Negative Correlations Between Cultivable Pyrene-degrading Sphingomonas and Active Soil Pyrene Degraders Explain Bioaugmentation Postpone or Failure

Background: Bioaugmentation is an effective approach to remediate soils contaminated by polycyclic aromatic hydrocarbon (PAHs), but suffers from unsatisfactory performance in engineering practices. It is hypothetically explained by the complicated interactions between indigenous microbes and introduced degrading consortium. This study isolated a cultivable pyrene degrader (Sphingomonas sp. YT1005) and an active pyrene degrading consortium consisting of Gp16, Streptomyces, Pseudonocardia, Panacagrimonas, Methylotenera and Nitrospira by magnetic-nanoparticle mediated isolation (MMI) from soils. Results: Pyrene biodegradation was postponed in bioaugmentation with Sphingomonas sp. YT1005, explained by its negative correlations with the active pyrene degraders. In contrast, amendment with the active pyrene degrading consortium, pyrene degradation eciency increased by 30.17%. In addition, pyrene degradation eciency was positively correlated with the abundance of pyrene dioxygenase encoding genes (nidA, nidA3 and PAH-RHDα-GP), which signicantly increased in MMI-isolated consortium. Pyrene degradation by Sphingomonas sp. YT1005 only followed the phthalate pathway, whereas the MMI-isolated pyrene degrading consortium exhibited both phthalate and salicylate pathways. The results indicated that Sphingomonas sp. YT1005 was not the actual pyrene degrader in soils, and MMI could successfully isolate the active pyrene degraders that were suitable for bioaugmentation. Conclusion: This work revealed the microbial intra-correlations during the bioaugmentation process, uncovered the underlying mechanisms of bioaugmentation postpone with cultivable degraders, and provided a deeper insight into the actual pyrene degraders and degradation pathways in PAHs contaminated soils. Our ndings gave new explanations for bioaugmentation postpone or failure, and offered clues to enhance bioaugmentation performance by the active degraders using MMI.


Introduction
Polycyclic aromatic hydrocarbons (PAHs) are a typical group of persistent organic pollutants (POPs) containing two or more combined aromatic rings in linear, angular or cluster arrangements ( Bioremediation is known as a cost-e cient and environmentally friendly approach to degrade soil pollutants, exhibiting low risks in secondary pollution (Peng et (Takizawa et al 1999), pdo (Krivobok et al 2003), phd (Saito et al 1999) and phn (Izmalkova et al 2013). Particularly, pyrene is a typical PAH for its similar structure to several carcinogenic PAHs (Peng et al 2008) and its biodegradation has been intensively studied via either metabolism as carbon and energy source, or co-metabolism as a non-growth substrate ( Bioaugmentation is a practical strategy to improve bioremediation performance at contaminated sites by introducing competent strains or consortia capable of degrading target contaminants (Perelo 2010). For instance, bioaugmentation with autochthonous Acinetobacter tandoii LJ-5 signi cantly improved phenanthrene removal e ciency by changing the diversity of phenanthrene degraders .
Although several eld studies have demonstrated the feasibility of bioaugmentation to enhance bioremediation performance in groundwater (Major et al 2002), wastewater (Wu et al 2018) and aquifer (Dybas et al 2002), scaling up from laboratory to commercial-scale eld is always a challenge. Both abiotic (temperature, moisture, pH, organic matter, aeration, nutrient and soil type) (Heinaru et al 2005) and biotic factors (competition between indigenous and exogenous microorganisms for limited nutrients) (Sørensen et al 1999, Mrozik and Piotrowska-Seget 2010) play essential roles in bioaugmentation performance. Additionally, the mechanisms of bioaugmentation are argued recently (Thompson et al 2005), and little is known about the interactions of the re-introduced degraders with indigenous microbes during bioaugmentation process.
Although cultivation-dependent approaches have isolated many PAH degraders, less than 1% of soil microorganisms are cultivable under arti cial conditions (Kaeberlein et al 2002, Vartoukian et al 2010. Stable isotope probing (SIP) is a well-developed approach to identify functional-yet-uncultivable PAH degraders in situ (Jiang et al 2018b). However, SIP relies on the isotope incorporation in the active degraders, e.g., DNA, RNA and proteins. More importantly, SIP cannot isolate the living degraders and is not suitable to provide bioaugmentation strains for practices. Recently, a novel technique, magneticnanoparticle mediated isolation (MMI), is innovated and aims at separating the active degraders from inert bacteria by magnetic gradient . This isotope-independent approach can isolate the active degraders from complex environment in a cost-effective manner and the separated microbes remain viable to be used in bioaugmentation (Sun et al 2021).
In the present study, both cultivation-dependent and MMI approaches were applied to isolate pyrene degraders from pyrene-contaminated soils at an abandoned steel plant site. Using cultivable pyrene degrader Sphingomonas sp. YT1005 and MMI-enriched active pyrene-degrading consortium in bioaugmentation, we analyzed the pyrene degradation e ciency, metabolic pathway, bacterial community structure and pyrene dioxygenase encoding genes. This work attempted to unravel the active pyrene degraders in soils, explore suitable inoculates in bioaugmentation for enhanced pyrene degradation, understand the bacterial interactions during bioaugmentation process, and explain the restriction factors causing bioaugmentation postpone or failure. As the rst study to separate and introduce the active pyrene-degrading microbes in soil bioaugmentation, this work can broaden our knowledge on the in uential factors affecting bioaugmentation performance and provide an effective approach isolating and applying the active degraders to improve bioaugmentation performance at PAHscontaminated sites.

Soil sample collection and analysis
Soil samples were collected at an abandoned site of Capital Steel Plant located in Shijingshan District, Beijing. This site was severely contaminated by PAHs and the PAHs concentration ranged from several to over 500 mg/kg. One kilogram of soils were obtained from the surface layer (0-20 cm) in the absence of PAHs contamination, stored at 4ºC and transferred to laboratory. After homogenization, stones and plant debris were removed. Soil geochemical properties were analyzed and listed in Table S1 (Electronic Supporting Information, ESI). There was no detectable PAHs in the collected soils.

Magnetic nanoparticle synthesis and functionalization
The synthesis of magnetic nanoparticles (MNPs) followed our published method (Zhang et al 2011). Brie y, 1 mL of FeCl 2 (1 mol/L in 2.0 M of HCl) and 2 mL of FeCl 3 (2 mol/L in 2.0 M of HCl) were gently mixed. Afterwards, 25 mL of NaOH (2 mol/L) was added drop by drop with vortex constant mixing for 30 min. The synthesized dark nanoparticles were then harvested by an external magnet for 10 min and washed with 30 mL of deionized water for several times until pH was neutral. The concentration of synthesized MNPs was 9.1 g/L. According to the optimal dosage for soil magnetic functionalization (Wang et al 2016), 0.91 mg of synthesized MNPs (0.1 mL) were added to 500 mg soils (dry weight) to prepare MNP-functionalized soils to isolate the active pyrene degrading consortium.

Isolation of cultivable pyrene degraders
Four grams of original soils were added into 100 mL of enrichment medium supplemented with pyrene (100 mg/L). The enrichment medium contained 5 g of beef extract, 10 g of peptone, 5 g of NaCl, 0.5 g of NaH 2 PO 4 and 1.5 g of Na 2 HPO 4 in 1.0 L of deionized water. After 7-day incubation with continuous shaking (180 rpm) at 28ºC in the dark, 1 mL of the suspension was transferred into 100 mL of fresh enrichment medium with increasing pyrene concentration (200, 300, 400 and 500 mg/L). Afterwards, the enriched suspension was diluted and spread onto a mineral medium agar plate supplemented with pyrene (500 mg/L) as the sole carbon source. One litre of mineral medium contained 0.5 g of KH 2 PO 4 , 0.5 g of NaCl, 0.1 g of CaCl 2 , 0.2 g of MgSO 4 ·7H 2 O, 0.5 g of FeSO 4 ·7H 2 O, 0.5 g of MnSO 4 and 1.5 g of (NH 4 ) 2 SO 4 (Zhang et al 2012). After incubation at 28ºC for 2 days in the dark, single colonies were picked from the plate and spread twice on fresh mineral medium agar plates supplemented with pyrene (500 mg/L) for puri cation.
The 16S rRNA genes of the isolated pyrene degraders were ampli ed by polymerase chain reaction (PCR) with a primer pair of 27F (5'-AGAGTTTGATCCTGGCTCAG-3') and 1492R (5'-GCTACCTTGTTACGACTT-3') (Guo et al 2010). PCR ampli cation consisted of 1 cycle of 95ºC for 5 min, 35 cycles of 95ºC for 30 s, 56ºC for 30 s and 72ºC for 1 min, and a nal extension at 72ºC for 10 min. The puri ed PCR products were used for sequence, and sequence similarity searches and alignments were performed using BLASTN at http://www.ncbi.nlm.nih.gov/BLAST/. All isolates have identical sequence sharing 98% similarity with Sphingomonas echinoides strain DSM 1805 (NCBI Accession No. MK934417), named as Sphingomonas sp. YT1005.

Pyrene degradation microcosms
To explore pyrene biodegradation performance in soils, ve treatments were carried out as follows: original soils treated with HgCl 2 (0.1%) and supplemented with pyrene as sterile control (OS_CK), original soils without pyrene (OS_NC), original soils supplemented with pyrene (OS_Pyr), MNP-functionalized soils without pyrene (MMI_NC) and MNP-functionalized soils supplemented with pyrene (MMI_Pyr). For OS_CK, OS_Pyr and MMI_Pyr treatments, pyrene was set at a nal concentration of 100 mg/kg, meeting with the average contamination level at the contaminated site. Each treatment was conducted in triplicates and incubated at room temperature for 30 days.  Table S2 (ESI) and PCR program was set as follows: an initial denaturation at 95ºC for 5 min, followed by 35 cycles of 95ºC denaturation for 30 s, annealing at different temperatures for each primer set (Table S2) for 30 s and extension at 72ºC for 1 min, and a nal extension at 72ºC for 10 min. The standard curves of Ct value and copy numbers for each gene were obtained by generating a ten-fold serial dilution of plasmids containing corresponding genes. The relative abundance of pyrene dioxygenase encoding genes was calculated as the ratio of the copy numbers of pyrene dioxygenase encoding genes to those of 16S rRNA genes.

Chemical analysis
To analyze soil pyrene contents, 5.0 g of soils were blended with 3 g of anhydrous Na 2 SO 4 , spiked with 10 µg of surrogate standards (phenanthrene-d10, AccuStandard®, Inc.), and added with 15 mL of extraction solvent (hexane:acetone = 1:1, v/v). After ultrasonication-assisted extraction for 20 min, the suspension was centrifuged for 3 min at 10,000 rpm. The extraction process was repeated twice, and the three fractions of supernatants were mixed and concentrated using a rotary evaporator to a nal volume of approximately 2 mL. It was further puri ed by solid phase extraction cartridges which were preconditioned with 4 mL of dichloromethane and 10 mL of hexane, and washed by 5 mL of dichloromethane:hexane (1:9, v/v). After soaking for 2 min, another 5 mL of dichloromethane:hexane (1:9, v/v) were added to elute pyrene and the ltrates were concentrated to completely dry under a gentle stream of nitrogen gas. Internal standards ( uorene, 100 µg/mL) were added to each ltrate prior to instrumental analysis.
The quantitative analysis of pyrene was performed using a gas chromatography mass spectrometry (GC-MS, Shimadzu, QP2010SE) equipped with a DB-5MS capillary column (30 m in length, 0.25 mm in diameter, 0.25 µm thickness) and a mass spectrometric detector. A total of 1.0 µL sample was injected in the splitless mode with a 5-min solvent delay time. The carrier gas was helium (99%) at a rate of 1.0 mL/min and the injector temperature was 280ºC. The GC oven temperature was set at 80ºC for 2 min, raised to 180ºC at a rate of 20ºC/min and maintained at 180 ºC for 5 min, and nally raised to 290ºC at a rate of 10ºC/min and maintained at 290ºC for 5 min. Electron impact source and selected ion monitoring (SIM) mode were used to identify individual metabolite. The ion source temperature was 230℃ and scanning range was from 45 to 450 atomic mass units (amu).
Five pyrene standard concentrations (5-500 µg/mL) were used to derive the calibration curve for pyrene. Mean recoveries of surrogate standards (phenanthrene-d10) in the present study ranged from 90-105% and the nal concentrations of pyrene were corrected by surrogate recovery. For pyrene metabolites, the molecular mass of each metabolite was searched against previous literatures and the database of PAH metabolites (National Institute of Standards and Technology, NIST). The chemical structure of each possible metabolite was con rmed by the pattern of fragment ions in the mass spectrum.

Statistical analysis
Statistical analysis was performed by SPSS 20.0 software. One-way analysis of variance (ANOVA) was used to compare the difference in pyrene contents and the relative abundance of 16S rRNA and pyrene dioxygenase encoding genes (p < 0.05). The correlation matrix was calculated and visualized using R ( The logarithmic nonlinear regression between pyrene degradation e ciency (P D , %) and the abundance of pyrene dioxygenase encoding genes (Pyrene g , copies/g soil) followed Eq. (1): Here, P 0 and α represents pyrene degradation e ciency in abiotic treatments and correlation slope, respectively.

Biodegradation and bioaugmentation performance
Pyrene degradation e ciencies in different treatments are illustrated in Fig. 1. In abiotic treatment (NC_Pyr), pyrene content did not show signi cant decrease throughout the incubation period. An acceptable pyrene degradation was achieved in OS_Pyr treatment, and pyrene degradation e ciency was 30.18% on day 10 and 44.24% on day 20.
After introducing the cultivable pyrene degraders (Sphingomonas sp. YT1005 in BA_Sph and CK_Sph treatments and MMI-isolated active pyrene-degrading consortium in BA_MFC and CK_MFC treatments) in bioaugmentation, pyrene degradation performance varied remarkably across treatments (Fig. 1). An unexpected postpone of pyrene degradation was found in BA_Sph and CK_Sph treatments that pyrene degradation e ciency was less than 20% after 20-day incubation, exhibiting no signi cant difference with that in abiotic sterile treatment (OS_CK, p > 0.05  Figure S1).

Pyrene degraders revealed by cultivation-dependent and cultivation-independent methods
In total, 8 isolates were obtained from mineral medium agar plates supplemented with pyrene.  2C).

Intra-correlation within soil bacterial community
To explain the mechanisms of bioaugmentation postpone in BA_Sph and CK_Sph treatments, we analyzed the bacterial intra-correlations by molecular ecological network (Fig. 4,

Metabolites and degradation pathway of pyrene
To understand pyrene degradation pathways in bioaugmentation, metabolites were analyzed during pyrene biodegradation process in CK_Sph and CK_MFC treatments. In total, ten metabolites were identi ed, and two pyrene degradation pathways were proposed ( Fig. 6 and S3, Table S3).
In CK_Sph treatment, the metabolite cis-4,5-pyrene dihydrodiol (m/z = 364, t R =19.82 min) might be generated from an initial pyrene oxidation at C-4 and C-5 positions by dioxygenase, designated as metabolite . Although the following metabolite phenanthrene-4-carboxylic acid (by ortho-cleavage and decarboxylation) was not detectable, its downstream metabolite was detected as dihydroxyphenanthrene In CK_MFC treatments, all ten metabolites are identi ed ( Fig. 6 and S3, and metabolite (m/z = 206, t R =20.87 min) was trans-2'-carboxybenzalpyruvate and salicylic acid, respectively. They were two downstream metabolites of 1-hydroxy-2-naphthoic acid (metabolite ). As metabolites and were only detectable in CK_MFC treatment, MMI-enriched active pyrene degrading consortium exhibited a unique salicylate pathway for pyrene degradation.
In the phthalate pathway, metabolites , , and were detected in all treatments but only on Day 10. Metabolite was detected on all sampling days in all treatments, whereas metabolites was only detectable on Day 20. These results suggested metabolites , , and were gradually metabolized into downstream metabolites ( and ), which accumulated during pyrene degradation process. In the salicylate pathway, metabolite was detected on days 20 and 30, also suggesting a signi cant accumulation of downstream metabolites.

Discussion
In the present study, Sphingomonas sp. YT1005 was identi ed as a cultivable pyrene degrader from soils at an abandoned steel plant site. Sphingomonas is a typical PAH-degrading microorganism ( After pyrene degradation, the relative abundance of Pseudonocardia, Streptomyces, Lysobacter and Thermomonas signi cantly increased. They are reported to be responsible for PAHs biodegradation by many previous studies. Lysobacter is linked to pyrene degradation  and Thermomonas has the capacity of degrading anthracene (Nzila et al 2018). Our ndings were consistent with these reports and indicated that predominant bacterial taxa in pyrene-contaminated soils were related to pyrene tolerance or degradation.
Both bacterial α-diversity and community composition were of no signi cant difference between OS_NC and MMI_NC treatments, indicating that MNP-functionalization had little impacts on soil bacterial community and did not affect pyrene biodegradation process . During the pyrene degradation process, the active pyrene degraders utilized pyrene and lost magnetism gradually, resulting in their enrichment in the magnetic-free fractions of MFCs (Zhang et al 2015). Accordingly, the community structure of MFC_Pyr was different from MFC_NC and the enriched bacterial lineages were the potential active pyrene degraders (Fig. 2B).
The active pyrene degrading consortium isolated by MMI included Gp16, Streptomyces, Pseudonocardia, Panacagrimonas, Methylotenera and Nitrospira (Fig. 3C). . However, no study has ever reported their involvement in pyrene degradation, and this work broadened our understanding on pyrene-degrading microbes, suggesting that numerous soil microbes have the capability to degrade pyrene.
The relative abundance of all pyrene dioxygenase genes was higher in OS_Pyr than OS_NC treatments, hinting the enrichment of pyrene degraders harboring dioxygenase genes with pyrene amendment. Additionally, their relative abundance was around 10-200 folds higher in MFC_Pyr than OS_Pyr, proving the successful isolation of the active pyrene degrading microbes in MFC fractions (Fig. 5A). Each pyrene dioxygenase gene exhibited a positive correlation with pyrene degradation e ciency (Table 1 and Fig. 1), attributing to the complex bacterial intra-correlations within the bacterial community. Pyrene degradation e ciency was consistent between OS_Pyr and CK_MFC treatments, proving that MMIenriched bacterial consortium had the same pyrene degradation capability comparing to indigenous microbiota. Accordingly, the enriched bacterial consortium consisting of Gp16, Streptomyces, Pseudonocardia, Panacagrimonas, Methylotenera and Nitrospira was responsible for in-situ pyrene degradation. However, Sphingomonas sp. exhibited negative correlations with those active pyrene degraders (Fig. 4), and the introduction of Sphingomonas sp. in BA_Sph treatments therefore inhibited their activities and consequently resulted in bioaugmentation postpone. This explanation was consistent with some previous studies that the correlation between pure-cultivation isolated and indigenous degraders affected the bioaugmentation performance. Non-active degrader Marmoricola LJ-33 signi cantly enhanced biphenyl degradation e ciency in soils by changing bacterial diversity in biphenyl metabolism (Li et al 2020). Another bioaugmentation by PAH-degrader Acinetobacter tandoii LJ-5 produced a signi cant increased phenanthrene mineralization, attributing to the altered diversity of the active phenanthrene degraders, instead of Acinetobacter tandoii LJ-5 itself . We then propose the mechanism explaining bioaugmentation performance regarding intra-correlations between cultivable and active pyrene degraders (Fig. 7). Ehanced bioaugmentation is expected when these two groups of bacteria exhibit positive intra-correlations, whereas negative intra-correlations between cultivable and active pyrene degraders consequently result in bioagumentation postpone or failure. As for neutral intra-correlations, the effectiveness and performance of bioaugmentation is non-deterministric and potentially dependent on other environmental variables.

Conclusion
This study investigated soil pyrene degraders via both cultivation-dependent and cultivation-dependent approaches. Sphingomonas sp. YT1005 was isolated via pure cultivation, and the active pyrene degrading consortium consisting of Gp16, Streptomyces, Pseudonocardia, Panacagrimonas, Methylotenera and Nitrospira was isolated by MMI. An unexpected postpone of pyrene degradation was found in bioaugmentation with Sphingomonas sp. YT1005, explained by its negative intra-correlations with the active pyrene degraders. In contrast, bioaugmentation with MMI-isolated pyrene degrading consortium signi cantly accelerated pyrene degradation in soils. It was supported by the increasing relative abundance of pyrene dioxygenase encoding genes (nidA, nidA3 and PAH-RHDα-GP) along with pyrene degradation e ciency. Further analysis of pyrene degradation pathway suggested that Sphingomonas sp. YT1005 only exhibited the phthalate pathway, whereas MMI-isolated pyrene degraders possessed both phthalate and salicylate pathways. This work broadens our vision on the actual pyrene degradation process and mechanisms in soils, suggesting that intra-correlations between the introduced degraders and the indigenous active degraders are key factors determining bioaugmentation performance. Isolating and reintroducing the active indigenous active degraders is a more promising strategy for bioaugmentation.
Declarations bioaugmentation treatment with sterile soils, pyrene and Sphingomonas sp. YT1005. BA_MFC refers to bioaugmentation treatment with original soils, pyrene and MMI-enriched pyrene degrading consortium.CK_MFC refers to bioaugmentation treatment with sterile soils, pyrene and MMI-enriched pyrene degrading consortium.   The co-occurrence molecular ecological network constructed based on core OTUs with the abundance > 0.5% and occurrence in over 10 of all 19 samples. Blue and red edges represent positive or negative correlations, respectively. Each point (node) stands for one OTU and the edges represent the correlations between connected OTUs. Only the modules with member number > 5 and correlated with cultivable or active pyrene-degraders were kept. OTU33525 represents the cultivable pyrene degrader Sphingomonas sp. YT1005. Core OTUs representing the active pyrene degraders include Nitrospira (OTU23472), Streptomyces (OTU29458), Panacagrimonas (OTU4395) and Methylotenera (OTU583). The other two active pyrene degraders, Gp16 (OTU33542) and Pseudonocardia (OTU29851), are not core OTUs in the molecular ecological network and not illustrated.

Figure 6
Pyrene degradation metabolites and pathways by the cultivable degraders Sphingomonas sp. YT1005 and MMI-enriched active pyrene degrading consortium. Dash arrows represent multiple metabolic steps.
Bracketed compounds are hypothetical metabolites not identi ed in the present study. Compounds in black color are metabolites identi ed in both NC_Sph and NC_MFC treatments, following the phthalate pathway. Compounds in red color are metabolites identi ed only in NC_MFC treatments, following the salicylate pathway.