Postharvest Pathology and Mycotoxins Virulence and Cultural Characteristics of Two Aspergillus flavus Strains Pathogenic on Cotton

Cotty, P. J. 1989. Virulence and cultural characteristics of two Aspergillus jlavus strains pathogenic on cotton. Phytopathology 79:808-814. Seventy Aspergillus flavus isolates from Arizona desert valleys were sorted into two distinct strains on the basis of sclerotial size, cultural characteristics, and virulence to cotton. Strain L isolates produced large sclerotia (over 400 /lm in diameter), and strain S isolates produced small sclerotia (less than 400 /lm in diameter). Strain S isolates produced greater quantities of sclerotia in a wider assortment of media and a broader temperature range than strain L isolates. Isolates of both strains exhibited pH homeostasis in culture. However, the strains maintained different pH values. Strain S isolates produced more aflatoxin than strain L isolates in culture~ however, the strains produced similar aflatoxin levels in developing cottonseed. Strain L isolates were more aggressive than strain S isolates at deteriorating cotton boll locks (locules) and spreading within A.\jJergillus flavus is the primary cause of aflatoxin contamination of corn, cottonseed, and tree nuts (10). A. flavus populations consist of individuals that vary widely in their ability to prod uce sclerotia or aflatoxins in vitro (13,18,19), and isolates are often grouped on the basis of the quantities of aflatoxins and sclerotia produced (12,17). Aflatoxin production in vitro is thought to reflect the ability of an isolate to contaminate plant tissues (23,26). The relationship between A. flavus isolates and preharvest aflatoxin contamination of crops is complicated both by changes that occur during culture in the abilities of isolates to produce aflatoxins (9,23) and by a lack of experimental evidence on the relation of toxin prod uction in vitro to virulence on host plants. Several studies correlated sclerotial production in vitro with aflatoxin production (9,14,16,22), and a single study found that This article is in the public domain and not copyrightable. It may be freely reprinted with customary crediting of the source. The American Phytopatholog ical Society, 1989. 808 PHYTOPATHOLOGY bolls, and this tendency partly explains the difference between in vitro and in vivo production of aflatoxin. The aflatoxin level in infected seeds was not correlated with either the aflatoxin level in vitro, intraboll fungal spread, or lock deterioration, but it was correlated with the product of the aflatoxin level in vitro and the intraboll spread parameter (r = 0.75-0.91). The correlation increased slightly (r = 0.78-0.93) when lock deterioration was factored into the equation. These results indicate that pathogenic aggressiveness contributes to the ability of an isolate to contaminate cottonseed with aflatoxins. The lack of correlation between pathogenic aggression and aflatoxin production in vivo suggests aflatoxin does not enhance virulence in the cotton-A. jlavus interaction. cottonseed infected with A.flavus contained higher aflatoxin levels when sclerotia were present than when the seed contained no sclerotia (25). Furthermore, aflatoxin biosynthesis and sclerotial morphogenesis may have interrelated regulation (7). "Atypical" isolates of A. flavus from agricultural soils in Thailand have recently been reported (21). These isolates produced abundant small sclerotia and large quantities of aflatoxins. In my laboratory, similar A. flavus isolates were obtained from cottonseed and soil collected in Arizona, where aflatoxin contamination of cottonseed is a perennial problem. It was speculated that the atypical isolates represent a distinct A. flavus strain. The repeated association of aflatoxin production with sclerotial production led to further speculation that this strain may be an important causative agent of aflatoxin contamination of cottonseed. The purpose of this study was to determine if Arizona desert valleys bear two A ..[lavus strains distinguishable by cultural characteristics and virulence to cotton. The relationship of aflatoxin production to the virulence of A. flavus on cotton was also investigated. MATERIALS AND METHODS Cultures and strains. The origins of the isolates studied are described in Table 1. All isolates examined produced sclerotia in culture. For convenience, isolates producing numerous small sclerotia were designated strain S ("small") isolates, which are synonymous with the atypical isolates of Saito et al (21); isolates with any sclerotia greater than 400 Jlm in diameter were designated strain L ("large") isolates. Soil isolates were obtained on a dicloran-amended medium by the dilution plate technique (3). Distinct colonies were transferred to 5/2 agar (5% V-8 juice and 2% agar, adjusted to pH 5.2 prior to autoclaving) and utilized directly in the studies. Seed isolates were obtained on 5/2 agar either directly from the lint or from pieces of manually delinted seed disinfected for 2 min with 95% ethyl alcohol; stock cultures of seed isolates originated from single spores. The fungi were maintained in the dark on 5/2 agar at 25-30 C. For long-term storage, plugs of mature cultures (3 mm in diameter) were maintained in sterile distilled water (7). Sclerotial production. Conidial suspensions (10 Jll) were seeded into the centers of petri dishes (9 cm in diameter) containing 30-35 ml of Czapek solution agar (CZ) with 3% NaN03 and incubated at either 25, 30, or 38 C for 12 days in the dark. Temperatures were maintained within 1 • Subsequently, conidia were washed from the plates with 95% ethyl alcohol, and the number of sclerotia per plate was estimated: 0 == no sclerotia, and I == 1-20, 2 == 50-100, 3 == 200-400, 4 == 500-1,000, and M == more than 1,000 sclerotia. Each test evaluated five to 12 isolates, and isolates were tested two to four times. Sclerotial production by several typical isolates of each strain was evaluated at 30 C on 5/2 agar, potato-dextrose agar (PDA), and CZ with either 0, 2, 3, or 6% NaN03. Sclerotia in three portions (each 4 cm 2) of each plate were counted after 12 days. The two tests each contained two replicates and compared four to six isolates. Sclerotial size. Diameters of sclerotia produced on CZ with 3% NaNO) at 30 C were determined by video image analysis. Conidia were washed from the plates with 95% ethyl alcohol, and sclerotia were dislodged with a spatula and fixed in a mixture of ethanol and glacial acetic acid (I: I, v/ v). This procedure reduced human exposure to conidia. Sclerotia were blotted dry and stuck to clear plastic plates with two-sided tape. Sclerotial silhouette areas were determined by video image analysis, and diameters were calculated on the assumption that the sclerotia were spherical. Aflatoxin production in vitro. Aflatoxin production was assessed qualitatively by seeding conidial suspensions (10 Jll) in petri dishes (9 cm in diameter) containing 30-35 ml of solidified Adye and Mateles medium (A&M agar) (6,15). After 10 days of growth at 30 C agar cultures were transferred to 250-ml jars with 25 ml of acetone and shaken for I min. Methylene chloride (25 ml) was added, the mixture was shaken for 1 min and filtered through 25 g of sodium sulfate, and the filtrate was evaporated to dryness. Residues were solubilized in methylene chloride for thin-layer chromatography (TLC). Extracts and aflatoxin standards were separated on TLC plates (silica gel 60, 250 Jlm thick) by development with a mixture of diethyl ether, methanol, and water (96:3: I, v/ v) (24) and examined under ultraviolet light. Isolates negative for aflatoxin production in initial tests were grown on three plates containing A&M agar, which were combined prior to extraction. Quantitative estimates of aflatoxin production in vitro were made by the rapid fluorescence method (6). Culture tubes containing 5 ml of A&M agar (6,15) were seeded with 100-JlI spore suspensions containing approximately 100 spores per microliter. After 3 days of incubation at 30 C, agar fluorescence 5 mm beneath the mycelial mat was measured with a scanning densitometer. Agar fluorescence measured in this manner is directly correlated with the quantity of aflatoxin in the culture (6). Aflatoxin was extracted with solvents from representative tubes and quantified by TLC in order to construct a standard curve (fluorescence vs. toxin concentration) for each experiment (6). Three experiments were performed, containing eight, 24, and 44 isolates replicated six, seven, and four times, respectively. Fungal influence on culture pH. A. flavus modifies culture pH in an isolate-specific manner (12,20); this characteristic has been used to group A. flavus isolates (12). Therefore, potential differences in culture pH between strains were evaluated. Conidial suspensions (10 Jll) containing about 10,000 spores were seeded in the centers of culture plates (9 cm in diameter) containing 25 ml of CZ with 0.05 M Hepes buffer adjusted to pH 6, 7, or 8 prior to autoclaving. After 10 days of growth at 30 C in the dark, the agar pH was measured with a flat electrode in the center of the plate on the reverse side of the culture. Aflatoxin production in vivo. Infection of cotton bolls by A. flavus and the subsequent aflatoxin contamination of cottonseed occur in the field via pink bollworm exit holes (2). Therefore, the ability of isolates to infect simulated pink bollworm exit holes and contaminate cottonseed with aflatoxins was determined in the greenhouse, as previously described (8). Gossypium hirsutum 'Deltapine Acala 90' was grown in 3-L pots containing a I: I mixture of Pro-mix (Premier Brands, New Rochelle, NY) and sand. After 21 days, the plants were fertilized weekly with 100 ml of Miracle-Gro (Stern's Miracle-Gro Products, Port Washington, NY) at 2,000 ppm. Bolls 25-28 days old were wounded with a cork borer (3 mm in diameter) and inoculated with 10 JlI of a conidial suspension containing approximately 10,000 spores. Bolls were harvested after 4 wk and dried at 60 C for 72 hr. The aflatoxin content of intact inoculated locks (locules) was determined by a modification (8) of the method of the Association of Official Analytical Chemists (24). Intact locks were hammered to pulverize the seed and added to 200 ml of a mixture of acetone and water (85: 15). The mixture was shaken for 15 sec, set overnight, and filtered with number 4 Whatman paper. The filtrate (100 ml) was mixed with 20 ml of a solution of zinc acetate and aluminum chloride (1.1 M (CHJCOOhZn and 0.04 M AICIJ), together with 80 ml of water and 5 g of diatomaceous earth. The mixture was shaken, left to settle for 1-2 hr, and passed through number 4 Whatman filter paper. The filtrate (100 ml) was extracted twice with 25 ml of methylene chloride. Fractions were pooled and concentrated to dryness, and the residues were solubilized in methylene chloride for TLC. Aflatoxins were separated and quantified by TLC (24). The plants were maintained at all times in complete randomized blocks. Each experiment evaluated eight isolates, which were replicated eight times. The replicates consisted of one plant with one or two bolls in the first experiment and two plants with two to five bolls in the second experiment. Evaluation of pathogenic aggression. Intralock fungal spread was used as one measure of pathogenic aggression. During growth on cotton lint, A. flavus produces kojic acid, which is converted by host peroxidase into a compound with bright green-yellow fluorescence (BGYF) (I). The presence of BGYF on cotton lint is a reliable indicator of the activity of A. flavus (1). BG YF on lint of uninoculated locks, therefore, indicates fungal spread from inoculated to uninoculated locks. Uninoculated locks were examined under ultraviolet light after drying, and BG YF on lint of each lock was rated as follows: 0 == no BGYF; I == BGYF on less than 50% of the lint; 2 == BGYF on more than 50% of the lint. The intraboll spread parameter was the BG YF value; for correlations, the values were transformed (0.1 was added to each value) in order to compensate for the low sensitivity of the technique and to avoid multiplying by O. After the locks were dried, the weight per lock of inoculated and uninoculated locks was determined. The weight of each lock was divided by the number of seeds within the lock, because the number of seeds per lock varied within bolls; the inoculated lock weights were then divided by the uninoculated lock weights to compensate for variance among bolls. For correlations, the resulting value was subtracted from I and squared; this produced a value positively related to aggression. Statistical analysis. Analyses were performed with the Statistical Analysis System (SAS Institute, Cary, NC). All multiple comparisons were first subjected to analysis of variance and then to Fischer's least significant difference test. Toxin and fluorescence Vol. 79, No.7, 1989 809 TABLE I. Origins and sclerotial characteristics of Aspergillus flavus isolates Origin Sclerotial diameter (J,tm)Y Sclerotial productionZ Sample Average Percentage Isolate Cropw Substrate Locale x size ±SD over 400 J,tm 25 C 30 C 38 C 24 S Alf Soil NGV 192 250 ± 42 0 M M M 69 S 132 235 ± 49 0 M M M 37 S SGV 104 212 ± 55 0 M M M 61 L 91 481 ± 109 82 2 4 0 10 S YV 117 217 ± 51 0 M M M 18 S 90 199 ± 40 0 M M M 33 S 219 196 ± 43 0 M M M 19 L Bg Soil YV 125 505 ± 119 82 1 3 0 5S 274 211 ± 47 0 M M M 21 L 20 325 ± 64 10 0 1 0 23 L 62 543 ± 101 92 1 2 0 26 S 405 184 ± 37 0 M M M 60 L 54 514 ± 82 94 2 3 0 65 S 163 191 ± 38 0 M M M 66 S 167 189 ± 41 0 M M M 20 S Cit Soil NGV 141 210 ± 54 0 M M M 25 S 480 238 ± 62 0 M M M 27 S 190 243 ± 43 0 M M M 39 L 61 436 ± 67 75 2 M 0 7S YV 107 200 ± 45 0 M M M 13 L 50 548 ± 103 94 0 2 0 15 S On Soil NGV 268 216 ± 51 0 M M M 59 S 148 209 ± 41 0 M M M 31 S YV 127 239 ± 43 0 M M M 46 S 220 203 ± 44 0 M M M 63 L 44 670 ± 110 98 0 4 0 64 L 80 486 ± 79 86 0 2 0 68 L 54 554 ± 108 91 2 M 0 lL PC Soil YV 10 650 ± 92 100 1 1 0 6L 61 603 ± 108 93 1 2 0 55 L 46 451 ± 89 67 2 3 0 70 S 159 163 ± 31 0 M M M 35 L UC Seed MC 77 396 ± 93 49 3 M 0 16 L 86 542 ± 108 92 0 1 0 17 L 47 691 ± 103 96 0 3 0 38 L 63 433 ± 57 65 3 M 0 41 L 101 447 ± 82 73 1 M 0 43 L 87 471 ± 76 82 4 M 0 45 L 47 447 ± 121 62 2 4 0 47 L 50 549 ± 61 98 2 4 0 48 L 88 431 ± 75 70 4 M 0 49 L 57 549 ± 77 98 0 4 0 51 L 2 480 ± 21 100 0 1 0 52 L 67 493 ± 96 84 2 4 0 53 L 73 409 ± 75 58 0 1 0 56 L 58 548 ± 81 97 1 3 0 57 L 60 628 ± 82 98 0 1 0 28 L YV 3 821 ± 91 100 0 1 0 29 S 134 211 ± 43 0 M M M 36 L 16 503 ± 117 75 1 4 0 40 L 70 506 ± 126 80 2 4 0 42 S 170 281 ± 36 0 M M M 50 L 74 521 ± 71 95 4 M 0 54 L 25 692 ± 130 92 1 2 0 58 S 342 263 ± 47 0 M M M 2L Soil NGV 125 515 ± 103 89 2 4 0 3S YV 127 193 ± 59 0 M M M 4L 131 433 ± 52 73 2 2 0 8L 90 466 ± 83 79 2 1 0 9L 36 595 ± 130 94 0 3 0 11 L 52 642 ± 102 98 0 1 0 12 S 156 217 ± 49 0 M M M 22 L 147 488 ± 147 80 0 1 0 34 L 55 430 ± 121 53 2 4 0 30 S 115 200 ± 53 0 M M M v Isolate numbers are followed by letters denoting the strain. Strain S isolates produce numerous small sclerotia (average diameter less than 300 J,tm); strain L isolates produce larger sclerotia, some of which have a diameter exceeding 400 J,tm. wAif == alfalfa; Bg == bermudagrass; Cit == citrus; On == onion; PC == Pima cotton; DC == upland cotton. xNGV == North Gila Valley; SGV == South Gila Valley; YV == Yuma Valley; MC == Maricopa County. YSilhouette areas were measured by video image analysis; average diameters were calculated from the areas on the assumption that the sclerotia were spherical. ZSclerotia per plate after 12 days on Czapek solution agar with 3% NaN03: I == 1-50, 2 == 50-100, 3 == 200-400, 4 == 500-1,000, and M == more than 1,000 sclerotia per plate.