIC Coffee Conference. Pp. 239 247. Helsinki, Finland. Bucheli, P., Meyer, I., Pittet, A., Vuataz, G. & Viani, R. (1998). Industrial storage of Robusta coffee under International Journal of Food Science and Technology 2004, 39, 501 507

Ochratoxigenic strains in coffee processing M. Suárez-Quiroz et al. 507 tropical conditions and its impact on raw material quality and ochratoxin A content. Journal of Agricultural and Food Chemistry, 46, 4507 4511. Bucheli, P., Kanchanomai, C., Meyer, I. & Pittet, A. (2000). Development of ochratoxin A duringrobusta (Coffea canefora) coffee cherry drying. Journal of Agricultural and Food Chemistry, 48, 1358 1362. Frank, M. (1999). Le Système HACCP et son roˆ le possible dans la maîtrise des mycotoxines: e valuation de l ochratoxine A dans la production de café. In: Troisie`me Confe renece Internationale Mixte FAO/OMS/PNUE sur les Mycotoxines. Pp. 1 13. Tunis, Tunisie, 3 6 mars. Frank, J.M. (2001). On the activity of fungi in coffee in relation to ochratoxin A production. In: 19th ASIC Coffee Conference, Trieste, Italy, 14 18 May. Guiraud, J.P. (1998). Microbiologie Alimentaire. Pp. 321 333. Paris: Dunod Ed. Hocking, A.D. & Pitt, J.I. (1980). Dichloran-glycerol medium for enumeration of xerophilic fungi from low moisture foods. Applied and Environmental Microbiology, 39, 488 492. Ho hler, D. (1998). Ochratoxin A in food and feed: occurrence, legislation and mode of action. Zeitschrift fu r Erna hrungswissenschaft, 37, 2 12. IARC (1993). Ochratoxin A. Monographs on the Evaluation of Carcinogenic Risk to Humans. Pp. 250 277. Lyon, France, IARC WorkingGroup, World Health Organization. Joosten, H.M.L.J., Goetz, J., Pittet, A., Schellenberg, M. & Bucheli, P. (2001). Production of ochratoxin A by Aspergillus carbonarius on coffee cherries. International Journal of Food Microbiology, 65, 39 44. Le-Bars, J. & Le-Bars, P. (2000). Mycotoxigenesis in Grains Application to Mycotoxic Prevention in Coffee. Coffee Biotechnology and Quality. Pp. 355 368. Sera, Netherlands: Kluwer Academic Publishers. Mantle, P.G. & Chow, A.M. (2000). Ochratoxin formation in Aspergillus ochraceus with particular reference to spoilage of coffee. International Journal of Food Microbiology, 56, 105 109. Nakajima, M., Tsubouchi, H., Miyabe, M. & Ueno, Y. (1997). Survey of aflatoxin B 1 and occhratoxin A in commercial green coffee beans by high-performance liquid chromatography linked with immunoaffinity chromatography. Food and Agricultural Immunology, 9, 77 83. Pitt, J.I. (2000). Toxigenic fungi: which are important? Medical Mycology, 38, 17 22. Romani, S., Sacchetti, G., Lo pez, C.C., Pinnavia, G.G. & Rosa, M.D. (2000). Screeningon the occurrence of ochratoxin A in green coffee beans of different origins and types. Journal of Agricultural and Food Chemistry, 48, 3616 3619. Samson, R.A. (1991). Identification of food-borne Penicillium, Aspergillus and Fusarium species. ACIAR Proceedings, No. 36. Pp. 39 45. Bangkok, Thailand, 23 26 April. Samson, R.A., Hoekstra, E.S., Frisval, J.C. & Filtenborg, O. (1995). Introduction to Food-Borne Fungi. P. 322. Delft, Netherlands: Centraalbureau voor Schimmelcultures Baarn. Silva, C.F., Schwan, R.F., Sousa-Diaz, E. & Wheals, A.E. (2000). Microbial diversity duringmaturation and natural processingof coffee cherries of Coffea arabica in Brazil. International Journal of Food Microbiology, 60, 251 260. Taniwaki, M.H., Pitt, J.I., Teixeira, A.A. & Iamanaka, B.T. (2003). The source of ochratoxin A in Brazilian coffee and its formation in relation to processingmethods. International Journal of Food Microbiology, 82, 173 179. Tournas, V., Stack, M.E., Mislivec, P.V., Koch, H.A. & Bandler, R. (2001). Bacteriological Analytical Manual Online. US Food & DrugAdministration (http:// www.cfsan.fda.gov/ebam/bam-18.html). Urbano, G.R., Taniwaki, M.H., Leitao, M.F.D.F. & Vicentini, M.C. (2001). Occurrence of ochratoxin A- producingfungi in raw Brazilian coffee. Journal of Food Protection, 68, 1226 1230. International Journal of Food Science and Technology 2004, 39, 501 507