*Article* **Reductive Soil Disinfestation Enhances Microbial Network Complexity and Function in Intensively Cropped Greenhouse Soil**

**Yuanyuan Yan 1,†, Ruini Wu 1,†, Shu Li 1, Zhe Su 1, Qin Shao 2, Zucong Cai 1,3,4, Xinqi Huang 1,3,4,5 and Liangliang Liu 1,2,\***


**Abstract:** Reductive soil disinfestation (RSD) is an effective practice to eliminate plant pathogens and improve the soil microbial community. However, little is known about how RSD treatment affects microbial interactions and functions. Previous study has shown that RSD-regulated microbiomes may degenerate after re-planting with former crops, while the effect of planting with different crops is still unclear. Here, the effects of both RSD treatment and succession planting with different crops on microbial community composition, interactions, and functions were investigated. Results showed that RSD treatment improves the soil microbial community, decreases the relative abundance of plant pathogens, and effectively enhances microbial interactions and functions. The microbial network associated with RSD treatment was more complex and connected. The functions of hydrocarbon (C, H), nitrogen (N), and sulfur (S) cycling were significantly increased in RSD-treated soil, while the functions of bacterial and fungal plant pathogens were decreased. Furthermore, the bacterial and fungal communities present in the RSD-treated soil, and soil succession planted with different crops, were found to be significantly different compared to untreated soil. In summary, we report that RSD treatment can improve soil quality by regulating the interactions of microbial communities and multifunctionality.

**Keywords:** reductive soil disinfestation; microbial community; microbial function; co-occurrence network; soil ecosystem

#### **1. Introduction**

Plastic shed production systems (PSPSs) play an important role in meeting the increasing food demands of a growing population [1]. However, primarily due to economic factors, PSPSs tend to be dominated by both continuous mono-cropping and overfertilization [2]. As a result of these poor management practices, soil quality in PSPSs can become severely degraded, and PSPS soils often suffer from imbalanced soil microbiota and elevated plant pathogen populations, resulting in increased plant disease pressure and economic losses [3,4].

**Citation:** Yan, Y.; Wu, R.; Li, S.; Su, Z.; Shao, Q.; Cai, Z.; Huang, X.; Liu, L. Reductive Soil Disinfestation Enhances Microbial Network Complexity and Function in Intensively Cropped Greenhouse Soil. *Horticulturae* **2022**, *8*, 476. https:// doi.org/10.3390/horticulturae 8060476

Academic Editors: Xiaohui Hu, Shiwei Song and Xun Li

Received: 2 May 2022 Accepted: 25 May 2022 Published: 27 May 2022

**Publisher's Note:** MDPI stays neutral with regard to jurisdictional claims in published maps and institutional affiliations.

**Copyright:** © 2022 by the authors. Licensee MDPI, Basel, Switzerland. This article is an open access article distributed under the terms and conditions of the Creative Commons Attribution (CC BY) license (https:// creativecommons.org/licenses/by/ 4.0/).

Microbial communities and their functional properties are critical for maintaining soil health [5], as microbes are the dominant contributors to both soil ecosystem succession and plant disease suppression [6,7]. For example, soil microbes are heavily involved in carbon, nitrogen, and sulfur cycling, helping to maintain soil nutrient balance and promote healthy plant growth [8]. Furthermore, a highly-connected microbial community network is a key indicator of soil ecosystem stability and functions to guard against invasion by soil-borne pathogens [9]. Microbial communities in healthy soils are often dominated by beneficial microbes which can release a variety of antimicrobial compounds to suppress plant pathogens [10–12]. Additionally, microbial network interactions in healthy soils are more complex than in diseased soils, suggesting that microbes in healthy soils can limit pathogen invasion through niche competition [13,14]. Therefore, how to improve microbial network interactions and functions is a major question in agronomic soil ecology.

Reductive soil disinfestation (RSD), also known as anaerobic soil disinfestation (ASD) or biological soil disinfestation (BSD), is a soil management practice consisting of (1) incorporating organic materials (e.g., crop residues, fresh cover crops, manure, molasses, etc.), (2) irrigating to maximum field capacity, and (3) covering the soil surface with plastic film to produce a reductive and anaerobic soil environment [14,15]. Previous studies have demonstrated that the combination of organic acids, ammonia, hydrogen sulfide, and metal ions produced during RSD treatment can successfully eliminate a broad spectrum of soilborne pathogens [16–18]. Moreover, in practice, RSD has been shown to produce positive impacts on both soil physicochemical properties and microbial communities, including alleviating soil acidification and salinization, and optimizing the soil microbiome [14,18].

Despite such promising results, we still lack a holistic understanding of how and why RSD results in soil improvement. Most studies on RSD have focused on the effect of this management practice on the above-mentioned soil characteristics, leaving many unanswered questions regarding its effect on soil microbial network interactions and functions. Alarmingly, the results of several studies indicated that the microbiomes of soils treated by RSD can degenerate to their previously diseases state when subjected to prior management practices, such as replanting with the previous crops [19–21]. Liu et al. [20,22] have suggested that this phenomenon can be explained by further degradation of the soil abiotic environment and the reintroduction of root exudates produced by the former crops. However, whether and how disparate crops differentially drive soil microbial community succession remains unclear.

Here, we used RSD to treat diseased, *Fusarium*-infected strawberry cultivation soil. Additionally, we succession-planted this soil with both cabbage and tomato. We sought to answer the following questions: (1) What is the influence of RSD treatment on microbial community interactions? (2) How the microbial community composition and function change when RSD-treated soil is succession planted with different crops?

#### **2. Materials and Methods**

#### *2.1. Field Site Description*

The experimental field is located at a plastic shed greenhouse in Suzhou City, Jiangsu Province, China (31◦17- N, 120◦49- E). This region has a marine subtropical monsoon climate, with an average annual temperature of 15.7 ◦C and precipitation of 1100 mm. Strawberry has been planted continuously in this greenhouse for nearly 5 years, and the plant disease incidence has exceeded 30% in recent seasons. The initial soil properties were as follows: pH, 5.26; electrical conductivity (EC), 0.16 mS cm−1; NH4 +-N, 28.68 mg kg−1, and available potassium (AK), 256.0 mg kg<sup>−</sup>1.

#### *2.2. Field Experiment Design*

Two treatments, control ("CTL") and RSD, were performed in the field. CTL was the soil without any treatment, except for its moisture content maintained at 15–20% during the incubation. RSD was the soil amended with 10 t ha−<sup>1</sup> molasses (TOC, 347.8 g kg<sup>−</sup>1; TN, 16.8 g kg<sup>−</sup>1; C/N, 20.7), irrigated to saturation, and covered with transparent plastic film (0.08 mm). The molasses was diluted 20 times before adding to the soil. These treatments were performed for three weeks in September 2019, with the temperature ranging from 25 to 40 ◦C. Each treatment had three replicates and each replicate covered an area of 80 m2. After treatment ("AT"), the plastic film was removed and the soil was drained. Subsequently, the soil was planted with cabbage during the first season ("FS", from October to November 2019) and tomato during the second season ("SS", from December 2019 to May 2020). We collected soil samples from each replicate at all three time points (AT, FS, and SS) using the 5–point sampling method [23]. Soil samples were stored at 4 ◦C for physicochemical analyses and −20 ◦C for DNA extraction.

#### *2.3. Analysis of Soil Physicochemical Properties*

Soil pH and electrical conductivity (EC) were determined at a water/soil (*v:w*) of 2.5:1 and 5:1 using the S220 and S230 metre (Mettler, Greifense, Switzerland), respectively. Soil NH4 +-N was extracted using 2 mol L−<sup>1</sup> KCl solution and determined by a continuous flow analyser (San + +; Skalar, Breda, The Netherlands). Available potassium (AK) was extracted with 1 mol L−<sup>1</sup> ammonium acetate and determined by flame photometry.

#### *2.4. Microbial DNA Extraction and Quantification*

Microbial DNA was extracted from each 0.50 g soil replicate sample using the FastDNA SPIN Kit (MP Biomedicals, Santa Ana, CA, USA). Both the concentration and purity of DNA were assessed using a DS-11 spectrophotometer (Denovix, Wilmington, DE, USA). Bacterial and fungal abundances were detected using the CFX96TM Real-Time System (Bio-Rad Laboratories Inc., Hercules, CA, USA). The PCR amplification mixtures were prepared using 10 μL of SYBR Green Premix Ex Taq™ (2×, TaKaRa, Kyoto, Japan), 1 μL of each primer (Eub338 ACTCCTACGGGAGGCAGCAG and Eub518 ATTACCGCGGCTGCTGG for bacteria, ITS1f CTTGGTCATTTAGAGGAAGTAA and ITS2R GCTGCGTTCTTCATC-GATGC for fungi), 2 μL of the DNA template, and 6 μL of ddH2O. The amplification conditions and standard curves were established as previously described [24].

#### *2.5. Illumina MiSeq Sequencing and Data Processing*

Illumina MiSeq sequencing was used to investigate the compositions, networks, and functions of soil microbial communities. The primers 515F (5- -adapter-MID-GTGCCAGCM GCCGCGG-3- ) and 907R (5- -adapter-MID-CCGTCAATTCMTTTRAGTTT-3- ) were used to amplify the V4-V5 region of the bacterial 16s rRNA gene [25]. The nuclear ribosomal internal transcribed spacer (ITS) region of fungi was amplified using the primers ITS1F (5- -adapter-MID-CTTGGTCATTTAGAGGAAGTAA-3- ) and ITS2 (5- -adapter-MID-GCTGCGTTCTTCATCGATGC-3- ) [14]. The reaction mixtures, amplification conditions, and PCR product purification methods were performed as previously described [13]. After quantification, the equimolar-concentration PCR products were sequenced on an Illumina MiSeq Benchtop Sequencer (Illumina Inc., San Diego, CA, USA) at Tianhao Biotechnologies, Inc. (Shanghai, China).

FASTQ sequence data were processed using the QIIME software(version 1.9.1, Colorado, USA) that developed by Caporaso et al [26]. Briefly, the paired-end sequences were merged and quality-controlled using the default arguments implemented in multiple\_joined\_paired\_ends.py and multiple\_split\_libraries\_fastq.py, respectively. The qualitycontrolled bacterial and fungal sequences were clustered to operational taxonomic units (OTUs) at a similarity of 97%, and then classified and annotated according to the Greengenes 13\_8 database (for bacteria) and UNITE database (for fungi), respectively [27,28]. Finally, the bacterial and fungal sequences were rarified to 137,666 and 134,430 in each soil, respectively. Microbial alpha-diversity was analyzed using the default arguments of alpha\_diversity.py.

#### *2.6. Microbial Functional Prediction and Data Analysis*

The functional compositions of bacterial and fungal communities were predicted using the FAPROTAX (script version 1.2.1, developed by Louca et al., Vancouver, Canada) and FUNGuild (v1.0, developed by Nguyen et al., Minnesota, USA) databases, respectively [29,30]. In order to avoid over-interpreting the fungal functional groups, we retained only those functional groups marked with the "probable" and "highly probable" confidence levels, and deleted those marked only as "possible". In order to compare microbial community and functional dissimilarities, principal coordinates analyses (PCoAs) of the distribution of OTUs, based on the Bray–Curtis distance matrices, were performed using the R 'phyloseq' package [31]. The relative contributions of treatment type and time point on microbial community dissimilarities were analyzed by variance partitioning analysis (VPA) and permutational multivariate analysis of variance (PERMANOVA) using the R 'vegan' package [32]. Co-occurrence network analysis was conducted to investigate the network interactions of microbial communities and functions in different treatments using R 'psych' package, and then visualized with Gephi (v0.9.2, Paris, France) that developed by Bastian et al. [33]. The relative abundances of the OTUs present at greater than 0.15% for bacteria and 0.05% for fungi were retained. Correlation coefficients |r| < 0.6 and *p*-values > 0.05 of the correlation R matrix were removed.

Statistically significant differences in microbial abundances (transformed by log10) between CTL and RSD soils at the same time point were determined using independent-samples *t*-test, and between different time points using LSD's test. The relationships between dominant genera and microbial functions were visualized by heatmap correlation analysis.

#### **3. Results**

#### *3.1. Soil Physicochemical Properties*

After treatment, soil NH4 +-N content was significantly (*p* < 0.05) increased in RSDtreated soil, whereas soil pH, EC, and AK were not significantly different (*p* > 0.05) between RSD and CTL soils (Table S1). During the crops cultivation, except that soil NH4 +-N and AK contents in RSD-treated soil were significantly higher (*p* < 0.05) than those in CTL soil after FS, other physicochemical properties after FS and all physicochemical properties after SS were not significant different (*p* > 0.05) between RSD and CTL soils (Table S1).

#### *3.2. Quantification of Soil Microbes*

After treatment, the abundance of bacteria in RSD-treated soil (1.07 × <sup>10</sup><sup>10</sup> 16S rDNA copies g−<sup>1</sup> dry soil) was significantly (*p* < 0.05) increased by 23.3% compared to CTL soil (8.31 × 109 16S rDNA copies g−<sup>1</sup> dry soil), while the abundance of fungi and the ratio of fungi to bacteria in RSD-treated soil were considerably (*p* < 0.05) decreased by 75.3% and 80.6%, respectively (Figure 1a–c). When soil was succession-planted, the abundance of bacteria in CTL soil and the abundance of fungi and the ratio of fungi to bacteria in RSD and CTL soils increased during the FS of succession planting, but decreased during the SS, with no significant (*p* > 0.05) differences between CTL and RSD-treated soils (Figure 1a–c).

**Figure 1.** Populations of bacteria (**a**) and fungi (**b**) and the ratio of fungi to bacteria (**c**). Error bars represent the standard errors (SES). CTL, untreated soil; RSD, soil incorporated with molasses, irrigated to saturation, and covered with plastic film. AT, FS, and SS indicate the soils collected from after treatment, the first season, and the second season, respectively.

#### *3.3. Soil Microbial Community and Functional Diversities*

Overall, a total of 3,926,128 16S and 3,732,518 ITS sequences passed quality control. After rarefaction, 17,166 bacterial OTUs and 5447 fungal OTUs clustered at 97% similarity.

In RSD-treated soil, the fungal Shannon diversity during the AT and the species richness during the SS were significantly (*p* < 0.05) decreased in comparison to CTL soil. All other measures of bacterial and fungal diversity showed no significant (*p* > 0.05) differences between CTL and RSD-treated soils (Table S2).

PCoA plots illustrated that the microbial communities and functions were considerably distinct between CTL and RSD-treated soils across the three time points (Figure 2a). VPA and PERMANNOVA indicated that microbial communities and functions were significantly influenced (*p* < 0.001) by both treatment and time point, and the relative contribution of time point (i.e., succession planting with different crops) (24.1~47.5%) was greater than that of treatment (7.7~23.1%) (Figure 2b).

**Figure 2.** Dissimilarities in soil microbial communities and functions and their contributors. Principal coordinates analyses (PCoAs) (**a**) of soil microbial communities and functions were determined based on Bray−Curtis distances in different treatments and time points. The relative contributions and significant effects of treatments and time points to the dissimilarities in microbial communities and functions were calculated using VPA and PERMANOVA analyses (**b**). CTL, untreated soil; RSD, soil incorporated with molasses, irrigated to saturation, and covered with plastic film. AT, FS, and SS indicate the soils collected from after treatment, the first season, and the second season, respectively.

#### *3.4. Co-Occurrence Networks of Microbial Communities*

We found that clear differences in the bacterial and fungal community networks, and their topological characteristics, between CTL and RSD-treated soil (Figure 3, Table 1). For example, the number of nodes and edges, average connectivity, and clustering coefficient of both bacterial and fungal networks were greater in the RSD-treated soil, while modularity and average path length were lessened (Table 1). Besides, the different modules contained distinct microbial compositions, and the phylum compositions of each module were also different between CTL and RSD-treated soils (Figure S1).

**Figure 3.** Co-occurrence networks of microbial communities and functions in CTL and RSD soils. The keystone taxa were marked in each network. CTL, untreated soil; RSD, soil incorporated with molasses, irrigated to saturation, and covered with plastic film.

**Table 1.** Topological characteristics of microbial community and function networks in the CTL and RSD soils.


In addition, the identities of the top ten bacterial and fungal keystone taxa between CTL and RSD-treated soils were strikingly different. For example, the bacterial keystone taxa *Comamonadaceae*, *Flavobacteriaceae*, *Rhizobiales*, *Bradyrhizobiaceae*, and *Xanthomonadaceae* were found in CTL soil, whereas *Rhizomicrobium*, *Amycolatopsis*, *Rhodococcus*, *Candidatus\_Koribacter*, *Achromobacter*, and *Arachidicoccus* were found in RSD-treated soil (Table S3). The fungal keystone taxa *Aspergillus*, *Rhodotorula*, *Talaromyces*, *Pseudozyma*, and *Simplicillium* were found in CTL soil, while *Penicillium*, *Chaetomium*, *Purpureocillium*, and *Acremonium* were found in RSD-treated soil (Table S3).

#### *3.5. Microbial Communities and Their Functional Compositions*

The relative abundances of most microbial genera were altered by both RSD treatment and succession planting (Figure S2). After RSD treatment, the relative abundances of dominant bacterial genera *Rhizomicrobium*, *Paenibacillus*, *Ramlibacter*, *Flavisolibacter*, *Hy-* *drogenispora*, *Geobacter*, and *Ruminiclostridium*, and dominant fungal genus *Zopfiella* were significantly increased (*p* < 0.05), whereas the relative abundances of dominant bacterial genera *Bacillus*, *Mizugakiibacter*, *Cladosporium*, *Aspergillus*, *Alternaria*, *Fusarium*, and *Acremonium*, and dominant fungal genus *Gibellulopsis* were significantly decreased (*p* < 0.05) compared to CTL (Figure 4a,b). After the FS of succession planting, the relative abundances of dominant bacterial genera *Rhcrobium*, *Ramlibacter*, *Flavisolibacter*, *Geobacter*, *Ruminiclostridium*, and *Zopfiella*, and dominant fungal genus *Conlarium* were significantly increased (*p* < 0.05), whereas after the SS, the relative abundances of dominant bacterial genera *Streptomyces*, *Paenibacillus*, *Ramlibacter*, and *Flavisolibacter*, and dominant fungal genus *Simplicillium* were significantly increased (*p* < 0.05), compared to CTL (Figure 4a,b).

**Figure 4.** Heatmap displaying the relative abundances of top 20 dominant genera and functional groups in bacteria (**a**) and fungi (**b**). \* (*p* < 0.05), \*\* (*p* < 0.01), and \*\*\* (*p* < 0.001) represent significant differences between CTL and RSD in different time points according to independent-samples *t*-test. CTL, untreated soil; RSD, soil incorporated with molasses, irrigated to saturation, and covered with plastic film. AT, FS, and SS indicate the soils collected from after treatment, the first season, and the second season, respectively.

Overall, RSD treatment had a stronger effect on bacterial functional groups than on fungal functional groups. A total of 69 bacterial function groups and 3 fungal trophic modes (pathotroph, symbiotroph, and saprotroph) were found. In RSD-treated soil, the relative abundances of bacterial functional groups related to hydrocarbon (C, H), nitrogen (N), and sulfur (S) cycling, such as hydrocarbon degradation, nitrate reduction, nitrate respiration, nitrogen fixation, sulfate respiration, and fungal functional groups associated with dung saprotrophy, ectomycorrhizal, and undefined saprotrophy were significantly (*p* < 0.05) enriched, compared to CTL. Notably, the relative abundances of bacterial aerobic ammonia oxidation, and both bacterial and fungal plant pathogens were decreased in RSD-treated soils (Figure 4a,b). After the FS of succession planting, the relative abundances of bacterial

fumarate respiration, nitrate reduction, nitrate respiration, and fungal parasite, and after the SS, bacterial aerobic chemoheterotrophy, chemoheterotrophy, sulfate respiration, and fungal parasite and lichenized were significantly (*p* < 0.05) increased compared to CTL (Figure 4a,b).

#### *3.6. Correlations between Dominant Genera and Functional Groups*

Most of the dominant genera which were increased in RSD-treated soil were significantly (*p* < 0.05) correlated with functional groups and microbial β-diversity indices. For bacteria (Figure 5a), the relative abundance of *Streptomyces* was significantly (*p* < 0.05) and positively correlated with carbon and nitrogen cycling, including hydrocarbon degradation, aromatic hydrocarbon degradation, and nitrate reduction. The relative abundances of *Ramlibacter*, *Flavisolibacter*, *Geobacter*, *Ruminiclostridium*, and *Hydrogenispora* were significantly (*p* < 0.05) and positively correlated with both iron and fumarate respiration. The relative abundances of *Geobacter*, *Ruminiclostridium*, and *Hydrogenispora* were significantly (*p* < 0.05) and positively correlated with both nitrogen fixation and sulfate respiration. For fungi (Figure 5b), the relative abundance of *Zopfiella* was significantly (*p* < 0.05) and positively correlated with ectomycorrhizals, and those of *Penicillium*, *Conlarium*, *Chrysosporium*, *Alternaria*, and *Agaricus* were significantly (*p* < 0.05) and positively correlated with fungal parasites. Additionally, most of the dominant microbial genera were significantly (*p* < 0.05) and positively correlated with microbial β-diversity indices, such as *Rhizomicrobium*, *Streptomyces*, *Paenibacillus*, *Amycolatopsis*, *Zopfiella*, and *Conlarium*.

**Figure 5.** Pairwise spearman's correlations between the relative abundances of dominant genera and microbial functions and β diversity. (**a**) Relationships between the relative abundances of bacterial dominant genera and functional compositions. (**b**) Relationships between the relative abundances of fungal dominant genera and functional compositions. \* *p* < 0.05, \*\* *p* < 0.01, \*\*\* *p* < 0.001, and the blank indicates *p* > 0.05. Blue and red colors indicate the negative and positive correlations, respectively.

#### **4. Discussion**

Many studies have found that RSD treatment can lead to significant changes in soil microbial communities and help resist invasion by soil-borne pathogens [14,15,34]. In agreement with these studies, we found that the relative abundances of known soil-borne pathogens, such as *Fusarium*, *Aspergillus*, *Alternaria*, and *Cladosporium*, were significantly reduced by RSD treatment. Both the bacterial and fungal communities were considerably improved by RSD treatment, showing remarkable increases in the relative abundances of known disease-suppressive agents, such as *Rhizomicrobium*, *Paenibacillus*, *Flavisolibacter*, *Geobacter*, *Hydrogenispora*, and *Zopfiella*. For example, members of the genera *Rhizomicrobium* and *Paenibacillus* can inhibit a broad spectrum of plant pathogens by producing lipopeptides, polypeptides, and bacteriocins [35,36]. Additionally, *Zopfiella* and *Flavisolibacter* have been found to successfully control soil-borne diseases caused by *Fusarium oxysporum*, *Rhizoctonia solani*, and *Cladosporium* in cucumber, Sanqi ginseng, and watermelon [14,20,37].

We also observed that succession planting with different crops (cabbage and tomato) produced similarly striking changes in microbial communities. Specifically, the relative abundances of *Rhizomicrobium*, *Ramlibacter*, *Flavisolibacter*, *Hydrogenispora*, *Geobacter*, *Ruminiclostridium*, *Zopfiella*, *Chrysosporium*, and *Conlarium* after cabbage cultivation, and the relative abundances of *Paenibacillus*, *Streptomyces*, *Ramlibacter*, *Flavisolibacter*, and *Simplicillium* after tomato cultivation, were still significantly enriched in RSD-treated soil. However, the relative abundances of most dominant genera in RSD-treated soil showed a decreased trend during crops cultivation. The trends we observed in the relative abundances were found to be similar to previous studies [19,20,38,39]. That is, these previous studies found that the relative abundances of dominant microbes in RSD-treated soil could return to the same levels of untreated soil after re-planting the former season crops.

Previous studies have suggested that the microbial networks of healthy soils are more complex than those of diseased soils, indicating that the characteristics of the microbial network play an important role in predicting plant health status [9,10,12]. In the present study, we found that both the complexity and connectivity of microbial networks were greater in RSD-treated soils, indicating that RSD treatment can effectively improve the stability of the soil microbial ecosystem. Keystone nodes are generally considered network initiating components, and keystone taxa may play a critical role in maintaining biotic connectivity [40]. We observed that the keystone taxa present in RSD-treated soils included *Amycolatopsis*, *Kribbella*, *Achromobacter*, *Chaetomium*, *Purpureocillium*, *Acremonium*, *Penicillium*, *Rhizomicrobium*, *Candidatus\_Koribacter*, *Achromobacter*, *Rhodococcus*, and *Arachidicoccus*. Many of these taxa have the capacity to control soil-borne pathogens and promote plant health, primarily through the production of a variety of antimicrobials. For example, members of *Amycolatopsis* and *Kribbella* produce antimicrobial compounds such as decaplanin and kribellosides to suppress a variety of pathogens [41,42]. Furthermore, *Chaetomium*, *Acremonium*, *Achromobacter*, and *Penicillium* have been widely reported to act as biological control agents effectively against plant pathogens [4,13,43,44]. Collectively, our results illustrate that RSD treatment can promote a highly connected and more complex microbial network.

Although it is clear that RSD can regulate the soil microbial community and suppress soil-borne pathogens, its effects on other soil functions are still not well-characterized. We found that RSD significantly increased microbial functions associated with nutrient cycling, ectomycorrhizal, and dung saprotrophy, and decreased both bacterial and fungal plant pathogens. These increased nutrient cycling functions, including fumarate, nitrate, sulfate, and iron respiration, nitrogen fixation, nitrate reduction, and hydrocarbon degradation, may be responsible for the decrease of NO3 − and SO4 <sup>2</sup><sup>−</sup> and the increase of organic carbon, NH4 +-N, and Fe2+, previously reported as occurring during RSD treatment [45,46]. The application of molasses is likely responsible for the enrichment in hydrocarbon degradation and fumarate respiration noted after RSD treatment, as molasses may promote the proliferation of carbon source decomposers [4,34]. It has been reported that the removal of NO3 − and SO4 <sup>2</sup><sup>−</sup> can decrease aluminum and iron toxicity, as well as suppress pathogenic growth, promoting the accumulation of plant biomass and increasing crop yield [38]. We found that RSD treatment enriched iron respiration functions, likely due to the increased growth of iron-reducing microbes under anaerobic conditions [47]. Additionally, ectomycorrhizal fungi, found to be enriched after RSD treatment, are able to form a large mycelial network which enhances plants tolerance to biotic and abiotic stresses [48,49]. Most of these functions were changed significantly during succession planting. For instance, the functions of hydrocarbon degradation, iron respiration, and nitrogen fixation were significantly decreased after succession planting, which may be related to the change from reducing to oxidizing soil conditions. After cabbage cultivation in particular, the

abundances of fumarate, nitrate, and sulfate respiration, and nitrate reduction were still significantly higher in RSD-treated soil. This might be due to the root exudates released by and the fertilization patterns applied to distinct crops, resulting in the selective enrichment of microbiomes [50].

Most of the dominant genera which were significantly increased in RSD-treated soils were closely correlated with the above-mentioned nutrient cycling functions. For example, *Hydrogenispora*, *Ruminiclostridium*, *Geobacter*, *Flavisolibacter*, and *Ramlibacter* were associated with both iron and fumarate respiration, and *Streptomyces* was associated with both nitrate respiration and hydrocarbon degradation. Previous studies have found that *Flavisolibacter* and *Geobacter* are important iron-reducing agents, both directly and indirectly converting Fe3+ to Fe2+ under anaerobic conditions [51,52], and that the accumulation of Fe2+ during RSD treatment is significantly inhibitive of the growth of plant pathogens [45]. In addition, *Hydrogenispora*, *Ruminiclostridium*, and *Ramlibacter* are known to be important decomposers of various carbohydrates, and thus stimulate the soil carbon cycle [53–55]. Members of *Streptomyces* are generally considered important nitrate reducers, effectively removing soil nitrate by releasing reductases [56]. Moreover, members of *Streptomyces* have been shown to promote crop growth by decomposing organic phosphorus and producing indole acetic acid [57,58]. These results suggest that the dominant microbial taxa regulated by RSD can effectively improve soil functions, such as organic carbon degradation, nitrate and sulfate removal, and iron reduction.

#### **5. Conclusions**

Our results show that RSD treatment significantly alters the soil microbial community and improves microbial function. The ammonium and available potassium contents in RSD-treated soil were maintained at a high level, even after the planting of two crops. The microbial network associated with RSD treatment was more complex and connected. RSD treatment was able to increase microbial functions associated with hydrocarbon (C, H), nitrogen (N), sulfur (S), and iron (Fe) cycling, and reduce the functions of plant pathogens. Moreover, the improved community functions were closely associated with most of the enriched genera, including *Hydrogenispora*, *Ruminiclostridium*, *Geobacter*, *Flavisolibacter*, *Ramlibacter*, and *Streptomyces*, many of which are also known to be disease-suppressive agents. Combined with succession planting with different crops, RSD can delay the degradation of microbial communities through the recruitment of beneficial microbes to some extent. Here, we suggest that after two consecutive crops are planted in RSD-regulated soil, a new RSD treatment should be re-applied to maintain soil health.

**Supplementary Materials:** The following supporting information can be downloaded at: https:// www.mdpi.com/article/10.3390/horticulturae8060476/s1, Table S1: Changes in soil physicochemical properties after treatment and plants cultivation; Table S2: Changes in the bacterial and fungal α diversities after treatment and plants cultivation; Table S3: Top ten keystone taxa of microbial community networks in CTL and RSD soils; Figure S1: The bacterial (a,b) and fungal (c,d) phylum composition for each module of microbial community networks in CTL and RSD soils. The relative abundance of each phylum was calculated by the average relative abundance of this phylum across all soils divided by the average total relative abundance of all phyla in each module; Figure S2: Relative abundances of bacterial (a) and fungal (b) genera. CTL, untreated soil; RSD, soil incorporated with molasses, irrigated to saturation, and covered with plastic film. AT, FS, and SS indicate the soils collected from after treatment, the first season, and the second season, respectively. "UC-" indicates that the given taxon could not be classified to genus level.

**Author Contributions:** Conceptualization, Y.Y.; Data curation, Y.Y. and R.W.; Funding acquisition, L.L.; Investigation, Z.S.; Methodology, S.L.; Project administration, Z.C. and L.L.; Software, S.L. and Z.S.; Supervision, Z.C.; Validation, Z.C.; Visualization, Q.S.; Writing—original draft, Y.Y. and R.W.; Writing—review and editing, X.H. and L.L. All authors have read and agreed to the published version of the manuscript.

**Funding:** This research was funded by the Key-Area Research and Development Program of Guangdong Province (2020B0202010006), the National Natural Science Foundation of China (Grant No. 32160748, U21A20226, 42090065), the China Postdoctoral Science Foundation (2021M691625), and the Key Research and Development Project (Agriculture) of Yichun City, Jiangxi Province (20211YFN4240).

**Institutional Review Board Statement:** Not applicable.

**Informed Consent Statement:** Not applicable.

**Data Availability Statement:** The raw sequencing data were deposited at the NCBI Sequence Read Archive database with the accession number of PRJNA738280.

**Conflicts of Interest:** The authors declare no conflict of interest.

#### **References**

