CONDITIONS NECESSARY FOR INACTIVATION OF AFLATOXIN B1 AND LOSS OF MUTAGENICITY IN COPRA MEAL ON CHLORINE GAS TREATMENT

Copra meal spiked with aflatoxin B1 (AFB1) at a concentration of 10 ~g/g was treated with chlorine gas in evacuated chlorination apparatus. More than 75% degradation of AFB 1 occurred on exposure of copra meal at 6.2% moisture content to chlorine gas for 2.5 hr. at a dosage of 16 mg chlorine gaslg copra meal. No visible changes in colour occurred in treated copra meal. Prolonged exposure of copra meal to chlorine gas for 24 hr. with or without continuous stirring did not increase the percentage degradation of AFB1. Chlorination of copra meal with increased moisture percentages up to 52 caused decreased percent degradatioh of AFB1. The mutagenicity of extracts of chlorine treated copra meal was tested with the Ames Salmonella mutagenicity assay. The percentage decrease in mutagenicity of the extracts showed a high correlation (r = 0.8856; p < 0.01) with percentage degradation of aflatoxin. No new mutagenic compounds were generated during chlorine induced degrada- tion of AFB 1.


Introduction
Copra meal, the residue remaining after expulsion of oil from kiln dried coconut kernels (copra) is a nutritionally valuable raw material for animal feed. The aflatoxins produced by fungi AspergillusJlavus and A. parasiticus in improperly dried copra, is retained partly in the copra meal (poonac) during expulsion of oil.7 Aflatoxin contaminated meal poses problem to animal health and has no export value.
Inactivation of pure aflatoxin B 1 (AFB1) on exposure to chlorine gas has already been reported.'' The degradation products were non-toxic and nonmutagenic. The possibility of using chlorine gas to inactivate AFBl in spiked corn meal, s iked copra meal and artificially contaminated peanut has already t been shown. In this study the conditions necessary to degrade AFBl in spiked copra meal and the changes in mutagenicity of the toxic meal on chlorine gas treatment were examined.

Copra Meal
Copra meal collected from a commercial oil mill in Sri Lanka was ground in a Waring blender to pass through a 2 mrn sieve. The powder was tested for the absence of AFB 1. The powdered meal was wetted with acetone, and spiked with pure AFB 1 in acetone (1 'pglml) while stirring magnetically to a concentration of 10 pglg of AFBl per gram of meal. Stirring was continued with slight warming to evaporate the solvents as assessed by drop in weight of the contents to the original weight.

Estimation of Moisture
Moisture was estimated by drying powdered copra meal to constant weight at a dial setting of 40 for 10-20 min. in a I. R. moisture balance (Cenco CSC Scientific Co., Fairfax VA, USA)

Estimation of Aflatoxins
The copra meal (log) was blended with 90 ml of 70% aqueous acetone, and filtered. The filtrate was extracted with chloroform, dried in a stream of nitrogen and made up to 500 p.L6 Separation of AFBl was done by spotting the samples in benzene-acetonitrile (98:2) on TLC plates (Fisher Rediplate; 250 nm, silica gel G) and developing in chloroform:acetone:propanol (85:15:2.5). The plates were observed under a uv source at 362 nm and the individual A m 1 spots were marked. The estimations were done by scanning the fluorescent spots using a Kratos model SD 3000 spectrodensitometer (Kratos Scientific, Westwood, NJ, USA) at 360 nm and comparing the integrated peak areas for spots of AFBl with a standard curve.

Generation of Chlorine Gas
Chlorine gas was generated by reacting potassium perrnanganate with 3M Hydrochloric acid in a closed system described earlier for chlorination of aflatoxins. lo

Estimation of chlorine
Chlorine gas (10ml) at standard temperature and pressure were passed into 50 ml of 0.01 N acidic potassium iodide solution kept under reduced pressure of 40 mm mercury in the chlorination apparatus. The mass of chlorine in this volume was estimated by titrating the liberated iodine with 0.01 N sodium thiosulphate with starch as i n d i c a t~r .~

Chlorination of Copra Meal
Copra meal (log) was placed in round bottom flasks of total capacity 120 ml and evacuated to a reduced pressure of 40 mm of mercury. Required amounts of chlorine gas was introduced at standard temperature and pressure through a syringe from the chlorination system to copra meal in the flasks. The flasks sealed by closing the gas adapter taps were disconnected from the chlorination system and stored for required durations in dark at 25 f 2°C.

Mutagenicity Testing
Aliquot of the extracts used to estimate AFBl in treated and control copra meal were dried under a nitrogen stream and dissolved in dimethyl sulfoxide. Ames Salmonella/rnicrosome assay was performed in the presence of a rat liver S-9 mix.'*4 The S-9 metabolic activation system was prepared from Arochlor 1256 induced rats. For plate incorporation tests, 0.1 ml of fresh overnight culture of Salntonella typhiinurium strain TA 98 was added along with the test chemical, and 0.5 ml of S-9 mix to 2.5 ml top agar containing small quantities of histidine and biotin. 2-amino fluorene was used as the positive control. All estimations were done at AFB 1 concentrations of 0.4,0.2 and 0.1 pg per petridish. The extracts from copra meal (chlorinated and unchlorinated) contained equivalent amount of AFB 1 originally. The top agar mixture was poured into minimal media agar plates. They were incubated at 37OC for 2 days and the resultant colonies were counted.

~e c o v e & of AFBl
The recovery of spiked AFBl under our experimental conditions was 77%. In all calculations the AFB 1 estimated in extracts from untreated copra meal was considered 100%.

Dose of Chlorine Gas
The dose of chlorine gas required for more than 75% degradation of AFB 1 was 160 mg per log of copra meal under our experimental conditions. The doseresporse plot for degradation showed a linear relationship (r = 0.980; p < 0.001) (Figure 1). The chlorine requirement for degrading AFBl in copra meal is slightly higher than the requirement for corn meal and is about ' 1, for peanuts. No changes in the colour of copra meal was observed on treatment at the suggested concentration of chlorine gas indicating low or negligible visible deleterious effects.

Duration of Exposure and Stirring
Attempts to increase the interactions by prolonged exposure of copra meal to chlorine gas, as well as by continuous stirring durifig exposure did not increase the percentage degradation of AFB 1 markedly (Table 1). Insufficient interac-  tion between the gas and AFBl did not appear to be an important factor at the ratio of copra meal to chlorine gas used during chlorination.

Effect of Moisture
Hypochlorous acid produced by the reaction of chlorine with moisture is the active entity responsible in releasin Clf for electro hillic reaction with A F B~.~

IIowever, increased moisture content in copra ma? during chlorination d i f n o t increase, but reduced. the percentage degradation of AFB1, indicating that chlorine'is made less available or unavailable at the sites of AFBl with in-
creased moisture concentrations ( Table 2). The storage moisture content of 6.2% in copra meal appeared to be sufficient for the chlorine induced degradation of AFBI. * Initial moisture content 6.2%. ** As estimated in extracts from untreated samples.

Mutagenicity
One of the major concerns in the use of chlorine gas in inactivating AFBl in foods and feeds is the possible interactions of chlorine with food components producing toxic, mutagenic or carcinogenic compounds. The percentage decrease in mutagenicity observed in extracts from chlorinated copra meal showed a high correlation with percentage degradation of AFBl detected chemically (r = 0.8856; p <0.01) (Figure 2) This observation suggests not only a loss of mutagenicity with chlorine induced degradation of AFB 1, but also that there is no generation of new mutagenic compounds. However, these observations need to be confirmed with other biological assays.

Discussion.
The chlorine gas treatment carry several advantages over ammoniation5 already established for detoxification of animal feeds. The reaction of chlorine gas is instantaneous, and hence the treatment is faster compared with ammoniation which needed exposure for several days or weeks under ambient conditions or a few hours at elevated temperatures and under high pressure. 9 The chlorine treatment done under reduced pressure avoid the risks associated with a high pressure ammoniation system. Chlorine gas is already permitted as a bleach, disinfectant, and in chlorination of water in the food industry." Chlorination of AFB 1 yield two products, 8, Pdichloro-AFB 1 and 8, Pdihydroxy-AFB1. The former is carcinogenic, but possesses a half life of a few minutes and is converted to the latter which is non-mutagenic.10 Thus the identified chlorinated products of AFBl appear to be more acceptable from a toxicological point of view than the aflatoxin D l produced during ammoniation of AFB1. Aflatoxin D l is shown to be mutagenic at a dosage 450 times that of AFBl and toxic to chick at a dosage 18 times AFBl whereas the chlorination products of AFBl are non-toxic and non-mutagenic. (See Figures 1 & 2 , pages 30,31).

Conclusions
More than 75% of degradation of AFBl could be attained with 16 mg chlorine gas per g copra meal at 6.2% moisture on exposure for 2.5 hr. Increased duration of exposure, stirring or higher moisture content did not increase degradation of AFB1. Chlorine gas treatment of toxic copra meal did not produce any mutagenic products. Further experiments on chlorination on an expanded scale and investigations using other bioassay methods are suggested. % Degradation of. AFB 1 Figure 2: Relationship between percent decrease in mutagenicity and percent degradation of aflatoxin B, in extracts of chlorine treated copra meal.
Chlorine gas (rng/lOg copra meal) Figure 1: Relationship between percent degration of Aflatoxin B, in copra meal and the dose of chlorine gas when exposed for 2.5 hr. The values are means of triplicate samples.