Streptomyces strains modulate dynamics of soil bacterial communities and their efficacy in disease suppression caused by Phytophthora capsici

The responses of rhizosphere bacterial communities of Streptomyces (SS14 and IT20 stains) treated-pepper plants following inoculation by Phytophthora capsici (PC) was investigated using Illumina MiSeq sequencing. Distinct modulation of the bacteriome composition was found for PC samples with the highest relative abundance (RA) of Chitinophaga (22 ± 0.03%). The RA of several bacterial operational taxonomic units (OTUs) was affected and caused changes in alpha and beta-diversity measures. In IT20, the RA of Cyanobacteria was enriched compared to SS14 (72%) and control samples (47%). Phylotypes belonging to Devosia, Promicromonospora, Kribbella, Microbacterium, Amylocolatopsis, and Pseudomonas genera in the rhizosphere were positively responding against the pathogen. Our findings show that the phosphate solubilizing strain IT20 has higher microbial community responders than the melanin-producing strain SS14. Also, positive interactions were identified by comparing bacterial community profiles between treatments that might allow designing synthetic bio-inoculants to solve agronomic problems in an eco-friendly way.


Molecular and phenotypical characterizations of the superior isolates.
On the medium ISP2, SS14, IT8, and IT20 were differentiated from each other according to the color of spore chains. On ISP3, SS14 and IT20 were distinct according to the color of aerial hyphae. SS14 and IT8 were different from IT20 based on melanin production. Physiological tests showed that IT20 and IT8 had the potential to grow at 42 °C (Table 2). These strains were able to grow on NaCl 6%. Analysis of the 16S rRNA gene sequences showed that IT20 and SS14 were closely related to the genus Streptomyces with more than 98% sequence similarity to S. rochei and S. vinaceusdrappus, respectively (Table 2).
Culturable rhizosphere microbiome. Inoculations of the pathogen and/or Streptomyces strains induced changes in the fungal and bacterial colonies. The number of fungal colonies significantly increased in Streptomyces treatments compared to control, especially after pathogen inoculation (P < 0.05) (Fig. 2). Pathogen inoculation significantly increased the colonies of Penicillium in soil (P < 0.05). In inoculated plants, IT20 caused more variation in culturable soil bacterial colonies than SS14 (Fig. 2S). Penicillium type colonies were significantly reduced in SS14 treated-plants (Fig. 3S). Table 1. In vitro growth inhibition of Phytophthora capsici and hydrolytic enzymes production by Streptomyces isolates. Values are the means (averaged from three replicates) ± SE. *Same letters represent nonsignificant difference according to Duncan's Multiple Range Test (P < 0.05). +: Producing −: non-producing.   Diversity and structure of the rhizosphere bacterial communities. The levels of microbial diversity were different among the treatments. Alpha rarefaction analysis indicated the phylogenetic diversity tree (PD whole tree), observed species, Shannon, and Simpson reciprocal indices were affected by the bacterial treatments. In inoculated samples, these indices were higher in IT20 than SS14 samples. The dominance index was lower in IT20 inoculated samples (0.007 ± 0.0). Specifically, Chao1 predicted a high value in both inoculated treatments of IT20 (1200.19 ± 20.28) and SS14 (1222.52 ± 40.08) compared to PC (1142.53 ± 5.74). IT20 exhibited a higher alpha-diversity than control samples (Table 3). PERMANOVA (function adonis) found a significant difference among PC and treated plants (PERMANOVA, p < 0.01; Fig. 3). Looking at beta-diversity, we found a separated clustering between IT20 and SS14 inoculated with the pathogen in PCoA plots (Fig. 4). By contrast, samples from control and IT20 clustered together (Fig. 4).

Discussion
The activity of hydrolytic enzymes is one of the effective mechanisms of Streptomyces to inhibit fungal growth 2,39 .
Our results are following a recent study 40 that showed the protease activity of Paenibacillus polymyxa is also involved in the antagonism against Phytophthora.
In this study, two strains S. vinaceusdrappus SS14 and S. rochei IT20 were different in terms of P solubilizing ability and melanin production. There is growing evidence that rhizosphere-microbe interactions are modulated by nutrient availability in the soil that bacterial communities act somewhat independently of plants 41 . In particular, the concentration of available P in the soil adjusted the establishment of them 42 or induced plant immunity through PR1 and WRKY40 gene expression which induced salicylic acid-dependent responses improving resistance against the pathogen 43 . Recent studies also showed that plants with a superior defense and nutrient acquisition, possibly, have specific microbial communities in the rhizosphere, suggesting a close link between plant growth parameters and rhizosphere microbiota functions. A fine example was recently provided that P starvation response 1 (PHR1) in Arabidopsis regulates P stress responses regulated a functionally appropriate set of immune-related genes and contributed to the assembly of root microbiomes 44 .
About 40% of Streptomyces species produce melanin pigments, dark-brown to black, on tyrosine-containing agar media 45 that are not essential for the growth and development of them but play a vital role in their survival and competitiveness. They have multiple functions including antioxidant and antimicrobial activities, tolerance to extreme conditions, and UV radio-protective 45 to prepare light stable bio-pesticides 46 . Conversely, melanin production and tyrosinase activity of soil bacteria (e.g. Rhizobium and Azospirillum) play a role in their symbiotic relationship with plants 45 . In this study, a melanin-producing strain SS14 suppressed the frequency of some Penicillium type colonies and the relative abundance of some bacterial OTUs in the rhizosphere (Fig. 2). Some species of Penicillium have P solubilizing ability and are involved in plant growth promotion 47 . Interestingly, SS14 increased the abundance of Rhizobium indicating the responses of Rhizobium to biocontrol species Under the non-sterile condition, IT20 showed an increased capacity to suppress Phytophthora blight and promote plant growth. The higher number of total bacterial and fungal colonies associated with IT20 may be correlated with the increased plant biomass which was not observed for SS14. Illumina amplicon sequencing analysis of 16S rRNA gene revealed that IT20 differently manipulated soil prokaryote communities compared to SS14. IT20 had higher microbial community responders than the melanin-producing strain SS14. Therefore, inorganic P solubilizing capacity and P compounds have significant effects in interactions between soil beneficial microbes. Hence, melanin production provided less priority than P solubilizing to select biocontrol strains of Streptomyces in the soil applications. Community differences between IT20 and control samples were less pronounced but phylum Cyanobacteria was differentially abundant. The community of Cyanobacteria improve soil fertility through nitrogen fixation and balancing mineral nutrition in the soil. Many members of cyanobacteria are known to release various kinds of biologically active components like phytohormones that act as an elicitor to promote plant growth 48 . Hence, increasing growth parameters like shoot length and plant biomass had been exposed to be positively linked with the various phytochemical components promoted by Cyanobacteria 48 .
The prokaryotic communities shaped with phylogenetically diverse OTUs that relative abundance of them increased or decreased compared to PC (Table 4). There is a positive correlation in the interaction between Streptomyces strains with rhizospheric bacteria that resulted in a lower abundance of some OTUs and lower disease prevalence (Tables 5, 6). Some phylotypes of these genera could correspond with pathogenic interactions or respond to pathogen inoculation. The different mechanisms could increase the abundance of specific plantassociated microbes. The stress condition modulates the root exudate secretion, which consequently attracts specific microbes. For example, Arabidopsis roots in response to a leaf pathogen infection attract Bacillus subtilis into the rhizosphere 49 . In the current study, the most enrichment in IT20 samples was recorded for the genus Devosia affiliated to Gammaproteobacteria, it was previously described that this genus is increased in response to B. velezensis and P. fluorescens, involved in biocontrol activity against R. solanacearum on tomato 50 . Another increase in relative abundance was recorded for the genus Gallionella that was previously described as a member of the core microbiome of the wheat healthy plant 23 . The plant-microbe interactions and plant ability to select neighbors may potentially benefit the plant's growth or defense 32 . The abundance of beneficial microbes is enriched to compete for space and resources using antimicrobial compounds that prevent pathogen growth and virulence 51 . Similar trends were observed for Dokdonella and Sphingobium that were enriched. Some strains of Sphingomonas produce indole acetic acid (IAA) and have protective effects that could be a member of the microbiome in disease-suppressive soils 52 . Dokdonella is an aerobic, non-spore-forming, gram-negative soil bacteria reported as an active root colonizing agent 18 . The other enrichment was observed in Achromobacter. As previously reported, A. xylosoxydans exhibited an antifungal effect and significantly reduced Fusarium wilt disease of tomato plants 22 . Therefore, this reveals that the population of some rhizospheric bacteria can increase as a response to different soil-borne fungal pathogens. The responders shared between both Streptomyces strains IT20 and SS14 (Table 5) shape the microbiota to inhibit pathogen growth, which consequently ameliorated disease suppression. www.nature.com/scientificreports/ There were notable differences in the community pattern of Actinobacteria among inoculated samples. Spearman's rank correlation coefficient showed a clear positive correlation in the interaction of IT20 with the members of Actinobacteria resulted in a higher abundance of corresponding OTUs and lower disease severity (Table 6). These bacteria might play beneficial roles in pepper plants such as supplying nutrients, conferring resistance against pathogens, and anti-oomycete. In contrast, OTUs affiliated to Bacilli including Peanibacillus, Sporosarcina, and Luteolibacter decreased in two Streptomyces treatments. A similar trend was reported by Araujo et al. 53 indicating that the application of biocontrol Streptomyces strains promoted wheat plant growth and modulated the root microbiome by decreasing Paenibacillus and increasing other beneficial bacterial OTUs. Interestingly, Guo et al. 54 applied a consortium of three PGPR strains (B. cereus, B. subtilis, and Serratia sp.) to suppress Phytophthora blight disease resulted in a negative association between Phytophthora disease prevalence and the relative abundance of Sporichthya. Therefore, these results indicate shifting in bacterial community composition induced by biocontrol species of Bacillus could be different from Streptomyces strains to suppress the same pathogen. Therefore, this proposes a possible cross-talk pathway that occurs between bacterial biocontrol agents to manipulate and shape the microbiome. Most importantly, Sporichthya was closely suppressed under the presence of IT20, undoubtedly proving the antagonistic mode of interaction between IT20 and Sporichthya (Fig. 6). www.nature.com/scientificreports/ The co-occurrence of microbes is linked to nutritional interrelationship 55 . In this situation, metabolites of one microbe can be utilized by other community members, then can cause a higher enrichment of microbial species in response to the pathogen 56 . The community pattern of Proteobacteria and Actinobacteria displayed different relationships with two Streptomyces strains and disease suppression. Cooperative relationships are the selective perceptions driving specific rhizospheric bacterial assemblages with plants 57 . The occurrence of these bacteria is being for the first time reported against P. capsici.
Our current understanding of microbiota-mediated plant protection provides an opportunity to recognize and characterize the positive plant microbial interactions for plant growth and survival under stress conditions. The potent biocontrol taxa, helper communities predicted through correlation analysis, would allow designing and constructing synthetic microbial communities (SynComs) for developing efficient inoculants. Overall, constructed communities provide a model to hypothesize and optimize targeted plant disease management and plant growth promotion. A better understanding of the microbiome between plant species and genotypes will increase our ability to efficiently manipulate plant-microbe systems for stable and predictable results in the open fields.

Conclusion
Plant growth-promoting Streptomyces species are used as natural alternatives to synthetic fungicides. Using high throughput sequencing method and microbiome profiling for the first time the dynamics of rhizosphere bacterial communities manipulated by phosphate solubilizing Streptomyces strain was explored and correlated with higher plant growth promotion and disease suppression. Actinobacteria were enriched following pathogen inoculation. In addition to enzyme activities, investigations on other characteristics of superior biocontrol strains of Streptomyces such as secondary metabolite profile and how they impact the assembly of the rhizospheric bacterial communities subsequent pathogen attack could be valuable to optimally design and develop SynComs of Actinobacteria for improving agricultural productivity and environmental sustainability.

Materials and methods
Microorganisms. Fourteen isolates were selected from the Agricultural Biotechnology Research Institute of Iran Culture collection (ABRIICC) based on the plant growth-promoting (PGP) and antifungal activities 2 . PGP traits including siderophore production, phosphate solubilizing ability, indole-3-acetic acid production, and enzyme activities including chitinase, protease, and cellulase were evaluated in a previous study 2 . The Oomycete For the second experiment, the seedlings were placed in pots (15 × 20 cm) filled with a mixture of non-sterile field soil (bulk soil) and peat moss (2:1 v/v). Two selected strains (IT20 and SS14) were evaluated into pathogen inoculated or non-inoculated treatments compared to control (C) and positive control (PC) with five replicates. Rhizosphere was sampled from each pot. After 15 days of inoculation, the plants were harvested and plant traits (shoot length, shoot, root fresh, and dry weight) were measured. Disease incidence (DI), disease severity (DS), and disease suppression (1-DS) were assessed. DS was conducted on a scale from 0 to 5: 0 = no symptoms = 0%, 1 = leaf yellowing = 25%, 2 = minor stem necrosis = 50%, 3 = moderate stem necrosis and some leaf wilt = 75%, 4 = severe stem necrosis and severe wilt, 5 = plant death = 100% 59 .
Molecular and morphological characterizations of the superior isolates. The potent antagonist isolates were characterized by differential morphological traits on ISP2, ISP3, and ISP4 media, melanin forma- Dynamics of the culturable microbiome in the rhizosphere. Fifteen days after pathogen inoculation, the plants were removed carefully and shaken gently. Soil adhering to the roots was considered as the rhizosphere. Rhizosphere samples were collected in sterile zip-lock polyethylene bags. Serial dilutions 1/100 (10 -2 ) for counting the most abundant and common soil fungi (e.g. Penicillium type colonies), 1/10,000 (10 -4 ), and 100,000 (10 -5 ) of each sample were prepared after soil suspension in the sterile saline serum (NaCl 0.9%). To obtain the total number of bacterial and fungal colonies respectively, 100 µl of each dilution was spread on the surface of TSB agar and PDA media supplemented with chloramphenicol (250 mg/l) to avoid bacterial contamination. Bioinformatics analysis of 16S rRNA gene diversity. The forward and reverse sequences (R1 and R2) were assembled using PEAR 66 . The quality checks were conducted using the QIIME pipeline 67 and short sequences were discarded (< 400 bp). Reference-based and de novo chimera detection and clustering (the identity thresholds 94%) of operational taxonomic units (OTUs) were performed using VSEARCH 68 based on reference databases (GREENGENES). The sequences of each OTU were aligned using PYNAST 69 . Taxonomic assignment was done using UCLUST 70  Statistical analysis. Statistical analysis was performed using analysis of variance (ANOVA) by SPSS version 22.0 (SPSS INC. Chicago, IL USA) packages. CFU data were presented on a log scale. The significant difference between treatments was evaluated using Duncan test at the level of P < 0.05. Welch's t-tests applied to compare abundance data of the top genera significantly differed between two bacterial treatments 74 . Spearman's rank correlation coefficient was used to evaluate the correlations between selected rhizosphere genera and disease suppression. The redundancy analysis (RDA) was done to evaluate the relationships between treatments, disease suppression, and microbial genera. Permutation multivariate analysis (PERMANOVA) was performed using the Bray-Curtis distance with the function "adonis" within vegan package of R software (version 3.6.1).