Aflatoxin in Arizona Cottonseed: Increase in Toxin Formation during Field Drying of Bolls

Aflatoxin assays and moisture determinations were made on locks from bolls inoculated with Aspergillus flavus 30-32 days from flower and harvested after additional periods of 5, 7, 9, 11, 13 and 15 days. Inoculated locks were always tight but had moisture contents comparable to those in noninoculated locks on the same bolls. Lowest toxin concentrations were in seeds from bolls still green at harvest with moisture contents >50%; highest concentrations were in bolls with fully fluffed locks and moisture contents <10%. The greatest increase in toxin concentrations occurred in bolls following suture opening, at the initiation of boll dry-down. Toxin concentrations were comparable for bolls that fluffed in 11 days from inoculation and those requiring 15 days for fluffing. A boll fluffing 11 days from inoculation had the highest level of toxin detected, ca 400 f-l-g/g. The ripening and drying processes rather than the duration of fungal/plant interaction per se seems critical for maximum toxin formation. Field studies in the valley regions of the Southwest have established that aflatoxin formation in cottonseed occurs before harvest following field infection of developing cotton bolls by Aspergillus flavus (Ashworth et at. 1969). A later study correlated fungal contamination and subsequent toxin formation in seed with insect damage (Ashworth et at. 1971). Renewed interest in the aflatoxin problem prompted a model study of insect-vectored A. flavus contamination (Lee et at. 1987). The model showed lower toxin concentrations in seeds from bolls that were harvested before boll dehiscence than were detected in bolls with some fully fluffed locks at harvest. One sample from that study, harvested before boll opening was frozen immediately and examined microscopically. Scanning electron microscopy (SEM) showed no fungal contamination under the seed coat, yet SEM of bolls harvested at the same time, but air-dried prior to examination, indicated fungal penetration into the chalazal area. The observation implied that the drying process is involved in fungal entry and indicated the necessity for further study. The present study sought to explore the relationship between field-drying of boBs and toxin formation. Hopefully, results would delineate the exact stage in the boll ripening/drying process at which toxin formation commences. Experimental Procedure Experiments were done in a field in Maricopa County, south of Phoenix, AZ during July and August, 1986. Bolls from flowers tagged in July were inoculated and harvested in August. Experiments were replicated three times in a modified randomized block design. Approximately 150 flowers were tagged on about 50 plants in each replicate and approximately 100 bolls were inoculated per replicate. Inoculation was with spores of A. jlavus SRRC 2002, a toxin-producing isolate obtained from Arizona cotton. Holes (3 mm diam) were drilled in the carpel walls of bolls 30-32 days from flower, as previous studies (Lee et al. 1987) indicated maximum toxin formation following inoculation of green bolls ca. 15 days prior to opening. Spores were dusted onto the injury site with a small artists' brush. Drilling was close to sutures so inoculum entered the lock and not the intercarpellary division separating one lock from another. Control bolls were drilled but not inoculated. Bolls harvested 5, 7, 9, 11, and 13 days following treatment were still closed, but bolls harvested 15 days after treatment had at least three fully fluffed locks. Ten unopened bolls were harvested for each replicate at each time period from 5 through 11 days post-inoculation and after bolls Aflatoxin in Cottonseed were fully fluffed (15 days post-inoculation). The 13th day postinoculation sample was smaller than the others because some of these bolls had already dehisced and could not be used. Prompted by the observation that bolls were not all opening at the same rate, some bolls were harvested based on development stage rather than days following inoculation per £e. These stages included bolls at suture opening, bolls partially fluffed, and bolls fully fluffed prior to 15 days from inoculation. Fewer bolls (3 to 7 per replicate) opened at these critical developmental stages of the maturing process and necessitated a smaller sampling. Excised bolls were sealed in individual plastic bags and transported in an ice-chest to the Cotton Research Center in Phoenix where they were refrigerated until processed. Bolls (60 each day) were processed over a 2-day period. To insure that aflatoxin formation did not occur in storage, extractions were initiated within an hour of harvest. The single inoculated lock was used for each toxin assay and the 3 or 4 other locks from each boll were combined for the moisture determination. Samples for moisture determinations were weighed within a three-hour period after return of bolls to the Center. Forty-five bolls not used for toxin assays were selected randomly for comparison of moisture contents of inoculated and noninoculated locks. Thirty of these bolls were harvested before opening and 15 bolls were harvested following suture opening. Moisture determinations followed a modification of the AOCS method (Official Methods 1986); seeds and fibers were treated as a composite, heated for 3 hr in a forced draft oven at BOcC in individual paper bags. For toxin assays, no attempt was made to completely separate seeds from the wet stringy lint. Long lint fibers were pulled from seeds by hand leaving a generous supply of wet linters on seeds. Even with linters it was possible to distinguish between white seeds and those that had darkened seed coats. Observations on seed coat color were made on seed from each lock. Lintered seed were also examined in ultra-violet light for brightest-green-yellow-fluorescence (BGYF) (Marsh et al. 1955). Aflatoxin assays were conducted, using an adaptation of a published procedure (Lee et al. 1987). Seeds (7 or 8 per lock) were weighed and soaked in 70% acetone over night. Sample clean-up and partition into methylene chloride were completed on the following day. Final extracts were evaporated to dryness in 4 dram (4 cc) vials and returned to New Orleans for densitometric quantitation on thin layer chromatography (TLC) plates against standards (Official Methods 1986). Aflatoxin values were reported on a dry weight basis, calculated on the moisture content of each boll. All results were analyzed statistically. Results and Discussion Analysis of variance and a Students' t test made on moisture data from the 45 bolls showed no statistically significant difference between the moisture content of inoculated locks and that of adjacent non-inoculated counterparts (p = 0.42 for unopened bolls and p = 0.26 for opened bolls). Statistical analysis merely confirmed the observation on moisture data that prompted the decision to use non-inoculated portions of bolls for moisture determinations (Table 1) and inoculated locks for aflatoxin assays (Table 2). This procedure enabled 417 comparisons of the properties on a boll to boll basis. In general, inoculated bolls opened slightly before control bolls. No bolls from the controls (drilled but not inoculated) opened before 13 days post inoculation (DPI). The act of carpel wall injury foUowed by fungal inoculation triggered a physiological change in the boH that stimulates boll opening. Moisture levels decreased gradually prior to boll opening. Closed boBs 5 DPI had 57·-58% moisture while closed bolls 13 DPI had 50-51% moisture. In a two day period, moisture content decreased to 8-9% in bolls fully fluffed 15 DPI. AU bolls remaining on plants for IS days following inoculation had fully fluffed locks. Funy fluffed describes the non-inoculated locks; inoculated locks were never fluffed, remaining tight and easily distinguishable from their fluffed counterparts. The carpel wall pulls back from these tight locks allowing them to dry even though fungal hyphae intermingled with fibers prevented fluffing. Moisture content of all fluffed bolls was low, between 8 and 10%. The hot dry desert air dries bolls quickly once they open. Fluffing requires about a day after initiation of opening when sutures crack (Lee 1988). Inoculation had little effect on boH moisture; percent moisture in control bolls was comparable to that of inoculated bolls. Variation in moisture contents of inoculated bolls harvested at all stages of maturity was less between replicates (± 0.4 to ±2.8) than within replicates (± 1.1 to ±5.4), indicating a wide boll to boll variation that averaged out when a larger population was examined. The same trend existed in control, non-inoculated bolls. Some seeds from each inoculated lock had BGYF linters. Fewer seeds (5%) from locks harvested 5 DPI had BGYF linters, while 50% from 7 DPI seeds, and all seeds maturing for 9 through 15 DPI had BGYF linters. Such BGYF linters, caused by the action of peroxidase from developing fibers on kojic acid produced by the fungus (Marsh and Simpson 1984) are an indication of fungal penetration from the inoculation site to seed surfaces. Neither lint nor seed linters from adjacent non-inoculated locks exhibited BGYF. Seed-coats darken as seeds mature; oxidized tannins cause the darkening but non-oxidized precursors are presenteven when seedcoats are white (Halloin 1982). Five days from inoculation all seedcoats were white. After an additional two days of maturation, 75% of the seedcoats were dark, and from that period on all seedcoats were dark. Our studies did not address the role of potential fungistatic action of tannin precursors but do indicate little such action by oxidized tannins. Further research is required to correlate physiologTable 1. Percent moisture of control cotton bolls and bolls inoculated with Aspergillus flavus 30-32 days from flower and harvested at various times following inoculation Boll age at harvest: DP!' Inoculated bolls Control Bollsb Replicatesc Replicates Variation Variation between between II III replicates I II III replicates 5 Closed Avg. 57.8 :': 2.4 (l0)d 58.3 :': 1.3 (10) 58.6 :': 1.4 (10) :':0.4 Avg. 57.4 ± 1.3 (10) 56.6 ± 2.0 (10) 56.5 :': 1.8 (10) ±0.5 Bolls Range 56-62 56-60 56-61 Range 56-60 52-59 53-59 7 Closed Avg. 56.3 ± 2.9 (10) 57.5 ± 3.9 (10) 52 ± 4.1 (10) ±2.8 Avg. 56.6 ± 2.9 (10) 56.2 ± 2.2 (10) 55.6 ± 1.6(10) ±0.5 Bolls Range 53-60 53-59 46-56 Range 53-60 52-59 54-59 9 Closed Avg. 51.6 ± 4.3 (10) 46.4 ± 4.7 (10) 53.4 ± 5.4 (10) ±0.4 Avg. 56.5 ± 2.5 (10) 56.8 ± 1.8 (10) 56.0 ± 1.9 (10) ±0.4 Bolls Range 46-58 40-53 46-58 Range 51-59 55-60 55-59 II Closed Avg. 50.3 :': 1.9 (10) 51.4 :': 1.8 (10) 51.0 :': 1.7 (10) ±0.6 Avg. 51.2 ± 1.8 (10) 52.1 ± 1.1 (10) 52 ± 1.3 (10) ±0.5 Bolls Range 48-51 49-54 48-52 Range 47-53 50-54 50-54 13 Closed Avg. 50.4 ± 3.1 (4) 51.7 ± 1.5 (3) 50.1 :': 1.8 (4) ±0.8 Avg.46 ± 1.5 (10) 45.6 ± 1.4 (10) 45.4 ± 2.2 (10) ±0.3 Bolls Range 41-53 50-53 48-51 Range 44-48 44-49 41-49 15 Avg. 9.2 ± 1.2 (10) 8.9 ± 1.1 (10) 8.8 ± 1.2 (10) :':0.2 Avg. 9.4:': 1.26 (10) 9.1 ± 0.9 (10) 8.2 ± 1.2 (10) :,:0.6