γ-Polyglutamic Acid Production, Biocontrol, and Stress Tolerance: Multifunction of Bacillus subtilis A-5 and the Complete Genome Analysis

Bacillus subtilis A-5 has the capabilities of high-molecular-weight γ-PGA production, antagonism to plant pathogenic fungi, and salt/alkaline tolerance. This multifunctional bacterium has great potential for enhancing soil fertility and plant security in agricultural ecosystem. The genome size of B. subtilis A-5 was 4,190,775 bp, containing 1 Chr and 2 plasmids (pA and pB) with 43.37% guanine-cytosine content and 4605 coding sequences. The γ-PGA synthase gene cluster was predicted to consist of pgsBCA and factor (pgsE). The γ-PGA-degrading enzymes were mainly pgdS, GGT, and cwlO. Nine gene clusters producing secondary metabolite substances, namely, four unknown function gene clusters and five antibiotic synthesis gene clusters (surfactin, fengycin, bacillibactin, subtilosin_A, and bacilysin), were predicted in the genome of B. subtilis A-5 using antiSMASH. In addition, B. subtilis A-5 contained genes related to carbohydrate and protein decomposition, proline synthesis, pyruvate kinase, and stress-resistant proteins. This affords significant insights into the survival and application of B. subtilis A-5 in adverse agricultural environmental conditions.


Introduction
The excessive application of traditional chemical inputs (e.g., fertilizers and pesticides) in agricultural environments leads to a series of problems, including crop losses, environmental pollution, and pathogen resistance. The application of multifunctional microorganisms is of great importance to the synchronous reduction of chemical fertilizers and pesticides, soil restoration, food security, agricultural productivity, and ecological agriculture development [1]. Microorganisms capable of promoting plant growth as a new fertilizer, as well as controlling pathogens as a novel pesticide, have become an emerging research target worldwide as they have the potential to improve the scientific and technological innovation of green agricultural inputs [2].
Bacillus can produce many kinds of metabolites, such as amino acids, antibiotics, vitamins, nucleosides, sugars, and various enzymes. It, thus, acts as an important resource strain for the utilization of microbial metabolites [3]. B. subtilis has been widely applied in novel fertilizers, biological feeds, green pesticides, and other products, because it is the B. subtilis A-5 was inoculated in Luria-Bertani (LB) medium at 37 • C and 200 rpm overnight. Then, 2% inocula were transferred to the optimized fermentation (FF) medium at 37 • C and 200 rpm for 48-96 h. The FF medium consisted of the following components (g/L): glucose, 40.0; sodium glutamate, 40.0; yeast extract, 5.0; MgSO 4 , 0.25; K 2 HPO 4 , 2.0. The pH was set to 7.0-7.2. The γ-PGA yield was measured using a cetyltrimethylammonium bromide-dependent spectrophotometer method and a UV-vis spectrophotometer (Puxi T6, Beijing, China). Its characteristics (number-average molecular weight, Mn; weight-average molecular weight, Mw; and polydispersity index, PDI) were measured by a gel permeation chromatography system (Shimadzu LC-20A, Kyoto, Japan) [10].

Analysis of Antagonistic Activity against Plant Pathogens
Plant pathogens were placed in the middle of a potato dextrose agar (PDA) medium plate using an Oxford cup (d = 0.80 cm) (Hopebio, Qingdao, China) and strain A-5 was inoculated 1.0 cm away from the plate edge. Plates only containing inoculated pathogen were used as the control treatment (CK). B. velezensis SS-20 was set as a positive inhibitor for comparison [11]. Furthermore, a glass slide was inserted into the PDA plate at a 45 • angle to collect the growing mycelia. Treatments were performed in triplicate and cultured at 28 • C for 5-7 d. Mycelial morphology changes were viewed by an optics microscope (Olympus BX43, Center Valley, Japan) and a Cryo-scanning electron microscope (Cryo-SEM) (FEI Quanta 450, Hillsboro, OR, USA) after gradient ethanol dehydration. The inhibition rate = [1 − (mycelial diameter in experimental group/mycelial diameter in control group)] × 100% [12].

Salinity and Alkalinity Tolerance of B. subtilis A-5
To determine the salinity and alkalinity tolerance of B. subtilis A-5, the strain was inoculated in LB medium with a pH and a NaCl concentration ranging from 5.5 to 10.5 and 1% to 12% (w/v), respectively. Then, treatments were cultured at 37 • C for 2-4 d. Growth biomass was used as the criteria and measured spectrophotometrically at an optical density (OD) of 600 nm. Positive and negative control strains (B. velezensis SS-20 and B. subtilis FD4-1) were also included to compare the statistical significance. Furthermore, the tolerance of strain A-5 to salinity and alkalinity combined was determined by increasing the salt content and pH.

Whole Genome Sequencing and Analysis of B. subtilis A-5
After inoculation for 10-12 h (late logarithmic stage) in LB medium at 37 • C and 200 rpm, strain A-5 was collected to extract the genomic DNA using the Wizard ® Genomic DNA Purification Kit (Promega, Madison, WI, USA) according to the manufacture's protocol. After purification, genomic DNA was quantified by a TBS-380 fluorometer (Turner, CA, USA). High-quality DNA (OD 260/280 = 1.8-2.0, >20 µg) was used for further experiments.
Genomic DNA was sequenced using a combination of PacBio RSII single molecule real-time (SMRT) and Illumina HiSeq PE150 sequencing platforms. For Pacific Biosciences sequencing, an aliquot of 15 µg of DNA was spun in a Covaris g-TUBE (Covaris, Woburn, MA, USA) at 6000 rpm for 60 s by a centrifuge (Eppendorf 5424, Hauppauge, NY, USA). Then, DNA fragments were purified, end-repaired, and ligated with SMRTbell sequencing adapters (Pacific Biosciences, Menlo Park, CA, USA). The resulting sequencing library was purified three times with 0.45× the volume of Agencourt AMPure XP beads (Beckman Coulter Genomics, Newton, MA, USA). Next, a~10 kb insert library was prepared and sequenced on one SMRT cell using standard methods. The Illumina data were used to evaluate and elevate genome completeness. The clean data reads were assembled into a contig using a hierarchical genome assembly process and canu [13]. The last circular step was checked and finished manually, thus generating a complete genome with seamless chromosomes and plasmids. Finally, error correction of the PacBio assembly results was performed using the Illumina reads provided by Pilon. Genome analysis was mainly carried out using the Majorbio Cloud Platform (https://cloud.majorbio.com, assessed on 3 March 2020). Glimmer [14], tRNA-scan-SE [15], and Barrnap were used for the prediction of coding sequence (CDS), tRNA, and rRNA, respectively. The predicted CDSs were annotated from non-redundant protein sequence (NR), Swiss-Prot, Pfam, gene ontology (GO), clusters of orthologous groups of proteins (COG), and Kyoto Encyclopedia of Genes and Genomes (KEGG) databases using alignment tools such as BLAST, Diamond, and HMMER. Briefly, each set of query proteins was aligned with the databases, and annotations of the best-matched subjects (E-value < 10 −5 ) were obtained for gene annotation. The carbohydrate-active enzyme (CAZy) database was used to annotate carbohydrate active enzymes. AntiSMASH 6.0.1 was applied to analyze the secondary metabolite synthesis gene clusters, and TMHMM 2.0 and PSORTb 3.0.3 were used to predict the transmembrane helices in the proteins and subcellular locations of proteins, respectively.

γ-PGA Fermentation by B. subtilis A-5
B. subtilis A-5 yielded 34.21 g/L γ-PGA with an ultra-high molecular weight (Mw = 4700 kDa) in the FF medium at 48 h ( Table 1). The yield and Mw of γ-PGA decreased to 24.58 g/L and 1106 kDa, respectively, when the cultured time reached 96 h, indicating that strain A-5 contained γ-PGA degradation-related enzymes [16]. PDI refers to the distribution breadth of a polymer's molecular weight; the larger the PDI is, the wider the molecular weight distribution. The PDI value of polymers produced via step-growth polymerization is usually greater than 2.0 [17]. In this study, the PDI value of γ-PGA produced by strain A-5 varied from 1.860 to 2.663, verifying that this γ-PGA belonged to a mixture with certain molecular weight ranges formed via stepwise polymerization.

Antagonistic Activity against Fungal Plant Pathogens
The inhibition rates of M. grisea, R. solani, and F. oxysporum on mycelia growth were 77.93%, 22.09%, and 6.66%, respectively ( Figure 1). Compared with the positive strain B. velezensis SS-20, strain A-5 demonstrated a significant decrease (32.05%; p < 0.01) in the inhibition efficiency on R. solani, which was not observed in the other fungal plant pathogens. The inhibitory efficiency of strain A-5 on F. oxysporum was considerably low compared to that of pathogens of rice (M. grisea and R. solani). Microscopy revealed the growth and morphological changes of M. grisea with or without strain A-5 treatment ( Figure 2). The mycelia of M. grisea showed severe morphological abnormalities in the interaction zones in situ as compared with those of the control at the 100 µm scale of observation (Figure 2a,b). Optical microscope and Cryo-SEM identified healthy, dense, and branched mycelium in the control treatment, with no abnormality or disruption (Figure 2c,e); however, the hyphae of M. grisea were disrupted, flattened, and shriveled when challenged with B. subtilis A-5 (Figure 2d,f). The damage to the mycelium structure eventually inhibited further growth of M. grisea. Notably, γ-PGA presented no effect on pathogen growth whether it was obtained from B. subtilis A-5 or from commercial ones (data not shown).
The inhibition rates of M. grisea, R. solani, and F. oxysporum on mycelia 77.93%, 22.09%, and 6.66%, respectively ( Figure 1). Compared with the posi velezensis SS-20, strain A-5 demonstrated a significant decrease (32.05%; p < inhibition efficiency on R. solani, which was not observed in the other fungal gens. The inhibitory efficiency of strain A-5 on F. oxysporum was considerab pared to that of pathogens of rice (M. grisea and R. solani). Microscopy reveale and morphological changes of M. grisea with or without strain A-5 treatmen The mycelia of M. grisea showed severe morphological abnormalities in th zones in situ as compared with those of the control at the 100 μm scale of (Figure 2a,b). Optical microscope and Cryo-SEM identified healthy, dense, a mycelium in the control treatment, with no abnormality or disruption (Figur ever, the hyphae of M. grisea were disrupted, flattened, and shriveled whe with B. subtilis A-5 (Figure 2d,f). The damage to the mycelium structure even ited further growth of M. grisea. Notably, γ-PGA presented no effect on path whether it was obtained from B. subtilis A-5 or from commercial ones (data n

Analysis of the Salinity and Alkalinity Tolerance of B. subtilis A-5
Compared to the positive strain (B. velezensis SS-20) and negative strain (B. subtilis FD4-1), B. subtilis A-5 conferred a high-level tolerance to salinity and alkalinity. Strain A-5 exhibited stress resistance under 12% NaCl or pH 10.5 condition, but the OD 600 was only 0.15 at the 4th d ( Figure 3). Under 1-10% contents of NaCl, strain A-5 grew well, with OD 600 ranging from 0.44 to 1.45. The OD 600 of strain A-5 exceeded 0.70 in acidic condition (pH = 5.5). As the alkalinity increased, the bacterial biomass increased to OD 600 = 1.48 (pH = 7.0) and then significantly decreased (p < 0.05). Furthermore, strain A-5 grew well in LB medium with up to 8-10% NaCl and pH 8.5-9.5 (data not shown).

Analysis of the Salinity and Alkalinity Tolerance of B. subtilis A-5
Compared to the positive strain (B. velezensis SS-20) and negative strain (B. subtilis FD4-1), B. subtilis A-5 conferred a high-level tolerance to salinity and alkalinity. Strain A-5 exhibited stress resistance under 12% NaCl or pH 10.5 condition, but the OD600 was only 0.15 at the 4th d ( Figure 3). Under 1-10% contents of NaCl, strain A-5 grew well, with OD600 ranging from 0.44 to 1.45. The OD600 of strain A-5 exceeded 0.70 in acidic condition (pH = 5.5). As the alkalinity increased, the bacterial biomass increased to OD600 = 1.48 (pH = 7.0) and then significantly decreased (p < 0.05). Furthermore, strain A-5 grew well in LB medium with up to 8-10% NaCl and pH 8.5-9.5 (data not shown).

General Features of the B. subtilis A-5 Genome
The complete genome of B. subtilis A-5 consisted of one chromosome (Chr) and two plasmids (pA and pB) with a length of 4,190,775 bp, 43.37% guanine-cytosine (GC) content, and 4721 genes ( Table 2). A total of 4605 CDSs, which make up 3,664,629 bp of the

General Features of the B. subtilis A-5 Genome
The complete genome of B. subtilis A-5 consisted of one chromosome (Chr) and two plasmids (pA and pB) with a length of 4,190,775 bp, 43.37% guanine-cytosine (GC) content, and 4721 genes ( Table 2). A total of 4605 CDSs, which make up 3,664,629 bp of the total length, accounted for 87.45% of the entire genome. For Chr, pA, and pB, the total nucleotide lengths were 4,120,646, 64,309, and 5820 bp, respectively, and their CDSs were 4512, 87, and 6, respectively. In addition, the genome encoded 86 tRNAs, 30 rRNAs, and 108 sRNAs. A circular plot of the genome (Figure 4) showed the number of bases, GC skew, GC content, and location of all annotated CDSs. A total of 345 virulence factor genes were identified in the B. subtilis A-5 genome. Virulence factors of bacterial pathogens (VFDB) database annotation showed that these were mainly offensive virulence factor, defensive virulence factor, nonspecific virulence factor, and regulation of virulence-associated genes ( Figure S1). Twenty-nine types of drug-resistant genes (239 in number) were identified in the B. subtilis A-5 genome. The dominated resistant drugs were macrolide antibiotic, fluoroquinolone antibiotic, penicillin, tetracycline antibiotic, cephalosporin, aminoglycoside antibiotic, and cephamycin (gene number > 15) ( Table S1). The genome sequencing data of B. subtilis A-5 have been submitted to GenBank (PRJNA824974).

Functional Gene Annotation in B. subtilis A-5
In general, 4605, 3947, 3655, 3358, 3337, and 2291 genes were annotated in NR, Swiss-Prot, Pfam, COG, GO, and KEGG databases, respectively. A total of 3358 genes were annotated in COG database, accounting for 72.92% of the whole genome. There were 3333, 22, and 3 genes located in Chr, pA, and pB, respectively. The genes were mainly distributed in four categories: information storage and processing, metabolism, cellular pro-

Functional Gene Annotation in B. subtilis A-5
In general, 4605, 3947, 3655, 3358, 3337, and 2291 genes were annotated in NR, Swiss-Prot, Pfam, COG, GO, and KEGG databases, respectively. A total of 3358 genes were annotated in COG database, accounting for 72.92% of the whole genome. There were 3333, 22, and 3 genes located in Chr, pA, and pB, respectively. The genes were mainly distributed in four categories: information storage and processing, metabolism, cellular processes/signaling, and poorly characterized. Notably, the "function unknown" type had the highest proportion (27.04%) ( Figure 5), indicating that there is great potential to investigate new genes and functions in B. subtilis A-5. The number of genes related to amino acid transport and metabolism, carbohydrate transport and metabolism, transcription, replication-recombination and repair, and cell wall/membrane/envelope biogenesis all exceeded 200, i.e., 298, 260, 259, 227, and 200, respectively. Among all the 25 functions, 5 functions were not annotated, namely: RNA processing and modification, general function prediction only, extracellular structures, nuclear structure, and cytoskeleton related genes ( Figure 5). In B. subtilis A-5, 3337 genes were annotated in the GO database, accounting for 72.46% of the entire genome ( Figure 6). There were 3296, 36, and 5 genes located in Chr, pA, and pB, respectively. Functional genes were distributed in biological process, molecular function, and cellular component categories. Genes related to the integrated component of membrane, plasma membrane, and cytoplasm exhibited the highest abundances among the sub-functions (i.e., 1001, 567, and 387 genes, respectively). Furthermore, integral component of membrane (68.29%) and DNA binding (21.95%) were the two functions that dominated in the plasmid. The reaction of γ-PGA biosynthesis is located in the cell membrane, and secondary metabolites also need to be transported and secreted into the extracellular matrix through the cell membrane [18]. Thus, it is necessary to explore indepth the importance of membrane-associated proteins in B. subtilis A-5, especially those in the plasmid. In B. subtilis A-5, 3337 genes were annotated in the GO database, accounting for 72.46% of the entire genome ( Figure 6). There were 3296, 36, and 5 genes located in Chr, pA, and pB, respectively. Functional genes were distributed in biological process, molecular function, and cellular component categories. Genes related to the integrated component of membrane, plasma membrane, and cytoplasm exhibited the highest abundances among the sub-functions (i.e., 1001, 567, and 387 genes, respectively). Furthermore, integral component of membrane (68.29%) and DNA binding (21.95%) were the two functions that dominated in the plasmid. The reaction of γ-PGA biosynthesis is located in the cell membrane, and secondary metabolites also need to be transported and secreted into the extracellular matrix through the cell membrane [18]. Thus, it is necessary to explore in-depth the importance of membrane-associated proteins in B. subtilis A-5, especially those in the plasmid.
A total of 2291 genes were annotated to the KEGG pathway, accounting for 49.75% of the total genes, and were divided into 6 metabolic pathways (Figure 7). Accordingly, 2282, 8, and 1 genes were predicted to be located in Chr, pA, and pB, respectively. Most of genes (56.67%) were distributed in the metabolism, followed by environmental information processing (14.49%) and genetic information processing (8.29%). In environmental information processing, membrane transport was the main metabolic pathway and contained 179 gene annotations, indicating the importance of intracellular and extracellular communication.
among the sub-functions (i.e., 1001, 567, and 387 genes, respectively). Furthermore, integral component of membrane (68.29%) and DNA binding (21.95%) were the two functions that dominated in the plasmid. The reaction of γ-PGA biosynthesis is located in the cell membrane, and secondary metabolites also need to be transported and secreted into the extracellular matrix through the cell membrane [18]. Thus, it is necessary to explore indepth the importance of membrane-associated proteins in B. subtilis A-5, especially those in the plasmid. A total of 2291 genes were annotated to the KEGG pathway, accounting for 49.75% of the total genes, and were divided into 6 metabolic pathways (Figure 7). Accordingly, 2282, 8, and 1 genes were predicted to be located in Chr, pA, and pB, respectively. Most of genes (56.67%) were distributed in the metabolism, followed by environmental information processing (14.49%) and genetic information processing (8.29%). In environmental information processing, membrane transport was the main metabolic pathway and contained 179 gene annotations, indicating the importance of intracellular and extracellular communication.

Analysis of Carbohydrate Active Enzyme Annotation Using the CAZy Database
A total of 145 genes encoding proteins were analyzed by CAZy database: 54 glycoside hydrolases (GHs), 45 glycosyl transferases (GTs), 8 polysaccharide lyases (PLs), 29 carbohydrate esterases (CEs), 2 carbohydrate-binding modules (CBMs), and 7 auxiliary activities (AAs) ( Table 3). γ-PGA is first identified in the extracellular capsule, where it combines with polysaccharides, and is also secreted into the extracellular matrix [19]; thus, PLs may be related to the synthesis or secretion of γ-PGA in B. subtilis A-5. Saccharides and proteins are the main components of the residues of plant pathogens in soil [1]. The genes encoding chitinase, lysozyme, peptidoglycan hydrolase, glycosidases, dextran, mannose, lactose, and galactose, were successfully annotated in B. subtilis A-5, which was also beneficial for the biocontrol mechanism analysis. In addition, genes encoding pyruvate kinase (pyk, A-5_gene 3131), glutamate synthesis complex enzyme (proABC, A-5_gene 2206-2208), catalase (katE, A-5_gene 1035/1484/4340/4383), and stress-resistant protein genes were also identified in the B. subtilis A-5 genome. The existence of these genes is suggested to be one of the underlying mechanisms of the salinity and alkalinity tolerance of B. subtilis A-5.

Evolution and Comparative Genome Analysis of B. subtilis A-5
Evolutionary analysis of 30 housekeeping genes and the phylogenetic tree subsequently constructed ( Figure S2) verified that strain A-5 belongs to Bacillus subtilis. B. subtilis A-5 and KH2, γ-PGA-producing strains, contained two and one plasmids, in addition to the chromosome, respectively, while B. subtilis 168 and XF-1 consisted of only one chromosome (Table 4). Strain A-5 had the highest number of CDSs in the genome (4605). The genome of XF-1 was the smallest (4.06 Mb), but its GC content was the highest (43.90%) among the four strains (B. subtilis A-5, 168, KH2, and XF-1). In addition, collinearity analysis of the genome of strain A-5 was performed by comparing with that of B. subtilis 168, B. subtilis KH2 [20], and B. subtilis XF-1 [21]. High similarity was observed in the genomes of the four B. subtilis strains (Figure 8). However, they significantly differed in terms of sequence insertion, turnover, translocation, and deletion.    In B. subtilis A-5, pgsBCA and the required factor (pgsE) were essential for γ-PGA synthesis (Figure 9), which functioned as a synthase complex. The transmembrane and subcellular localization of related proteins were predicted by TMHMM and PSORTb. pgsB (gene 4031) possessed a full length of 1182 bp nucleotides encoding 393 bp of amino acids (aa), which could catalyze γ-PGA synthesis intracellularly as a non-transmembrane protein by anchoring to the plasma membrane through an anchor hook. pgsC (gene 4030) contained 450 bp nucleotides encoding 149 aa; pgsC was predicated to have four transmembrane helixes and to be fixed on the plasma membrane to connect pgsB and pgsA. Protein pgsA (gene 4029) was predicted to contain one transmembrane helix in the 380 aa and to be anchored to the plasma membrane through the N-terminal transmembrane region and anchor hooks. Therefore, pgsA is also responsible for the extracellular transport of γ-PGA. pgsE (gene 4028) was annotated to encode the factor required for γ-PGA synthesis and one transmembrane helix was predicted in the membrane with 27-55 aa.

Characteristics of Genes Related to γ-PGA Synthesis
In B. subtilis A-5, pgsBCA and the required factor (pgsE) were essential for γ-PGA synthesis (Figure 9), which functioned as a synthase complex. The transmembrane and subcellular localization of related proteins were predicted by TMHMM and PSORTb. pgsB (gene 4031) possessed a full length of 1182 bp nucleotides encoding 393 bp of amino acids (aa), which could catalyze γ-PGA synthesis intracellularly as a non-transmembrane protein by anchoring to the plasma membrane through an anchor hook. pgsC (gene 4030) contained 450 bp nucleotides encoding 149 aa; pgsC was predicated to have four transmembrane helixes and to be fixed on the plasma membrane to connect pgsB and pgsA. Protein pgsA (gene 4029) was predicted to contain one transmembrane helix in the 380 aa and to be anchored to the plasma membrane through the N-terminal transmembrane region and anchor hooks. Therefore, pgsA is also responsible for the extracellular transport of γ-PGA. pgsE (gene 4028) was annotated to encode the factor required for γ-PGA synthesis and one transmembrane helix was predicted in the membrane with 27-55 aa.

Analysis of the Secondary Metabolite Gene Clusters
Secondary metabolites synthesized by Bacillus have good antimicrobial properties, which can directly or indirectly stimulate plant systemic resistance and avoid pathogenic microbes [21]. In this study, nine secondary metabolite gene clusters were predicted and located in the Chr of B. subtilis A-5 by antiSMASH analysis. Surfactin, fengycin, bacillibactin, subtilosin_A, and bacilysin were annotated and are listed in Table 6. Four gene clusters (Cluster2, 4, 5, and 9) were predicted with unknown functions, including two terpenes, one T3PKS and one bacteriocin, suggesting that B. subtilis A-5 might synthesize new antimicrobial substances. Strain A-5 may have the potential for development and application in the agriculture and bio-pharmaceutical industries. All four strains (B. subtilis A-5, 168, KH2, and XF-1) contained antimicrobial gene clusters (e.g., surfactin, fengycin, bacillibactin, subtilosin, and bacilysin). Fourteen secondary antimicrobial gene clusters were annotated in B. subtilis 168; among them, bacillaene, sublancin 168, and thailanstatin A were only identified in strain 168. Note: NRPS, non-ribosomal peptide synthetase; T3PKS, type III polyketide synthase. "-" refers to no annotation.

Discussion
Excessive application of chemical fertilizers and pesticides causes serious environmental pollution [10]. Green agricultural input, characterized by their safety, eco-friendliness, and high efficiency, has become a research hotspot. Among them, multifunctional microorganisms have gained much attention because of their synergy with fertilizers and biocontrol [5,6]. Bacillus has been widely studied in terms of its roles in plant growth stimulant production, biological control, pollutant degradation, and the synthesis of degradable biomaterials [25].
In this study, the molecular weight of γ-PGA changed dynamically when fermented in the FF medium. The maximum Mw was 4700 kDa, which then gradually decreased to 1106 kDa with a cultivation time of 96 h ( Table 1). During polymerization, glutamic acid was gradually added to the tail of γ-PGA n-1 for longer γ-PGA synthesis, thus leading to viscosity increase and dissolved oxygen decrease in the fermentation system. Continuous consumption of nutrients in the culture medium and the presence of γ-PGA degradationrelated enzymes (pgdS, GGT, and cwlO) jointly affected the ultimate yield and molecular weight of γ-PGA [26]. Additionally, B. subtilis A-5 inhibited the mycelial growth of M. grisea, R. solani, and F. oxysporum (6.66-77.93%); it also conferred a high tolerance of salinity and alkalinity (NaCl 1-10% and pH 5.5-9.5) (Figures 1-3). Therefore, it can be hypothesized that B. subtilis A-5 has potential as a valuable bio-organic fertilizer.
The functional genes related to γ-PGA metabolism, pathogen control, and salt/alkaline stress resistance have been successfully explored in the whole genome of B. subtilis A-5. Annotation by the COG database showed that, with the exception of the unknown functions, amino acid transport and metabolism were the main functions, which explained the production and secretion of γ-PGA and secondary substances in B. subtilis A-5. Similarly, in the genome of B. licheniformis WX-02, a γ-PGA producer, the protein functions were significantly enriched in unknown functions, carbohydrate transport/metabolism, and replication/recombination/repair [27]. The main functions annotated by KEGG were carbohydrate metabolism, global and overview maps, and amino acid metabolism, which was similar to the metabolic pathways analyzed in the γ-PGA-synthesizing strain B. amyloliquefaciens LL3 [28]. The proteins related to carbohydrate and protein decomposition, proline synthesis, and stress resistance in B. subtilis A-5 may be the contributory factors to biocontrol and stress resistance. For example, proline has strong hydration ability to maintain osmotic balance, prevent cell deformation, and remove reactive oxygen species; Bacillus can accumulate proline by de novo synthesis to cope with osmotic changes and ensure survival and growth [29]. The dominating functions annotated in COG database were predicted to be "FUNCTION UNKNOWN" in pA and pB, suggesting that the unknown functional genes in B. subtilis A-5 may participate in important biological processes [30] and are thus worthy of further research.
The γ-PGA synthase complex was composed of pgsBCA and factor (pgsE), where pgsB was the catalytic component and could provide ATP for catalysis. pgsC formed a catalytic site with pgsB and then polymerized the substrate (glutamate acid). pgsA, located on the cell surface with a membrane-anchoring region, assisted γ-PGA transportation and played an important role in γ-PGA prolongation [31]. The pgsBCA complexes are different to a certain degree in the glutamate-dependent/independent strains, especially pgsA [32], since it is vital for γ-PGA yield and molecular weight. Enzyme cwlO was the only peptidoglycan hydrolase with extracellular activity and was not regulated by γ-PGA degradation system [23]. In theory, knockout of pgdS, GGT, or cwlO may shut down the γ-PGA degradation pathway, thus facilitating the expression of γ-PGA synthase cluster and product accumulation. For example, after knockout of the degradation enzymes, γ-PGA production of strain NK-E11 significantly increased from 4.15 g/L to 9.18 g/L in a previous study [33]. However, in contrast, deletion of pgdS and ggt inhibited and increased γ-PGA production, respectively, in B. licheniformis RK14-46 [24]. Feng et al. [22] found that the γ-PGA yield decreased to only 2.69 g/L when pgdS, GGT, and cwlO were synchronously knocked out. Generally, these results indicate the different intrinsic interactions of pgdS, GGT, and cwlO in γ-PGA production and decomposition. Metal ions, such as Fe 3+ , K + , and Mn 2+ have also been reported to influence synthesis and accumulation of γ-PGA [34,35]. The yield and molecular weight distribution of γ-PGA can be regulated by enhancing pgsBCA expression, knocking out γ-PGA-degrading enzymes, and regulating the content of elements in fermentation system.
Whole genome analysis revealed that B. subtilis A-5 also possessed the genes encoding secondary antimicrobial substances (Table 6). Surfactin is a cyclic lipopeptide biosurfactant synthesized by the non-ribosomal peptide synthetase (NRPS) pathway, which can reduce the surface tension of water from 72 mN/m to 27 mN/m at 20 µm. It has various biological functions, such as hemolysis, antibiosis, anti-mycoplasma/chlamydia, antiviral action, and tumor cell proliferation inhibition [36]. Fengycin is another cyclic lipopeptide antibiotic produced by Bacillus, which consists of 10 aa residues and 1 β-hydroxy fatty acid. It exerts antibacterial effects by disrupting the fungal cell membrane and damaging its structure [37]. Chung et al. [38] found that fengycin inhibited Staphylococcus aureus colonization by inhibiting quorum sensing. Bacillibactin is a kind of catechol siderophore with high affinity to competitively bind to the soluble irons necessary for the growth and activity of pathogens. Subtilosin_A is another uncommon lantibiotic with a special structure composed of 35 aa, which has inhibitory activity to both gram-positive and gram-negative bacteria. Bacilysin, a peptide antibacterial substance, belongs to the ribosomal synthesis pathway. It has a wide spectrum against bacteria and strong inhibitory effects on some fungi (e.g., Candida albicans) [39]. Furthermore, alkaline serine protease (gene 2005) was found in B. subtilis A-5 with a similarity rate of 99.4% (data not shown), which may reduce root-knot nematodes [40]. Therefore, strain A-5 may achieve antimicrobial effects through the combined actions of lipopeptide antibiotics, functional proteins, and competition effects. As for the internal relationship between γ-PGA metabolism, biocontrol efficiency, and stress tolerance, Wang et al. [41] concluded that pgsB deletion impaired the bacterial colonizing capacity and lowered the biocontrol rate from 62.08% to 14.22% against Fusarium root rot. They also found that knockout of pgdS and ggt led to a considerable improvement in biocontrol efficiency. Furthermore, the expression level of stress-related genes was notably changed by pgsB deletion [41], demonstrating the importance of the interconnectedness of γ-PGA production, biocontrol, and stress tolerance. Based on the linkage of amide bonds between α-NH 2 and γ-COOH groups in the γ-PGA backbone and polyanionic characteristics, it may adsorb cations to form dissolved complexes for slow release. Thus, the presence of γ-PGA may facilitate to alleviate salt/alkaline stress [42]. In the future, the mechanism of the joint effect should be studied in-depth and a field-scale investigation and verification need to be carried out.

Conclusions
The multifunctional bacterium B. subtilis A-5 can produce γ-PGA (maximum Mw: 4700 kDa), antagonize plant fungal pathogens (6.66-77.93%), and tolerate salinity and alkalinity stress (NaCl 1-10% and pH 5.5-9.5). The genome size of B. subtilis A-5 was 4,190,775 bp, containing 1 Chr and 2 plasmids (pA and pB) with 43.37% GC content and 4605 CDSs. The γ-PGA synthase gene cluster, pgsBCA and factor (pgsE), were annotated and the enzyme complex was predicted to work in the cell membrane. γ-PGA-degrading related enzymes were mainly pgdS, GGT, and cwlO, located in the cell wall or extracellular matrix. Nine secondary metabolite gene clusters, including four unknown function gene clusters and five antibiotic synthesis gene clusters (surfactin, fengycin, bacillibactin, subtilosin_A, and bacilysin), were predicted in the genome of B. subtilis A-5. In addition, B. subtilis A-5 contained genes related to carbohydrate and protein decomposition, proline synthesis, pyruvate kinase, and stress-resistant proteins, which could facilitate biocontrol and stress resistance. In conclusion, B. subtilis A-5 has great potential for application in sustainable agriculture systems. This study provides theoretical support to the investigation of the underlying metabolic mechanism of B. subtilis A-5 and impetus to construct engineered strains. In the future, the interior mechanism of the different effects of this strain should be studied and field-scale application should be carried out.
Supplementary Materials: The following supporting information can be downloaded at: https:// www.mdpi.com/article/10.3390/ijerph19137630/s1, Table S1: Analysis of the drug resistance genes in B. subtilis A-5; Figure S1: Analysis of the virulence factors in B. subtilis A-5 using VFDB database; Figure S2: Phylogenetic tree of 30 house-keeping genes using the neighbor-joining method.