Variations in Kojic Acid Production and Corn Infection Among Aspergillus flavus Isolates Suggest a Potential Role as a Virulence Factor

Abstract

Kojic acid is a secondary metabolite with strong chelating and antioxidant properties produced by Aspergillus flavus and A. oryzae. Although antioxidants and chelators are important virulence factors for plant pathogens, the ecological role of kojic acid remains unclear. We previously observed a greater gene expression of antioxidants, especially kojic acid, by non-aflatoxigenic A. flavus when co-cultured with aflatoxigenic A. flavus. Aflatoxin production was also reduced. In this study, we investigated kojic acid production in 22 A. flavus isolates from Louisiana and compared them to four common A. flavus strains in liquid medium and on corn kernels. Corn kernel infection was assessed by quantifying the maize beta tubulin DNA content of the kernels using drop digital PCR (ddPCR). Maize beta tubulin DNA content decreased with increased corn kernel infection. Greater kojic acid production by A. flavus isolates coincided with greater levels of corn kernel infection. All isolates produced 60 and 700 times more kojic acid than aflatoxin and cyclopiazonic acid (a known virulence factor), respectively, which varied among sclerotial size categories. A. flavus strains with small sclerotia, which were rarely isolated from corn, produced the least kojic acid and infected corn kernels the least, while medium and large sclerotia strains-mainly isolated from corn-produced the most kojic acid and were more infectious. Non-aflatoxigenic isolates from Louisiana produced the most kojic acid. These results suggest that kojic acid is a potential virulence factor and may increase the pathogenic success of medium and large sclerotia-producing A. flavus, which could ultimately lead to more effective A. flavus biocontrol strains. Further studies are required to determine the effects that kojic acid has on the redox environment during corn infection and how the altered redox environment decreases aflatoxin production.

Keywords: aflatoxin; biological control; ddPCR; mycotoxins; pathogenicity; plant–fungal interactions; secondary metabolism.

Figure 1

Kojic acid production by Aspergillus flavus isolates in (a) standard medium and (b) corn kernels. Kojic acid was measured for isolates belonging to different vegetative compatibility groups (VCGs) and with different sclerotia sizes (small (S), mixed (M), and large (L)) grown in (a) standard medium and (b) corn kernels for five and six days, respectively. The VCG determined in Sweany et al. 2011 [23] is denoted as the first number in the isolate name followed by a dash and the SRRC collection number. Isolates are ordered based on their phylogenetic similarities reported in Sweany et al. 2024 [28]. Within (a,b), means and standard error followed by the same letters are statistically similar based on preplanned comparisons of LS-means (α < 0.05) implemented in linear models. N.D. indicates that data are missing for an isolate.

Figure 2

Aflatoxin B₁ production by Aspergillus flavus isolates in (a) standard medium and (b) corn kernels. Aflatoxin B₁ was measured for isolates belonging to different vegetative compatibility groups (VCGs) and with different sclerotia sizes (small (S), mixed (M), and large (L)) grown in (a) standard medium and (b) corn kernels for five and six days, respectively. The VCG determined by Sweany et al. 2011 [23] is denoted as the first number in the isolate name followed by a dash and the SRRC collection number. Isolates are ordered based on their phylogenetic similarities reported by Sweany et al. 2024 [28]. Means and standard error bars followed by the same letters are statistically similar based on preplanned comparisons of LS-means (α < 0.05) implemented in a single linear model, which included both standard medium and corn kernel aflatoxin values. N.D. indicates that data are missing for an isolate. (c) Kojic acid production values of each isolate are plotted against aflatoxin production values in both substrates, which did not have a linear relationship in either standard medium (F_1,33.8 = 1.55, p = 0.221) or corn kernels (F_1,41.3 = 0.00, p = 0.960).

Figure 3

Cyclopiazonic acid production (CPA) by Aspergillus flavus in corn kernels. (a) CPA was measured for isolates belonging to different vegetative compatibility groups (VCGs) and with different sclerotia sizes (small (S), mixed (M), and large (L)) grown in corn kernels for six days. The VCG determined by Sweany et al. 2011 [23] is denoted as the first number in the isolate name followed by a dash and the SRRC collection number. Isolates are ordered based on their phylogenetic similarities reported by Sweany et al. 2024 [28]. Means and standard error followed by the same letters are statistically similar based on comparisons of LS-means (α < 0.05) implemented in linear models. N.D. indicates that data are missing for an isolate. (b) In the top panel, CPA values are plotted against aflatoxin for each isolate. Estimated linear relationships between CPA and aflatoxin are depicted by lines. In the bottom panel, CPA is plotted against kojic acid for individual isolates, and since there is no significant linear relationship between CPA and kojic acid, no lines are depicted.

Figure 4

Conidiospore production of Aspergillus flavus isolates on corn kernels. Green conidia and black sclerotia on corn inoculated with isolates belonging to different vegetative compatibility groups (VCGs) and with different sclerotia sizes (small (S), mixed (M), and large (L)). The VCG determined by Sweany et al. 2011 [23] is denoted as the first number in the isolate name followed by a dash and the SRRC collection number. The same letter behind the isolate name refers to the mean conidial production that is statistically similar based on preplanned comparisons of LS-means (α < 0.05) implemented in linear models. Isolates are ordered in columns from top to bottom followed by rows from left to right based on their phylogenetic similarities reported by Sweany et al. 2024 [28].

Figure 5

Relative levels of genomic DNA in corn kernels inoculated with A. flavus over an 11-day time course. Droplet digital PCR was used to amplify the target genomic sequence (beta tubulin), and the ratio (i.e., proportion (p)) of positive droplets that were amplified to total droplets formed was determined. The same letters reported above average ratios (p) with standard error bars are statistically similar based on comparisons of LS-means (α < 0.05) implemented in generalized linear models.

Figure 6

Ratio of maize gene content (p) is reduced during infection with isolates that produced more kojic acid, CPA, aflatoxin, and conidia. (a) The ratio of maize beta tubulin gene droplets to total droplets (p) was measured with droplet digital PCR for total DNA extracted from kernels inoculated with large (L), mixed (M), and small (S) sclerotia isolates. Smaller ratios (p) indicate less corn DNA in a sample and a greater level of corn infection. (b) The ratio of beta tubulin (p) was plotted against kojic acid, CPA, conidiospore, and aflatoxin production for each isolate. Estimated lines were also plotted if there was a statistically significant linear relationship between kojic acid, CPA, conidiospore, aflatoxin production, and maize beta tubulin within sclerotial size groups.