Abstract
The aim of this study was to evaluate the impact and contribution of various drying surfaces (soil, cement, and tarpaulin) and raking frequencies (1 and 4/day) on the incidence of toxigenic molds, ochratoxin A (OTA) production, and on the overall cup quality during preparation of arabica and robusta coffee cherry in India. Two individual experimental batches (run 1 at the begin of harvest and run 2 at the end of harvest) were set up for the study. Results showed high incidence of molds in coffee dried on soil surface compared with that on cement and tarpaulin surfaces. In both arabica and robusta, OTA could be detected in Aspergillus ochraceus contaminated samples at the end of harvest. Raking of the cherries 4 times/day showed lower fungal incidence with no OTA levels detected. Overall, coffee cherry prepared by drying on tarpaulin surface with 4 rakings/day showed lower OTA and fungal incidence with good and acceptable cup quality, and this is recommended to be practiced at the farm level.
Introduction
C
Ochratoxin A (OTA) is considered as a potent nephrotoxic, immunosuppressive, carcinogenic and a teratogenic compound, which has been classified as possibly carcinogenic to humans (Group 2B) (IARC, 1987; Bhat et al., 2010). OTA is reported to be produced by Penicillium verrucosum, Aspergillus ochraceus, Aspergillus niger, and Aspergillus carbonarius and has been routinely isolated from coffee beans and its products. OTA has been reported in coffee beans processed by both wet and dry methods, wherein the concentration ranged between 0.2 and 360 μg/kg (Levi et al., 1974; Mantle, 1998; Bucheli et al., 2000; Taniwaki, 2006; Batista et al., 2009). Apart from coffee, OTA has also been detected in cereal grains, nuts, meat, spices, dried fruits, fruit juices, and fermented beverages such as beer and table wine (Heenan et al., 1998; Pateraki et al., 2007; Bhat et al., 2010).
OTA contamination can possibly occur in coffee cherries (bean processed by dry method) before harvest or might be formed as a result of recontamination after harvest. Frank (2001) reported that OTA could be primarily formed in the pulp and husk when cherries are sun dried. It is envisaged that some of the routinely employed processing methods for coffee cherry preparation such as raking frequency, surface employed for drying, and the particular stage of harvest can contribute tremendously to the growth of toxigenic molds and OTA production. Since the exact stage of OTA contamination during coffee cherry preparation is not completely understood, it is highly imperative to explore the various stages during on-farm processing.
In this view, the main aim of the present study was to evaluate the effects of drying surfaces, raking frequencies, and stage of harvest on toxigenic mold incidence and OTA contamination of arabica and robusta cherry harvested in India. The baseline information generated from this study is envisaged to be useful for accomplishing good agricultural practice (GAP) and good manufacturing practice (GMP) in the coffee growing regions of the world. To our knowledge, this is the first detailed study being reported from India.
Materials and Methods
Experimental layout and coffee sampling
For the present study, two independent trials were undertaken for arabica and robusta varieties. The first trial was conducted at the beginning of the harvest whereas the second trial was carried out at the end of the harvest to compare the exact stage of harvest on the incidence of toxigenic molds. Along with this, the impact of drying surface and raking frequencies were also considered.
Generally, in the plantations, coffee cherries are hand plucked at three different stages of maturity, which includes: ripened, half-ripened, and greens. Coffee cherries collected at the beginning of the harvest mainly comprise of hand-picked ripened cherries (60%–80%), whereas end of the harvest samples include those cherries wherein majority are greens, half-ripe, and tree dried that are generally stripped instead of hand picking.
For the present study, coffee samples were obtained from three different sources (blocks/varieties) from the coffee research farm (at Chettalli, Coorg District, Karnataka, India), and each of the individual sources of coffee was considered as an individual replication (unit). Cherries (beans not subjected to fermentation) at a rate of 25 kg/m2 were loaded into hallow wooden frames (1 × 2 m) on different drying surfaces, that is, concrete, tarpaulin (plastic sheets used for coffee picking), and on soil, the procedure of which is routinely employed for drying coffee beans in estates. For each individual day, four rakings were done until the desired moisture level was attained. The coffee cherry was monitored daily for the loss in moisture and aw. The fungal communities were assessed only for the initial and final samples. Since undertaking mycological studies for each and every sample on a day-to-day basis is highly impossible when carrying out a pilot study at the plantation level similar to that of the present one, we assessed for fungal occurrence only for the initial and final samples. Additionally, the OTA contamination in different treatments was analyzed in the finally obtained clean coffee samples. This is because where economical value is concerned (for marketing of the coffee in terms of quality and safety), results of the final samples presume more importance rather than the initial samples.
Moisture and aw estimation
The moisture content was estimated from the percentage loss in weight of sample heated in an oven at 105°C ± 1°C for 24 h (or until a constant weight was attained). The aw of the samples was measured using Rotronic Hydro palm aw meter (Rotronic). The measurements were done in replicates of three for the samples taken from three different blocks wherein the same varieties of coffee were grown.
Mycological analysis
The coffee bean samples (freshly hulled and dried) were surface sterilized with 1% sodium hypochlorite for about 10 min, followed by washing (three times) with sterile distilled water, and wiping with sterile filter paper. This was later aseptically plated on the sterilized DG-18 medium for identification of molds (Hocking and Pitt, 1980). For each sample, a total of 70 beans (7 beans/plate) were plated on the media to study the extent of toxigenic mold infection. The inoculated plates were later incubated at 25°C and observations were made after 5–7 days of incubation. The mold infection was expressed as bean infection rate (%) (all the values represent frequency, not growth; infections were considered independently so the sum of infections, taxon-by-taxon, can exceed 100% overall bean) and the representative colonies of fungal species based on the manual of Klich and Pitt (1988).
OTA analysis
OTA analysis was carried out by high performance liquid chromatography (HPLC) method. The extraction and purification of samples were carried out according to the method of Patel et al. (1997) with some modifications. In brief, 10 g of finely powdered coffee bean samples (subsampled from 500 g) were extracted with chloroform (50 mL) and 0.1 M orthophosphoric acid (5 mL) for 30 min in a mechanical shaker. The extract was filtered using Whatman No. 1 filter paper, and the filtrate was evaporated under vacuum in rotary evaporator (below 40°C). Clean up was performed on silica gel (60–120 mesh) column (5 g of silica gel was slurred in chloroform and packed in a glass column of 1.5 × 30 cm with a stopper. After silica gel settled, 1 g of anhydrous sodium sulfate was added.). Further, the evaporated residue was dissolved in 2 mL chloroform (twice) and loaded on to a silica gel column and drained by gravity. The column was washed with 50 mL chloroform-methanol mixture (97:3), and the wash was discarded.
OTA was eluted with 50 mL of toluene:acetic acid (9:1), evaporated under vacuum (at 40°C), transferred to a 5 mL test tube with toluene:acetic acid (9:1), and evaporated under nitrogen. The residue was dissolved in the HPLC solvent and used for HPLC analysis. HPLC analysis was performed on a 15 cm Bandaclone RPC-18 Spherisorb column using acetonitrile (55%) water (45%) and acetic acid (2%) as mobile phase with a flow rate of 1 mL/min, excitation maxima at 330 nm, and emission at 460 nm. The OTA was detected and quantified by comparison of peak areas of samples and standard OTA. Presence of OTA was confirmed with Boron trifluride reaction with samples extracts (AOAC, 1995). Fifty microliters of Boron trifluride was added to dried extract, reacted at 60°C for 60 min, evaporated, dissolved in mobile phase (100 μL), and injected into HPLC column. The OTA peaks were confirmed by changes in the retention time of OTA in samples and standards.
OTA detection limits were calculated based on the lower limit of detection (i.e., 0.1 ppb). The blank sample (reference sample) without any traces of OTA was used for the recovery of OTA. The samples were spiked with different levels of OTA (5 and 10 ppb), followed by extraction and calculation, the recovery of which was 85% for 5 ppb and 80% for 10 ppb.
Cup quality evaluation
After the completion of drying, the coffee bean samples were hulled in a mechanical huller, cleaned, graded, and packed. Further, all the experimental coffee samples were sent to “Coffee quality evaluation centre,” Bangalore, India, where the quality was evaluated based on the physical and cup quality attributes. Cup quality was assessed by experienced and professional coffee tasters (3–5 numbers). The cup quality was evaluated based on Hedonic scale with rating from 1 to 6 (6, good; 5, above average; 4, average; 3, below average; 2, falling off; 1, poor). These ratings are a measure of cup quality such as acidity, body, and aroma or flavor. The rating presented in the current study is based on the coffee quality evaluation generally employed for marketing purposes in India and elsewhere.
Results and Discussion
In India, coffee cherry is generally dried on soil in the plantations (bigger in size, in large hectares) or at the farm level (considered to be of a lesser area than plantation). This is mainly due to the lack of good drying yard and huge volume of coffee that needs to be processed in a shorter time after harvest. During the drying process, thickness of spreading, stirring, and drying time are considered as significant factors that contribute to the overall quality of coffee. In India, except for the corporate sector and large growers, majority of the farming community (marginal or small coffee growers) do not give importance to vital factors such as GAP or GMP during drying. This is mainly due to the lack of knowledge as well as less premium price that the small growers get. Additionally, in India, due to high cost involvement in construction of concrete drying yard and space requirement, farmers generally prefer mud (soil) surface for drying. This type of drying generally leads to the deterioration in the quality of coffee. Additionally, in majority of the cases, after the final stage of processing, the dried coffee beans obtained from various growers are pooled into a single lot and sold to the curers.
Based on these facets, in the present study, we tried to use low cost tarpaulin sheets (picking mats of coffee) and cement surfaces as an alternative to soil, to evaluate whether these surfaces can enhance the drying process as well as inhibit the proliferation of molds and mold infection.
Additionally, this is the first detailed report on various drying experiments carried out in India. Though coffee is produced since a long time, quality aspects were not given due importance due to the marketing system (pooled quota sale system) that prevailed in India. Previously, the “Coffee Board of India” used to pool all the coffee processed in a single season and market it locally or internationally. However, this scenario changed after the introduction of “Free Sale Quota” (FSQ), according to which a planter (grower) can market the grown coffee directly without the intervention of the Coffee Board. After FSQ, planters started giving importance to quality aspects, as they fetched a higher price in the local and intentional market. In addition to this, during 1980s and early 1990s the Research wing of the Indian Coffee Board mainly dealt with the production and processing aspects and the research was more concentrated toward fermentation, enzymes (for washing), and washing systems. However, during the past few years, under the FAO-CFC-ICO Global mould project, “Improvement of coffee quality through prevention of mold formation,” detailed studies on improvement of coffee qualities were undertaken. Unraveling the route of OTA in coffee bean and ways to prevent it in the final product was given high importance. Hence, accordingly the present work is the outcome of such a detailed study carried out exclusively in India, which mainly deals with drying.
With regard to stirring, in India, generally farmers adopt a maximum of 1–2 stirs/day. However, in the present study, we compared the control samples with 4 stirrings/day, which is also the standard recommendation by the Indian Coffee Board. Here, control means the farmers' practice, wherein stirring of coffee cherry is performed only for 1–2 times/day. From our results, it was evident that cement and tarpaulin (with 4 stir/day) show faster drying rate compared with soil; hence this method is recommended to be adapted at the field level during preparation of coffee cherry.
The dry processing of coffee (cherry drying) has been identified as one of the important steps (apart from inappropriate storage or transport conditions) during which OTA formation can occur under humid tropical conditions (Bucheli et al., 2000; Bucheli and Taniwaki, 2002). Tables 1 and 2 show the results on the effects of raking, days of drying, the percent of mold recovered at the beginning and end of harvest of arabica and robusta cherry, and the levels of OTA detected. As observed in Tables 1 and 2, the drying days varied. This is because of the variations in the drying rates of individual coffee cherry in different blocks. For example, 12–13 days indicate coffee from one block that has dried on the 12th day itself, whereas in the other two blocks it might have dried on the 13th day. Generally, the drying rate of coffee varies due to the quality of coffee cherry lot (batch/group), that is, the coffee lot is unsorted and contains all the stages of ripeness that posses different drying rates. In the present study, we adopted the method followed by the farmers for cherry preparation, which is, drying unsorted lot of coffee after harvest.
All the values are average of three blocks.
OTA (ppb), positive samples which showed the presence of A. ochraceus have been analyzed for OTA.
OTA, ochratoxin A; SD, standard deviation.
All the values are average of three blocks.
OTA (ppb), positive samples which showed the presence of A. ochraceus have been analyzed for OTA.
BDL, below detection level (detection limit 0.1 ppb); ND, not determined.
In the present study, in arabica samples, the initial and final moisture content was 55% and 8.5% and 64% and 10% in run 1 and 2, respectively, whereas the corresponding aw levels in the initial and final samples were 0.98 and 0.66 aw and 0.99 and 0.72 aw in run 1 and 2, respectively. In the robusta samples, the moisture level in run 1 for the initial and final samples was 53% and 7.0%; and in run 2, it was 68.5% and 13%, respectively. With regard to aw, the recorded levels were 0.91 and 0.83 aw for run 1 and 0.97 and 0.83 aw for run 2, respectively.
Irrespective of arabica or robusta, the dominant fungal flora isolated from the coffee cherry belonged to either A. niger or A. ochraceus species. The fungi A. ochraceus and A. niger in coffee are considered to be a potent OTA producers (Taniwaki et al., 1999; Mantle and Chow, 2000). In arabica, coffee cherry samples analyzed for the first run of samples (begin of harvest) showed higher percent mold contamination for 1 rakings/day than for 4 rakings/day. The higher incidence of molds was in soil followed by cement surface and tarpaulin. The fungi A. niger could be isolated as the major contaminant in all the samples spread for drying on various surfaces (Table 1). However, in the first run (begin of harvest), A. ochraceus was completely absent. This might be possibly attributed to the conditions required for growth and proliferation of the individual fungi. For example, A. niger can grow at a varied temperature range of 6°C–47°C, with the optimal growth conditions being 35°C–37°C (0.77 aw), whereas for A. ochraceus the normal growth can occur between 8°C and 37°C, with an optimum growth level being 24°C–31°C (0.95–0.99 aw) (Pitt and Hocking, 1997). Additionally, factors such as surrounding humidity, temperature, and stage of ripeness of cherries (ripe, semi-ripe, tree dried, etc.) play a vital role in fungal infection and proliferation. The environmental condition at the study site of our present work had an average humidity of 85%–90%, and the temperature ranged between 20°C and 22°C.
With regard to run 2 of the experimental setup (end of harvest), cherry dried on soil surface showed highest mold contamination, irrespective of raking period compared with cement or tarpaulin surface. Also, the potential OTA producing mold A. ochraceus could be isolated after the completion of run 2, along with the detection of OTA in the samples (see Table 1). Interestingly, cherries dried on cement and tarpaulin with 4 stir/day showed the absence of OTA in the samples. Batista et al. (2009) have made similar observations wherein higher incidence of filamentous fungi was observed in the coffee swept from ground and in “floating” coffee samples.
In robusta, compared with arabica, the incidence of mold contamination was higher with A. niger being the dominant fungi in the first run (begin of harvest) (Table 2). The coffee cherry spread on soil as well as on cement surface for drying showed the presence of OTA (1 stir/day). In the experimental samples lot of robusta cherry, in run 1 (begin of harvest–4 stir/day), A. niger or A. ochraceus was absent and only yeast could be isolated. In run 2 (end of harvest), mold contamination could be observed in almost all the coffee samples, irrespective of the spreading surface. The samples that were stirred for 1 time/day also showed the presence of OTA. These results highlight that unlike arabica, the drying surface might have not played a major role in prevention of mold attack. In general, based on the results, undertaking 4 stir/day was found to be beneficial with regard to the presence of OTA in the samples. Favorable growth conditions (moisture, aw, humidity, stage of harvest, etc.) might have contributed significantly to the formation of OTA by molds in certain coffee samples used in the present study. This observation is supported by some of the earlier reports on OTA formation in coffee and the favorable conditions (Mantle and Chow, 2000; Joosten et al., 2001).
Even though some of the common toxigenic molds such as Aspergillus carbonarius and A. niger have been stated to have a role in the OTA contamination of green coffee beans, A. ochraceus has been reported to be the principal OTA producing mold in coffee bean (Frank, 1999; Urbano et al., 2001; Taniwaki et al., 2003). Additionally, in one of the studies by Taniwaki et al. (2003) on the distribution of OTA producing Aspergilli species, out of 408 coffee samples from four coffee producing regions of Brazil, it was evident that nearly 269 of A. ochraceus isolates (out of 872 isolates) are capable of producing OTA. Also, in tropical regions such as India, A. ochraceus can contaminate coffee beans more rapidly than other molds. Supporting our view, Frisvad et al. (2004) have reported the most important OTA producing species in coffee bean to be A. ochraceus. Hence, based on these facets, in the present study, OTA concentration was analyzed only for A. ochraceus contaminated samples and not for others.
Results on the cup quality have been provided in Table 3. From the results, it could be observed that coffee cherry prepared by drying on cement and tarpaulin surface provides better cup quality than those dried on soil. Coffee dried by raking for 4 times/day on tarpaulin surface showed good cup quality results. Overall, except for the coffee dried on soil, which showed poor cup quality, not much difference was observed between arabica and robusta varieties dried on cement or tarpaulin surfaces.
Hedonic scale: 6, good; 5, above average; 4, average; 3, below average; 2, falling off; 1, poor.
In the present scenario of quality and health awareness among consumers, GAP and GMP play a major role in all the commodities including coffee. Since India exports nearly 80% of the coffee produced, quality plays paramount importance in the competitive market, especially under the new World Trade Organization (WTO) regime. Stringent regulation for quality and OTA limits in coffee, especially in the importing countries, has rendered it a mandatory to produce and export high quality coffee with minimal contaminants. In this scenario, the drying step in coffee is one of the major factors that decides coffee quality and mold infection. This paper reports the vital data on coffee drying practice in India and how it can be modified to improve coffee quality by prevention of toxigenic molds. Additionally, in India, due to the dominance of small and marginal farmers (constituting for nearly 98%) in the coffee sector, the adoption of GAP and GMP at farm level is highly difficult due to two major factors, namely (1) finance and (2) labor availability. Generally, a coffee planter can have a good yield or grow coffee organically; but when it comes to drying, usually the coffee drying yard is not of good standard, and drying (raking) of the cherries (up to desirable moisture level) is not adequate. This is the stage where fungal contamination occurs, thereby leading to deterioration in quality.
Based on the results obtained from the present study, it is concluded that better quality of coffee with minimal mold and OTA contamination could be obtained by drying the coffee cherry on tarpaulin and cement surfaces with a minimum of 4 rakings/day rather than drying directly on soil surface. This study formulates one of the GMP in the coffee chain. This method is highly recommended to be practiced at the plantation or farm level in all the coffee growing regions of the world.
Footnotes
Acknowledgments
The present work was carried out under the ICO-CFC-FAO Global mould project GCP/INT/743/CFC and the CFC/ICO/06 “Enhancement of Coffee Quality through Prevention of Mould Formation” projects at the Coffee Research Substation Chettalli, Coffee Board, India. We gratefully acknowledge Dr(s). Raghuramulu Y and Jayarama, Central Coffee Research Institute, Chikmagalur, India for their keen interest and encouragement. We also thank the anonymous referees for comments and constructive suggestions provided for improving this manuscript.
Disclosure Statement
No competing financial interest exists.