The strain P. eryngii is a strain cultivated by Lvshengyuan Biotechnology Co., Ltd (Zhangzhou, Fujian Province, China). The strain was maintained on potato dextrose agar medium (potato 200 g, glucose 20 g, peptone 3 g, yeast powder 3 g, KH₂PO₄ 1.5 g, MgSO₄⋅7H₂O 1.5 g, agar 20 g, H₂O 1 L) at 25°C. The bacteria were cultivated in LB medium (yeast powder 5 g, beef extract 5 g, peptone 10 g, NaCl 10 g, agar 20 g, pH 7.2∼7.4). All these media were prepared by homemade raw materials.

Cultivation bags of P. eryngii were collected from Lvshengyuan Biotechnology Co., Ltd (Zhangzhou, Fujian Province, China). The growing substrate formulation (all ingredients based on dry substrate weight, w/w) consisted of 25% sawdust, 27% corncob powder, 13% bagasse, 15% bran, 10% bean pulp, 8% corn starch, 1% light calcium carbonate, and 1% calclime. All mixed substrates were sterilized with high-pressure steam under constant monitoring for 2.5 h at 135^°C. Cultivating substrates at different hyphal growth stages of P. eryngii were sampled from the cultivation bags. Three different growth stages are PEBH (the half-bag of P. eryngii hypha stage), PEBF (the full-bag of P. eryngii hypha stage), and PEBM (the post-ripe stage) because these three different growth stages are extremely representative in the life history of P. eryngii. At these three different growth stages, a lot of MIAB in the cultivation bags can be detected and classified systematically and the diverse functions of different bacterial communities can be investigated and clarified in detail. Most of all, the mushroom growth promoting bacteria (MGPB) can be isolated and the growth promotion effects can be systematically analyzed at these three different growth stages. During the experiment, samples (5 g each) were collected from the top, middle, and bottom of each bag using aseptic techniques to make the samples representative of the microorganism population. Sampling was repeated in three independent bags. Three samples from the same part of three independent bags were immediately mixed and frozen by liquid nitrogen and then used for DNA extraction and microbial diversity analysis.

The 16S rDNA library preparations and Illumina HiSeq 2500 sequencing were performed at Biomaker, Inc. (Beijing, China). According to the manufacturer’s instruction, DNA was extracted from 0.4 g of samples from cultivation bags using EasyPure Genomic DNA Kit (TransGen, Beijing, China). The bacterial 16S rDNA gene was amplified with the forward primer 338 F (5′-ACTCCTACGGGAGGCAGCA-3′) and the reverse primer 806 R (5′-GGACTACHVGGGTWTCTAAT-3) targeting the V3 and V4 regions (Sun et al., 2020a). When amplifying the bacterial 16S rDNA gene, PCR thermal cycle profile was as follows: 2 min at 98°C, 30 cycles of 30 s at 98°C, 30 s at 50°C, and 1 min at 72°C, a final extension at 72°C for 5 min, and hold 4°C.

Amplicons were sequenced using a paired-end method by Illumina HiSeq. Raw data generated from the high-throughput sequencing run were joined using the FLASH v1.2.11, filtered using the Trimmomatic software (V 0.33), and removed the chimera sequences using the UCHIME software (V8.1). Only the effective tags classified into OTUs will be pre-clustered to 97% sequence identity using the software QIIME v1.8.0 (Morey et al., 2013).

The OTUs were assigned into taxonomic categories using the ribosomal database program (RDP) with a confidence threshold of 0.8 and predicted to the kingdom, phylum, class, order, family, genus, and species level using the Silva database. The alpha diversity index including ACE value, Chao1, Simpson index, and Good’s coverage were calculated using “vegan” R package. One-way analysis of variance (ANOVA) and subsequent post hoc Tukey’s honestly significant difference (HSD) tests were used to compare differences in OTU richness among different periods. Chord diagram and heatmap were plotted to investigate abundant taxa at phylum and genus levels using “circlize” and “pheatmap” R packages, respectively. Then, principal coordinate analysis (PCoA) based on Bray–Curtis distance was performed to investigate the patterns of the bacterial community structure. In addition, analysis of composition of microbiomes (ANCOM) test (Mandal et al., 2015; Kaul et al., 2017) were used to explore the significant differences in the relative abundance at the genus level among the different groups. We generated a table grouping the sequences taxonomically assigned to the same genus according to the report (Ramirez et al., 2021) before the implementation of ANCOM.

Pleurotus eryngii hyphae were sampled from cultivation bags, washed by sterilized water, and ground in a sterile environment. Serial dilutions of the hyphae were prepared with sterilized pure water. A total of 0.1 mL of the dilutions was spread on LB agar media plates. The samples were incubated at 35°C for 48 h to enumerate and isolate the microbic colonies. The bacterial 16S rRNA gene was amplified with the forward primer 16S-F (AGAGTTTGATCCTGGCTCAG) and the reverse primer 16S-R (GGTTACCTTGTTACGACTT) (Sun et al., 2020a).

The culture of isolated bacteria was grown in LB liquid medium and incubated on an incubator shaker at 37°C and 180 r⋅min^–1 until the OD₆₀₀ reached 0.6∼0.7 after dilution two-fold. The culture then was centrifuged (at 10 000 r⋅min^–1 and 4°C for 20 min) and filtered (Φ = 0.45 μm) to obtain a cell-free fermentation broth.

The modified media were prepared by mixing different volumes of cell-free fermentation broth into sterilized PDA enrichment media. Then the media poured into a petri dish. The additive amounts of culture and fermentation broth were 1, 3, 5, 7, 9, and 11% (v/v). Discs of P. eryngii hyphae (Φ = 6.0 mm) were inoculated in the middle of modified media and grew at 25°C. The diameters of P. eryngii hyphae and their growth rates were measured when they grew over the entire plate (Zhao et al., 2007). The average of three diameter measurements was used to estimate mycelial growth rate, and the mycelial growth rate was expressed as a mean diameter of the increase in mycelia length. The measurement and calculation were repeated in triplicate. When submerged fermentation was carried out, the same additives of cell-free fermentation broth were added to sterilized liquid PDA enrichment media. Seven mycelial agar discs were cut from P. eryngii hyphae (Φ = 5.0 mm) into each flask. The strain was cultivated in a rotary shaker at 120 r⋅min^–1, at 25°C, and was sampled each day for 7 days. The density of P. eryngii hyphae pellets and activity of laccase and amylase in the fermentation broth was measured as follows: 4.0 mL of P. eryngii mycelial broth was transferred to a glass Petri dish, where the number of mycelial pellets was counted. The experiment was repeated three times. The average of the three replicates was used to represent the density of P. eryngii hyphae pellets. According to the method established by Sun et al. (2020b), laccase activity was determined by monitoring the rate of 5 mM 2,2′-azino-bis(3-ethylbenzthiazoline-6-sulfonic acid) (ABTS) oxidizing to its cation radical (ABTS⁺) at 436 nm (ε₄₃₆ = 29,300 M^–1 cm^–1) in 0.1 M sodium acetate buffer (pH 5.0) at 30°C. Briefly, the 3 mL reaction system consisted of 2.4 mL of HAc-NaAc buffer solution, 0.2 mL of ABTS solution, and 0.4 mL of crude enzyme. The mixture was preheated in a 30°C water bath for 5 min, and the changes in absorbance at 436 nm for the first 3 min were recorded after zeroing the spectrophotometer. One unit (U) of enzyme activity was defined as the amount of enzyme required to oxidize 1 μmol of ABTS per minute at 30°C.

The amylase activity assay was conducted using the 3,5-dinitrosalicylic acid (DNS) method (Miller, 1959). Briefly, 1.0 mL of sample and an equal amount of substrate (1.0% w/v soluble starch) were mixed thoroughly, and test tubes were incubated at 37°C for 15 min in a water bath. After 10 min, the reaction was stopped by adding 2.0 mL of DNS reagent, and tubes were kept in a boiling water bath for 5 min. Tubes were cooled at room temperature, and absorbance was measured at 540 nm against substrate and enzyme blank. One unit (U) of amylase activity was defined as the amount of enzyme that releases 1 μmol of reducing sugar as D-glucose per min under the assay conditions. Extracellular protein concentration was estimated using the Lowry method and bovine serum albumin as the standard.

All experiments were performed in triplicate to ensure the precision of the results. Data analysis was performed with SPSS software. The difference between samples was estimated using a one-way analysis of variance (ANOVA) to test for the significant differences between mean ± SD of treatments (n = 3) at p < 0.01 significance level.

In this study, Illumina HiSeq 2500 was used for its incredible speed and throughput, unprecedented flexibility, exceptional data quality, and complete end-to-end sequencing solutions. A total of 457,104 pairs of reads were generated from the high-throughput sequencing runs. After quality control, an average of 42,374 effective tags was recovered from nine samples (three samples per stage). We removed OTUs that do not belong to bacterial taxa and have fewer than 0.005% relative abundance. Then, we rarefied each sample to the minimum size (17,948), resulting in a normalized dataset comprising 1,240 bacterial OTUs. As shown in Figure 1A, the average OTUs numbers varied at different growth periods of P. eryngii hyphae. The maximum OTUs were generated at the half-bag stage of the P. eryngii hyphae (PEBH), which was significantly more than OTUs of PEBF (the full-bag stage) and PEBM (the post-ripe stage). PEBM has more OTUs numbers than PEBF, but not statistical significance. It is interesting to note that 111 OTUs were shared among three growth stages of P. eryngii (Figure 1B). When the growth shifted from the half-bag to the full-bag stage, 218 OTUs were preserved and shared among groups, and 359 new OTUs were generated. When the growth shifted from the full-bag stage to the post-ripe stage, the numbers of shared OTUs and newly generated OTUs were changed to 169 and 237, respectively. These results indicated that the cultivation bags have different bacterial communities coexisting with P. eryngii hyphae at various growth periods.

The overall perception of “microecologics” has been around. Currently, edible fungi can no longer be considered an independent population because most evidence pointed out the symbiosis between bacteria and mushrooms (Li et al., 2015; Siyoum et al., 2016; Gong et al., 2018; Benucci et al., 2019). It is well known that the edible mushrooms displaying above-ground fruit bodies and under-ground hyphae assets can be colonized with diverse bacteria due to the biodiversity of air and soil (Rousk et al., 2010; Zagriadskaia et al., 2013; Deveau, 2016). However, little research has been done on interacting bacteria in cultivation bags full of culture substrates of P. eryngii. Therefore, the high-throughput sequencing technologies were used to better understand the whole micro-ecosystem of the cultivation bags during growth periods of P. eryngii hyphae. Although only 3 samples per stage were used in the high-throughput sequencing step, we still got pretty good results because all four reversible terminator-bound dNTPs are present in each sequencing cycle and this feature minimizes bias of incorporation and greatly reduces raw error rates in HiSeq 2500 System.

The average coverage of bacteria in the samples from cultivation bags was between 0.9974 and 0.9997 (Table 1), indicating that the sequencing depth was sufficient and the sampling was reasonable. The results could reflect the bacterial community composition in the cultivation bags. The microbial species richness was estimated and compared by Chao1, ACE, and Simpson indices. The order of the ACE and Chao indices at the half-bag period (PEBH) was significantly higher than those at the full-bag period (PEBF) and post-ripe period (PEBM), and PEBH showed the lowest Simpson value among the three periods (Table 1). These results indicated that the bacterial diversity at the half-bag period of P. eryngii hyphae was higher than those at the other periods, suggesting that the bacterial diversity in the samples was closely related with the growth process of P. eryngii hyphae.

Our previous study demonstrated a correlation between the community structure of interacting bacteria and the growth and development of edible fungi (Sun et al., 2020a). On this basis, a novel strategy was developed to improve the industrial production of edible fungi using interacting microbial resources. Therefore, the detailed analysis of the microbiome of P. eryngii in production bags will provide an important theoretical basis for a deeper understanding of the existence of the abundant interacting bacteria and the knowledge of interactions of the fungus with other microorganisms. And on top of that, the fungus production technology can be improved by regulating microbial communities after confirming their growth-promoting effects. In the present study, the bacterial community in the samples from cultivation bags was examined by RDP. The 1,240 OTUs generated from the samples were classified into 37 phyla, 74 classes, 151 orders, 228 families, and 410 genera. As shown in Figure 2A, the top 10 abundant phyla were Proteobacteria (34.65∼46.28%), Firmicutes (4.58∼22.91%), Epsilonbacteraeota (0.46∼44.30%), Actinobacteria (2.52∼19.66%), Bacteroidetes (3.50∼8.81%), Tenericutes (0.10∼3.91%), Acidobacteria (0.48∼2.28%), Patescibacteria (0.23∼2.15%), Chloroflexi (0.44∼1.63%) and Synergistetes (0.09∼1.15%), accounting for 96.51% of the total bacteria in three samples. Except for the two phyla Firmicutes and Actinobacteria, other dominant bacteria phyla included Gram-negative bacteria. Throughout the three periods of hyphal growth, phylum Proteobacteria was the second and the first dominant phyla at PEBH and the latter two growth periods of P. eryngii hyphae, respectively, mainly including classes Alphaproteobacteria, Gammaproteobacteria and Deltaproteobacteria. The class Alphaproteobacteria was significantly (P < 0.01) increased from 2.61 to 12.86% when the growth of P. eryngii hyphae switched from PEBH period to PEBF period. However, Campylobacteria, the dominant class in the phylum Epsilonbacteraeota was significantly (P < 0.001) decreased from 44.39% to 1.39%. Besides, Firmicutes, Actinobacteria, and Acidobacteria levels were significantly increased up to 5.00, 5.13, and 4.76-folds from PEBH to PEBF. The fourth-highest abundant phyla Actinobacteria was significantly increased from 12.96% in the PEBF sample to 19.67% in the PEBM sample.

A total of 410 genera were identified across the samples, which revealed that bacteria in the cultivation bags were plentiful and varied. The dominant genera (relative abundances > 1%) in the samples was shown in Figure 2B. Those 14 abundant genera were shared by the samples at the three growth periods. Arcobacter was the most abundant genus, whose relative abundance was harply decreased to less than 0.15% at PEBM, from 40.31% at PEBH. Four genera (Pseudomonas, Escherichia shigella, Acinetobacter and Bacteroide) were dominant in PEBF samples with a relative abundance higher than 3.0%. Whereas, the other five genera (Acinetobacter, Ochrobactrum, Escherichia-Shigella, Corynebacterium, and Lactobacillus) were dominant in PEBM samples. The relative abundance of some genera such as Acinetobacter, Ochrobactrum, and Corynebacterium was significantly increased up to 2.6∼3.1 folds from PEBF to PEBM. Additionally, ANCOM results showed that some abundant genera including Acinetobacter, Escherichia-Shigella, Corynebacterium, Achromobacter, Cutibacterium, and Brevibacterium were also the important differential generic indicators structuring the bacterial composition among the different periods (Table 2).

Heatmap (Figure 3) showed that the bacteria genera differed in the cultivation bags collected at different growth periods of the P. eryngii hyphae. The relative abundances of nine bacteria genera in P. eryngii hyphae bags, including Sulfurimonas, Gluconobacter, Desulfuromonas, Actinobacillus, Fibrisoma, Gaiella, Sphingopyxis, Acholeplasma, and Ameyamaea, were decreased sharply or even disappeared. Figure 3 also showed a high degree of similarity between the PEBF and PEBM samples. For example, 11 genera including Oceanobacter, Denitrovibrio, Acholeplasma, Thiovirga, Thiomicrorhabdus, Sulfurovum, Desulfuromonas, Thiomicrospira, Sulfurimonas, Arcobacter, and Chrysiogenes were obviously more abundant in PEBH than other two periods (Figure 3).

Regarding prokaryotic communities, Pseudomonas, Acinetobacter, Escherichia-Shigella, Ochrobactrum, and Lactobacillus were dominant in the microbiome of P. eryngii hyphae bags. Still, the bacterial communities varied at different growth periods, and the relative abundance of Pseudomonas (Proteobacteria) was raised during the PEBF period. The genus Pseudomonas is known as “fungiphills” (Warmink et al., 2009; Zagriadskaia et al., 2013). Warmink et al. (2009) demonstrated that the selection of fungiphills depends on their ability to utilize organic substrates from the fungi exudates. Moreover, the growth of the pathogenic genus Arcobacter (Epsilonbacteraeota) could be inhibited by organic acids, such as citric acid and lactic acid (Phillips, 1999). At PEBF, the acid-forming bacteria, such as the genera Pseudomonas, Acinetobacter, Escherichia-Shigella, Bacteroides, Ochrobactrum, Lactobacillus, and so on, increased significantly and became the dominant genera. As a result, the genus Arcobacter with a significant abundance (40.31%) at PEBH was almost disappeared during the PEBF period. Besides, the identified species of Pseudomonas, including Pseudomonas putida, which is a growth-promoting inoculant for A. bisporus (Zarenejad et al., 2012; Siyoum et al., 2016), may also contribute to promoting the growth of P. eryngii hyphae.

After the growth of P. eryngii hyphae to fill the cultivation bags, Proteobacteria (41.85%) was still one of the dominant phyla. In comparison, Firmicutes (22.91%), Actinobacteria (12.96%), and Bacteroidetes (8.81%) replaced Epsilonbacteraeota (44.30%) and became the dominant phyla. Bacteria communities in P. eryngii hyphae bags are different from those in the soil beneath other edible fungi. Actinobacteria, Chloroflexi, and Proteobacteria were found to comprise the core microbiome from the soils beneath Morchella sextelata (Benucci et al., 2019). Three bacteria phyla, Proteobacteria, Acidobacteria, and Actinobacteria, were dominant in the soil samples from the forestland where Stropharia rugosoannulata (wine-cap mushroom) (Gong et al., 2018) and other 16 species of mushrooms were cultivated (Pent et al., 2017). Bacterial communities found in the casing layer of A. bisporus were Proteobacteria, Firmicutes, Bacteroidetes, and Actinobacteria (Siyoum et al., 2016), similar to phyla found in cultivation bags of P. eryngii hyphae. The abundance of Proteobacteria was stimulated by the high nutritional status of soil, while Acidobacteria preferred low nutritional status (Siyoum et al., 2016; Pent et al., 2017; Gong et al., 2018). Although most of the bacterial phyla observed in cultivation bags of P. eryngii hyphae and casing layer were very similar, their relative abundances in genera differed significantly between these two habitats. The bacterial population varied even in the casing layers of A. bisporus from different mushroom farms (Siyoum et al., 2016; Vieira and Pecchia, 2018; Li et al., 2019). But a general agreement of the distribution of functional categories and different bacterial communities among different composts has also been confirmed (Martins et al., 2013).

The beta diversity was analyzed by PCoA, and the percent variability for each principal component is PCoA1: 56% and PCoA2: 18%. As shown in Figure 4, ordinations based on Bray-Curtis metric demonstrated a clear separation among the three periods. Thereinto, the distance between the samples PEBH and PEBF was longer than that between the samples of PEBF and PEBM, indicating that the bacterial communities in the PEBF samples were more similar to those at the period of PEBM.

The well-grown bacteria were isolated from P. eryngii hyphae and identified by 16S rRNA sequencing and physiological and biochemical indexes to explore promoting effects of the eight most dominating bacterial genera on the growth of P. eryngii hyphae. The results showed that Bacillus species became the dominant microbiota, and the isolated strain belonged to the Bacillus cereus group with a similarity of 100%. The strain was labeled as Bacillus cereus Bac 1 and then was subjected to submerged fermentation.

As shown in Figures 5, 6, the different volumes of B. cereus Bac 1 fermentation broth could promote the growth of P. eryngii hyphae. The mycelia growth rate was significantly faster in the modified media containing the 5-mL fermentation broth of B. cereus Bac1 than the control ones (P < 0.01). The mycelia growth rate reached 0.46 mm/h, which is 1.15-fold higher than the corresponding control. The mycelia cultured in the modified media were stronger and thicker than the control ones. The results showed that the addition of B. cereus Bac1 fermentation broth to culture media was beneficial for the growth of P. eryngii hyphae. The composition analysis of B. cereus Bac1 fermentation broth performed by UHPLC-QTOF-MS technology showed that the extracellular metabolites were mainly classified into antagonistic substances, promoting materials and cyclopeptides, including carboxylic acids and derivatives, organic oxides, benzene, and substituted derivatives, hydrazine and its derivatives, pyrimidine nucleotides, phenols, imidazopyridines, and fatty acyl. These compounds are associated with the growth and development of P. eryngii.

Interestingly, the cyclopeptides in the cell-free fermentation broth of B. cereus Bac1 promoted the mycelium development of P. eryngii as supplemental nitrogen. In addition, the indole acetic acid in the fermentation broth of B. cereus could also increase the mycelium growth of P. eryngii. However, when a volume of 7–11 ml was used, the inhibiting effects of antagonistic substances became prominent, which decreased the positive effect of the B. cereus cell-free broth on the growth of P. eryngii.

The effects of B. cereus Bac 1 fermentation broth on the mycelial pellet density of P. eryngii were analyzed. The results indicated that the mycelial pellet density was increased in the modified media containing less than 9-mL fermentation broth (Table 3). The mycelial pellet density is expressed as the number of mycelial pellets (NMP) per milliliter. The mycelial pellet density reached 44.33 NMP/mL in the modified media containing 5-mL fermentation broth of B. cereus Bac1, significantly higher than in the control groups (P < 0.01). The higher pellet density indicates that the mycelia grow well with more growing points. The addition of 5-mL B. cereus Bac1 fermentation broth to culture media could effectively promote the growth of P. eryngii hyphae.

According to our former research (Sun et al., 2011), we found that the activity of laccase and amylase in edible fungi and their growing cycles were closely related. The edible fungi strains with short growing cycles originated from their high secreted laccase and amylase activity levels. Therefore, the amylase and laccase produced by P. eryngii were investigated to demonstrate the effects of B. cereus Bac 1 fermentation broth on the enzyme activity. As shown in Figure 7, the enzymatic activities of laccase and amylase were significantly increased by adding Bac 1 fermentation broth. The laccase activity reached a maximum of 144.17 U/mL, 1.27-fold higher than that of the corresponding control. Compared to the control group (113.61 U/mL), The amylase activity was enhanced by 43.83% and reached 38.16 U/mL. These results demonstrated that the Bac1 fermentation broth could significantly increase two extracellular enzymatic activities and further promote the mycelial growth of P. eryngii. In general, decreasing the carbon-to-nitrogen ratio in the growth substrate by adding nitrogen sources can promote fungi to produce ligninolytic enzymes, especially laccase (Brezani et al., 2019). The composition analysis of B. cereus Bac1 fermentation broth showed that the extracellular metabolites were mainly classified into antagonistic substances, promoting materials, and cyclopeptides. The cyclopeptides in the cell-free fermentation broth of B. cereus Bac1 increased nitrogen supply, thus decreased carbon-to-nitrogen ratio in the growth substrate, which stimulated two extracellular enzymatic secretions and activity.

Although fungi have coexisted and interacted with bacteria since the earliest stages of fungal evolution, we still understand relatively little about their interactions (Caporaso et al., 2010). In this study, a Bacillus cereus strain was isolated from the cultivation bags of P. eryngii. B. cereus strain was reported to improve the heavy metal stress tolerance of its hosts (Azcón, 2010), promote growth (Wu et al., 2012), and inhibit pathogenic bacteria and fungi, such as Rhizoctonia solani (Pleban et al., 1997), Fusarium oxysporum, Fusarium solani, and Pythium ultimum (Chang et al., 2007). The results showed that the Bac1 fermentation broth could promote the mycelial growth and extracellular enzyme secretion of P. eryngii, demonstrating the mutualistic symbiosis relationship of the fungal host P. eryngii and microorganisms during cultivation of P. eryngii (Caporaso et al., 2010).