Streptomyces Application Triggers Reassembly and Optimization of the Rhizosphere Microbiome of Cucumber

: Streptomyces partum Act12 and Streptomyces roche D74 are biocontrol strains that can promote plant growth and enhance stress resistance in different crops. However, their effects on the rhizosphere microbiome and the role of the reassembled microbiome in plant growth promotion and stress resistance enhancement remain unclear. This study investigated the variation in the rhizosphere microbiome induced by Streptomyces application through a cucumber ( Cucumis sativus L. cv. “Youliang”) pot experiment. The bacterial and fungal communities of rhizosphere soils inoculated with and without Streptomyces were, respectively, compared based on 16S rRNA and internal transcribed spacer rRNA gene sequences. Following Streptomyces application, the bacterial alpha diversity increased signiﬁcantly, while the fungal alpha diversity exhibited the opposite trend. The bacterial and fungal communities’ compositions clearly shifted in the inoculated soil. Compared with the uninoculated control, the relative abundance of the genus Streptomyces increased by 68.3%, and the bacterial co-occurrence network in the rhizosphere soil was enriched signiﬁcantly. The relative abundance of bacteria associated with nitrogen ﬁxation was increased by 7.5% following Streptomyces application. Based on the results of this study, we conclude that the application of Streptomyces Act12 and D74 can be used to reassemble and optimize the rhizosphere microbiome of cucumber, which is conducive to plant survival.


Introduction
Microbiomes are considered functional drivers in their eukaryotic hosts. Plant microbiomes can expand the genomic and metabolic capabilities of their hosts, providing or facilitating a range of essential life-support functions, including nutrient acquisition, immunity modulation, and stress tolerance [1]. In addition, the composition and functions of plant microbiomes influence their beneficial effects on plant growth and health.
The rhizosphere is the soil zone near the plant roots, where microbial activity is affected by the availability of plant-derived nutrients and oxygen. It is not a defined zone with a specific size or shape, but rather a zone with gradients of physical, chemical, and biological properties, which change along the radial and vertical directions of the roots [2]. According to Edward et al. [3], the root-associated microbiome extends from the rhizosphere to the rhizoplane (root surface) and to the endosphere (root interior) of rice. The soil can be considered a "seed bank" for root microbiota, the rhizosphere a "selective growth chamber," the rhizoplane a specific habitat or a transitional boundary, by Streptomyces application, and the effect of this variation on cucumber productivity based on biomass.

Pot Experiment
Experimental design: The pot experiment was conducted from 28 February to 30 May 2020 in a solar greenhouse at Northwest A&F University in Yangling, Shaanxi Province, China (34.26 N, 108.07 E'). A two-group experimental design was used, in which one group of pots contained soil without Streptomyces application (control) and the other group of pots contained soil with Streptomyces application (treatment). Each group had nine pots.
Streptomyces agent preparation: The Streptomyces agent consisted of S. partum Act12 (accession number: MH542148) and S. rochei D74 (accession number: KJ145878). Solid fermentation powder from strains Act12 and D74 was mixed at a 1:1 (w/w) ratio to formulate the Streptomyces agent with a total viable count of 8 × 10 9 colony-forming units g -1 , counted by the spread plate method. Streptomyces has been shown to have long-term viability in field experiments and biocontrol strain colonization tests [26].
Soil preparation: The experimental soil was a silty loam obtained from the topsoil (0-20 cm) in the farmland at the experimental site in Yangling. The soil was derived from loess material and classified as Eum-Orthic Anthrosol (Cumulic Haplustalf in the USDA system). The collected soil was air-dried, ground, passed through a 2 mm sieve, and mixed with an organic fertilizer (125 g kg -1 ; Dadi Agricultural Organic Fertilizer Company, Baoji, China). Subsequently, the soil was divided into two portions for the preparation of treatment and control, respectively.
Inoculation of the treatment soil: Before the pot experiment, the Streptomyces agent was inoculated by mixing it thoroughly with the reserved soil (2 g kg -1 dry soil). There was no inoculation during the pot experiment.
Pot preparation: Each pot was filled with 5 kg of the prepared soil according to the experimental design.
Cucumber seeding: Cucumber seeds (Cucumis sativus L. cv. "Youliang") were soaked with 1% sodium hypochlorite solution for 15 min and then washed using distilled water. Germination was carried out at 25 • C over soaked cotton cloth in disinfected Petri dishes. The most robust germinated seeds were selected and transplanted into pots (three seeds per pot). During the experimental period, all the pots were placed randomly in the greenhouse and transposed every day. The field capacity of the soil was maintained at 20% by watering every two days.

Sampling and Biomass Estimation
Three-month-old cucumber plants were dug out using a sampling shovel to obtain the complete root system along with the root-zone soil. The loosely attached soil was shaken off, and only the tightly attached soil from the roots was collected. The roots from each plant were placed in a sterile Ziplock bag, kept on ice, and immediately transferred to the laboratory. The root-adhering soil was gently removed using fine forceps. The soil retained on the root surface was considered the rhizosphere soil. The three roots from the same pot were washed with sterile water to form one composite slurry sample directly. The slurry was centrifuged (at 1.1 × 10 4 × g), the supernatant was removed, and the precipitate was stored as rhizosphere soil at -80 • C for further study. Each fresh plant was washed and dried with absorbent paper. Plant biomass was measured in terms of fresh shoot weight, fresh fruit weight, and dry fruit weight (after oven-drying at 60 • C).

DNA Extraction, PCR Amplification, and High-Throughput Sequencing
Total genomic DNA was extracted from the 18 soil precipitates using the hexadecyl trimethyl ammonium bromide (CTAB) method. The extracted DNA was checked on 1% (w/v) agarose gel, quantified using a UV-VIS spectrophotometer (ND-1000; Nan-oDrop Technologies, Wilmington, DE, USA), and diluted to 1 ng µL -1 . The univer-  (forward: 5 -GTGCCAGCMGCCGCGGTAA-3 /reverse: 5 -GGACTACHVGGGTWTCTAAT-3 ) and ITS-1737F/ITS-2043R (forward: 5 -TCCGTAGGTG  AACCTGCGG-3 /reverse: 5 -GCTGCGTTC-TTCATCGATGC-3 ) were used to amplify the V3-V4 region of the bacterial 16S rRNA gene and the ITS region of the fungal 18S rRNA gene, respectively [30]. DNA fragments of the target genes were amplified from template DNA with Phusion Hot Start Polymerase (New England Biolabs, Ipswich, MA, USA) and Phusion ® High-Fidelity PCR Master Mix with GC buffer according to the manufacturer's instructions. The final amplicon libraries were subjected to the Illumina NovaSeq6000 platform (Illumina Inc., San Diego, CA, USA) for sequencing at Novogene Bioinformatics Technology Co. Ltd (Beijing, China). All the resulting sequences were deposited in the NCBI Sequence Read Archive database (https://www.ncbi.nlm.nih.gov/sra (accessed on 29 January 2021)) under BioProject number PRJNA695869 for bacteria and PRJNA698281 for fungi.

Bioinformatics Analysis
Raw reads of the 16S rRNA gene and ITS region were processed using self-R scripts and USEARCH v.10.0 as follows [31]: assemble paired-end Illumina reads; relabel sequencing names (-fastq_mergepairs); remove barcodes and primers (-fastx_truncate); filter of low-quality reads (expected error per base greater than 1% were discarded, and sequence records with abundance less than 0.0001% of total amplicons were filtered out); find nonredundancy reads (-fastx_uniques); denoise using the command "unoise3" in USEARCH; obtain amplicon sequence variants (ASVs); remove plastid and non-bacterial sequences (for bacteria); and generate the feature table by USEARCH.
All samples were rarefied to the same number of reads as the sample with the lowest number of reads. For each representative ASV, the SILVA 138.1 (for bacteria) and RDP ITS (WARCUP Training Set V2, Jun-2016, for fungi) [32] databases were used to annotate the taxonomic information. Alpha-diversity analysis was carried out using the richness index and the Shannon index [33]. Functional annotation of fungal taxa was carried out using FUNGuild (https://github.com/UMNFuN/FUNGuild (accessed on 17 December 2020), update by zewei Song) [34].

Data Analysis and Statistics
Biomass analysis: As plant growth differences were observed within one pot, cucumber plants within one pot were divided into three groups based on plant height (high, medium, and low), and then the biomass data were analyzed by grouping, according to the model of a Randomized Block Design.
One-way analysis of variance was performed to determine differences in taxa abundance, and microbial alpha-diversity indices between the control and Streptomyces treatments. Statistically significant differences were defined at the level of p < 0.05. Principal coordinate analysis (PCoA) based on the Bray-Curtis dissimilarity algorithm was used to assess variation in microbial community beta diversity between the control and Streptomyces treatments.
Co-occurrence network analysis was performed to assess the connections within bacterial and fungal communities at the genus level. The bacterial and fungal genera with average relative abundances higher than 0.1% were selected. The Spearman correlation matrix between the relative abundance of two genera was calculated. Both the correlation matrix (r matrix) and the significance matrix (P matrix) were estimated using the "Hmisc" package in R, v.4.0.3 (http://www.r-project.org/ (accessed on 10 Oct 2020)). Correlation data were filtered with a cut-off of |r| = 0.6-0.93 and significant correlation at p < 0.05. The data were loaded into Gephi software version 0.9.2 (https://gephi.org/ (accessed on 26 Sep 2017)) to construct co-occurrence networks and visualized using the Fruchterman-Reingold layout algorithm [7]. The following topological features were calculated: average connectivity, average path length, average clustering coefficient, and modularity [15]. The STAMP software v2. 1 2015)) was used to identify and visualize the phyla and genera with significant differences in relative abundance between the control and the Streptomyces treatments (t-test, p < 0.05).

Biomass of Cucumber Plants
The weights of the cucumber shoots and fruits are shown in Table 1. Regardless of the size of the cucumber plant, the fresh shoot weight of cucumber had no significant differences between control and Streptomyces-treated plants. Streptomyces application resulted in weight increases of 58% in the fresh fruit and 65% in the dry fruit for the high group (p < 0.05); however, these weights were not affected by Streptomyces application in the middle group. The potted plants in the low group did not bear cucumber fruit. 3.16 ± 0.63 The duplicates within one pot were divided into three groups with high, middle, and low plant height, respectively. The control was treated without Streptomyces. Values are means ± standard deviation (n = 3). Different letters indicate statistically significant difference between the control and Streptomyces treatments, at the level of p < 0.05.

Distinctive Microbial Community Diversity
The bacterial alpha diversity (richness index and Shannon index) of rhizosphere soil treated with Streptomyces was significantly higher than that of the control, while the alpha diversity of the fungal community exhibited the opposite trend ( (e) (f)

Increased Streptomyces Abundance and Enriched Bacterial Networks
After Streptomyces application, the relative abundances of the phylum Actinobacteria and the genus Streptomyces increased considerably, by 50.4% and 69.9%, respectively, in the rhizosphere soil of cucumber (t-test, p < 0.05; Figure 4). The microbial co-occurrence networks of bacterial and fungal genera were compared between the rhizosphere soils inoculated with Streptomyces and without inoculation ( Figure 5; Table S1). The bacterial In addition, the relative abundances of bacterial genera Arenimonas, R. sphaerophysae, RB41, and Hydrogenophaga decreased by 39.7%, 25.7%, 42.2%, and 30.1%, respectively, following Streptomyces application (p < 0.05) compared to the control. In contrast, the relative abundances of all 11 of the other bacterial genera increased following Streptomyces application (p < 0.05; Figure 3b).

Increased Streptomyces Abundance and Enriched Bacterial Networks
After Streptomyces application, the relative abundances of the phylum Actinobacteria and the genus Streptomyces increased considerably, by 50.4% and 69.9%, respectively, in the rhizosphere soil of cucumber (t-test, p < 0.05; Figure 4). The microbial co-occurrence networks of bacterial and fungal genera were compared between the rhizosphere soils inoculated with Streptomyces and without inoculation ( Figure 5; Table S1). The bacterial network of the control soil had 296 nodes and 2342 links (Table S1). In the Streptomycestreated soil (Figure 5b) there were 313 nodes in the bacterial network, which was similar to the number in the bacterial network of the control (Figure 5a). However, the number of links in the bacterial network of the Streptomyces-treated soil increased markedly to 3536, which was 51.0% higher than the number in the control treatment. In particular, the links connected with Streptomyces increased from 20 to 41, which was 105.0% higher than that in the control. The fungal network in the control soil ( Figure 5c) had 62 nodes and 405 links, which changed slightly to 69 nodes and 399 links in the fungal network in the Streptomyces-treated soil (Figure 5d).  (Table S1). In the Streptomycestreated soil (Figure 5b) there were 313 nodes in the bacterial network, which was similar to the number in the bacterial network of the control (Figure 5a). However, the number of links in the bacterial network of the Streptomyces-treated soil increased markedly to 3536, which was 51.0% higher than the number in the control treatment. In particular, the links connected with Streptomyces increased from 20 to 41, which was 105.0% higher than that in the control. The fungal network in the control soil ( Figure 5c) had 62 nodes and 405 links, which changed slightly to 69 nodes and 399 links in the fungal network in the Streptomyces-treated soil (Figure 5d).

Potential Functions of Rhizosphere Microbial Communities
Streptomyces application significantly altered the potential functions of the rhizosphere bacterial communities associated with cucumber plants (Figure 6). The relative abundances of bacteria associated with carbon cycling were higher in Streptomyces-treated soil than in the control soil ( Figure 6). This was especially true for bacterial taxa associated with aromatic compound degradation, hydrocarbon degradation, and xylanolysis, which increased by 12.1%, 23.0%, and 85.1%, respectively (t-test, p < 0.05) after Streptomyces application. Furthermore, the relative abundances of bacteria associated with nitrogen fixation increased by 6.9% following Streptomyces application (t-test, p < 0.05). After Streptomyces application, the relative abundances of nitrification-associated bacteria increased by 39.2%, in contrast to a 2.3% reduction in the relative abundance of ammonification-associated bacteria.

Potential Functions of Rhizosphere Microbial Communities
Streptomyces application significantly altered the potential functions of the rhizosphere bacterial communities associated with cucumber plants (Figure 6). The relative abundances of bacteria associated with carbon cycling were higher in Streptomyces-treated soil than in the control soil ( Figure 6). This was especially true for bacterial taxa associated with aromatic compound degradation, hydrocarbon degradation, and xylanolysis, which increased by 12.1%, 23.0%, and 85.1%, respectively (t-test, p < 0.05) after Streptomyces application. Furthermore, the relative abundances of bacteria associated with nitrogen fixation increased by 6.9% following Streptomyces application (t-test, p < 0.05). After Streptomyces application, the relative abundances of nitrification-associated bacteria increased by 39.2%, in contrast to a 2.3% reduction in the relative abundance of ammonification-associated bacteria.
with aromatic compound degradation, hydrocarbon degradation, and xylanolysis, whi increased by 12.1%, 23.0%, and 85.1%, respectively (t-test, p < 0.05) after Streptomyces a plication. Furthermore, the relative abundances of bacteria associated with nitrogen fix tion increased by 6.9% following Streptomyces application (t-test, p < 0.05). After Strep myces application, the relative abundances of nitrification-associated bacteria increased 39.2%, in contrast to a 2.3% reduction in the relative abundance of ammonification-ass ciated bacteria. Figure 6. Heatmap of the potential functional groups of rhizosphere bacteria communities. T and the color purple represent Streptomyces treatment; CK and the color blue represent the control without Streptomyces. C and the color green represent bacteria abundance associated with the carbon cycling function; N and the color green represent bacteria abundance associated with the nitrogen cycling function. * p < 0.05; ** p < 0.01.

Discussion
The Streptomyces strains Act12 and D74 have been proved to have strong positive effects on plant growth and stress resistance in different crops, such as cucumber (Tables  S2-S5), tomato [15], wheat [16], and corn [14]. Consistently, the biomass of cucumber plants was remarkably increased by the application of these two Streptomyces strains in the present study (Table 1). However, it remains unclear whether the effects of Streptomyces application on the rhizosphere microbiome contribute to plant growth. Therefore, we compared the structure and function of the rhizosphere microbiome of cucumber between soils inoculated with Streptomyces and those without inoculation.

Rhizosphere Microbiome Diversity and Structure Influenced by Streptomyces Application
In the present study, Streptomyces application increased the alpha diversity of the rhizosphere bacterial community but reduced the alpha diversity of the rhizosphere fungal community associated with cucumber plants. In the Streptomyces-treated rhizosphere soil, the increased richness index of bacteria and the reduced richness index of fungi are consistent with the findings of a previous culture-dependent study that concluded that the rhizosphere microbial community shifts from a fungal type to a bacterial type [33]. Furthermore, the beta diversity of both rhizosphere bacteria and fungi varied considerably between Streptomyces treatments and control treatments. The results are consistent with the findings of a previous study on monkshood [26], indicating that Streptomyces could considerably alter the diversity of rhizosphere microbial communities.
Among the top five bacterial phyla detected in the rhizosphere soil of cucumber, the relative abundances of phyla Proteobacteria, Bacteroidetes, and Acidobacteria decreased non-significantly following Streptomyces application. The relative abundances of Actinobacteria and Gemmatimonadetes increased following Streptomyces treatment, which could facilitate the healthy growth of plants. The phylum Actinobacteria comprises many plant-promoting rhizobacteria, including both Streptomyces and non-Streptomyces [35]. The phylum Gemmatimonadetes was reported to show an adaptation to drier soils [36], which may be correlated to plant drought resistance by Streptomyces application. On the other hand, Firmicutes also increased following Streptomyces treatment, which is consistent with a previous study that the enhanced Firmicutes was also observed when the same Streptomyces treatment was applied on tomato [15]. The phylum Firmicutes comprises the class Bacillus, and the function of growth promotion and disease resistance has been reported in the genus Bacillus aryabhattai [37,38], Bacillus subtilis [39,40], and Bacillus amyloliquefaciens [41,42]. The enhancement of this phylum in the present study indicates that the stress resistance of the microbiota in the rhizosphere was enhanced by Streptomyces. Both Firmicutes and Actinobacteria comprise Gram-positive bacteria, which means that there was an increase in the abundance of Gram-positive bacteria in this study. This phenomenon may have been induced by the selection of antibiotics (unpublished results) secreted by the inoculated Streptomyces. All the above discussions were based on the relative abundance; the richness index of bacteria increased by 5.3% following Streptomyces application.
Compared with the annotation of bacteria, only 50% and 75% of fungi in the control and treatment groups were annotated at the phyla level, respectively. This may have been caused by the following reasons: (1) The fungal species are less abundant than bacterial species in the environment; usually, the true OUT or ASV is less than 500. Then, the same variation of absolute abundance would give higher variation in relative abundance in fungi than in bacteria. (2) The fungus database is not as comprehensive as the bacteria database. (3) In most pot experiments, the potted habitat is isolated with a fixed narrow space, compared with field experiments. The ecosystem stability is lower than that of field experiments, which is conducive to forming some species with high relative abundance. If the dominant species cannot be annotated, then the fungal annotation rates could be very low. Among the fungal phyla identified in the rhizosphere soil of cucumber, the relative abundances of Ascomycota, Zygomycota, and Basidiomycota decreased after Streptomyces application. Ascomycota includes plant-harming microorganisms such as Myrothecium roridum, Monographella cucumerina, Humicola fuscoatra, and Verticillium dahlia [26]. Such changes support the notion that plant pathogens were inhibited by Streptomyces application. All the above discussions were based on the relative abundance; the richness index of fungi decreased by 19.8% following Streptomyces application.
Compared with the previous results following monkshood treatment with a Streptomyces agent [26]: (i) the results of both studies support that all quantitative changes in specific bacteria and fungi in the rhizosphere soil explain the positive effects of Streptomyces on cucumber plants; (ii) however, in contrast to the previous study, increases in the relative abundances of Proteobacteria, Bacteroidetes, and Ascomycota were not observed in the present study. These results imply that Streptomyces application has a positive effect on different plants, while specific variations on the soil rhizosphere microbiome may be closely associated with the soil type or the plant type.
Microbiomes in the rhizosphere are mainly analyzed with the relative abundance, which may introduce a negative correlation bias and display spurious correlations [43]. In recent years, the more accurate methods of ANCOM and Aldex2 were increasingly used to avoid this bias from compositional data, especially in the studies for human microbiome data [43,44]. To improve the accuracy of our data analysis, the community variation induced by Streptomyces application was also analyzed with ANCOM in ASV level (Supplementary Figure S2). The ANCOM results showed trends similar to our analysis based on traditional relative abundance data, in both bacteria and fungi community. This phenomenon can be explained by the similar volume observed in both treatment and control samples.

Bacterial Co-Occurrence Enhanced by Streptomyces in the Rhizosphere
In the present study, we observed increases in the relative abundances of the phylum Actinobacteria and the genus Streptomyces in the Streptomyces-treated rhizosphere soil of cucumber when compared with those in the control. A similar result was observed in the rhizosphere of monkshood, where the colonization of strain D74 on the root surface of monkshood 7 days after inoculation was verified using the green fluorescent protein labeling technique [26]. Another study found that Streptomyces were predominant in wheat roots during the first weeks of inoculation [45]. Although these studies did not last as long as 3 months, their results are consistent with and support our study, where the inoculation of Streptomyces increased the relative abundances of the phylum Actinobacteria and the genus Streptomyces.
Here, network analysis results revealed that Streptomyces application altered the cooccurrence patterns of bacteria in the rhizosphere soil of cucumber. Following Streptomyces application, as the number of links connected with Streptomyces increased, the total links in the bacterial co-occurrence network increased. However, changes in bacterial co-occurrence networks could have benefits related to plant growth promotion and pathogen resistance. Indeed, Li and coworkers observed that pathogen invasion was influenced by interactions (competition or facilitation) among microbial communities [46]. Therefore, the establishment of the complex bacterial network also implies that the possibility of pathogen invasion was changed. Additionally, it has been shown that Streptomyces can promote plant growth indirectly by stimulating the growth of other plant-beneficial microbes. For example, the application of strain Act12 stimulated an increase in the population of P. koreensis in the rhizosphere of ginseng [27], which was consistent with our results that the genus Pseudomonas was promoted by 33.82% following Streptomyces application. Therefore, the changes in bacterial co-occurrence networks observed here also imply altered interactions between the plant and the rhizosphere microbiome.
Moreover, to avoid the bias from compositional data, more rigorous methods were invented for the network construction, such as SParcc and SPIEC-EASI [47]. However, due to the large calculation volume, these methods are still mainly used in human data analysis, and rarely used to analyze samples from environments and soils.

Reassembled Rhizosphere Microbiome Function May Be More Conducive to Plant Survival
In the present study, Streptomyces application increased the relative abundances of bacteria associated with the carbon cycling function in the rhizosphere soil of cucumber, which implies that more carbon sources-in both type and amount-were provided in the rhizosphere as substrates, via which the host could acquire the capacity to reassemble a preferred rhizosphere microbiome.
With regard to nitrogen cycling, there was a remarkable increase in the relative abundance of bacteria associated with nitrogen fixation after Streptomyces application, which suggests that the reassembled bacterial community could import more nitrogen into the rhizosphere zone. In addition, the relative abundances of bacteria associated with nitrification and ammonification increased and reduced, respectively, following Streptomyces application. Such conflicting changes suggest that more nitrogen is transformed into nitrate after Streptomyces application, which can be more easily assimilated by plant roots. Therefore, the rhizosphere microbiome reassembled by Streptomyces application can facilitate the acquisition of more nitrogen (in both quantity and quality) in plant roots.
In summary, the function of the rhizosphere microbiome of cucumber was optimized by Streptomyces application, which offers the plants an opportunity to recruit potentially beneficial microbes and reassemble a rhizosphere microbiome that is more conducive to plant survival.

Conclusions
The results of the pot experiments in the present study showed that the rhizosphere microbiome diversity and structure shift remarkably after Streptomyces application. In particular, the abundances of rhizosphere bacteria increase, while the abundances of fungi decrease. In addition, Streptomyces application enriches the bacterial co-occurrence network substantially; however, the effect on the fungal co-occurrence network remains unclear. The functions of the reassembled rhizosphere microbiome are also altered by Streptomyces application, leading to an enhancement in the nitrogen use efficiency. Therefore, the application of Streptomyces could facilitate the recruitment of beneficial bacteria by cucumber plants. Furthermore, the reassembled and optimized rhizosphere microbiome is more conducive to plant survival.