Penicillium Species Isolated From Cocoa, Coffee Beans, and Dried Cassava in Yogyakarta Indonesia and Their Ochratoxin Production

The presence of Penicillium in cocoa and coffee beans, and dried cassava are detrimental due to its ability in ochratoxin A (OTA) production which carcinogenic and mutagenic to human. Objectives of this study were to isolate and identify Penicillium from cocoa, coffee beans and dried cassava in Yogyakarta by morphological and molecular characteristics, as well as to observe the ability of these isolates in OTA production on Yeast Extract Sucrose Agar (YES) medium. In this study, morphological characteristics were mainly based on the growth of isolates on identification media, while molecular characteristics were based on the similarity of PCR products using ITS4 and ITS5 as primers. OTA was detected by ELISA and UPLC methods. The result showed that 15 of 16 representative isolates obtained during this study were identified as Penicillium citrinum, one of the representative isolate from cocoa beans was identified as Penicillium paneum. Surprisingly, 13 among 15 of the obtained P. citrinum isolates from cocoa and coffee beans were positive in the production of OTA in YES medium, at the concentration of 4.64 to 25.26 μg/g media, while OTA was not detected in YES grown media by P. paneum and two isolates of P. citrinum from dried cassava. Conclusion of this study, the most found species Penicillium in cocoa and coffee beans were belong to P. citrinum which likely have a capability in the production of OTA.


Introduction
Cocoa and coffee beans, and cassava have an important role in Indonesia, especially in Yogyakarta, as a source of economic income. Mostly, drying and storage of cocoa, coffee beans and cassava are still using traditional method, such as drying the products out in the yard under the sun and storing in a humid warehouse condition. These suboptimal drying and storage processes render the products to be easily contaminated by filamentous fungi.
The presences of toxicogenic Penicillium species are generally responsible for the toxin content, even when fungal contamination were not observable anymore. Fungal contamination is suspected due to quite high moisture content (> 8%) and the slow sun drying process of the commodities. Several species of Penicillium are known to contaminate foods such as P. verrucosum on dried foods and cereals, P. nordicum on fermentation of meat and cheese which potential in OTA production (Larsen et al., 2001;Castella et al., 2002;Cabanes et al., 2010).
International Agency for Research on Cancer (IARC, 1982) mentioned that some Penicillium species capable of producing ochratoxin A (OTA) are dangerous because carcinogenic and mutagenic effects to humans and animals that consumed the contaminated food or feed. Recently the European Commission stated 2 g and 1 mg for the maximum level of OTA in raw cacao and the processed products (Martinez-Culebras et al., 2008).
Study on OTA contamination during fermentation and drying of cocoa beans at Ivory Coast, Dano et al., (2013) reported that OTA was detected in beans at all stages of post-harvesting operations, such as pod-opening, fermentation, drying, and storage. Our previous paper showed that black Aspergillus was the dominant species of Aspergillus in cocoa and coffee beans, and we found that A. carbonarius was the OTA producer among the black Aspergillus (Nugroho et al., 2013). Our concern in this study was to characterize the species of Penicillium that also contaminate cacao and coffee beans, and dried cassava in Yogyakarta and to determine their potency in the production of OTA.

Isolation of Penicillium
Filamentous fungi Penicillium were isolated from cocoa and coffee beans, and cassava that have undergone drying process. Samples were collected from several farmers, 1-3 kg of samples were collected each of them. Cocoa and coffee beans samples were collected from Kulonprogo, while the dried cassava samples from Gunung Kidul, Yogyakarta. All of the samples were plated by direct and dilution methods as described by Samson et al., (2004). Media for plating were Dichloran Rose Bengal Chloramphenicol Agar (DRBC, Oxoid) and Dichloran Gliserol-18 Agar (DG-18, Oxoid). The plates were incubated for 7 days in room temperature. Pure culture of Penicillium were transferred to Malt Extract Agar (MEA) slant tubes, incubated for 7 days, and were stored at 4 °C for further study.

Identification of Fungi
Primary identification of Penicillium isolates was done based on macroscopic and microscopic characteristics . Single spore of the isolates were inoculated at three points of identification media, and incubated for 7 days at 25°C. Identification used media were Malt Extract Agar (MEA), Czapek Yeast Extract Agar (CYA), and Yeast Extract Sucrose agar (YES). Isolates were then grouped according to their similarity of macroscopic and microscopic characteristics, and one was selected from each group as representative isolates for further study.

Extraction of DNA and PCR Amplification
Molecular analysis for species identification of Penicillium isolates was based on the similarity of amplicon using ITS4 and ITS5 as primers. Isolates were grown in Potato Dextrose (PD) broth for 7 days to obtain genomic DNA. The mycelia were harvested and grinded into fine powder after freezing with liquid nitrogen. DNA was extracted with DNeasy Plant Qiagen, Valencia (followed manufacturer's directions). DNA was then amplified by PCR using ITS4 and ITS5 as primers (Bisbal et al., 2009). PCR amplification was performed with a volume of 25 μL consisting of GoTaq® 5x buffer (Promega), GoTaq® DNA polymerase (5 μ/μL) (Promega), 25 mM MgCl2, dNTP (Promega) for each primer and DNA isolates.
PCR analysis was performed using a thermal cycler (Peltier PTC-100) with the parameters of 1 cycle of pre-denaturation 5 min at 95°C; 30 cycles consisting of 30 s at 95°C for denaturation, 30 s at 58°C for annealing, and 1 min at 72°C for extension; followed by 1 cycle of final extension 5 min at 72°C (Bisbal et al., 2009). PCR products were separated by electrophoresis using 1% agarose; containing Ethidium Bromide (EtBr) for staining. DNA bands were visualized under ultraviolet (UV) light, and sizes were estimated by comparison with a DNA standard length (TM GeneRuler 1kbp DNA ladder, MBI Fermentas, Lithuania).

Sequencing Analysis
PCR products were purified using the QIAquick PCR Purification Kit, Qiagen, Valencia (followed the manufacturer's instruction). Nitrogenous bases sequencing of purified products were performed in the Center of Chemical Biology, Universiti Sains of Malaysia (USM). All of the DNA sequence of the isolates were matched for similarity using BLAST to NCBI. All the isolates were also analyzed for network joining (NJ) using bootstrap method and the maximum composite likelihood with Mega version 4.0 software. Penicillium maleagrinum Angg-PBB6-13 was used as an out-group.

Production of OTA
Isolates were grown in YES medium and incubated for at 25 °C for 7 days to determine the ability of Penicillium isolates in producing OTA (Accensi et al., 2004). Three agar plugs were removed from different points of the colony (about 5 g) from each culture. Samples were analyzed for OTA using RIDASCREEN ® OTA ELISA test kits (r-BioPharm). All sample preparation and test procedures followed the manufacturer's instructions. Limit of detection is about 2.5 ppb. For double check, OTA was also detected by UPLC. Three agar plugs of YES medium grown by isolates were removed and extracted by 0.5 mL of methanol. After centrifugation, the supernatant was analyzed by utilizing UPLC.

Results and Discussion Isolation of Penicillium
This study was succeeded in obtaining 67 isolates of the filamentous fungi of Penicillium in cacao and coffee beans, and cassava samples. The isolates were grouped according to their colony appearance in MEA, CYA and YES, and from each group isolates were taken as their representative for further study.
Fungal contamination in cacao and coffee beans, and cassava were further characterized by a w value of samples at a range between 0.81 and 0.94. The a w values were optimum for the growth of species of the genera Aspergillus and Penicillium which have minimum a w requirements of 0.75 to 0.86 (Pitt and Hocking, 2009;Esteban et al., 2006).
Isolate Y-PC-12 has a very rapid growth on MEA medium (37-47 mm diameter), dull-green colored colonies, and pale reverse. In CYA medium the colony diameter was 28-29 mm, pale green to dark blue green with velvety surface, biverticillate conidiophores, rough stipa, rough-walled phialide, smooth-walled globose shaped conidia. Isolate Y-PC-12 was suspected as Penicillium paneum. Odserve and reverse appearance of Penicillium colonies and Penicillium maleagrinum AngPBB6-13 was used as control (negative OTA producers) in this study as presented in Fig. 1.
Variation in size of the amplified fragment was caused by the rDNA length of the ITS regions. Variation of ITS sequences in this region allows the use of this region for phylogenetic study of many different organisms (Ahmed et al., 1999). Some very significant variations can occur because of deletions, insertions, or substitutions between species in the ITS (Peay et al., 2008). The similarity for species identification was done by comparing the sequence of nucleotide bases in the ITS region of isolates with ITS region of Penicillium species that exist in the GenBank, then the sequence of nucleotide bases in http://blast.ncbi.nlm.nih.gov/Blast were read using BLAST. BLAST results and comparison of isolates of Penicillium nitrogenous base sequences in the ITS region (% of similarity) can be seen in To support the results of morphology and BLAST analysis, the species similarity analysis can be done by analyzing the sequence of nitrogenous bases with a multi-alignment of ITS regions. The results of sequence alignment Penicillium that based on the nucleotide bases Multalin can be seen in Fig. 4. Base sequence of isolate Y-PK-6, Y-PK-9, Y-PK-14, Y-PG-4, Y-PK-10, Y-PK-24, Y-PK-20, Y-PK-16 , Y-PG-9, Y-PK-17, Y-PK-31, Y-PK-7, Y-PC-3, Y-PC-5 were different from Y-PC-12 and Ang-PBB6-13 in sequence number 64, 121-123, and 222-260 (Fig. 3). Fadıloglu and Erkmen (1999) suggested the possibility that differences in amino acid affect the shape, structure, and reactivity of these proteins.
The alignment performed with the aim to determine the degree of homology of DNA base sequences analysis, to find out the differences and similarities as well as the sequence of nucleotide base composition (Whitford, 2005). The alignment results showed a high degree of homology between the samples studied. Due to the nature of ITS4 and ITS5 variation; this juxtaposition appears in the gap (indicated by dashed lines). The existence of this gap indicates a high occurrence of mutations either in the form of insertions and deletions (Hidayat et al., 2008). Nevertheless the results of Clustal-W alignment previously demonstrated a high degree of homology (96-99%) between genes in the cluster.

Ochratoxin A Producing Ability
All 16 selected isolates were tested for their ability to produce OTA. The results are shown in Table 2. Ochratoxin analysis using ELISA (LOD = 2.5 μg/g) from YES agar medium show that 13 isolates of isolated Penicillium from cocoa and coffee beans were able to produce OTA from 4.65 to 25.26 µg/g media. These isolates were supposed to be P. citrinum. Two isolates also supposed to be P. citrinum (Y-PG-4 and Y-PG-9) from dried cassava, while one isolates Y-PC-12 that supposed to be P. paneum from cocoa beans have no capability in producing OTA. P. paneum known to be non OTA producer (Mounjouenpou et al., 2008), however, there is no publication at the moment that mentioned P. citrinum has a capability in producing OTA, only two species of Penicillium, i.e., P. verrucosum and P. nordicum in this study, detection of OTA was double checked with UPLC, and the result was similar to ELISA methods.
The used method for determination of OTA was ELISA. ELISA is a reliable and rapid method for OTA determination in various kinds of commodities (Fujii et al., 2007;Matrella et al., 2006;Zheng et al., 2005). ELISA was preferred because it requires less sample volume and sample preparation is faster than other conventional methods such as TLC and HPLC (Zheng et al., 2005). This method is very sensitive to mycotoxins that was extracted from the natural matrix (White and Johnson, 2003).  Y-PK-6 Coffee beans 4,72 5.
Ang PBB6-13 as control < LOD However, it should be noted that ELISA method uses enzymes, which is quite complex because its activity is influenced by various factors that can interfere with the binding of toxins and mycotoxins in samples, causing overor underestimation (Asihara and Kasahara, 2001). Visconti et al., (1999) mentioned that problems of OTA analysis in red wine may be due to the presence of several anthocyanin and other pigments that interfere with OTA-binding to the antibody so as to provide false positive results. In this study the above problems were not found as components of YES media avoid the potential of crosslinking.

Conclusion
Penicillium citrinum was the dominant contaminant in cacao and coffee beans, and cassava. Macroscopic, microscopic, and molecular level data were required to be able to accurately identify the species of Penicillium. Nitrogen base sequences of ITS region proved specific and sensitive enough to distinguish species of Penicillium.
In this study 13 among 15 of P. citrinum isolates obtained from cocoa and coffee beans were positive in the production of OTA in YES medium, at the concentration of 4.64 to 25.26 µg/g media, while OTA was not detected in YES grown media by P. paneum and two isolates of P. citrinum from dried cassava.