Inhibition of Aflatoxin Production in Aspergillus flavus by a Klebsiella sp. and Its Metabolite Cyclo(l-Ala-Gly)

During an experiment where we were cultivating aflatoxigenic Aspergillus flavus on peanuts, we accidentally discovered that a bacterium adhering to the peanut strongly inhibited aflatoxin (AF) production by A. flavus. The bacterium, isolated and identified as Klebsiella aerogenes, was found to produce an AF production inhibitor. Cyclo(l-Ala-Gly), isolated from the bacterial culture supernatant, was the main active component. The aflatoxin production-inhibitory activity of cyclo(l-Ala-Gly) has not been reported. Cyclo(l-Ala-Gly) inhibited AF production in A. flavus without affecting its fungal growth in a liquid medium with stronger potency than cyclo(l-Ala-l-Pro). Cyclo(l-Ala-Gly) has the strongest AF production-inhibitory activity among known AF production-inhibitory diketopiperazines. Related compounds in which the methyl moiety in cyclo(l-Ala-Gly) is replaced by ethyl, propyl, or isopropyl have shown much stronger activity than cyclo(l-Ala-Gly). Cyclo(l-Ala-Gly) did not inhibit recombinant glutathione-S-transferase (GST) in A. flavus, unlike (l-Ala-l-Pro), which showed that the inhibition of GST was not responsible for the AF production-inhibition of cyclo(l-Ala-Gly). When A. flavus was cultured on peanuts dipped for a short period of time in a dilution series bacterial culture broth, AF production in the peanuts was strongly inhibited, even at a 1 × 104-fold dilution. This strong inhibitory activity suggests that the bacterium is a candidate for an effective biocontrol agent for AF control.


Introduction
Some Aspergillus sp., such as Aspergillus flavus and Aspergillus parasiticus, infect food products and contaminate them with aflatoxins (AFs), which are fungal secondary metabolites with high toxicity and strong carcinogenicity [1,2].AF contamination in food and feed affects human and animal health, and it has a serious impact on the agricultural economy [3].Although AF contamination in crops occurs in tropical and subtropical areas, 4.5 billion people are potentially exposed to AFs through global food distribution [4,5].As a large amount of food contaminated with AFs is discarded because of regulations in many countries, preventing AF contamination is necessary for protecting human health and resolving food shortages [6].However, few methods are currently available for preventing AF contamination; therefore, it is critical to develop effective and practical methods for this purpose.
Although the use of antifungal agents is a possible method for controlling AF contamination [7], their use can lead to the emergence of strains resistant to antifungal agents used in AF control and medicine.Because AFs are fungal secondary metabolites that are not necessary for fungal growth, AF production inhibitors that do not affect fungal growth can be used for AF control without incurring the rapid spread of resistant strains [8].Inhibitors that specifically target AF production are also useful biological probes for investigating the AF production mechanism in fungi, which is important for developing novel AF control methods.Therefore, we have been studying AF production inhibitors obtained from microbial metabolites, essential oils, pesticides, and food additives and investigating their modes of action in inhibiting AF production [8,9].
Fungi and bacteria have been assessed as biocontrol agents which are friendlier to the environment than chemical pesticides in preventing AF contamination.Non-aflatoxigenic strains of Aspergillus that can competitively exclude aflatoxigenic strains from crops are used in practice for AF control [10,11].However, with this method, safety issues, such as the production of other toxic metabolites by the non-aflatoxigenic strain, have not been eliminated [12,13].Many microbes have displayed potential for AF control, and they can be classified into the following groups: group 1, which can inhibit fungal growth and, consequently, AF production; group 2, which can inhibit AF production without affecting fungal growth; group 3, which can inhibit AF production and produce an AF production inhibitor; group 4, which can degrade AFs; and group 5, which can absorb AFs on their cells.Excellent review articles on these AF control microbes have recently been published [14,15].We have been studying the group 3 microbes and found that Stenotrophomonas sp. can inhibit AF production by A. flavus and A. parasiticus and produce the AF production-inhibiting diketopiperazines cyclo(L-Val-L-Pro) and cyclo(L-Ala-L-Pro) [16].Cyclo(L-Leu-L-Pro), a similar diketopiperazine that inhibits AF production, was isolated from the group 3 microbe Achromobacter xylosoxidans [17].The AF production-inhibiting diketopiperazines isolated from group 3 microbes have commonly contained L-proline residues.
Recently, during an experiment in which aflatoxigenic A. flavus was cultivated on peanuts, a bacterium adhering to the peanuts was found to inhibit AF production by the fungus.The bacterium, designated strain KTTM, was found to produce an inhibitor of AF production.Here, we describe the isolation and identification of the strain KTTM and the AF production inhibitor produced by the strain, as well as the effects of the inhibitor, its related compounds, and the bacterial cells on AF production.The inhibitor was identified as a diketopiperazine without the L-proline residue, and the cells exhibited a strong inhibitory activity of AF production by A. flavus cultivated on peanuts.

Isolation and Identification of the Strain KTTM
The strain KTTM was discovered from peanut paste in an experiment conducted in another study.To investigate the AF production of A. flavus in peanut paste, peanuts were ground in a grinder, dispensed into the wells of a 12-well plate, and inoculated with A. flavus.After several days, we found that fungal growth in one well was severely inhibited.A dilution series with water was prepared from the paste of this well and spread on agar plates.Several single colonies were obtained, and each bacterium involved in a colony was tested for AF production-inhibitory activity by co-culturing A. flavus with the bacterium in a liquid culture.A bacterium that inhibited the AF production was designated as strain KTTM, which was a rod-shaped, Gram-negative bacterium, and it was identified as Klebsiella aerogenes by morphological and biochemical analysis as well as by a comparison of its 16S rDNA sequence with those in a database (99.6% identity with Klebsiella aerogenes NBRC 13534 [AB680425]; Figure S1).

Isolation and Identification of the AF Production Inhibitor Produced by the Strain KTTM
The culture supernatant of the KTTM strain was applied to a charcoal column packed with water and eluted with 10% ethanol after washing the column with water.AF pro-duction inhibition, which was tested in a liquid medium inoculated with A. flavus, was observed in the 10% ethanol eluate.The 10% ethanol eluate fraction was purified twice by HPLC using a C 18 column and further purified by HPLC using a C 22 column to obtain the active component.
The 1 H and 13 C NMR spectra of the active component (Figures S2 and S3) suggested the presence of one residue each of Ala and Gly in it, and its molecular formula was determined as C 5 H 8 N 2 O 2 based on the HRESI-Q/TOF mass spectrum.These findings strongly indicated that the active component was a diketopiperazine consisting of Ala and Gly.A comparison of the optical rotation value and the retention time on HPLC with those of commercial cyclo(L-Ala-Gly) identified the active component as cyclo(L-Ala-Gly) (Figure 1).

Isolation and Identification of the AF Production Inhibitor Produced by the Strain KT
The culture supernatant of the KTTM strain was applied to a charcoal column with water and eluted with 10% ethanol after washing the column with water.AF p tion inhibition, which was tested in a liquid medium inoculated with A. flavus, w served in the 10% ethanol eluate.The 10% ethanol eluate fraction was purified tw HPLC using a C18 column and further purified by HPLC using a C22 column to ob active component.
The 1 H and 13 C NMR spectra of the active component (Figures S2 and S3) sug the presence of one residue each of Ala and Gly in it, and its molecular formula w termined as C5H8N2O2 based on the HRESI-Q/TOF mass spectrum.These fi strongly indicated that the active component was a diketopiperazine consisting of A Gly.A comparison of the optical rotation value and the retention time on HPLC wit of commercial cyclo(L-Ala-Gly) identified the active component as cyclo(L-Ala-Gly) 1).

AF Production Inhibition by Cyclo(L-Ala-Gly) and Related Compounds
Cyclo(L-Ala-Gly) inhibited aflatoxins B1 (AFB1) production by A. flavus with value of 0.75 mM in a liquid culture (Table 1).The compound did not affect fungal m weight even at a concentration of 10 mM (Figure 2).The inhibitory activity of cyclo Gly) was stronger than that of cyclo(L-Ala-L-Pro), which displayed the strongest among the three known L-Pro-containing diketopiperazines.Cyclo(D-Ala-Gly) ex no inhibitory activity at 10 mM, indicating the importance of the L stereochemistr Ala residue to the inhibition activity.Five compounds similar to cyclo(L-Ala-Gly), namely, cyclo(L-Abu(2)-Gly), cyc Gly), cyclo(L-Val-Gly), cyclo(L-Nva-Gly), and cyclo(L-Leu-Gly) (Figure 1), were pr

AF Production Inhibition by Cyclo(L-Ala-Gly) and Related Compounds
Cyclo(L-Ala-Gly) inhibited aflatoxins B 1 (AFB 1 ) production by A. flavus with an IC 50 value of 0.75 mM in a liquid culture (Table 1).The compound did not affect fungal mycelial weight even at a concentration of 10 mM (Figure 2).The inhibitory activity of cyclo(L-Ala-Gly) was stronger than that of cyclo(L-Ala-L-Pro), which displayed the strongest activity among the three known L-Pro-containing diketopiperazines.Cyclo(D-Ala-Gly) exhibited no inhibitory activity at 10 mM, indicating the importance of the L stereochemistry of the Ala residue to the inhibition activity.
Table 1.IC 50 values of cyclo(L-Ala-Gly) and related compounds for inhibiting AF production.

Inhibition of AF Production by the Strain KTTM
The effect of the KTTM strain on AF production was tested by cultivating A. flavus on peanuts.After autoclaving law peanuts lacking shells and skins, each peanut was shortly dipped in the strain KTTM culture broth or a dilution of the culture broth and inoculated with A. flavus spores.After 30 days of incubation at 25 • C, the amount of AF in each peanut was measured.AF production was significantly reduced even upon exposure to 1 × 10 4 -fold diluted broth (6.2 × 10 4 cells/mL) without affecting the growth of A. flavus (Figures 4 and S4).

AF Degradation Activity of the Strain KTTM
As a report has demonstrated that Klebsiella sp.degrades AF [18], the AF degradation activity of the strain KTTM was examined.When AFB 1 was added to a liquid medium and the strain KTTM was cultured in the medium for 3 days, the concentration of AFB 1 in the culture broth was consistent with that of the control in which AFB 1 was kept in the liquid medium without inoculation of the bacterium for 3 days (Figure 5).By contrast, when AFB 1 was incubated in the KTTM strain supernatant of the culture broth, its concentration was slightly lower than that of the control (Figure 5).
the structure of the side chain.Cyclo(L-Abu(2)-Gly), cyclo(L-Val-Gly), and cyclo(L-Nva-Gly), in which the methyl group of cyclo(L-Ala-Gly) was replaced by ethyl, isopropyl, and propyl, respectively, more strongly inhibited AF production (75-, 18.8-, and 8.3-fold stronger, respectively) than cyclo(L-Ala-Gly).However, cyclo(L-Leu-Gly), possessing an isobutyl group, exhibited weaker activity than cyclo(L-Ala-Gly).Cyclo(Gly-Gly), lacking side chains, displayed weak activity at 5 mM (Figure 3).dipped in the strain KTTM culture broth or a dilution of the culture broth and ino with A. flavus spores.After 30 days of incubation at 25 °C, the amount of AF in each was measured.AF production was significantly reduced even upon exposure to fold diluted broth (6.2 × 10 4 cells/mL) without affecting the growth of A. flavus (F and S4).

AF Degradation Activity of the Strain KTTM
As a report has demonstrated that Klebsiella sp.degrades AF [18], the AF degr activity of the strain KTTM was examined.When AFB1 was added to a liquid med the strain KTTM was cultured in the medium for 3 days, the concentration of AF culture broth was consistent with that of the control in which AFB1 was kept in th medium without inoculation of the bacterium for 3 days (Figure 5).By contras Toxins 2024, 16, x FOR PEER REVIEW AFB1 was incubated in the KTTM strain supernatant of the culture broth, its conce was slightly lower than that of the control (Figure 5).

Discussion
We found a bacterium, named strain KTTM, with strong AF production-inhibitory activity by chance.The strain was identified as K. aerogenes, and it had a high 16S rDNA sequence homology with Klebsiella aerogenes NBRC 13534.The KTTM strain culture broth effectively inhibited AF production by A. flavus even after being diluted 1 × 10 4 -fold, suggesting that a mass number of bacterial cells is not necessary for AF control when the strain is used in practice as a biocontrol agent.
The diketopiperazine cyclo(L-Ala-Gly) was isolated from the culture supernatant of the KTTM strain as an AF production inhibitor.The diketopiperazines produced by microbes can be classified into two groups, namely, one group with simple structures consisting of two amino acids and another group with relatively complex structures biosynthesized by modifying the simple diketopiperazines of the first group [20].The latter group includes many secondary metabolites with a variety of bioactivities, and they are produced mainly by fungi and Actinobacteria.Concerning the former group, many diketopiperazines, such as cyclo(L-Ala-Gly), are metabolites of a wide range of microbes [21,22].Although some diketopiperazines of the former group exhibit bioactivities such as antimicrobial activity [23], their bioactivities are less specific and weaker than those of the latter group.Therefore, there are few in-depth studies on the bioactivities of the former group of diketopiperazines, excluding those that inhibit AF production.Cyclo(L-Ala-Gly) and related compounds, which were found to inhibit AF production in this study, could be useful as both lead compounds for developing AF control agents and as biological probes for investigating the regulatory mechanism of AF production in aflatoxigenic fungi.
The culture supernatant of two Klebsiella sp.strains was reported to degrade AF [18], as the amount of AF decreased by approximately 70% after incubation with the culture supernatant for 3 days.Compared to the activity of the reported strains, that of the strain KTTM was weak.As the amount of AF did not decrease upon culture with the strain KTTM, the relationship between the AF degradation in the culture supernatant of the strain KTTM and the AF production-inhibitory activity of the strain appears weak.It might be important to clarify the relationship between cyclo(L-Ala-Gly) production and the AF production-inhibitory activity of KTTM to determine the mechanism underlying the strain's strong inhibitory activity.
We previously found that cyclo(L-Ala-L-Pro) inhibited AfGST activity [19].As fungal GST is believed to play a role in the response to oxidative stress, a key factor for AF production by aflatoxigenic fungi, we speculated that cyclo(L-Ala-L-Pro) would inhibit AF production by inhibiting AfGST activity.Although cyclo(L-Ala-Gly) and cyclo(L-Ala-L-Pro) commonly possess an L-Ala-containing diketopiperazine skeleton and inhibit AF production, cyclo(L-Ala-Gly) did not inhibit AfGST activity at the concentrations tested, suggesting that the mechanism by which cyclo(L-Ala-Gly) inhibits AF production does not involve AfGST inhibition.Research to identify the target molecule of cyclo(L-Ala-Gly) for AF production inhibition to clarify the mode of action of the compound and its congeners is currently ongoing.

Conclusions
The strain KTTM, identified as K. aerogenes, strongly inhibited the AF production by A. flavus, suggesting that the strain is a candidate agent for AF control.Cyclo(L-Ala-Gly) was isolated from the KTTM strain culture broth as an AF production inhibitor.The simple structure and strong AF production-inhibitory activity of cyclo(L-Ala-Gly) and its related compounds highlight their utility as AF control agents and biological probes for investigating the mechanisms of AF production by aflatoxigenic fungi.

Strains and Culture
A. flavus IFM 47798 was used as a producer of AFB 1 and aflatoxin B 2 (AFB 2 ).The strain was cultured on potato dextrose agar medium (BD, Franklin Lakes, NJ, USA) at Toxins 2024, 16, 141 8 of 12 25 • C for 2 weeks.A spore suspension was prepared from the culture at a concentration of 1.1 × 10 5 CFU/µL and used as the stock and inoculum.
The strain KTTM was isolated as explained in Section 2.1.by a dilution method using a Bennet liquid medium consisting of glucose 1%, peptone 0.2%, meat extract 0.1%, and yeast extract 0.1%, with a pH of 7.2.KTTM was identified as Krebsiella aerogenes by TechnoSuruga Laboratory (Shizuoka, Japan).The strain was cultured at 27.5 • C on a rotary shaker (150 rpm) for 3 days in a Bennet liquid medium, and the culture broth was stored as a 20% glycerol stock at −80 • C. The number of bacterial cells was determined from the number of colony-forming units.

Assay Method in Liquid Culture
A sample water solution (100 µL), passed through a 0.25 µm filter, or a bacterial culture broth (5 µL), cultured in a Bennet liquid medium at 27.5 • C for 3 days, was added to the potato dextrose liquid medium (1.9 or 2.0 mL) in a well of a microplate (24 wells).A spore suspension of A. flavus (5 µL) was inoculated into the medium and incubated statically for 4 days at 25 • C. AFB 1 was produced mainly under these culture conditions.A mixture of 100 µL of culture broth and 400 µL of water:acetonitrile (9:1, v/v) was filtered (MinisartRC4, Sartorius, Gottingen, Germany), and the amount of AFB 1 involved in the filtrate was analyzed by HPLC on a 250 mm × 4.6 mm inner diameter Capcell pak C 18 UG 120 column (Osaka Soda, Osaka, Japan) via the isocratic elution of acetonitrile:methanol:water (1:3:6, v/v/v) over 20 min.at a flow rate of 1.0 mL/min with fluorescence detection at 450 nm (excitation: 365 nm).

Assay Method Using Peanuts
Peanuts (Chiba-handachi) were grown from seedlings, harvested, and stored in their shells at −25 • C.After removing the shells and skins from the raw peanuts (each weighing approximately 1.2 g), the peanuts were autoclaved.After being dipped in the strain KTTM culture broth, cultured in a Bennet medium at 27.5 • C for 3 days, or dipped in a dilution series of the culture broth for a few seconds, each peanut was put into a well of a microplate (12 wells), inoculated with A. flavus spores (5 µL), and incubated statically for 30 days at 25 • C. AFB 1 and AFB 2 were produced under these culture conditions.Each peanut, including the fungal mycelia grown on the peanut, was crushed and extracted with 5 mL of water: acetonitrile (1:9, v/v).The supernatant, obtained by centrifugation, was passed through a cartridge for AF purification (Autoprep MF-A 1000, Resonac Co., Tokyo, Japan), and 1 mL of the pass-through solution was lyophilized.The residue was dissolved in 400 µL of water:acetonitrile (9:1, v/v) and filtered (MinisartRC4), and then the AFB 1 and AFB 2 contents in the solution were analyzed by HPLC as previously explained.

Isolation of the Active Component from the Strain KTTM
The bacterial glycerol stock (210 µL) was inoculated into a Bennett medium (7 mL) in a test tube and incubated at 27.5 • C for 2 days at 150 rpm.This preculture (3 mL) was transferred into a Bennett medium (100 mL) in a 500 mL Erlenmeyer flask, which was incubated at 27.5 • C for 5 days at 150 rpm.
The cyclo(D-Ala-Gly) was identified as follows

AF Degradation Activity of the Strain KTTM
The AF acetonitrile solution (50 or 100 µL, 2.5 ppm of AFB 1 ) was placed in each well of the microplate (12 wells), and the plate was left in a safety cabinet for 1 h to remove the acetonitrile.A liquid Bennet medium (2 mL) with or without inoculation with the KTTM strain (5 µL of the culture broth cultured in a liquid Bennet medium for 3 days) or the KTTM strain supernatant of the culture broth (1 mL), centrifuged and passed through a 0.25 µm filter, was placed in each well (to a final concentration of AFB 1 : 0.125 ppm).After 3 days of incubation at 27.5 • C, the AFB 1 amount in each well was measured according to the method described in Section 5.2.

Measurement of the GST Activity of AfGST-FLAG
The GST activity was measured according to a previously reported method [19], with some modifications.Briefly, AfGST-FLAG (2 µg) in 100 µL of 50 mM Tris-HCl (pH 8.0) was incubated with ethacrynic acid or diketopiperazines and glutathione (GSH) at room temperature for 30 min.After incubation, 100 µL of 50 mM Tris-HCl (pH 8.0) containing 1-chloro-2,4-dinitrobenzene (CDNB) was added to the mixture, and the absorbance measurement at 340 nm was started immediately (0 min).The final concentrations of GSH and CDNB were 5 and 1 mM, respectively.The absorbance at 0 min was subtracted from that at 3 min to estimate the ∆A 340 /min.The ratio of the ∆A 340 /min was calculated using the control without chemicals as 1.

Statistical Analysis
Boxplots were generated using the ggplot2 package of R (https://www.r-project.org,accessed on 16 June 2023).The statistical differences of the treated groups compared to the controls were determined by ordinary one-way ANOVA followed by Dunnett's test when the variances of all groups could be assumed to be equal and by Brown-Forsythe and Welch ANOVA followed by Dunnett's T3 test when the variances could not be assumed to be equal.Statistical analyses were performed with GraphPad Prism 10 ver.10.1.1 (GraphPad Software, La Jolla, CA, USA).

Figure 2 .
Figure 2. Effect of cyclo(L-Ala-Gly) on the mycelial weight of A. flavus.Boxplots of the dried lial weight.The colored dots indicate individual values.n = 4.

Figure 2 .
Figure 2. Effect of cyclo(L-Ala-Gly) on the mycelial weight of A. flavus.Boxplots of the drie lial weight.The colored dots indicate individual values.n = 4.

Figure 2 .
Figure 2. Effect of cyclo(L-Ala-Gly) on the mycelial weight of A. flavus.Boxplots of the dried mycelial weight.The colored dots indicate individual values.n = 4.

Figure 4 .
Figure 4. Effects of the strain KTTM on AF production by A. flavus grown on peanuts.Bo the relative amounts of aflatoxins B1 and B2.The mean value of AF amounts in the 0 (contro was set as 100.The colored dots indicate individual values.n = 4 (treated) or 13 (control).versus control, Brown-Forsythe and Welch ANOVA followed by Dunnett's T3 test.

Figure 4 .
Figure 4. Effects of the strain KTTM on AF production by A. flavus grown on peanuts.Boxplots of the relative amounts of aflatoxins B 1 and B 2 .The mean value of AF amounts in the 0 (control) group was set as 100.The colored dots indicate individual values.n = 4 (treated) or 13 (control).* p < 0.05 versus control, Brown-Forsythe and Welch ANOVA followed by Dunnett's T3 test.

Figure 5 .
Figure 5. AF degradation activity of the strain KTTM.Boxplots of the AFB1 amounts.Th dots indicate individual values.n = 4. * p < 0.05 versus control, ordinary one-way ANOVA by Dunnett's test.

Figure 5 .
Figure 5. AF degradation activity of the strain KTTM.Boxplots of the AFB 1 amounts.The colored dots indicate individual values.n = 4. * p < 0.05 versus control, ordinary one-way ANOVA followed by Dunnett's test.