The Stimulation of Indigenous Bacterial Antagonists by γ-Glutamyl-S-Allyl-l-Cysteine Increases Soil Suppressiveness to Fusarium Wilt

ABSTRACT

The development of suppressive soil is an ideal strategy to sustainably combat soilborne diseases. Previously, the cultivation of Allium plants increased antagonistic bacteria populations in soil, alleviating Fusarium wilt of different crops. This study aimed to identify a compound produced by Allium plants that can induce bacteria-mediated soil suppressiveness toward Fusarium wilt. The amendment of soils with γ-glutamyl-S-allyl-l-cysteine (GSAC), a unique dipeptide abundantly detected in the root extract of Welsh onion (Allium fistulosum), significantly suppressed Fusarium wilt diseases, whereas three other commercial dipeptides had no such effects. GSAC application did not suppress the disease in sterilized soil. Furthermore, the suppressiveness of soil amended with GSAC could be transferred to sterilized soil via soil microflora transplantation. This suppressiveness was eliminated by pretreating GSAC-amended soil microflora with antibacterial antibiotics, indicating that the suppressiveness of GSAC-amended soil is generated by the activity of antagonistic bacteria. Amplicon sequencing of the 16S rRNA gene revealed that GSAC application significantly increased the relative abundance of Pseudomonas (OTU224), Burkholderia-Caballeronia-Paraburkholderia (OTU387), and Bdellovibrio (OTU1259) in soils. Surprisingly, the relative abundance of OTU224 was significantly greater in Welsh onion rhizospheres than in noncultivated soil. Pseudomonas strains corresponding to OTU224, isolated from Welsh onion rhizospheres, displayed a remarkable suppressive effect against cucumber Fusarium wilt, implying that OTU224 was involved in GSAC-mediated suppressiveness. This is the first study on the potential of GSAC as a soil microflora-manipulating agent that can enhance soil suppressiveness to Fusarium wilt.

IMPORTANCE Methods for increasing soil suppressiveness via soil microflora manipulation have long been explored as an ideal strategy to protect plants from soilborne pathogens. However, viable methods offering consistent disease control effects have not yet been developed. Previously, the cultivation of Allium plants was demonstrated to induce bacteria-mediated soil suppressiveness to Fusarium wilt of different crop plants. This study discovered that the application of γ-glutamyl-S-allyl-l-cysteine, a unique dipeptide synthesized by Welsh onion, to soil enhances Fusarium wilt suppressiveness by increasing the relative abundance of indigenous antagonistic bacteria irrespective of the soil type. This finding will facilitate research supporting the development of environmentally friendly control measures for soilborne diseases.

INTRODUCTION

Soilborne pathogens pose a great threat to crop production. Therefore, crop losses due to soilborne diseases must be minimized to guarantee a sustainable food supply. In parallel, chemical pesticide use to control soilborne diseases must be limited because of their negative effects on the environment and human health (1), and therefore, alternative environmentally friendly and safer disease management strategies must be implemented. Recent studies revealed that soil microflora has a non-negligible influence on the occurrence and development of soilborne diseases (2). Soil microflora, in general, can limit the growth and activity of soil-dwelling pathogens to a certain level through a combination of competitive and antimicrobial actions, also termed general suppression (3). However, when general soil suppressiveness declines, attributable mostly to a decrease in the diversity or activity of soil microflora, soils become conducive to pathogens, resulting in a high prevalence of soilborne diseases (4). This suggests that the prevalence of soilborne diseases would be minimized if the antagonism between soil microbes and pathogens is properly manipulated (5).

Suppressive soils are ideal examples of soil microflora-mediated suppression of soilborne diseases (6). In suppressive soils, the disease incidence or severity of susceptible host plants remains low despite the presence of pathogens (7). To date, suppressive soils have been described for various soilborne pathogens globally, such as Fusarium oxysporum, Rhizoctonia solani, Thielaviopsis basicola, and Ralstonia solanacearum (8–11). Many studies indicated that the accumulation of a particular population of antagonistic microbes is a crucial factor in soil suppressiveness (7, 8, 10). Therefore, manipulating populations of specific indigenous microbial antagonists might represent a dependable and durable technique for combating soilborne diseases. However, the mechanisms promoting the accumulation of antagonistic microbes are largely unknown because suppressive soils may have developed because of the long-term and complex interplay of many factors, such as soil properties, climate, and cropping history (12). Thus, it is extremely difficult to establish a strategy for manipulating indigenous antagonistic populations to create suppressive soils with the current understanding.

Crop rotation or intercropping with particular plant species has long been known to reduce the incidence of soilborne diseases (13). In Asian countries, crop rotation and/or intercropping with Allium plants, such as Welsh onion (Allium fistulosum), onion (Allium cepa), and Chinese chive (Allium tuberosum), have been practiced to suppress Fusarium wilt of cucurbitaceous crops (e.g., cucumber, watermelon, bottle gourd) caused by the destructive soilborne fungus F. oxysporum (14, 15). Furthermore, recent studies illustrated that crop rotation and/or intercropping with Allium plants could effectively reduce Fusarium wilt of several other crops, including tomato, banana, and spinach (16–18), implying that these cultivation practices are highly versatile strategies for preventing F. oxysporum infection in a wide range of plants. In previous studies, the cultivation of Welsh onion and onion induced the accumulation of antagonistic soil bacteria, such as Flavobacterium and Pseudomonas spp., resulting in the conversion of conducive soil to suppressive soil against cucumber Fusarium wilt (19, 20). Plants can profoundly modify soil microflora through root exudation (10, 21). Therefore, it was postulated that certain unique compounds produced by Allium roots are responsible for the accumulation of antagonistic bacteria in soils. To develop a safe and eco-friendly technique for controlling Fusarium wilt by manipulating the soil microflora, this study attempted to identify the key compounds produced by an Allium plant (Welsh onion) that induce the accumulation of antagonistic soil bacteria.

RESULTS

Soil suppressiveness induced by amendment with Allium root extract.

Amendment of “GF” soil (soil collected from a research field at Gifu University) with the aqueous root extract of Welsh onion significantly reduced the severity of cucumber Fusarium wilt. Conversely, amendment with cucumber and tomato root extracts had no such effect (Fig. 1A). This result suggested that certain compounds present in the root extract of Allium plants can confer Fusarium wilt suppressiveness to the soil.

FIG 1.

Soil suppressiveness to cucumber Fusarium wilt induced by amendment with an aqueous root extract of Welsh onion. (A) Comparison of Fusarium wilt suppressiveness among soils amended with the root extracts of different plants. (B) Fusarium wilt suppressiveness of Welsh onion root extract-amended soils after different periods of incubation. Statistical analysis of disease severity was performed using the mean of three repeated experiments. Bars represent the means ± SD, and the different letters above the bars indicate statistically significant differences (P < 0.01, Tukey’s test). SDW, sterile distilled water.

Figure 1B presents the relationship between the incubation period after amendment with Welsh onion root extract and the severity of cucumber Fusarium wilt. Disease severity was marginally, but not significantly, reduced in soils incubated for <3 days after extract amendment compared to the findings in control soil. In contrast, the disease was significantly suppressed in soils incubated for >7 days after root extract amendment.

Metabolites in root extracts.

Liquid chromatography-tandem mass spectrometry (LC-MS/MS) identified seven compounds (adenosine, adenine, 2′-deoxyadenosine, γ-glutamyl-S-allyl-l-cysteine, 5′-S-methyl-5′- thioadenosine, azobenzene, and testosterone cypionate) abundant in Welsh onion root extract. Table 1 shows the retention times and peak areas for each compound from the Welsh onion root extract as calculated in selected ion monitoring (SIM) mode, as well as the areas of the peaks with the same m/z values detected at around the same retention times from the cucumber and tomato root extracts. The area of each peak was significantly larger in Welsh onion root extract than in the other extracts (both P < 0.01; Table 1).

TABLE 1.

Peak area of each compound in the aqueous root extracts of Welsh onion, cucumber, and tomato plants^a

Compound	Retention time (min)b	Peak area
		Welsh onion	Cucumber	Tomato
Adenosine	2.00	7.8 ± 0.0a	5.5 ± 0.0c	6.1 ± 0.0b
Adenine	2.01	7.2 ± 0.0a	4.5 ± 0.2c	5.3 ± 0.0b
2′-Deoxyadenosine	2.02	7.1 ± 0.0a	3.8 ± 0.1c	4.8 ± 0.1b
γ-Glutamyl-S-allyl-cysteine	4.53	6.9 ± 0.0a	3.8 ± 0.0c	4.0 ± 0.1b
5′-S-Methyl-5′-thioadenosine	4.95	6.0 ± 0.0a	3.4 ± 0.0c	5.0 ± 0.0b
Azobenzene	5.66	6.6 ± 0.0a	5.5 ± 0.0b	3.8 ± 0.1c
Testosterone cypionate	10.80	5.4 ± 0.1a	3.8 ± 0.0b	3.9 ± 0.0b

^aThe area of the peaks detected at the same m/z values from each root extracts was calculated in selected ion monitoring (SIM) mode. The values were converted to logarithmic value (log 10 peak area). The data are presented as the means ± SD (n = 3). For each compound, numbers followed by different letters indicate significant differences (P < 0.01, Tukey’s test).

^bRetention time for each compound detected in the Welsh onion root extract.

Among them, two compounds (i.e., azobenzene, testosterone cypionate) were classified as hazardous compounds in the Globally Harmonized System of Classification and Labeling of Chemicals. The two compounds were excluded because they were considered unsuitable as disease control agents. Of the remaining five compounds, four compounds (i.e., adenosine, adenine, 2′-deoxyadenosine, 5′-S-methyl-5′-thioadenosine) have been found in a variety of plants (22–24), and γ-glutamyl-S-allyl-l-cysteine (GSAC) has been detected only in Allium plants (25, 26). Since small peaks with the same m/z values as GSAC were detected from cucumber and tomato root extracts (with area values of about 1 in 1,300 and 1 in 800 of that detected from Welsh onion root extract, respectively), we analyzed the theoretical MS² spectra of those peaks and found that it did not match with that of GSAC. In this study, therefore, GSAC was selected as the specific Allium root-synthesized compound. From the quantitative analysis based on peak areas using external standards, the GSAC concentration in the root extract of Welsh onion was approximately 1,037 mg l⁻¹.

Soil suppressiveness induced by GSAC amendment.

The severity of cucumber Fusarium wilt was significantly decreased in GF and MK (soil collected from a research field at the Minokamo Farm of Gifu University) soils amended with GSAC (termed “GSAC soil”) compared to the findings in their respective negative-control soils amended with sterile distilled water (SDW; termed “SDW soil”; Fig. 2). In contrast, disease severity was not reduced in positive-control soils amended with three different dipeptides (dl-alanyl-dl-leucine, d-leucyl-l-tyrosine hydrate, and glycyl-l-valine). Furthermore, GSAC amendment significantly decreased Fusarium wilt severity in spinach plants (Fig. S1). These results suggested that GSAC is a unique dipeptide with the capacity to induce suppressiveness to Fusarium wilt diseases in soils. Conversely, the application of GSAC to sterilized soil did not alter the severity of cucumber Fusarium wilt (Fig. S2), suggesting that Fusarium wilt suppression in GSAC soil is mediated by the direct or indirect effects of soil microbes.

FIG 2.

Induction of cucumber Fusarium wilt suppressiveness in two different soils (GF soil collected from a research field at Gifu University; and MK, (soil collected from a research field at the Minokamo Farm of Gifu University) by dipeptide amendment. (A) Example of Fusarium wilt suppressive effect provided by GSAC amendment. Cucumber seedlings grown in the GSAC-amended GF soil (right) show lesser symptoms than that grown on the SDW-amended GF soil (left). Photos were taken at 21 days after planting. (B) Disease severity of cucumber seedlings grown in GF and MK soils amended with SDW and dipeptides. Disease assessment was done at 21 days after planting. SDW, sterile distilled water; GSAC, γ-glutamyl-S-allyl-l-cysteine; AL, dl-alanyl-dl-leucine; LT, d-leucyl-l-tyrosine-hydrate; GV, glycyl-l-valine. Statistical analysis of disease severity was performed using the means of three repeated experiments. Bars represent the means ± SD, and the different letters above the bars indicate statistically significant differences between GF soil (a, b) and MK soil (x, y) (P < 0.01, Tukey’s test).

Soil suppressiveness by transplanting the GSAC soil microflora.

When the microflora (filtrate) of GSAC soil was transplanted into sterilized soils, the severity of cucumber Fusarium wilt was significantly reduced compared to the findings for sterilized soils transplanted with the microflora of SDW soil (Fig. 3A and B). However, this phenomenon was not observed when the filtrates of GSAC soil were pretreated with antibacterial antibiotics. These results suggested that soil microbes, particularly bacteria, accumulated by GSAC amendment cause Fusarium wilt suppressiveness.

FIG 3.

Fusarium wilt suppression induced by transplanting microflora from GSAC soil with and without antibacterial treatment for GF soil (A) and MK soil (B). Statistical analysis of disease severity was performed using the mean of three repeated experiments. The different letters above the bars indicate statistically significant differences (P < 0.05, Tukey’s test). Bars represent the means ± SD. Control, SDW soil, and GSAC soil indicate that the double-autoclaved soil mixture was treated with sterile distilled water (SDW), SDW soil microflora, and γ-glutamyl-S-allyl-l-cysteine (GSAC) soil microflora, respectively.

Bacterial community of GSAC soil.

Sequencing resulted in 4,212,776 raw reads detected from 18 soil DNA samples (Table S1). After merging forward and reverse reads using dada2 and removing operational taxonomic units (OTUs) classified as chloroplasts, mitochondria, or Archaea and singletons, the number of merged reads for each sample ranged from 16,144 to 130,821. The number of OTUs at 95% similarity varied from 418 to 2,144. The rarefaction curve suggested that the number of reads was sufficient to assess the bacterial community diversity of each soil (Fig. S3).

Principal coordinate analysis (PCoA) based on OTUs according to the weighted UniFrac index indicated that the bacterial community structures of GF and MK soils were significantly separated (R = 1.0; Fig. S4). Therefore, to evaluate the effect of GSAC amendment on the soil bacterial community, PCoA was performed on each GF and MK soil sample. PCoA revealed a significant difference in bacterial community structures between GSAC and SDW soils for each soil type (R > 0.75; Fig. 4A and B). Furthermore, the community structures of Gram-negative bacteria in GSAC soil and SDW soil were significantly separated (R > 0.75; Fig. 4C and D) but not those of Gram-positive bacteria (R < 0.50; Fig. 4E and F). On the contrary, there were no significant differences in bacterial α-diversity (Shannon index) between GSAC soil and SDW soil (Table S1). These results suggested that GSAC amendment altered the structure of the Gram-negative bacterial community in soils without significantly changing the bacterial community diversity.

FIG 4.

Principal coordinate analysis (PCoA) of the bacterial communities of GSAC soil and SDW soil based on the operational taxonomic unit (OTU) composition according to the weighted UniFrac index. (A) Bacterial communities in the GF soil samples. (B) Bacterial communities in the MK soil samples. (C) Gram-negative bacterial communities in the GF soil samples. (D) Gram-negative bacterial communities in the MK soil samples. (E) Gram-positive bacterial communities in the GF soil samples. (F) Gram-positive bacterial communities in the MK soil samples. SDW soil samples are represented by blue circles, whereas GSAC soil samples are represented by red circles. Gram-positive bacteria communities included phyla Actinobacteriota and Firmicutes, and Gram-negative bacteria communities include all bacterial phyla other than Actinobacteriota and Firmicutes. Analysis of similarities (ANOSIM) was performed between the bacterial communities of SDW soil and GSAC soil samples. R > 0.75 indicates that groups are well separated, R between 0.5 and 0.75 suggests overlapping but distinguishable groups, and R < 0.25 describes the groups that can barely be separated (60). SDW, sterile distilled water; GSAC, γ-glutamyl-S-allyl-l-cysteine.

The relative abundance of 14 and 12 OTUs was significantly higher in GSAC-amended GF and MK soils, respectively, than in their respective control soils amended with SDW (linear discriminant analysis [LDA] score > 2.5; Fig. 5A). The common OTUs increased by GSAC amendment in GF and MK soils included three OTUs affiliated with genera Pseudomonas spp. (OTU224), Burkholderia-Caballeronia-Paraburkholderia (OTU387), and Bdellovibrio (OTU1259; Table S2). OTU224 was also accumulated significantly in Welsh onion rhizosphere soil (LDA score > 2.5; Fig. 5A; Table S2). A previous study isolated 27 and 6 Pseudomonas strains from Welsh onion rhizosphere soil and noncultivated soil, respectively (20). Therefore, this study analyzed the phylogenetic relationship between those strains and Pseudomonas OTUs (OTU224 and OTU945) that accumulated in GF and/or MK soil after GSAC amendment. In total, 27 strains isolated from Welsh onion rhizosphere soil were grouped into the clade with OTU224 (designated OTU224 clade), and six strains isolated from noncultivated soil were placed in the clade containing OTU945 (designated OTU945 clade) that accumulated only in GSAC-amended GF soil (Fig. 5B; Table S2). Furthermore, Pseudomonas strains isolated from Welsh onion rhizosphere soil showed significantly higher suppression of cucumber Fusarium wilt than Pseudomonas strains isolated from noncultivated soil (P < 0.01; Fig. S5).

FIG 5.

Bacterial OTUs accumulated in GSAC soil and Welsh onion rhizosphere soil. (A) OTUs that were significantly more abundant (circle A) in GSAC soil than in SDW soil for soil GF, (circle B) in GSAC soil than in SDW soil for soil MK, and (circle C) in the rhizosphere soil of Welsh onion than in noncultivated soil. The significance of differences in the abundance of each OTU between GSAC soil and SDW soil or between the rhizosphere soil of Welsh onion and noncultivated soil was checked by linear discriminant analysis effect size (LEfSe) with a minimum linear discriminant analysis (LDA) score of 2.5. The numbers of OTUs in each subset are presented within parentheses. The taxonomy of OTUs in circles A, B, and D to G is presented in Table S2. (B) Phylogenetic tree of the 16S rRNA gene V3 to V4 regions of OTU224 corresponding to circle G. OTU945 assigned to genus Pseudomonas corresponding to circle A, 27 strains recovered from Welsh onion rhizosphere soil (antagonistic Pseudomonas strains), and 6 strains from noncultivated soil (nonantagonistic Pseudomonas strains), and type strains of Pseudomonas spp. were used as references. Thiopseudomonas denitrificans was used as an outgroup. Maximum likelihood trees were bootstrapped using RAxML-NG (version 0.9.0), and tree visualization was performed using Figtree (http://tree.bio.ed.ac.uk/software/figtree/). ML bootstrap values were obtained using 100 bootstrap replicates and are shown on branches if above 60%. OTU, operational taxonomic unit; GSAC, γ-glutamyl-S-allyl-l-cysteine; SDW, sterile distilled water.

DISCUSSION

Reducing the negative environmental impact of chemical pesticides while preventing agricultural losses caused by soilborne diseases represents a significant global challenge. To attain this goal, there is an urgent need to develop eco-friendly strategies that can enhance the soil’s potential resistance or resilience to soilborne pathogens. This study explored Allium root-synthesized compounds that can convert conducive soil into soil that suppresses Fusarium wilt. Soil amendment with the aqueous root extract of Welsh onion suppressed cucumber Fusarium wilt, whereas the amendment with root extracts of cucumber and tomato did not. Allium plants are well known to produce various antimicrobial compounds, including thiosulfinates, flavonoids, and alcohols (27, 28). Zhang et al. reported that several of these compounds, such as 2-methyl-2-pentenal, dimethyl trisulfide, dimethyl disulfide, dipropyl disulfide, and dipropyl trisulfide, might play an important role in Fusarium wilt suppression conferred by intercropping or rotation with Allium plants (29). However, significant reduction in disease severity was observed in soils incubated for more than 7 days after root extract amendment but not within 3 days, indicating that disease suppression in soil amended with Welsh onion root extract was primarily attributable to the induction of soil suppressiveness rather than the direct suppression of the pathogen by the antimicrobial compounds present in the extract.

Notably, GSAC, which is abundant in the root extract of Welsh onion, has remarkable potential to induce suppressiveness to cucumber Fusarium wilt in soils regardless of the soil type. Furthermore, GSAC soil suppressed Fusarium wilt in not only cucumber but also spinach (Fig. S1). GSAC is an Allium-specific γ-glutamyl dipeptide of sulfur-containing amino acids (30). Although we did not test the effects of other γ-glutamyl dipeptides in this study, soil amendment with three commercial dipeptides (dl-alanyl-dl-leucine, d-leucyl-l-tyrosine-hydrate, and glycyl-l-valine) did not affect the severity of cucumber Fusarium wilt, suggesting that GSAC is a unique dipeptide with the potential to confer suppressiveness to Fusarium wilt diseases to soil. To date, there have been no reports on the toxicity or environmental hazards of GSAC. Rather, it has attracted attention among researchers as a useful compound with blood pressure-lowering and cholesterol-lowering effects (31, 32) and various products containing this compound have been commercialized as functional foods and health supplements in Japan. Therefore, although the adverse effects of GSAC on nontarget organisms and the environment need to be carefully studied in the future, this compound has the potential to be used as a safe, less environmentally harmful, and viable agent for managing Fusarium wilt of various plants grown in different soils.

Because a recent study reported the antimicrobial activity of GSAC (33), it was possible that direct inhibition of F. oxysporum by GSAC was involved in the Fusarium wilt suppressiveness of GSAC soil. However, sterilized soil amended with GSAC did not suppress cucumber Fusarium wilt, suggesting that the suppressiveness of GSAC soil was conferred by the activity of soil microbes rather than the antimicrobial activity of GSAC. Generally, two types of microbial-mediated soil suppressiveness, namely, general and specific, have been described (6). General suppression is attributed to the activity of the collective microbial flora, and it is not transferrable between soils. Conversely, specific suppression is caused by the concerted activities of specific groups of indigenous microorganisms, and it is transferrable. GSAC-mediated soil suppressiveness could be transferred via soil microflora transplantation to sterilized soil. This transferable suppressiveness was fully eliminated by pretreating the soil microflora with antibacterial antibiotics, indicating the importance of the soil bacterial flora in GSAC-induced soil suppressiveness. Amplicon sequencing of the 16S rRNA gene revealed that the bacterial community structures of GSAC soils differed significantly from those of SDW soils. Several genera of Gram-positive (e.g., Streptomyces, Paenibacillus) and Gram-negative bacteria (e.g., Pseudomonas, Chryseolinea, Terrimonas, Ohtaekwangia) were recently identified as the underlying microbes governing the suppressive effect of Fusarium wilt-suppressive soils (7, 8, 34, 35). The present findings illustrated that GSAC amendment significantly affected the community structure of Gram-negative bacteria but not Gram-positive bacteria, implying that Gram-negative bacteria are the main drivers of GSAC-induced suppressiveness. This finding aligned with previous data indicating that Fusarium wilt suppressiveness induced by Allium cultivation is primarily attributable to the accumulation of antagonistic Gram-negative bacteria (19). Among Gram-negative bacteria, Pseudomonas (OTU224), Burkholderia-Caballeronia-Paraburkholderia (OTU387), and Bdellovibrio (OTU1259) were more abundant in both GF and MK soils amended with GSAC than in their respective control SDW-amended soils. OTU224 was affiliated with Pseudomonas fluorescens complex, which includes Pseudomonas brassicacearum, Pseudomonas thivervalensis, Pseudomonas mandelii, Pseudomonas umsongensis, Pseudomonas silesiensis, and Pseudomonas frederiksbergensis (36, 37). Several of these microbes have been reported to produce the antifungal compound 2,4-diacetylphloroglucinol, which accumulates in disease-suppressive soils (10, 38, 39). OTU224 accumulated significantly in both GSAC soil and Welsh onion rhizosphere soil. Pseudomonas strains previously isolated from Welsh onion rhizospheres exerted a remarkable suppressive effect on cucumber Fusarium wilt. Surprisingly, all of these strains were grouped into the OTU224 clade in phylogenetic analysis. Therefore, it was conceivable that Pseudomonas spp. corresponding to OTU224 that accumulated in GSAC soil are also antagonists with a strong disease-controlling effect, and they might play a key role in GSAC-induced soil suppressiveness. Various species of Burkholderia-Caballeronia-Paraburkholderia (formerly genus Burkholderia) have been recognized as biocontrol agents for plant pathogens, including F. oxysporum, because of their antifungal properties (40–42); thus, bacterial species corresponding to OTU387 may also be directly involved in GSAC-induced soil suppressiveness. Bdellovibrio (OTU1259) is an obligate predator of Gram-negative bacteria that is expected to be used as a biocontrol agent for Gram-negative pathogens (43). However, because Bdellovibrio does not prey on fungi, this genus is unlikely to directly suppress the growth or activity of Fusarium wilt pathogen. On the other hand, bacterivorous protists and predatory bacteria have been shown to affect bacterial community structure and dynamics (44, 45). Ye et al. demonstrated that the predatory myxobacterium Corallococcus sp. strain EGB modified the soil microflora and decreased the abundance of F. oxysporum f. sp. cucumerinum in the soil, resulting in a reduction of cucumber Fusarium wilt (46). Therefore, it is possible that Bdellovibrio affected the microflora and thus indirectly helped to increase the populations of the aforementioned antagonistic bacteria in the GSAC soil. Taken together, it was considered that the application of GSAC conferred Fusarium wilt suppressiveness to soils by promoting the accumulation of specific Gram-negative antagonistic bacteria.

The mechanism by which GSAC induces the accumulation of antagonistic bacteria is currently unknown. It has been reported that soil amendment with various nutrient compounds can increase the population or activity of antagonistic microbes, resulting in disease suppression (47–50). Chitin and its derivative chitosan are representative examples of nutrient compounds for antagonistic microbes. It has been reported that the application of these molecules to soil induces the accumulation of chitin-degrading bacteria such as Streptomyces and Chitinophaga spp. and suppresses soilborne diseases, including Fusarium wilt (47–49). In addition, the application of certain amino acids (e.g., lysine, serine, proline, arginine, glutamic acid) or sugars (e.g., glucose) has been reported to increase soil microbial activity and consequently induce soil suppressiveness to bacterial wilt of tomato (50). In addition, recent studies illustrated that plants also release certain nutrient compounds, such as long-chain organic acids (e.g., pentadecanoic acid, palmitic acid) and amino acids (e.g., isoleucine, proline, glutamic acid) and accumulate specific antagonistic bacterial groups with plant-protective properties in the rhizosphere (51, 52). Furthermore, although unrelated to plant-microbe interactions, several peptides have been demonstrated to have the ability to stimulate the growth of certain bacterial groups. For example, soybean conglycinin peptides have been found to selectively stimulate the growth of bifidobacteria, even though the bacteria do not utilize them as nutrients (53). Similarly, a tripeptide isolated from the hydrolysates of broad bean protein was reported to exert growth-stimulating effects on Lactobacillus delbrueckii (54). In addition, a pentapeptide found in a bovine hemoglobin hydrolysate stimulated the growth and activity of several enteric bacteria and Pseudomonas aeruginosa (55). Further studies are needed to determine whether GSAC functions as a specific nutrient or signal with growth-stimulating activity for Fusarium-antagonistic bacteria.

The accumulation of common antagonistic bacteria (OTU224) in GSAC soils and Welsh onion rhizosphere soils suggests that GSAC produced by Allium roots is responsible for the accumulation of OTU224 in their rhizospheres. To clarify this hypothesis, further studies are needed to confirm whether Welsh onion secretes GSAC in the rhizosphere soil. Furthermore, as previously reported, Welsh onion cultivation enriches the populations of antagonistic Pseudomonas and Flavobacterium spp. in the soil (19, 20). However, GSAC did not induce the accumulation of Flavobacterium spp. Therefore, compounds other than GSAC produced by Allium plants may be involved in the enrichment of Flavobacterium spp. Identifying compounds capable of triggering Flavobacterium accumulation will be an interesting subject for future studies to further clarify the mechanisms of microbe-mediated soil suppressiveness induced by Allium cultivation.

In summary, we identified GSAC as a new type of disease control agent that suppresses Fusarium wilt by stimulating the populations of indigenous antagonistic bacteria. To the best of our knowledge, this is the first study to report a specific dipeptide that can induce microbe-based soil suppressiveness. We believe that this research will provide new strategies for the development of environmentally friendly control measures for soilborne diseases.

MATERIALS AND METHODS

Preparation of pathogen inoculum.

F. oxysporum f. sp. cucumerinum strain GUS77 was used as the challenge pathogen in this study. GUS77 was cultured in potato sucrose broth (200 g potato and 20 g sucrose per 1 liter of distilled water) at 25°C with shaking at 120 rpm. After 4 days of incubation, the culture broth was filtered through three layers of sterile gauze to remove hyphae, and the filtrate was centrifuged at 3,000 rpm for 10 min to harvest spores. The precipitated spores were suspended in SDW, and the spore suspension was adjusted to 1.5 × 10⁴ or 3.0 × 10⁴ spores mL⁻¹ using a hemocytometer.

Soils.

Two different soils were used in this study. “GF” soil was collected from a grassy area of a research field at Gifu University (Yanagido, Gifu City, Gifu Prefecture, Japan). “MK” soil was collected from a research field at the Minokamo Farm of Gifu University (Makino, Minokamo City, Gifu Prefecture, Japan) that had been planted with Italian ryegrass (Lolium multiflorum Lam.) for >5 years. Soil GF was classified as sandy clay loam (pH 6.7; 5 mg kg⁻¹ ammonium nitrogen, 24 mg kg⁻¹ nitrate-nitrogen, 840 mg kg⁻¹ available phosphoric acid, and exchangeable potassium 390 mg kg⁻¹), whereas MK soil was sandy loam (pH 5.2; 8 mg kg⁻¹ ammonium nitrogen, 19 mg kg⁻¹ nitrate-nitrogen, available 720 mg kg⁻¹ phosphoric acid, and 360 mg kg⁻¹ exchangeable potassium). Soils were sieved with a 2-mm mesh to remove larger particles, such as stones and plant debris.

Preparation of aqueous root extracts.

Seeds of Welsh onion (A. fistulosum L. cv. Kujohoso), cucumber (Cucumis sativus L. cv. Tokiwa jibai), and tomato (Solanum lycopersicum L. cv. Momotaro) were sown in plug trays containing a commercial potting soil (Saika-ichiban; Ibigawa Kogyo, Gifu, Japan) and grown in a controlled environmental chamber (25°C, 12 h of daylight). Welsh onion (90 days old), tomato (40 days old), and cucumber seedlings (20 days old) were transplanted into a plastic pot (bottom, 7 × 7 cm²; top, 10 × 10 cm²; height, 13 cm) containing commercial potting soil (one plant per pot) and grown in a glasshouse under natural light with regular watering and fertilization. Seventy days after transplanting, the plants were uprooted from the pots, and the roots were gently washed free of soil, detached from the plants, and dried on paper towels. At the sampling time, soil moisture content measured with a soil moisture meter (DM-18; Takemura Electric Co., Ltd., Tokyo, Japan) was about 30% in all soils. The fresh root weight of each plant species and the wet weight of the potting soil remaining in each pot were recorded to determine the root weight density (RWD; grams of fresh root weight per kilogram of wet soil). In this experiment, 15 individuals of each plant species were sampled from the pots. Based on the calculation, the mean RWD of Welsh onion, cucumber, and tomato were 10.2, 9.7, and 12.8 g kg⁻¹, respectively. The roots (5 g fresh weight) of each plant were homogenized in liquid nitrogen using a mortar and pestle, and the homogenate was immediately transferred to a 50-mL tube. To extract aqueous compounds, 25 mL of SDW were added to the tube containing the root homogenate and mixed thoroughly with a vortex mixer for 1 min. Thereafter, the homogenate suspension was filtered using filter paper (No. 1; Advantec, Tokyo, Japan) twice and a microfilter (Vakuumfiltration 500 “rapid” Filtermax; pore size, 0.22 μm; Techno Plastic Products AG, Trasadingen, Switzerland) to remove plant debris and microorganisms. The aqueous root extracts were stored at −20°C until use.

Effect of soil amendment with root extracts on cucumber Fusarium wilt.

In the first experiment, Fusarium wilt suppressiveness of soils amended with the root extract of Welsh onion was compared to that of soils amended with the root extracts of cucumber and tomato. Each aqueous root extract was diluted with SDW, and a 5-mL aliquot of the diluted extract was mixed thoroughly with 30 g (wet weight) of GF soil in a plastic cup (75 mm in diameter and 38 mm in height). Root extracts were applied to each soil such that the quantity applied would correspond to an extraction from the quantity of roots that would be expected based on the aforementioned RWD. The control treatment was prepared using 5 mL of SDW instead of the root extract. These soils were incubated in a controlled environment chamber (25°C, 12 h of daylight) for 28 days with watering twice daily with 2.5 mL of SDW. After incubation, the cucumber Fusarium wilt suppressiveness of each soil was assessed as described previously (19) with slight modifications. Briefly, each soil was mixed with double-autoclaved potting soil (Ikubyou-baido; Takii Seed, Kyoto, Japan) and double-autoclaved river sand at a ratio of 2:1:1 (wt/wt/wt). Six grams of each type of soil mixture were placed in flat-bottomed glass tubes (3 cm in diameter and 12 cm in height; Iwaki Glass, Chiba, Japan). These soils were inoculated with 4 mL of GUS77 spore suspension (1.5 × 10⁴ spores mL⁻¹), and the soil surface was covered with 2 g of double-autoclaved vermiculite. Surface-sterilized pregerminated cucumber seeds were planted in this vermiculite layer (one seed per tube), covered with a small amount of sterile vermiculite, and fertilized with 2 mL of 500-fold diluted Hyponex solution (type 6-9-5; Hyponex Japan, Osaka, Japan). The tubes were placed in a controlled environment chamber (25°C, 12 h of daylight) for 21 days. After cultivation, the symptoms of cucumber plants were visually scored on a scale of 0 to 3 as described previously (19), and the results were expressed as the disease index (%), which was calculated using the following formula: (Σ[disease severity of cucumber seedlings]/[the number of replicated tubes × 3]) × 100. There were five replicated tubes (one seedling per tube) per treatment, and the experiment was repeated three times.

The second experiment investigated the influence of the incubation period after soil amendment with the Welsh onion root extract on the degree of soil suppressiveness to cucumber Fusarium wilt. Soil GF was amended with Welsh onion root extract as previously mentioned. The soil was incubated in a controlled environment chamber (25°C, 12 h of daylight) with watering twice daily with 2.5 mL of SDW to maintain soil moisture for 0, 3, 7, 14, 21, or 28 days. SDW-amended nonincubated soil served as the control. The suppressiveness of each soil to cucumber Fusarium wilt was examined as previously described. There were five replicated tubes (one seedling per tube) per treatment, and the experiment was repeated three times.

LC-MS/MS analysis of root extracts.

To explore compounds that are specifically present in large amounts in the root extract of Welsh onion used in the above cucumber seedling assay, water-soluble organic compounds in the aqueous root extracts of Welsh onion, cucumber, and tomato plants were identified by LC-MS/MS. The analysis was performed on a Q Exactive Focus MS coupled with UltiMate 3000 (Thermo Fisher Scientific, Waltham, MA, USA). Mass calibration was performed before the analysis according to the manufacturer’s recommendations every 7 days using external mass calibration. An Acclaim C18 reverse-phase column (150 × 2.1 mm; 2.2 μm; Thermo Fisher Scientific) was applied for LC separation. Each 10-fold diluted root extract was injected, and the column temperature was set at 30°C with a flow rate of 0.25 mL min⁻¹. The mobile phases consisted of (A) 0.1% acetic acid in water and (B) acetonitrile. The gradient program was as follows: 0 min with 10% B; 4 min with 4, 50% B, 16 min with 90% B, 24 min with 90% B, 24 min with 10% B, and 30 min with 10% B. The HESI and II source conditions were as follows: heater temperature, 298°C; sheath gas, 40 arbitrary units (AU); auxiliary gas, 10 AU; spray voltage, +2.01 kV; capillary temperature, 349°C; and S-lens radio frequency level, 60.0. MS was performed using the full-scan mode and a subsequent data-dependent acquisition (DDA) mode. The settings for the full-scan mode were as follows: polarity, switching; resolution, 70,000; scan range, m/z 120 to 1,000; automatic gain control (AGC) target, 1e6; maximum injection time, auto; microscans, 1; and spectrum data type, profile. The settings for the DDA mode were as follows: dd-MS², discovery; resolution, 17,500; isolation window, m/z 3.0; cell stepped normalized collision energy, 30.0%; AGC target, 5e4; maximum injection time, auto; loop count, 1; minimum AGC target, 8.0e3; dynamic exclusion, auto; exclude isotopes, off; and spectrum data type, profile. Retention time alignment, peak picking, and database search were carried out using Compound Discoverer 2.0 (Thermo Fisher Scientific) with default settings. The peak areas of compounds detected in the root extract of Welsh onion were calculated. Thereafter, we calculated the areas of the peaks with same m/z values detected at around same retention times from the root extracts of cucumber and tomato in SIM mode. To select compounds more abundant in the Welsh onion root extract, peak area for each compound was compared between extracts. In addition, the concentration of γ-glutamyl-S-allyl-l-cysteine (GSAC) in Welsh onion root extract was quantified using an external standard (commercial GSAC; US Pharmacopeia, Rockville, MD, USA).

Suppressive effect of soils amended with GSAC on cucumber Fusarium wilt.

Commercial GSAC was diluted with SDW to yield solutions with a concentration of 1.0 mg mL⁻¹ (equivalent to that of the diluted Welsh onion root extract applied to the soil in the cucumber seedling assay). A 5-mL aliquot of the diluted solution was added to 30 g (wet weight) of GF and MK soils (designated GSAC soil) in a plastic cup (75 mm in diameter and 38 mm in height). For comparison, soils amended with same concentration of one of three commercially available dipeptides, namely, dl-alanyl-dl-leucine, d-leucyl-l-tyrosine-hydrate, and glycyl-l-valine, were prepared as positive controls, whereas soils amended with SDW were used as negative controls (designated SDW soil). These soils were incubated in a controlled environment chamber (25°C, 12 h of daylight) for 7 days with watering twice daily with 2.5 mL of SDW. The suppressive effect of each soil against cucumber Fusarium wilt was evaluated at 21 days after planting as previously described. There were five replicated tubes (one seedling per tube) per treatment, and the experiment was repeated three times.

Fusarium wilt suppression by transplanting microflora from GSAC-amended soil.

Thirty grams (wet weight) of soils (GF and MK) in a plastic cup (75 mm in diameter and 38 mm in height) were amended with a 5-mL aliquot of 1.0 mg mL⁻¹ GSAC and incubated in a controlled environment chamber (25°C, 12 h of daylight) with watering twice daily with 2.5 mL of SDW for 7 days. Soils amended with 5 mL of SDW were also incubated under the same conditions. After incubation, 5 g of each soil were collected from the cup, suspended in 45 mL of SDW in a 200-mL Erlenmeyer flask, and filtered through a nylon net filter (Sterile 50-mL disposable vacuum filtration system; pore size, 41 μm; Merck Millipore, Carrigtwohill, Ireland) to remove soil particles. The obtained soil filtrates were defined as soil microflora. Each soil microflora (1.9 mL) and SDW (0.1 mL) were applied to a flat-bottomed glass tube (3 cm in diameter and 12 cm in height) containing 6 g (dry weight) of a double-autoclaved soil mixture (field soil/commercial potting soil/river sand = 2:1:1 [wt/wt/wt]). A double-autoclaved soil mixture treated with SDW instead of soil microflora served as a control. To clarify the importance of bacteria in Fusarium wilt suppressiveness, the effect of antibacterial treatment on the suppressive efficiency of soil microflora transplantation was examined. In this experiment, each soil microflora (19 mL) was mixed with 1 mL of a mixture of antibacterial antibiotics (500 ppm ampicillin, 500 ppm streptomycin sulfate, and 500 ppm tobramycin) in a 50-mL tube and slowly rotated for 24 h at 25°C to kill bacteria. Subsequently, 2 mL of this filtrate was applied to the soil mixture in a glass tube as previously explained. These soil mixtures were inoculated with 2 mL of GUS77 spore suspension (3.0 × 10⁴ spores mL⁻¹), planted with cucumber seeds, fertilized, and incubated in a controlled environmental chamber (25°C, 12 h of daylight) as previously described. After 14 days of incubation, the disease severity of the seedlings was assessed. There were five replicated tubes per treatment, and the experiment was repeated three times.

Bacterial community analysis by 16S rRNA gene amplicon sequencing.

We performed 16S rRNA gene amplicon sequencing to determine bacterial communities in GSAC- and SDW-amended GF and MK soils used for the above cucumber seedling assays. Soil DNA extraction was conducted using a FastDNA SPIN kit for soil (MP Biomedicals, Santa Ana, CA, USA) as described previously (19) with the modification that 0.35 g of each soil was used. DNA extraction was repeated three times. Partial 16S rRNA gene sequences were amplified using a primer set specific for the V3–V4 regions (56), and the 16S rRNA gene amplicon libraries were paired-end sequenced on an Illumina MiSeq platform (Illumina, CA, USA) using two 300-bp overlapping paired-end reads according to the manufacturer’s protocol.

Sequence processing was conducted using the Qiime2 pipeline (version 2019.4). The paired-end fastq files were demultiplexed with demux-summarize and processed by quality filtering, merging of the paired ends, and chimera removal with dada2 (57). In dada2 processing, this study used the following options: (i) primer sequences were removed, (ii) forward and reverse reads were truncated to 286 and 210 bp, respectively, and (iii) the reads containing bases with a quality score of ≤15 were truncated.

Taxonomy was assigned to each read using the SILVA database (silva_132_99_16S.fna; https://www.arb-silva.de/documentation/release-132/) and feature-classifier. Subsequently, reads classified as chloroplasts, mitochondria, or Archaea and singletons were removed. To compare bacterial communities at the genus level, each read was clustered into OTUs at 95% similarity using VSEARCH (58, 59). β-Diversity according to the weighted UniFrac index was calculated using the OTU data and statistically compared using analysis of similarities (ANOSIM) with 9,999 permutations in the Qiime2 pipeline. ANOSIM was performed between the bacterial communities of SDW soil and GSAC soil (60). PCoA plots were generated using ggplot2 with data extracted from Qiime2 artifacts using qiime2R (version 0.99.31; https://github.com/jbisanz/qiime2R/). An LDA effect size (LEfSe) with a minimum LDA of 2.5, was used to identify OTUs that were significantly increased in their abundance in GSAC soil compared to SDW soil.

Furthermore, changes in the relative abundances of specific taxa in the bacterial communities of GSAC-amended and Welsh onion rhizosphere soils were assessed. For this analysis, in addition to the data obtained in this study, abundance data from bacterial 16S rRNA gene sequencing data of Welsh onion rhizosphere soil and noncultivated soil obtained in a previous study (deposited in the Sequence Read Archive database under BioProject PRJDB6419) were used. These sequence sets were analyzed using Qiime2 and LEfSe as previously described to identify OTUs with significantly higher abundance in Welsh onion rhizosphere soils than in noncultivated soils. To select potential antagonistic bacterial taxa against F. oxysporum, OTUs with significantly higher abundance ratios in both GSAC soils and Welsh onion-rhizosphere were extracted. The phylogenetic tree was prepared using RAxML-NG (version 0.9.0) (61) and Figtree (http://tree.bio.ed.ac.uk/software/figtree/).

Statistical analysis.

Statistical analyses using Tukey’s test were performed using BellCurve for Excel (version 2.13; Social Survey Research Information, Tokyo, Japan). Statistical analysis of disease severity was performed using the mean of three repeated experiments. ANOSIM was performed using the Qiime2 pipeline. LEfSe analysis was conducted using Galaxy of the Huttenhower laboratory (https://huttenhower.sph.harvard.edu/galaxy/).

Data availability.

Sequence data were deposited in the Sequence Read Archive database under accession numbers DRR346241 to DRR346252.