J Reprinted from Journal of Food Protection. Vol. 54. No. 8 Pages 623-626 Copyright@ International Association of Milk. Food and Environrnemal Sanitarians

Reduction in Aflatoxin Content of Maize by Atoxigenic St rains of Aspergillus flavus

ROBERT L. BROWN*, PETER J. COTTY, and THOMAS E. CLEVELAND

Southern Regional Research Center, U. S. Department of Agriculture. Agricultural Research Service, New Orleans. Louisiana 70179

(Received for publication February 18, 1991)

ABSTRACT

In field plot experiments, an atoxigenic strain of Aspergillus flavus interfered with preharvest aflatoxin contamination of corn when applied either simultaneously with or one day prior to a toxigenic strain. The atoxigenic strain reduced preharvest afla- toxin contamination 80 to 95%. The atoxigenic strain was also effective in reducing postharvest aflatoxin contamination caused by both an introduckd toxigenic strain and by strains resident on the kernels. The results suggest that atoxigenic strains of A.flavus may have potential use as biological control agents directed at reducing both preharvest and postharvest aflatoxin contamination of corn.

Aflatoxins, toxic metabolites of the fungi Aspergillus jlavus Link: Fr. and A. parasiticus Speare, are potent carcinogens which pose serious health hazards to humans and domestic animals because they frequently contaminate agricultural commodities (4,8). Corn grown in the south- eastern United States is more frequently colonized by high populations of A. flavus than corn grown in the Midwest, where most U.S. corn is grown (19). However, the drought of 1988 created favorable conditions for A. j7avus in the Midwest, and the Wall Street Journal (February 23, 1989) reported one-third of the crop tested in Iowa and Illinois contained dangerous levels of aflatoxin. This, according to the article, caused greatly increased concern in the U.S. corn industry.

Biological control of several plant diseases has been demonstrated by utilizing certain strains of the causal or- ganism (12,14,15,20). Strains of A. flavus that do not produce aflatoxins have been selected from fungal popula- tions in cotton and corn fields (6,17). In greenhouse tests, atoxigenic strains of A.flavus significantly reduced produc- tion of aflatoxin B, in cottonseed coinoculated during development with toxigenic strains (7). When atoxigenic strains were introduced into wounded cotton bolls one day prior to the toxigenic strain, even greater control of afla- toxin was obtained (7). This study provided useful informa- tion but was camed out in a controlled environment. In the field, where A. j7avus first associates with the crop, greater biological and environmental complexities exist.

Also, it is not known if atoxigenic strains influence con- tamination of harvested crops which already possess com- plex microflora and may be previously infected with A. f laws (10). In the present study, an atoxigenic strain, previously identified and shown effective in greenhouse tests on cottonseed, was tested for efficacy on corn under field and storage conditions. A preliminary presentation of these studies has been made (3).

MATERIALS AND METHODS

F~rngal strains and growth conditions Strains of A. flavus utilized in this study were isolated from

agricultural soil and cottonseed in Arizona (6,7). Strain 13 pro- duced large quantities of aflatoxins both in culture and in devel- oping cottonseed. while strain 36 did not produce detectable levels (<I0 nglg) (6). Cultures were grown at 30°C in the dark on a 5% V-8 juice, 2% agar medium. Plugs (3 mm in diameter) of sporu- lating cultures were stored at 8OC on a long-term basis in Cdram vials containing 5 ml of distilled water (5.6). Conidia from 7- to- 10-d-old cultures suspended in deionized water served as inocula

(7).

lnoculotion of developing corn kernels Field corn (Pioneer Brand 3369A) planted on 76.2-cm cen-

ters in a silty-loam type soil in New Orleans, LA was utilized in these experiments. Ten d after the 50% silk stage, each devel- oping ear was wounded once with a cork borer (3 mm diameter) to a depth of 5 mm. Each ear was inoculated by applying 20 p1 of a spore suspension (4.0 x 106 conididml) to a wound.

The treatments were applications of either the toxigenic or atoxigenic strains alone, the toxigenic strain followed immediately by the atoxigenic strain or the toxigenic strain 24 h after the atoxigenic strain. Wounded, uninoculated ears were used as addi- tional controls. Treatments were replicated six times (3 ears/ replicate) and organized into randomized complete blocks. Ex- periments were performed twice (tests 1 and 2). Corn used in the first field experiment was planted in mid-April and harvested in midJuly (1989); corn for the second experiment was planted in early May and harvested in early August (1989).

In all experiments, ears were harvested at maturity and dried in a forced-air oven at 60°C for 2 d. After drying, ears were kept at room temperature in sealed plastic bags containing silica gel desiccant until aflatoxin analysis.

JOURNAL OF FOOD PROTECTION, VOL. 54, AUGUST 1991

624 BROWN. CO'ITY AND CLEVELAND

Inoculation of harvested corn kernels Ears harvested at maturity from uninoculated portions of the

above test plots were shelled and.the kernels dried at 60°C for 2 d in a forced-air oven. After dryins, kernels were kept at room temperature in sealed plastic bags containing silica gel desiccant until used. The initial moisture content of the kernels was deter- mined according to the official method of the American Oil Chemists Society (2).

Ten g of kernels in 50-ml Erlenmeyer flasks were either inoculated with a single fungal strain, or a mixture (1:l) of the toxigenic and atoxigenic strains or hydrated with sterile, deionized water alone. Conidia for each inoculation, approximately 4.0 x loh per strain tested, were suspended in the amount of deionized water needed to bring kernel moisture content to 22% upon application. Flasks were subsequently sealed with Styrofoam plugs, covered with aluminum foil, and incubated at 2 8 T for 12 d. The contents were then dried in a forced-air oven at 60°C for 2 d to halt fungal activity and prepare the sample for aflatoxin analyses. In a second set of postharvest experiments, the above protocol was altered to simulate aspects of postharvest storage. In these tests, all inoculum suspensions and water controls contained a surfactant (0.02% Tween 80) to improve seed coverage. After either inoculation with a single fungal strain or hydration to 22% moisture content with deionized water, kernels were incubated for 24 h at 28°C and then dried at 4S°C for 3 d in a forced-air oven. Dried kemels were stored 8 d at 28OC and then rehydrated to 22% with either a suspension of spores (4.0 x 106) of a toxigenic strain or distilled water. After rehydration, flasks were incubated for 12 d at 28°C. dried, and analyzed for aflatoxins.

All experiments were performed twice and replicated six times; each replicate consisted of one flask.

Aflatoxin analyses The aflatoxin B, content of replicates from both the preharvest

and the postharvest studies was determined with official methods of the American Oil Chemists Society ( I ) . Toxin was identified by thin layer chromatography (TLC) and quantified directly on the TLC plates with a scanning densitometer with a fluoromet~y attachment (Model CS-930; Shimadzu Scientific Instruments, Inc., Tokyo, Japan). In the case of the preharvest tests, each replicate1 sample to be analyzed contained 45 kernels (12 to 13 g), 15 from

each of three ears. The 15 kemels were chosen by first identifying the kernel which lined up with the wound in the husks. This "central" kernel was then used to identify the 14 adjacent kernels which consisted of one column to the left and right and two rows to the top and bottom.

Statistical malysis Analyses were performed with the Statistical Analysis Sys-

tem (SAS Institute Inc., Cary, NC). Treatment replicates from each experiment were ranked. and ranks were subjected to split- plot analyses where the test was the main plot and the treatment was the subplot. Differences among treatment means were deter- mined by the least significant difference test. All multiple com- parisons were first subjected to analysis of variance.

RESULTS

In preharvest experiments, the test variable was not significant (P=0.05) and did not interact with the treatment variable. This allowed data from both tests to be pooled (Table 1). Very high levels of aflatoxin B, were detected in kernels from ears inoculated with the toxigenic strain in the field. Coinoculation with the atoxigenic strain significantly reduced B, quantities, as did inoculation with the atoxigenic strain 24 h in advance.

Aflatoxin B, levels in kernels coinoculated with the toxigenic and atoxigenic strains after harvest were also significantly lower than levels detected in kernels inocu- lated with the toxigenic strain alone (Table 1). Also, as in the preharvest experiments, the test variable was not sig- nificant (P=0.05) and did not interact with the treatment variable, allowing data from both tests to be pooled. Sig- nificantly less B, was produced postharvest in kemels inoculated with the atoxigenic strain alone than in the uninoculated control.

In postharvest tests where harvested kernels were in- oculated with the atoxigenic strain before the toxigenic strain (Table I), a significant (P=0.05) interaction between

TABLE 1. Afatoxin content of corn kernels wound-inoculated in the field or inoculated after harvest \r.ith atoxigenic and toxigenic strains of A. flaws.

Atlatoxin B, content of kemels (pg/g)' Preharvest Postharvest coinoculation Posrhawest prior inoculationb

Treatment Test 1 Test 2 Combined Test 1 Test 2 Combined Test 1 Test 2

Toxigenic 3,045 2,146 2,595 a 16,044 13,534 14,789 a 30,385 a 23,976 a Atoxigenic 9 0 5 c 14 186 100 d 2 1 8 c 14,040b Control- 0 0 0 c 1,046 1,254 1 , 1 5 0 ~ 1 3 5 c 6,715b

uninoculated Atoxigenic" 101 179 140 b N A P A NA 5.1 18 b 7,844 b before toxigenic

Coinoculation 184 912 548 b 3,909 2,628 3,269 b N A N A

"Values followed by the same letter are not significantly different by the least significant difference test. bValues for replicates from tests 1 and 2 were not combined because a significant interaction occurred between the test and treatment variables. 'Kernels were inoculated with the atoxigenic strain 24 h before the toxigenic strain in preharvest prior inoculation experiments. In postharvest prior inoculation experiments, kernels were inoculated with the atoxigenic strain, incubated for 24 h, then dried and stored for 8 d, before inoculation with the toxigenic strain. "NA = Not applicable.

.lOURNAL OF FOOD PROTECTION, VOL. 54, AUGUST 1991 -

REDUCTION OF AFLATOXIN IN MAIZE 625

the test and treatment variables occurred. Thus, data from these two tests were not combined. Inoculation with the atoxigenic strain prior to the toxigenic strain significantly

I reduced B, in both tests when compared to kernels inocu- lated with the toxigenic strain alone. In test 1, levels of B, in the atoxigenic and uninoculated control treatments were significantly lower than in the kernels inoculated with the toxigenic strain after the atoxigenic strain. However, in test 2, the levels detected in these three treatments did not differ significantly.

DISCUSSION

Populations of A. flavus in agricultural fields are com- posed of strains that vary widely in aflatoxin-producing ability, and this ability is apparently unrelated to a strain's potential to infect and colonize host tissues ( 6 7 ) . These observations suggest that naturally occuning atoxigenic strains may be able to outcompete toxigenic strains during infection of developing crops and thereby prevent aflatoxin contamination (6). Several fungi can interfere with afla- toxin production on artificial media or other sterile sub- strates (9,13,16). However, the biocontrol potential of those fungi tested in developing corn, grown under con- trolled conditions, is inadequate in preventing aflatoxin production by A. flavus (18). Greenhouse experiments verified the biocontrol potential of atoxigenic A. j7avus strains on cotton (7). Part of the attraction of controlling aflatoxin contamination with strains of A. flavus relates to the theoretical ability of the atoxigenic strains to be active under the same environmental conditions where A. flavus toxigenic strains are active (7). This ability was tested in the field experiments presented here and the results suggest this speculation may have merit. The atoxigenic strain of A. flavus was very effective at preventing contamination of corn under field conditions favoring high levels of aflatoxin contamination.

The atoxigenic strain used in these studies was isolated from soil collected in a cotton field in Arizona. There may be some adaptation of A. flavus strains to particular hosts or regions. If adaptation does exist, even greater control could be provided by atoxigenic strains obtained through the screening of corn field populations of A. flavus from particular growing regions.

A. flavus comes in contact with corn kernels prior to harvest and remains with the crop throughout harvest, storage, and even use (11). The potential for aflatoxin contamination thus exists both before and after harvest (11) . The atoxigenic strain reduced contamination by 76 to 8 1 % when shelled kernels were coinoculated with toxi- genic and atoxigenic strains. Strain efficacy was also demonstrated in the experiment where dried kernels were exposed to a simulated breakdown in storage conditions. These observations indicate potential use of atoxigenic A. flavus strains to prevent postharvest infection and subse- quent contamination by toxigenic strains. In two out of four postharvest tests, there was a significant decrease in afla- toxin detected in kernels inoculated with the atoxigenic strain, when compared with the uninoculated control. This may indicate postharvest application of atoxigenic strains

has potential to prevent postharvest contamination of corn by strains associated with the crop in the field. Failure to see reductions in all four tests may be due to interference from uncontrolled microbial activity. Since experimental conditions were designed to exclude sterilization, kernel microflora may have differed considerably in each test.

Very low amounts of aflatoxins were detected in corn inoculated in the field with the atoxigenic strain alone. Toxin was present in only one replicate in one test. This contamination may be due to chance introduction of a toxigenic strain endemic to the field.

In test 2 of the postharvest prior inoculation experi- ment, atoxigenic strain efficacy was significant but not as great as in other tests (Table I). This apparent reduced efficacy is due to a single high value.

for

1.

2.

3.

4.

5.

6.

7.

8.

9.

10.

I I

12.

13.

14.

15.

16.

ACKNOWLEDGMENTS

We thank Herbert Holen for technical assistance and Brian Vinyard statistical assistance.

REFERENCES

Anonymous. 1988. Aflatoxins in corn. In Official methods and recommended practices, 3rd ed. American Oil Chemists Society, Champaign, IL. A13-87. Anonymous. 1989. Moisture and volatile matter. In Official rneth- ods and recommended practices, 4th ed. American Oil Chemists Society, Champaign, IL. AC2-41. Brown, R. L., P. J. Cotty, and T. E. Cleveland. 1990. Effect of atoxigenic strains of Aspergillrrsflovus on prehawest and postharvest aflatoxin contamination of maize kernels. Phytopathology 80:1020 (Abstr.). CAST. 1979. Aflatoxins and other mycotoxins: An agricultural perspective. Counc. Agric. Sci. Technol. Rep. 80. Arnes, IA. Cotty, P. J. 1989. Effects of cultivar and boll age on aflatoxin in cottonseed after inoculation with Aspergillusflavus at simulated exit holes of the pink bollworn. Plant Dis. 73:489-492. Cotty, P. J. 1989. Virulence and cultural characteristics of two Aspergilhs flovus strains pathogenic on cotton. Phytopathology 79: 808-8 14. Cotty, P. J. 1990. Effect of atoxigenic strains of Aspergrllusflovus on aflatoxin contamination of developing cottonseed. Plant Dis. 74: 233-236. Diener, U. L., R. J. Cole, T. H. Sanders, G. A. Payne, L. S. Lee, and M. A. Klich. 1987. Epidemiology of aflatoxin formation by Asperillusflavus. Annu. Rev. Phytopathol. 25249.270. Ehrlich, K., A. Ciegler, M. Klich, and L. Lee. 1985. Fungal compe- tition and mycotoxin production on corn. Experientia 41:691-693. Hanison. M. A.. J. C. Silas, and J. A. Carpenter. 1987. Incidence of aflatoxigenic isolates of Aspergillusflovuslpurusiricus obtained from Georgia corn processing plants. J. Food Qual. 10:101-105. Lillehoj, E. B. 1987. The aflatoxin-in-maize problem: The historical perspective. pp. 13-32. In M.S. Zuber (ed.), Aflatoxin in maize: Proceedings of the workshop. CIMMYT, Mexico, DF. Lindow, S. E. 1987. Competitive exclusion of epiphytic bacteria by ice-Pseudomonus Syringae. Appl. Environ. Microbiol. 53:2250- 2527. Roy, A. K, and H. K. Chourasia. 1990. Inhibition of aflatoxins production by microbial interaction. 3. Gen. Appl. Microbiol. 3659- 62. Schneider, R. W. 1984. Effects of nonpathogenic trains of Fusurium ox).sporum on celery root infection by F. oxysporum f. sp. upri and a novel use of the Lineweaver-Burk double reciprocal plot tech- nique. Phytopathology 74546-653. Trigalet, A., and D. Trigalet-Demery. 1990. Use of avirulent mu- tants of Pseudomonus solanucenrum for biological control of bacte- rial wilt of tomato plants. Physiol. Mol. Plant Pathol. 36:27-38. Wicklow, D. T., C. W. Hesseltine, 0 . L. Shotwell, and G. L. Adarns.

JOURNAL OF FOOD PROTECTION, VOL. 5b, AUGUST 1991 h

626 BROWN, COTTY AND CLEVELAND

1980. Interference, competition and aflatoxin levels in corn. Phyto- pathology 7O:76 1-764.

17. Wicklow, D. T., 0 . L. Shotwell, and G. L. Adams. 1981. Use of aflatoxin-producing ability medium to distinguish aflatoxin-produc- ing strains of Aspergill~~sflovits. Appl. Environ. Microbial. 41:697- 699.

18. Wicklow, D. T., B. W. Horn, 0. L. Shotwell, C. W. Hesseltine, and R. W. Caldwell. 1988. Fungal interference with Aspergillrrs

flavus infection and aflatoxin contamination of maize grown in a controlled environment. Phytopathology 78:68-74.

19. Wilson, D. M. 1989. Potential for biological control of Aspergillus flavus and aflatoxin contamination. pp. 55-66. Itr K G. Mukerit and K. L. Garg (ed.), Biocontrol of plant diseases, vol. 11. CRC Press, Inc. Boca Raton, FL.

20. Yeh, S.-D., D. Gonsalves, H.-L. Wang, R. Namba. and R.-J. Chiu. 1988. Control of papaya ringspot virus by cross protection. Plant Dis. 72:369-460.

JOURNAL OF FOOD PROTECTION, VOL. 54, AUGUST 1991