Effect of arbuscular mycorrhizal fungi (AMF) and plant growth-promoting rhizobacteria (PGPR) on microbial community composition of phenanthrene and pyrene contaminated soils using Illumina HiSeq sequencing

Abstract

In order to determine the influence of arbuscular mycorrhizal fungi (AMF, Glomus versiforme) and plant growth-promoting rhizobacteria (PGPR, Pseudomonas fluorescens, Ps2-6) on degradation of phenanthrene (PHE) and pyrene (PYR) and the change of microbial community composition in soils planted with tall fescue (Festuca elata), four treatments were set up in phenanthrene (PHE) and pyrene (PYR) contaminated soil: tall fescue (CK), AMF + tall fescue (GV), PGPR + tall fescue (PS), and AMF + PGPR + tall fescue (GVPS). Our results showed that the highest removal percentage of PHE and PYR in contaminated soil as well as biomass of tall fescue were observed in GVPS. PHE and PYR accumulation by tall fescue roots were higher than shoots, the mycorrhizal status was best manifested in the roots of tall fescue inoculated with GVPS, and GVPS significantly increased the number of PGPR proliferation in tall fescue rhizosphere soil. Paired-end Illumina HiSeq analysis of the 16S rRNA gene and Internal Transcribed Spacer (ITS) region sequences was used to assess changes in the bacterial and fungal communities composition of the four treatments. GVPS positively affected the species and abundance of bacteria and fungi in PHE and PYR contaminated soil, an average of 71,144 high quality bacterial 16S rDNA tags and 102,455 ITS tags were obtained in GVPS, and all of them were assigned to 6,327 and 825 operational taxonomic units (OTUs) at a 97% similarity, respectively. Sequence analysis revealed that Proteobacteria was the dominant bacterial phylum, Ascomycota was the dominant fungal phylum in all treatments, whereas Proteobacteria and Glomeromycota were the most prevalent bacterial and fungal phyla in GVPS, respectively. And in the generic level, Planctomyces and Meyerozyma were the richest bacterial and fungal genus in all treatments, respectively, whereas Sphingomonas and Fusarium were the dominant bacterial and fungi genus in GVPS. Overall, our findings revealed that application of Glomus versiforme and Pseudomonas fluorescens, Ps2-6 had an effective role in improving the growth characteristics, root infection of F. elata and soil microbial community composition in PHE and PYR contaminated soils, but no obvious in degradation efficiencies of PHE and PYR as compared to the control.

1 Introduction

Soil contaminated with polynuclear aromatic hydrocarbons (PAHs) is a major environmental problem worldwide, mainly caused by the incomplete combustion of organic macromolecule substances generally concerning industrial and urban activities ¹. PAHs such as phenanthrene (PHE) and pyrene (PYR) have become the most ubiquitous environmental pollutants ². Due to highly mutagenic and carcinogenic properties and being commonly found in soil at high concentrations in many countries, PHE and PYR soil contamination has attracted particular attention ³. 

In recent years, new technology employing microorganisms and plants to remove PAHs from contaminated soils has been proposed by many researchers ^4,5. In situ bioremediation with microorganisms has been recognized as the most cost-effective, reliable, and promising approach for restoration of PHE and PYR contaminated soil ⁶. Among these root associated microorganisms, arbuscular mycorrhizal fungi (AMF) are one of the most important rhizosphere microorganisms that participate in beneficial symbiosis with the root system of nearly 80% of terrestrial vascular plants. Besides improving host nutrition, AMF are also known to alleviate the plant host from biotic and abiotic stress ^7,8. Recently, it was found that AMF increased biodegradation of soil PAHs both by increasing the soil microbial population and via accumulation of PAHs in fungal tissues ^6,9-11, and affecting the uptake and translocation of PAHs in plants ¹², Also, AMF does interact with and modify the microbial communities that the extraradical hyphae encounter in soil, and in this manner they can affect the microbial degradation processes indirectly ¹³. All the above studies indicating the potential of AMF in bioremediation for PAHs contaminated soil. 

Meanwhile, plant growth-promoting rhizobacteria (PGPR), a kind of beneficial bacteria found in the rhizosphere, have also received a lot of interest in the field of phytoremediation, which were utilized to combine plants to remove contaminants from soil ^14,15. In recent years, PGPR have been recognized as one of the most effective technologies for decontaminating PAHs-polluted soils ^16,17. Rao et al. (2015) screened Bacillus cereus CPOU13 from soil samples of petroleum contaminated areas to effectively degrade PHE, anthracene and PYR in soil. Inoculation of PAH-degrading bacteria (Acinetobacter sp.) resulted in a much higher dissipation (43%–62%) of PYR in the rhizosphere of rice compared with control (6–15%) ¹⁸. 

Therefore, AMF, PGPR, and other soil microorganisms that establish mutual symbiosis with the majority of higher plants, can provide positive impacts on plant survival in contaminated soils. The composition of rhizosphere microbial community is the main limiting factor in the process of rhizoremediation. Therefore, effects of combined inoculation with AMF and PGPR to mitigate the adverse impacts of PAHs on soil and bioremediation are limited. By detecting the base sequence of specific genetic substances in soil microbial cells, the complexity and diversity of soil microbial communities can be revealed more comprehensively and accurately, which has been widely used in the study of soil microbial communities ^19,20. Traditional molecular fingerprint techniques, such as denatured gradient gel electrophoresis (DGGE) ²¹ and terminal restriction fragment length polymorphism (T-RFLP) ²² have great limitations in the analysis of complex microorganisms in PAHs contaminated soil. High-throughput sequencing has been widely used in rhizosphere microbial diversity research ^23,24. Recent next-generation sequencing (NGS) methods, such as Illumina sequencing techniques, may provide researchers a new way to detect the microbial taxa, especially those with low-abundant species changes ²⁵. Understanding the changes of microbial community composition or enrichment genera related to the biodegradation of PAHs is helpful to deepen the understanding of the theory of rhizoremediation of PAHs-contaminated soils. However, few studies have attempted to link PAH degradation to the interactive effects of AMF and PGPR on the microbial community composition of soil contaminated with PAHs. Thus the three objectives of this work were: (1) to investigate the effects of dual inoculation AMF and PGPR on microbial community for soils with PHE and PYR pollutants, and (2) to find the impacts of AMF and PGPR on plant uptake and accumulation of PHE and PYR in soils, and (3) to determine the influence of AMF and PGPR inoculation on the growth of Festuca elata in soil contaminated by PHE and PYR were also investigated. 

Materials And Methods

The soil, AMF, plant materials and experimental methods used in this study were full complied with relevant institutional, national, and international guidelines and legislation.

Soil

The soil used in this study was collected from natural wasteland harvested from non-farmland (total PAHs<0.2mg·kg^-1) in campus of Qingdao Agricultural University, China. The soil has the following basic characteristics: pH(1:2.5 water) 5.62, organic matter 8.6 g·kg^-1, total N 0.85 g·kg^-1, total P 0.40 g·kg^-1, total K 10.7 g·kg^-1, hydrolyzable N 44mg·kg^-1, available P 12.1mg·kg^-1, available K 76.3mg·kg^-1. 52.1% sand, 27% silt, 20.9% clay and 2.03% soil organic matter. Soil was then sieved and mixed with washed sand (1:1). The soil was air-dried and passed through 2mm sieve to remove stones and roots. Then appropriate concentrations of the mixtures of PHE and PYR (each at a final concentration of 100 mg·kg^-1) were spiked into soil samples to achieve certain PAH concentrations.

Microbial inocula and host plants

Tall fescue seeds (Festuca elata ‘Crossfire II’) purchased from Clover Group Co., Ltd., Beijing, China. The seeds were surface sterilized with 10% (v/v) hydrogen peroxide for 10 min and then rinsed with sterile distilled water. Mycorrhizal inoculums of a Glomus versiforme strain were the most popular AMF in this soil ²⁶. The Glomus versiforme used in this study is provided by the Institute of Mycorrhizal Biotechnology of Qingdao Agricultural University, and the AMF inoculums is consisted of a mixture of rhizospheric soil from trap cultures (Trifolium repens) containing spores, mycelium, sand and root fragments was sieved (<2 mm). The PGPR bacteria Pseudomonas fluorescens Ps2-6 was isolated, identified and preserved from the rhizosphere of clover and alfalfa, which was also provided by the Institute of Mycorrhizal Biotechnology of Qingdao Agricultural University, and Ps2-6 was cultured in beef extract peptone and inorganic salt medium for standby.

Surface sterilized seeds were sown in porcelain pots (20 cm in diameter and 25 cm in depth) containing 3 kg air-dried soil. The germinated seedlings were thinned to keep 200 uniform seedlings in each pot. followed by inoculation with 50 g AMF inoculum and/or 10 ml PGPR zymotic fluid (1×10⁸ CFU·ml^-1). In the non-inoculation treatments, an equivalent amount of sterilized inoculum was used to provide similar conditions. All the treatments were prepared in decuplicate.

All the pots were arranged randomly in a greenhouse, with natural light and day/night temperature of 30/25°C and humidity of 60%±2%. Quarter of the Hoagland solution was supplied regularly and the pots were weighed every week to adjust the water content. 

After 60 days of treatment, plant shoots and roots were harvested separately and rinsed with sterile distilled water. The entire soil in each pot was thoroughly homogenized, ground sufficiently to pass through a 100-mesh sieve, and divided into two sets. One was stored at −20°C for DNA extraction, and the other was stored at 4°C for PAHs analysis.

PAHs analysis

5 g soil sample was freezing-dried, ground, and homogenized, and extracted with 15 ml dichloromethane: acetone (1:1) and extracted for 20 min, then centrifugation at 3000 rpm for 10 min. The supernatant was collected and concentrated into about 2 ml in a rotary evaporator, dissolved in 10 ml n-hexane and loaded on to a column packed with layers of silica gel (200-300 mesh), neutral aluminum oxide (100-200 mesh) and Na₂SO₄, and then eluted with with 80 ml hexane and dichloromethane (7:3, v/v) mixture. The filtrate was re-concentrated to 2 ml and further carefully blown dry with nitrogen. The residue was dissolved in 100 μl of n-hexane and filtrated with 0.45 μm-Teflon filter to remove particles prior to analysis. 2 g plant sample was ground and homogenized, and then extracted and cleaned as described above. The PAHs analysis was performed using an Agilent 7890A gas chromatography equipped with a flame ionization detector, by using methanol and water (90:10) as the mobile phase at a flow rate of 1ml·min⁻¹. PHE and PYR were detected by absorbance at 220 and 234 nm, respectively.

Mycorrhizal infection rate and PGPR proliferation colonization amounts in rhizosphere

After a growth period of 60 days, shoots of tall fescue were harvested, and washed with sterile water. Parts of fresh roots were randomly collected from each pot to determine the mycorrhiza infection rate. Mycorrhizal infection rate was calculated according to the root segment frequency conventional method of ²⁷, using Eq. (1): C = Rc/Rt × 100, where C (%) is the infection  rate, Rc is the total number of root segments colonized, and Rt is the total number of root segments studies, photograph was taken with Olympus BX51 microscope. Relative mycorrhizal dependency was calculated using Eq. (2): RMD = [(PDWm – PDWn)/ PDWm] × 100%, where RMD is relative mycorrhizal dependency (%), PDWm is mycorrhizal plant dried weight, PDWn is non-inoculated plant dried weight. The PGPR proliferation amounts in rhizosphere is calculated according to the dilution plate method ²⁸, 10g rhizosphere soil was fully dissolved in 90ml sterile water, and then 10 μl supernatant was applied evenly on the basal mineral medium agar plates (g·L^-1: NH₄Cl 1.0, K₂HPO₄ 0.3, KH₂PO₄ 0.2, MgSO₄ 0.5, agar 15, pH 7.2) containing a mixture of PHE and PYR (each at a final concentration of 100 mg·kg^-1), and 50 mg·L^-1 of cycloheximide for suppression of fungal growth. There are three replicates for each dilution, and all plates were incubated at 28°C. The colonies formed were counted after 2 weeks of incubation, and the number is expressed as colony-forming units (CFU)·g^-1 dry soil ²⁸.

Soil DNA extraction

The total genomic DNA of the samples was extracted from 10 g of soil using the E.Z.N.A. stool DNA Kit (Omega Bio-tek, Norcross, GA, U.S.), according to the manufacturer’s protocols. The DNA quantity and quality were assessed by NanoDrop spectrophotometer (Thermo Fisher Scientific, USA) and agarose gel electrophoresis, respectively. Extracted DNA was diluted to 1 ng·μl^-1 and stored at -20°C until further processing.

PCR amplification and sequencing 

Diluted DNA from each sample was used as a template for PCR amplification of bacterial 16S rRNA gene and fungal ITS region sequences with barcoded primers and HiFi Hot Start Ready Mix (KAPA). To determine the composition of the bacterial communities in different samples, the universal primer set 341F (5’-CCTACGGGNGGCWGCAG-3’) and 806R (5’-GGACTACHVGGGTATCTAAT-3’) was used to amplify the V3-V4 regions of the 16S rRNA genes. And the universal primer set KYO2F (5’-GATGAAGAACGYAGYRAA-3’) and ITS4R (5’-TCCTCCGCTTATTGATATGC-3’) was used to amplify the ITS2 variable regions for fungal-diversity analysis. The barcode is an eight-base sequence unique to each sample. PCR reactions were performed in triplicate in a 50 μl mixture containing 5 μl of 10 × KOD Buffer, 5 μl of 2.5 mM dNTPs, 1.5 μl of each primer (5 μM), 1 μl of KOD Polymerase, and 100 ng of template DNA. Cycling conditions involved an initial 2 min denaturing step at 95°C, followed by 27 cycles of 10 s at 98°C, 30 s at 62°C and 30 s at 68°C, and a final extension phase of 10 min at 68°C. 

The amplicon quality was assessed by visualization after 2% agarose gels, followed by purification with the AxyPrep DNA Gel Extraction Kit (Axygen Biosciences, Union City, CA, U.S.) according to the instructions, the final amplicon was quantified using QuantiFluor^TM (Promega, U.S.). Purified amplicons were pooled in equimolar and paired-end sequenced (2×250bp) on an Illumina Hiseq2500 platform. ITS region and the 16S rRNA gene sequences obtained from this study were deposited to the NCBI Sequence Read Archive database under the accession no. SRR11873579- SRR11873590, and SRR11879883- SRR11879894. 

Processing and analyzing of sequencing data

Raw Illumina fastq files were de-multiplexed, quality filtered, and analysed using QIIME (version 1.17) ²⁹ with the following criteria: 1) Removing reads containing more than 10% of unknown nucleotides (N); 2) Removing reads containing less than 80% of bases with quality (Q-value)＞20; 3) Paired end clean reads were merged as raw tags using FLSAH (v 1.2.11)³⁰ with a minimum overlap of 10bp and mismatch error rates of 2%; 4) Noisy sequences of raw tags were filtered by QIIME (V1.9.1) ²⁹. Reads that could not be assembled were discarded. Operational taxonomic units (OTUs) with ≥ 97 % similarity using UPARSE (version 7.1), and chimeric sequences were identified and removed using UCHIME algorithm (http://www.drive5.com/usearch/manual/uchime_algo.html). Chao1, Simpson and Shannon diversity indices were calculation in QIIME. OTU rarefaction curve and Rank abundance curves was plotted in QIIME. Statistics of between group Alpha index comparison was calculated by a Welch's t-test and a Wilcoxon rank test in R. Alpha index comparing among groups was computed by a Tukey’s HSD test and a Kruskal-Wallis H test in R. The beta diversity analysis was performed using UniFrac ³¹. The principal component analysis (PCA), Venn diagrams and Heatmap figures were calculated and plotted in R.

Statistical analysis

All data were analyzed with Excel 2010 and SPSS11.5. All data were first analyzed by analysis of variance (ANOVA) to determine significant differences for the treatment effects (P = 0.05). Significant differences between individual means were determined using Duncan’s Multiple Range Test (P = 0.05). 

Results

Promotion of AMF and PGPR on the growth of F. elata and removal of PHE and PYR in soil

The effect of the inoculation of PGPR and/or AMF on the growth of different treatment soils was investigated after 60 days of experiments in greenhouse condition. Compared with the CK, F. elata inoculated with GV, PS, and GVPS produced larger fresh weight, dry weight and higher plant height under PHE and PYR stress conditions, however, the GV and PS did not have an effect on tiller number of F. elata. Moreover, the GVPS treatment produced the highest fresh weight (0.56 ± 0.03) g, dry weight (0.1177 ± 0.003) g, and plant height (36.7 ± 0.8) g (p<0.05) in the concentration of PHE and PYR at 100 mg·kg^-1. The fresh weight, dry weight, and plant height increased by 1.43-fold, 93.90% and 51.03% compared with that of CK, respectively (Table 1).

Table 1 

Growth indexes of F. elata in response to different treatments

The data represent the mean ± standard deviation of three replicates. FW means fresh weight, DW means dry weight, DW means dry weight, PH means plant height, and TN means tiller number. Values in each column followed with different lowercase letters (a, b, c and d) indicated significant differences between different treatments.

The mycorrhizal status was best manifested in the roots of plants inoculated with GVPS; the percentage of root mycorrhizal proliferation of GVPS treatment was 69% (Fig 1). GVPS treatment significantly increased the PGPR proliferation amounts in tall fescue rhizosphere soil. The PGPR proliferation amounts reached a maximum of 9.5×10⁷ CFU·g⁻¹ at 100 mg·kg⁻¹ of PHE and PYR. Meanwhile, the GVPS treatment significantly enhanced the hyphal density, entry point number, and mycorrhizal relative dependence (p < 0.05, Fig 1) by 75%, 73% and 383%, respectively. However, the treatment did not significantly (p < 0.05) increase the infection, spore density, vesicle number, or arbuscule infection of tall fescue (Fig 1). By the end of the experiment, the PHE concentrations decreased from the initial value of 100 mg·kg⁻¹ to 2.85, 2.85, 2.84, and 2.46 mg·kg⁻¹ dry soil in CK, GV, PS and GVPS, respectively, corresponding to degradation efficiencies of 97.10%, 97.10%, 97.10%, and 97.50%, respectively. The PYR concentrations decreased from the initial value of 100mg·kg⁻¹ to 10.12, 8.72, 8.92 and 5.11 mg·kg⁻¹ dry soil in treatments CK, GV, PS and GVPS, respectively, corresponding to removal efficiencies of 89.67%, 91.10%, 90.90% and 94.80%, respectively (Table 2). Residual concentrations of PHE and PYR in tall fescue shoots and roots are also shown in Table 2, and high concentrations of PHE and PYR were detected in tall fescue roots, root concentrations of PHE in tall fescue grown for 60 days in soils with GV, PS, and GVPS inoculation were 22.62%, 31.13%, 53.63% higher than those of CK, and root concentrations of PYR in tall fescue with GV, PS, and GVPS inoculation were 56.95%, 30.21%, and 67.11% higher than those of CK. Concurrently, PHE and PYR concentrations of shoots in contaminated soil inoculated with GV, PS, and GVPS were significantly (p < 0.05) higher than those of CK (Table 2). 

Table 2 The content of PHE and PYR in soil as well as roots and shoots of F. elata under different treatments 

The data represent the mean ± standard deviation of three replicates. RC means residual concentration, DE means degradation efficiency, Values in each column followed with different lowercase letters (a, b, and c) indicated significant differences between different treatments. The same letter within a column indicates no significant difference assessed by Duncan’s multiple range test (P ≤ 0.05) following analysis of variance.

Evaluation of sequencing results of soil microbial library

After sequencing the original data, the low-quality data or non-biologically meaningful data (such as chimeras) are removed to ensure the statistical reliability and biological validity of subsequent analysis. The sequencing run of 16S rRNA amplicons yielded an average of 68,534.67 ± 10,870.14, 69,611 ± 7,337.083, 72,296.333 ± 9,922.511, and 71,144 ± 1,136.746 clean tags (per sample), with 4,916, 5,164, 5,570 and 6,327 total OTUs from the CK, GV, PS and GVPS samples, respectively (Table 3). The sequencing run of ITS amplicons yielded 96023 ± 2098.435, 92420.67 ± 8008.382, 100643.3 ± 2112.008, and 102455 ± 6585.639 clean tags, with 595, 649, 783, and 825 total OTUs from the CK, GV, PS and GVPS samples, respectively (Table 3). 

Table 3 Summary of sequencing date, number of operational taxonomic units (OTUs), and alpha diversity in different treatment under the pollution of PHE and PYR.

The total number of OTUs detected at 97% that shared sequence similarity was very high in PHE and PYR contaminated soil, both in terms of bacteria and fungi, and the estimated α-diversities indicated abundant microbial diversity present in all samples. For bacteria, the number of different phylogenetic OTUs ranged from 1,639 to 2,109, with dual inoculation (GVPS) showing higher 16S rRNA gene diversity than single inoculation (GV, PS) and CK. For fungi, the number of different phylogenetic OTUs in all samples ranged from 198 to 275, with GVPS exhibiting higher diversity than CK, GV, and PS. The GVPS displayed the highest Shannon index and number of OTUs, whereas CK samples had the lowest (Table 3). 

Venn diagrams were created in R, based on the shared OTU tables from 4 different soil groups (Fig 2a). The total number of unique bacterial OTUs was 3,415, of which 119 OTUs were shared between PS and GVPS treatments, 187 were associated only with treatment of GV (GV, GVPS), and 1035 were shared by all samples (Fig 2a). Furthermore, in terms of fungi, 504 different OTUs were identified, both PS vs GVPS and GV vs GVPS groups, shared only 46 and 21 OTUs, respectively, and 92 were shared by all samples (Fig 2b).

Analysis of microbial community composition

All valid reads were classified from the phylum to the genus level using the default settings in QIIME. The bacterial and fungal communities from the 12 samples were analyzed at phylum, family, and genus levels. In total, all the bacteria and fungi identified were classified into 28 and 6 phyla, respectively. Proteobacteria, Saccharibacteria, and Parcubacteria were the dominant bacterial phyla, and there are three main phyla of fungi : Ascomycota, Chytridiomycota, and Basidiomycota. All the treatments shared similar bacterial and fungal communities. Most samples from the same group shared high similar bacterial communities at all classification levels.

At the phylum level, the CK, GV, PS and GVPS samples shared common phyla, Proteobacteria was the most prevalent bacteria phylum, while different proportions of valid reads from 33.80% to 41.73% were observed for all treatments. More Proteobacteria taxa (41.73%) were detected in GVPS than that in GV, PS and CK (Fig 3a). Fungal classification results showed that the dominant phylum was Ascomycota, accounting for 33.13–52.04% of all valid reads, with an average relative abundance of 43.56%. The next most dominant fungal phyla were Chytridiomycota (average abundance 12.13%) and Basidiomycota (average abundance 6.60%), and the abundance of Glomeromycota (0.27%) in GVPS was significantly higher than that in other treatments (Fig 3b).

The most prevalent bacterial families detected in all 12 groups included Xanthomonadaceae (7.40%-12.58%), Planctomycetaceae (average abundance 6.25%), and Sphingomonadaceae (average abundance 3.59%). The abundance of Xanthomonadaceae (12.58%), Phytophthoraceae (6.76%), and Sphingomycidae (4.43%) in GVPS was significantly higher than others (Fig 3c). At the family level, according to the classification of fungi, Debaryomycetaceae (average abundance 17.46%) is the richest fungus family in all samples, accounting for 9.94% - 27.67% of the total. Spizellomycetaceae is the second most abundant fungal family with an average abundance of 12.12%. The proportion of Nectriaceae (11.76%), Pseudoglobulaceae (4.44%), and Cladosporidae (1.08%) were significantly higher in GVPS samples compared to other samples (Fig 3d).

At the genus level, according to the results of bacterial taxonomy, Planctomyces is the richest genus in all samples, accounting for 3.0% - 3.39% of the total. Sphingomonas is the second most abundant bacteria genus with an average abundance of 2.36%. The other major bacterial genera were Mycobacterium (average abundance 2.31%), Arenimonas (average abundance 1.92%), Pseudomonas (average abundance 1.75%), and Pirellula (average abundance 1.53%). The abundance of Sphingomonas (3.17%), Pseudomonas (2.05%), and Piriformis (1.79%) in GVPS was significantly higher than that in other treatments (Fig 3e). Meanwhile, Meyerozyma is the richest fungi genus in all samples, accounting for 9.94% - 27.67% of the total. Spizellomyces is the second most abundant fungi genus with an average abundance of 12.12%. The other dominant fungal genera were Gibberella (average abundance 4.14%), Fusarium (average abundance 3.93%), Serendipita (average abundance 3.17%), Alternaria (average abundance 2.93%), Aspergillus (average abundance 2.09%), and Chalastospora (average abundance 0.88%). The abundance of Fusarium (8.65%), Alternaria (4.09%), and Cladosporium (1.07%) in GVPS treatment was significantly higher than other treatments (Fig 3f). Heatmap clustering analysis results revealed that Planctomyces, Mycobacterium bacterial genera had high abundance in CK, GV, and PS, while the abundance of Sphingomonas, Planctomyces, and Arenimonas genera were higher in  GVPS (Fig 4a). For fungi, heatmap clustering analysis showed that Meyerozyma and Spizellomyces fungus genera had relatively high abundance among all the treatments, while Fusarium had a high abundance in GVPS (Fig 4b). These findings were consistent with previous results (Fig 3).

Effects of AMF and PGPR on soil microbial community richness and diversity in the root zone of F. elata 

The rarefaction curve can evaluate whether the sequencing quantity is sufficient to cover all groups and indirectly reflect the species richness in the treatments. Rarefaction curves of four treatments (CK, PS, GV, GVPS) for bacteria and fungi are shown in (Fig A1). None of the rarefaction curves are parallel with the x-axis, the rarefaction curves of bacteria and fungi calculated at 97% levels showed that the order of OTUs numbers from high to low among samples both were GVPS > PS > GV > CK. The OTU densities of GVPS were higher than the other three treatments (Fig A1). The bacteria and fungi richness based on rarefaction curves were strongly supported by statistical diversity estimates, based on the abundance results of OTUs, the Alpha diversity of each treatments were calculated by QIIME software, including Chao 1 value, ACE value, Shannon index, and Simpson index (Table 3). The results showed that the values of Chao 1 and ACE of GVPS treatment were higher, which indicated that the richness of microbial community under GVPS treatment was higher. The Simpson diversity index of the four treatments had little difference, indicating that the uniformity of the four treatments and the dominant OTU of the community were similar. The Shannon diversity index was higher in GVPS treatment, indicated a richer microbial community in GVPS treatment, (Table 3). Based on the relative abundance of the genera from (Fig 5), the genera with an average abundance of >1 % in at least one group were defined as dominant.

In addition,principal component analysis (PCA) was used to identify the microbial community composition differences under different treatments (Fig 5). The data are presented as a 2-dimensional plot to better illustrate the relationship among treatments. At OTU level, PCA demonstrated that four treatments of 12 soil samples were clustered. In bacteria, except for CK-1 and GV-3, microorganism communities in most treatments gathered together, and different soil samples from CK and PS gathered together than others. In addition, the GV samples had a relatively higher PC1 value, followed by PS and GVPS treatment, whereas the CK samples had a higher PC2 value at OTU level (Fig 5a). In fungi, the GVPS groups had a relatively higher PC1 value, followed by PS and CK, while the samples from GV were closer than the other groups. Meanwhile, no significant gatherings were observed among four groups (Fig 5b). 

UPGMA clustering obtained a phylogenetic tree by using unweighted group averaging method (Fig A2). Results indicated that same type of samples showed high similarity of bacterial communities (Fig  A2-a), while similarity of fungal communities from the same treatment were relatively weaker (Fig A2-b). 

Discussion

Fungi, bacteria and other microorganisms are widely distributed in urban soil, among which beneficial microorganisms have the ability to promote plant growth, effective utilization and absorption of nutrients, and support plant health ³², and their symbiotic relationship with plants also provides an ecological basis for bioremediation technology. AMF and PGPR have been proved to be the two most influential microbial groups to improve soil health and productivity. A greenhouse experiment was conducted to evaluate the potential effect of symbiotic remediation of PHE and PYR polluted soil among tall fescue, AMF (Glomus versiforme, Gv), and PGPR (Pseudomonas fluorescens, PS 2-6). The results showed that compared with the control, the fresh weight, dry weight and plant height of tall fescue treated with GV, PS and GVPS were significantly increased, and an analogous pattern was reported in previous research ¹⁷, which showed higher fresh and dry weight of Avena sativa inoculated with Serratia marcescens BC-3 alone or mixed with Rhizophagus intraradices than those of the control in petroleum hydrocarbon polluted soil. Two wheat cultivars inoculated with the R. irregularis and the Pseudomonas putida KT2440 dramatically enhance plant growth and root shoot ratio ³³ This study further proved that AMF and PGPR could promote the growth of tall fescue in PHE and PYR contaminated soil to some extent. In addition, the relationship between G. versiforme and PS 2-6 is mutual promotion. Compared with G. versiforme single inoculation, GVPS treatment significantly increased the mycorrhizal infection rate, hyphal density, entry point number, and mycorrhizal relative dependence (Fig 1). Meanwhile, GVPS treatment significantly promoted the proliferation of PGPR in tall fescue rhizosphere soil. In previous studies, it was found that Some endophytic species of PGPR were known to excrete cellulase and pectinase ^34,35 and these enzymatic activities would no doubt aid in mycorrhizal infection. In the meantime, inoculation of G. versiforme significantly increased the populations of P. fluorescens in contaminated soil. This study further proved that AMF and PGPR can improve the growth of plants under polycyclic aromatic hydrocarbons (PAHs) contaminated conditions, among which G. versiforme and P. fluorescens Ps2-6 are mutually beneficial microorganisms and can promote each other's proliferation. 

Inoculation with GV, PS and GVPS significantly removed PHE and PYR in contaminated soil in the study, and promoted the accumulation of PHE and PYR in tall fescue. In previous studies, tall fescue has been commonly selected for phytoremediation of PAHs contaminated soils, and removal percentage of PAHs in tall fescue rhizosphere soil was 11% higher than that in unplanted soil³⁶, However, phytoremediation alone may not be the best choice, and at present, the synergetic elimination of PAHs by plants and rhizosphere microorganisms is considered to be a promising, economical and environmentally friendly soil remediation technology. PGPR enhanced phytoremediation has been widely used to remove petroleum hydrocarbons and other organic pollutants from contaminated soil ^37-39, and AMF can also promote the natural attenuation of host toxicity by producing high root surface area during symbiosis with plants ^40,41. In addition, AMF inoculation could increase the content of glomalin related protein (GRSP) in soil, and then improve the availability of PAHs in soil ⁴². Previous studies have shown that the degradation rate of total petroleum hydrocarbons with PGPR Serratia marcescens BC-3 and AMF Glomus intraradices co-inoculation treatment was up to 72.24 % ¹⁷, this study also supported the previous conclusions, and the degradation rates of PHE and PYR in rhizosphere soil were 97.5% and 94.8% in GVPS treatment group (Table 2), although the degradation characteristics of PAHs by AMF and PGPR are different, but they may complement each other in function. Recent studies have found that the liquid membrane on the surface of fungal mycelium can help bacteria to migrate in the soil environment, and at the same time, fungal mycelium can also transport PAHs in the soil to bacteria through cytoplasmic flow, which enhances the availability of pollutants and promotes their biodegradation⁴³. In addition, as AMF and PGPR are both microorganisms that promote plant growth, we speculate that AMF and PGPR can synergistically improve the growth status of plants in the polluted environment, further improve the pollutant removal efficiency and reduce the environmental stress suffered by plants. 

In addition, we also analyzed the microbial community composition of soil under the four treatment conditions. We found that the abundance and composition of bacteria and fungi were affected by GV, PS or GVPS. Moreover, the microbial community diversity of PHE and PYR contaminated soil for combined inoculation (GVPS) was higher than that of GV or PS single inoculation. Compared with CK, the total number of bacterial OTUs in GVPS samples increased from 4916 to 6327, and that of fungal OTUs increased from 595 to 825 (Table 3). Sang et al. (2012) showed that inoculating exogenous microorganisms in soil would directly or indirectly affect microbial community composition and soil microbial activity, and then affect the function of soil ecosystem.

For bacteria, we observed that Proteobacteria were the most abundant bacterial phyla in CK, GV, PS and GVPS samples. The abundance of Proteobacteria detected in GVPS increased from 33.80% of CK to 41.73% (Fig 3a). At the genus level, Sphingomonas (3.17%) was the highest in GVPS treatment group, and the abundance of Pseudomonas (2.05%) and Piriformis (1.79%) was also significantly higher than the other three treatments (Fig 3e, f). This is not consistent with previous studies, data show that Pseudomonas is dominant in crude oil degrading rhizobacteria ⁴⁴, and can use the high molecular weight PAHs (HMW-PAHs) as a source of carbon and energy ⁴⁵. However, some studies have also found that Sphingomonas may play a key role in the early degradation of PAHs ^46,47. Therefore, our results indicate that the increase of Pseudomonas and Sphingomonas is one of the reasons for the improvement of PHE and PYR degradation rate in soil, and it is speculated that PAHs degrading bacteria may also exist in Piriformis.

Ascomycota is the dominant phylum of fungi in all samples, with an average relative abundance of 43.56%, ant the second dominant phylum was chytrium (average abundance 12.13%) and Basidiomycota (average abundance 6.60%), while the abundance of Glomeromycota in GVPS (0.27%) was significantly higher than that in GV, PS and CK treatments (Fig 3b). Previous data showed that members of the Glomeromycota phylum depend on carbon and energy derived from plant synthesis to survive, and share a symbiotic relationship with the roots of plants ^48,49. Among the fungi identified in our samples, the abundance of Glomeromycota was greater in GVPS treatment, thus, we speculate that the Glomeromycota phylum may also affect symbiosis and interactions between tall fescue roots and soil microbes, For fungus in the genus level, the abundance of Fusarium (8.65%) in GVPS treatment was significantly higher than other treatments (Fig 3f), and the ability of Fusarium spp. to degrade some recalcitrant substances has also been reported, Fusarium sp. produced the most significant effect on degradation of HMW-PAHs, giving an overall removal rate of over 30% for 5- and 6-ring PAHs ⁵⁰, and bromegrass inoculated with Fusarium sp. ZH-H2 can effectively repair aged PAH-contaminated soil in coal mining areas ⁵¹. Therefore, we believe that higher abundance of Fusarium may be involved in the dissipation of PHE and PYR. Interestingly, diversity indices (Shannon and Simpson indices) indicated that the diversity of fungal and bacterial community in GVPS were significantly higher than GV, PS and CK, in addition, the PCA revealed that fungal community changes in the contaminated-soils are more complex than bacteria community changes in soil (Fig 2 and Table 1). It was speculated that the effect of rhizosphere soil environment on fungal community was more significant, compared with bacterial community, phenanthrene and pyrene might mainly affect the fungal activity and community diversity.

Therefore, we speculate that the ecological toxicity of PAHs can change the composition of soil microbial community and contribute to the enrichment of a large number of primary degradation bacteria. The change of microbial community diversity may be caused by the different reactions of different types of microorganisms to PAHs. In addition, the inoculation of different types of microorganisms will also affect the composition of soil microbial community of tall fescue rhizosphere soil polluted by PHE and PYR. To sum up, the diversity of soil microbial community is closely related to the interaction among the plants, soil environment and inoculated microorganisms, in order to better understand the impact of GV and PS on soil microorganisms, and we can consider the effects of plant species, root exudates, GV and PS secretions on soil microbial composition, and more systematic, detailed and accurate experimental results combined with big data analysis will be provided in the future. 

Conclusions

In conclusion, a detailed picture of bacterial and fungal community variations in PHE and PYR polluted soils under four treatments (CK, GV, PS and GVPS) were analyzed based on the high throughput Illumina sequencing method. The results reflected the significant contribution of GVPS in increasing the species and abundance of bacteria, whereas no significant differences were observed for fungi in PHE and PYR contaminated soil. Meanwhile, the highest dissipation rates of PHE and PYR as well as biomass of tall fescue in GVPS were observed. Tall fescue associated with GVPS significantly (p < 0.05) enhanced dissipation of PHE and PYR from soil; PHE and PYR accumulation by tall fescue roots were higher than shoots. Sequence analysis revealed that Proteobacteria and Glomeromycota were the most prevalent bacterial and fungal phyla in GVPS, respectively. In the generic level, Sphingomonas was the dominant  genus, while the dominant fungi in GVPS was Fusarium. GVPS had an effective role in improving the growth characteristics, root infection of F. elata, and soil microbial community composition in PHE and PYR contaminated soils, but there were no obvious degradation efficiencies of PHE and PYR as compared to the control.

Declarations

Competing interests

The authors declare no competing interests.

Author contributions

The project was designed and conceived by Shaoxia Guo. The experimental work was carried out by Wenbin Li, Wei Li, and Lijun Xing; Wei Li analysed the data. Wei Li wrote the manuscript. All authors have read and approved the final version of this manuscript.

Acknowledgement

This work was supported by High-level Science Foundation of Qingdao Agricultural University [663/1115006].

References

1. Eremina, N., Paschke, A., Mazlova, E. A. & Schüürmann, G. Distribution of polychlorinated biphenyls, phthalic acid esters, polycyclic aromatic hydrocarbons and organochlorine substances in the Moscow River, Russia. Environ. Pollut. 210, 409–418 (2016).
2. Anyanwu, I. N. et al. Impact of nitrogen-polycyclic aromatic hydrocarbons on phenanthrene and benzo[a]pyrene mineralisation in soil. Ecotox. Environ. Safe. 147, 594–601 (2018).
3. Calonne, M. et al. Polyaromatic hydrocarbons impair phosphorus transport by the arbuscular mycorrhizal fungus Rhizophagus irregularis. Chemosphere. 104, 97–104 (2014).
4. Haritash, A. K. & Kaushik, C. P. Biodegradation aspects of Polycyclic Aromatic Hydrocarbons (PAHs): A review. J. Hazard. Mater. 169, 1–15 (2009).
5. Khan, S., Afzal, M., Iqbal, S. & Khan, Q. M. Plant–bacteria partnerships for the remediation of hydrocarbon contaminated soils. Chemosphere. 90, 1317–1332 (2013).
6. Gao, Y., Li, Q., Ling, W. & Zhu, X. Arbuscular mycorrhizal phytoremediation of soils contaminated with phenanthrene and pyrene. J. Hazard. Mater. 185, 703–709 (2011).
7. Smith, S. E. & Read, D. J. Mycorrhizal. symbiosis(Academic press, 2010).
8. Cornejo, P. et al. Contribution of inoculation with arbuscular mycorrhizal fungi to the bioremediation of a copper contaminated soil using Oenothera picensis. J. Soil. Sci. Plant. Nut. 17, 14–21 (2017).
9. Cheema, S. A. et al. Degradation of phenanthrene and pyrene in spiked soils by single and combined plants cultivation. J. Hazard. Mater. 177, 384–389 (2010).
10. Wu, F. Y., Yu, X. Z., Wu, S. C., Lin, X. G. & Wong, M. H. Phenanthrene and pyrene uptake by arbuscular mycorrhizal maize and their dissipation in soil. J. Hazard. Mater. 187, 341–347 (2011).
11. Aranda, E. et al. Role of arbuscular mycorrhizal fungus Rhizophagus custos in the dissipation of PAHs under root-organ culture conditions. Environ. Pollut. 181, 182–189 (2013).
12. Wu, N. et al. Phenanthrene uptake by Medicago sativa L. under the influence of an arbuscular mycorrhizal fungus. Environ. Pollut. 157, 1613–1618 (2009).
13. Joner, E. J. & Corinne, L. Rhizosphere gradients of polycyclic aromatic hydrocarbon (PAH) dissipation in two industrial soils and the impact of arbuscular mycorrhiza. Int. J. Environ. Sci. Te. 37, 2371–2375 (2003).
14. Ahmad, F., Ahmad, I. & Khan, M. S. Screening of free-living rhizospheric bacteria for their multiple plant growth promoting activities. Microbiol. Res. 163, 173–181 (2008).
15. Ma, Y. et al. The hyperaccumulator Sedum plumbizincicola harbors metal-resistant endophytic bacteria that improve its phytoextraction capacity in multi-metal contaminated soil. J. Environ. Manage. 156, 62–69 (2015).
16. Yateem & Awatif. Rhizoremediation of oil-contaminated sites: a perspective on the Gulf;War environmental catastrophe on the State of Kuwait. Environ. Sci. Pollut. Res. Int. 20, 100–107 (2013).
17. Dong, R., Gu, L., Guo, C., Xun, F. & Liu, J. Effect of PGPR Serratia marcescens BC-3 and AMF Glomus intraradices on phytoremediation of petroleum contaminated soil. Ecotoxicology. 23, 674 (2014).
18. Gao, Y. et al. Interactions of rice (Oryza sativa L.) and PAH-degrading bacteria (Acinetobacter sp.) on enhanced dissipation of spiked phenanthrene and pyrene in waterlogged soil. Sci. Total. Environ. 372, 1–11 (2006).
19. Bokulich, N. A. & Mills, D. A. Improved selection of internal transcribed spacer-specific primers enables quantitative, ultra-high-throughput profiling of fungal communities. Appl. Environ. Microbiol. 79, 2519–2526 (2013).
20. Caporaso, J. G. et al. Ultra-high-throughput microbial community analysis on the Illumina HiSeq and MiSeq platforms. ISME. j. 6, 1621 (2012).
21. Pacwa-Płociniczak, M., Płaza, G. A. & Piotrowska-Seget, Z. Monitoring the changes in a bacterial community in petroleum-polluted soil bioaugmented with hydrocarbon-degrading strains. Appl. Soil. Ecol. 105, 76–85 (2016).
22. Grant, R., Muckian, L. & Clipson, N. Em. Microbial community changes during the bioremediation of creosote-contaminated soil. Ett. Appl. Microbiol. 44, 293–300 (2010).
23. Wang, L., Li, F., Zhan, Y. & Zhu, L. Shifts in microbial community structure during in situ surfactant-enhanced bioremediation of polycyclic aromatic hydrocarbon-contaminated soil. Environ. Sci. Pollut. R. 23, 14451–14461 (2016).
24. Sun, F. L., Fan, L. L. & Xie, G. J. Effect of copper on the performance and bacterial communities of activated sludge using Illumina MiSeq platforms. Chemosphere. 156, 212–219 (2016).
25. Uroz, S. et al. Functional assays and metagenomic analyses reveals differences between the microbial communities inhabiting the soil horizons of a Norway spruce plantation. Plos. One. 8, e55929 (2013).
26. Li, S. M., Wang, Y. Q. & Liu, R. J. [Diversity of arbuscular mycorrhizal fungi in special habitats: a review]. Ying. Yong. Sheng. Tai. Xue. Bao. 24, 3325–3332 (2013).
27. Biermann, B. & Linderman, R. G. Quantifying Vesicular-Arbuscular Mycorrhizae: A Proposed Method Towards Standardization. New. Phytol. 87, 63–67 (1981).
28. Noorasyikin, M. N. & Mohamed, Z. PGPR by bacillus in Vetiver Grass's Root enhance strength of soil-root matrix system toward slope stabilization in Malaysia. Journal. of. Engineering. and. Applied. Sciences. 12, 3051–3054 (2017).
29. Caporaso, J. G. et al. QIIME allows analysis of high-throughput community sequencing data. Nat. Methods. 7, 335–336 (2010).
30. Magoc, T. in IEEE International Conference on Fuzzy Systems.
31. Lozupone, C., Lladser, M. E., Knights, D., Stombaugh, J. & Knight, R. UniFrac: an effective distance metric for microbial community comparison. Isme. J. 5, 169–172 (2011).
32. Glick & Bernard, R. Bacteria with ACC deaminase can promote plant growth and help to feed the world. Microbiol. Res. 169, 30–39 (2014).
33. Pérezdeluque, A. et al. The interactive effects of arbuscular mycorrhiza and plant growth-promoting rhizobacteria synergistically enhance host plant defences against pathogens. Sci. Rep. 7, 16409 (2017).
34. Kovtunovych, G. et al. Correlation between pectate lyase activity and ability of diazotrophic Klebsiella oxytoca VN 13 to penetrate into plant tissues. Plant. Soil. 215, 1–6 (1999).
35. Verma, S. C., Ladha, J. K. & Tripathi, A. K. Evaluation of plant growth promoting and colonization ability of endophytic diazotrophs from deep water rice. J. Biotechnol. 91, 127–141 (2001).
36. Guo, M. et al. Enhanced polycyclic aromatic hydrocarbons degradation in rhizosphere soil planted with tall fescue: Bacterial community and functional gene expression mechanisms. Chemosphere. 212, 15–23 (2018).
37. Agarry, S. E., Aremu, M. O. & Aworanti, O. A. Kinetic Modelling and Half-Life Study on Bioremediation of Soil Co-Contaminated with Lubricating Motor Oil and Lead Using Different Bioremediation Strategies. Soil. Sediment. Contam. 22, 800–816 (2013).
38. Liu, W., Sun, J., Ding, L., Luo, Y. & Chen, M. Rhizobacteria (Pseudomonassp. SB) assist phytoremediation of oily-sludge-contaminated soil by tall fescue (Testuca arundinaceaL.). Plant. Soil. 371, 533–542 (2013).
39. Fester, T., Giebler, J., Wick, L. Y., Schlosser, D. & Kästner, M. Plant–microbe interactions as drivers of ecosystem functions relevant for the biodegradation of organic contaminants. Curr. Opin. Biotech. 27, 168–175 https://doi.org/10.1016/j.copbio.2014.01.017 (2014).
40. Eke, P. et al. Mycorrhiza consortia suppress the fusarium root rot (Fusarium solani f. sp. Phaseoli) in common bean (Phaseolus vulgaris L.). Biol. Control. 103, 240–250 (2016).
41. Lehmann, A., Veresoglou, S. D., Leifheit, E. F. & Rillig, M. C. Arbuscular mycorrhizal influence on zinc nutrition in crop plants – A meta-analysis. Soil. Biol. Biochem. 69, 123–131 (2014).
42. Shuang, C., Jian, W., Gatheru, W. M. & Yanzheng, G. Glomalin-related soil protein influences the accumulation of polycyclic aromatic hydrocarbons by plant roots. Sci. Total Environ. Sci. Total Environ. 644, 465–473 (2018).
43. Otto, S., Banitz, T., Thullner, M., Harms, H. & Wick, L. Y. Effects of Facilitated Bacterial Dispersal on the Degradation and Emission of a Desorbing Contaminant. Environ. Sci. Technol. 50, 6320–6326 (2016).
44. Yan et al. Oleispiralenta sp. nov., a novel marine bacterium isolated from Yellow sea coastal seawater in Qingdao, China. Antonie. Van. Leeuwenhoek. 101, 787–794 (2012).
45. Folwell, B. D., Mcgenity, T. J. & Whitby, C. Characterisation of biofilm and planktonic bacterial and fungal communities transforming high molecular weight polycyclic aromatic hydrocarbons. Appl. Environ. Microbiol. 82, 2288–2299 (2016).
46. Bacosa, H. P. & Inoue, C. Polycyclic aromatic hydrocarbons (PAHs) biodegradation potential and diversity of microbial consortia enriched from tsunami sediments in Miyagi, Japan. J. Hazard. Mater. 283, 689–697 (2015).
47. Gallego, S., Vila, J., Tauler, M., Nieto, J. M. & Grifoll, M. Community structure and PAH ring-hydroxylating dioxygenase genes of a marine pyrene-degrading microbial consortium. Biodegradation. 25, 543–556 (2014).
48. Hannula, S. E., Boschker, H. T. S., De Boer, W. & Veen, J. A. Van. 13C pulse-labeling assessment of the community structure of active fungi in the rhizosphere of a genetically starch-modified potato (Solanum tuberosum) cultivar and its parental isoline. New Phytol. 194, 784–799 (2012).
49. Lu, Y., Murase, J., Watanabe, A., Sugimoto, A. & Kimura, M. Linking microbial community dynamics to rhizosphere carbon flow in a wetland rice soil. Microb. Ecol. 48, 179–186 (2004).
50. Potin, O., Rafin, C. & Veignie, E. Bioremediation of an aged polycyclic aromatic hydrocarbons (PAHs)-contaminated soil by filamentous fungi isolated from the soil. Int. Biodeterior. Biodegradation. 54, 45–52 (2004).
51. Shi, W. et al. Effective remediation of aged HMW-PAHs polluted agricultural soil by the combination of Fusarium sp. and smooth bromegrass (Bromus inermis Leyss.). J. Integr. Agr. 16, 199–209 (2017).