Effect of T. asperellum M45a on watermelon health properties and FON content
The effect of T. asperellum M45a on watermelon health parameters was assessed using the FW disease incidence (DI) and the length of watermelon vine. FW disease occurs in the seedling stage and then erupts rapidly until it becomes stable in the flowering stage. Compared to CK, the biocontrol effects of T. asperellum M45a on watermelon FW disease were 89.65%, 72.62% and 66.72% at the seedling period (S2), the smoke trailing period (S3) and the blooming period (S4), respectively, which was significantly different from the CK group in the same period (P < 0.01) (Fig. 1a). Additionally, the vine lengths of watermelon inoculated with strain M45a also increased significantly by 29.44%, 26.43% and 49.15% in the S2, S3 and S4 periods, respectively (Fig. 1b).
To understand the dynamic colonization relationship between Trichoderma spp. and Fusarium oxysporum f. sp. niveum (FON) in the rhizosphere soil, qPCR technique was applied. In the study, we found that the number of FON in CK rapidly increased to 9543.66 cfu/g in the germination period (S1), which was significantly higher than T. asperellum M45a treatment. Except for the smoke trailing period (S2), FON in the control rhizosphere soil was significantly higher than that in M45a treatment (Fig. 1a). In addition, the contents of Trichoderma spp. in the rhizosphere soil of the M45a treatment were 13.89, 10.96, 8.47 and 15.4 times higher than CK treatment in the same stage, respectively, and remained at 2.32–4.25 ng/g for the onset of FW (Fig. 1b).
Effect of T. asperellum M45a on soil enzyme activities and properties in rhizosphere soil
In the study, the enzyme activities of CL, ACP, CAT and SC in rhizosphere soil were increased following treatment with M45a, except to the activities of SC in the seedling period (S2) (Fig. 2). For example, the maximum increases in the activities of SC (224.15%) and ACP (95.80%) were obtained in the smoke trailing period (S3). Additionally, the activities of UE in the control group decreased gradually with the occurrence of watermelon FW. In contrast, M45a treatment could effectively enhance the UE enzyme activity, and the highest activity (501.478 U/g) was observed in the S4 period. In addition, no significant difference was observed in OM and the available K, while the TN and available nutrients (NO3-N, and P) were under significantly different treatments (Table 1). Compared with the control (CK), treatments with inoculated M45a significantly increased the contents of TN, NO3-N and AP, and decreased the AK content.
Table 1
Effect of inoculating M45a on soil enzyme activities in the rhizosphere of watermelon. The application of T. asperellum M45a in continuous cropping soil (Trichoderma). The non-inoculated control (CK). Trichoderma1, Trichoderma2, Trichoderma3, Trichoderma4: the treatment with T. asperellum M45a at S1, S2, S3 and S4 period, respectively; CK1, CK2, CK3, CK4: the CK treatments at S1, S2, S3 and S4 period, respectively. S1: the germination period; S2: the seedling period; S3: the smoke trailing period; S4: the blooming period.
Treatment | TC (%) | TN (mg/kg-1) | NH4+ (mg/kg-1) | NO3- (mg/kg-1) | AP (mg/kg-1) | AK (mg/kg-1) |
CK 1 | 2.72 ± 0.09a | 2000.60 ± 6.86a | 25.24 ± 0.60a | 364.14 ± 15.50a | 67.43 ± 1.99a | 490.54 ± 9.46a |
Trichoderma 1 | 2.65 ± 0.05a | 2015.34 ± 3.40b | 24.88 ± 1.12a | 367.69 ± 4.67a | 76.85 ± 0.77b | 358.29 ± 6.64b |
CK 2 | 2.21 ± 0.09a | 1816.65 ± 4.10a | 38.66 ± 0.51a | 338.55 ± 11.20a | 65.44 ± 1.87a | 511.52 ± 4.64a |
Trichoderma 2 | 2.33 ± 0.12a | 2058.45 ± 5.76b | 48.43 ± 2.01b | 365.94 ± 6.55b | 68.46 ± 1.35a | 483.30 ± 8.31b |
CK 3 | 2.51 ± 0.09a | 1807.73 ± 7.46a | 14.60 ± 1.13a | 268.75 ± 8.21a | 69.88 ± 1.53a | 420.91 ± 5.63a |
Trichoderma 3 | 2.27 ± 0.08ab | 1920.03 ± 9.10b | 30.38 ± 1.17b | 283.62 ± 5.54ab | 63.26 ± 2.18b | 417.80 ± 5.93a |
CK 4 | 2.74 ± 0.09a | 1797.65 ± 13.09a | 10.08 ± 0.80a | 258.40 ± 10.04a | 56.85 ± 0.88a | 323.62 ± 17.40a |
Trichoderma 4 | 2.58 ± 0.10a | 1837.18 ± 6.58ab | 9.38 ± 0.78a | 287.23 ± 8.80ab | 71.84 ± 2.21b | 266.37 ± 6.75b |
Trichoderma 1 | 2.65 ± 0.05a | 2015.34 ± 3.40b | 24.88 ± 1.12a | 367.69 ± 4.67a | 76.85 ± 0.77b | 358.29 ± 6.64b |
Effect of T. asperellum M45a on microbial community structure in rhizosphere soil
In total, 3,607,435 and 3,417,512 high-quality 16S and ITS sequences were obtained from the rhizosphere soil samples in the four stages, respectively. In the present study, the significant difference of alpha diversity between the two groups is shown in Fig. 3 and Table S1. There was significant difference in Chao1 between the M45a-treated group and the CK group (ANOVA, p < 0.05), and the fungal Shannon index in the M45a group was significantly lower than those of the CK group at all stages (ANOVA, p < 0.01). With the occurrence of watermelon wilt, the fungal Chao1 values in the S1 and S2 groups were significantly higher than those of the S3 and S4 groups (Table S1, p < 0.05). Additionally, the principal coordinate analysis (PCoA) ordinations showed that M45a(T) had significantly effect on bacterial and fungal community composition (Fig. 3). The first coordinate (PCoA1) showed 41.1% and 50.4% difference in community variation, and PCoA2 explained 17.6% and 25.8% dissimilarity, Respectively. In addition, it was further verified by the PERANOVA dissimilarity tests based on Bray-Curtis distance among the two groups (Bacterial: R = 0.05787, p = 0.001; Fungal: R = 0.27598, p = 0.001).
Effect of T. asperellum M45a on the microbial community composition
In general, the dominant bacterial phyla between two groups are visualized in the Fig. 4a and Fig. S1. In brief, the most abundant bacterial phyla were Proteobacteria, Actinobacteria, Chloroflexi, Gemmatimonadetes, Saccharibacteria and Acidobacteria, which contributed almost 86.7–91.8% of the bacterial sequences. While the rhizosphere communities were compared at different growth stages, Saccharibacteria (2.0-14.1%) was present at a significantly increased proportion in the rhizosphere with the onset of FW. At the genus level, among the top 20 genera, the relative abundance levels of Pseudomonas, Sphingomonas, Actinomadura and Rhodanobacter in the M45a group were significantly higher than in the CK group (Fig. S1, p < 0.05).
To investigate the difference in the fungal classes between the two groups, as illustrated in Fig. 4b, we found that the most relative abundant class was Sordariomycetes (60.12–78.67%) in the M45a group, which was significantly higher than in the CK group. However, the second relative abundance of Eurotiomycetes was consistently lower in M45a treatments (0.89–24.71%) than in the CK group (32.26–55.62%). For the fungal genera, the relative abundance levels of Penicillium, Chaetomium, Aspergillus and Acremonium in the M45a-treated rhizosphere soil were significantly lower than in the CK group (Fig. S2, p < 0.05). However, Trichoderma was present at a significantly increased proportion during the growth stages in M45a treatment, consistent with the real-time PCR results.
Effect of T. asperellum M45a on potential functional composition diversities
For the bacterial community, amino acid metabolism (10.81%-10.98%), carbohydrate metabolism (10.45%-10.81%), and energy metabolism (5.53%-5.81%) were the main bacterial metabolic pathways in all treatments (Fig. S3). In this study, the enzyme families (2.02%-2.11%) were different between the M45a treatment and the control (CK). Compared with the control soil (CK), functional profiles with lower abundance were related to enzyme families in M45a-treated soil at different stages, but there were no significant differences in response to disease (Fig. 5a).
For the fungal community, the different functional profiles of trophic mode (symbiotroph, saprotroph, pathotroph) and guild (plant pathogen) of fungal communities were compared between M45a and CK treated soil at different stages. Compared with the CK treated soil, the relative abundance levels of Ectomycorrhizal, Endophyte, Animal pathotroph and saprotroph (Dung Saprotroph, Plant Saprotroph, Soil Saprotroph and Wood Saprotroph) in the fungal community were significantly lower in M45a-treated soil (Fig. 5b). Additionally, the relative abundance levels of pathogens and other trophic modes showed no significant differences in these treatments.
Relationships among soil enzyme activities, soil properties and microbial communities
The RDA analysis showed that enzyme activities (CL, ACP and CAT) and the properties (AP) could greatly affect the microbial community composition in rhizosphere soil (Fig. 6). In addition, a significant positive correlation was observed between the FW disease incidence (DI) and soil cellulase activities (CL). However, TN, NH4+-N and NO3-N were negatively correlated with the DI. Likewise, there were significant positive correlation between Sphingomonas, Rhodanobacter, Pseudomonas, Gemmatimonsa, Streptomyces and S-CL activities, and a significant negative correlation between Nocardioides and S-CL activities (p < 0.05) (Table S2). The correlations between the major fungal genera and the rhizosphere soil enzyme activities were then observed (Table S3). Fungi are likely to be more sensitive to ACP activities than bacteria; thus, the increased Trichoderma spp. in the M45a treatment soil are sufficient to impose a stress on fungi and thereby likely influence fungi species, such as Penicillium, Chaetomium, Aspergillus and Dendroclathra, which were significantly negatively related to the ACP and SC activities in the rhizosphere soil, while the opposite trend was observed for the UE activities (p < 0.05).