Journal of Hazardous Materials 283 (2015) 35–43
Contents lists available at ScienceDirect
Journal of Hazardous Materials journal homepage: www.elsevier.com/locate/jhazmat
Pyrosequencing reveals the effect of mobilizing agents and lignocellulosic substrate amendment on microbial community composition in a real industrial PAH-polluted soil ˜ d,∗ S. Lladó a,b , S. Covino b , A.M. Solanas a , M. Petruccioli c , A. D’annibale c , M. Vinas a
Department of Microbiology, University of Barcelona, Diagonal 645, E-08028 Barcelona, Spain Institute of Microbiology, Academy of Sciences of the Czech Republic, Vídenská 1083, 142 20 Prague 4, Czech Republic c Department for Innovation in Biological, Agro-Food and Forest Systems [DIBAF], University of Tuscia, Via S. Camillo de Lellis snc, 01100 Viterbo, Italy d GIRO Joint Research Unit IRTA-UPC, Institute of Research and Technology Food and Agriculture [IRTA], Torre Marimon, E-08140 Caldes de Montbui, Spain b
h i g h l i g h t s • • • • •
Soil microbial community assessment through classical (MPN) and molecular tools (DGGE and pyrosequencing) is provided. A failure of exogenous white rot fungi to colonize the polluted soil is shown by DGGE and pyrosequencing. Surfactant Brij 30 hampers 4-ring PAHs degradation due to toxicity over Actinobacteria and Bacteroidetes populations. A high prevalence of Fusarium and Scedosporium populations is revealed during soil bioremediation. Cupriavidus, Mycobacterium and Chithinophagaceae are potential HMW–PAH degraders in the soil.
a r t i c l e
i n f o
Article history: Received 3 June 2014 Received in revised form 19 August 2014 Accepted 21 August 2014 Available online 16 September 2014 Keywords: Soil bioremediation White rot fungi (WRF) Creosote HMW–PAH Non-ionic surfactant Next generation sequencing (NGS)
a b s t r a c t Bacterial and fungal biodiversity throughout different biostimulation and bioaugmentation treatments applied to an industrial creosote-polluted soil were analyzed by means of polyphasic approach in order to gain insight into the microbial community structure and dynamics. Pyrosequencing data obtained from initial creosote polluted soil (after a biopiling step) revealed that Alpha and Gammaproteobacteria were the most abundant bacterial groups, whereas Fusarium and Scedosporium were the main fungal genera in the contaminated soil. At the end of 60-days laboratory scale bioremediation assays, pyrosequencing and DGGE data showed that (i) major bacterial community shifts were caused by the type of mobilizing agent added to the soil and, to a lesser extent, by the addition of lignocellulosic substrate; and (ii) the presence of the non-ionic surfactant (Brij 30) hampered the proliferation of Actinobacteria (Mycobacteriaceae) and Bacteroidetes (Chitinophagaceae) and, in the absence of lignocellulosic substrate, also impeded polycyclic aromatic hydrocarbons (PAHs) degradation. The results show the importance of implementing bioremediation experiments combined with microbiome assessment to gain insight on the effect of crucial parameters (e.g. use of additives) over the potential functions of complex microbial communities harbored in polluted soils, essential for bioremediation success. © 2014 Elsevier B.V. All rights reserved.
1. Introduction Microbial biodegradation is the main process occurring in natural decontamination processes and bioremediation treatments
∗ Corresponding author. Tel.: +34 934674040; fax: +34 934674042. E-mail addresses: [email protected] (S. Lladó), [email protected] (S. Covino), [email protected] (A.M. Solanas), [email protected] (M. Petruccioli), ˜ [email protected] (A. D’annibale), [email protected] (M. Vinas). http://dx.doi.org/10.1016/j.jhazmat.2014.08.065 0304-3894/© 2014 Elsevier B.V. All rights reserved.
exploiting such microbial activities represent a valuable alternative for the clean-up of matrices contaminated by recalcitrant pollutants, such as polycyclic aromatic hydrocarbons (PAHs) [1]. Furthermore, bacteria, yeasts and filamentous fungi showing remarkable PAH-degrading capabilities are ubiquitous in the terrestrial environment. In spite of the importance of resident microbial communities for success of bioremediation treatments [2], little is known about the shifts occurring in bacterial and fungal populations during reclamation of real historically polluted soils [3–5]. Moreover, when allochthonous microorganisms such
36
S. Lladó et al. / Journal of Hazardous Materials 283 (2015) 35–43
as white rot fungi (WRF) are inoculated in the polluted matrix for mycoaugmentation treatments, an in-depth analysis of the cooperative or antagonistic relationships existing between exogenously added fungi and the native microflora is essential [6]. In our previous study [7], a creosote-polluted soil, previously biotreated by dynamic biopiling and characterized by the presence of a highly recalcitrant PAH fraction, was further treated by biostimulation of the indigenous soil microbiota and mycoaugmentation with two WRF. The impact of single or combined soil amendments on both PAH depletion and resident microbial community evolution in the proposed bioremediation approaches were comparatively evaluated [7]. The low WRF colonization rates as well as the effects of soil supplementation with (i) mobilizing agents (MAs) and (ii) a lignocellulosic substrate (LS) on the autochthonous microbial communities were also discussed [7]. On an overall basis, biostimulating the soil by adding sterile LS promoted a significantly higher total petroleum hydrocarbon (TPH) and high molecular weight-PAHs (HMW–PAHs) removal than that achieved by mycoaugmentation with the two WRF. This prompted us to investigate the overall microbial processes occurring in soil during different bioremediation approaches, in order to improve our knowledge about native microbial communities and their role in polluted sites. As far as we know, the present work report for the first time the use of pyrosequencing to study both fungal and eubacterial native populations throughout different bioremediation strategies in a real PAH-polluted soil. 2. Materials and methods 2.1. Soil, materials and microorganisms A composite sample of an aged creosote-contaminated soil was collected after four months of biopiling at pilot-scale [8]
and henceforth referred to as initial soil. Its main properties were as follows [7]: TPH content, 2815 mg kg−1 ; pH 7.5; waterholding capacity, 33.7%; cultivable heterotrophic bacteria (CHB) and hydrocarbon-degrading bacteria (CHDB) were 1.3 × 107 and 2.3 × 106 MPN g−1 , respectively with a CHDB vs CHB ratio of 21%. Brij 30 (dodecyl tetraethylene glycol ether), soybean oil and manganese (II) sulfate (MnSO4 ) were purchased from SigmaAldrich (Saint Louis, MO, USA). Two different WRF strains, Trametes versicolor ATCC 42530 and Lentinus tigrinus CBS 577.79, were cultivated to produce inocula as described previously [7]. 2.2. Microcosm set-up Sterile soybean oil (SO) and Brij 30 were used at a final concentration of 4.5% (w/w) for soil pre-treatment as reported elsewhere [9]. The lignocellulosic substrate (LS) composed of a wheat straw/wheat bran mixture (80:20, w/w) was added to 16 cm × 3.5 cm test tubes and sterilized (121 ◦ C for 45 min) [10]. Hence, LS moisture was adjusted to 70% (w/w) with sterile deionized water, while MnSO4 (20 mg kg−1 ) was added as a manganese source when the treatment required it. Pre-colonized wheat (Triticum aestivum) seeds were used as a fungal inoculum for the LS at 5% (w/w) in those treatments where T. versicolor and L. tigrinus were assayed. Both inoculated and non-inoculated test tubes, the latter being referred to as biostimulation experiments (BS), were incubated for 7 d at 28 ◦ C. Regardless of the treatment, the LS mass amounted to 10% of the total dry weight for each microcosm. Subsequently, a layer of either mobilizing agent (MA) supplemented or bulk soil (25 g), the moisture content of which had previously been adjusted to 60% of its water-holding capacity (w/w), was added to the test tubes. The soil had been previously
Table 1 Treatment description, molecular analyses carried out and summary of chemical and microbiological results [7]. Treatmenta
DGGE 16S MPNb
Pyrosequencing 16S/ITSb
TPHc
4-Ring PAHs
5-Ring PAHs
CHDBd
qPCRe 16S
qPCR ITS
Initial soil
+
+
2815 ± 233
272 ± 10
117 ± 4
6.37 ± 0.07 (21%)
6.45 ± 0.25
6.56
IC IC + SO IC + Br30
+ + +
+ + +
1439 ± 51 1395 ± 30 1515 ± 179
96 ± 10.2 157 ± 8.1 214 ± 8.4
80 ± 3 71 ± 2 77 ± 3
6.21 ± 0.01 (6.9%) 6.32 ± 0.02 (0.9%) 3.07 ± 0.06 (0.004%)
7.34 ± 0.23 6.83 6.80 ± 0.1
6.92 6.78 6.74 ± 0.12
BS-LS BS-LS + SO BS-LS + Br30
+ + +
− − −
1077 ± 242 1098 ± 207 1255 ± 68
76 ± 3.3 77 ± 12 115 ± 12
62 ± 5 55 ± 6 62 ± 3
6.76 ± 0.16 (2.2%) 6.07 ± 0.05 (0.2%) 5.36 ± 0.07 (0.08%)
9.12 ± 0.1 9.46 ± 0.07 9.19 ± 0.14
8.79 ± 0.27 9.04 ± 0.05 8.67 ± 0.5
BS-LS + Mn2+ BS-LS + SO + Mn2+ BS-LS + Br30 + Mn2+
+ + +
− + +
1106 ± 26 810 ± 27 766 ± 27
63 ± 3.9 56 ± 7.8 58 ± 3.3
54 ± 1 37 ± 2 33 ± 2
6.74 ± 0.26 (2.6%) 6.03 ± 0.01 (0.15%) 5.20 ± 0.22 (0.05%)
9.31 ± 0.1 9.32 ± 0.34 8.86 ± 0.22
8.71 ± 0.25 9.24 8.51 ± 0.57
TV-LS TV-LS + SO TV − LS + Br30
− − −
− − −
1545 ± 153 1338 ± 204 1552 ± 29
77 ± 5.6 81 ± 9 127 ± 10.7
78 ± 6 68 ± 0.1 70 ± 3
6.11 ± 0.02 (0.62%) 6.98 ± 0.18 (3.8%) 5.65 ± 0.09 (0.14%)
10.2 ± 0.21 9.70 ± 0.1 8.92 ± 0.52
8.64 ± 0.19 9.62 ± 0.64 9.41 ± 0.21
TV-LS + Mn2+ TV-LS + SO + Mn2+ TV-LS + Br30 + Mn2+
− + −
− + −
1417 ± 155 1449 ± 65 1436 ± 60
78 ± 12 94 ± 11 114 ± 10
70 ± 8 62 ± 3 67 ± 3
6.06 ± 0.21 (0.24%) 7.71 ± 0.06 (16.8%) 6.07 ± 0.05 (0.65%)
9.44 ± 0.35 9.44 ± 0.38 9.27 ± 0.29
8.88 ± 0.34 9.23 ± 0.59 9.00 ± 0.35
LT-LS LT-LS + SO LT-LS + Br30
− − −
− − −
1396 ± 15 1467 ± 170 1578 ± 42
79 ± 6.5 91 ± 14 131 ± 14
70 ± 2 64 ± 9 70 ± 3
6.19 ± 0.37 (1.05%) 7.66 ± 0.17 (12%) 5.72 ± 0.20 (0.14%)
9.38 ± 0.06 9.62 ± 0.19 9.45 ± 0.06
8.94 ± 0.08 9.51 ± 0.16 8.92 ± 0.28
LT-LS + Mn2+ LT-LS + SO + Mn2+ LT-LS + Br30 + Mn2+
− + −
− + −
1147 ± 53 1093 ± 71 1260 ± 2
61 ± 6.7 72 ± 13 112 ± 3.7
58 ± 2 51 ± 2 69 ± 0.2
5.60 ± 0.03 (0.16%) 7.62 ± 0.24 (27.6%) 6.11 ± 0.01 (0.19%)
9.38 ± 0.16 9.64 ± 0.24 9.21 ± 0.7
8.96 ± 0.06 9.55 ± 0.35 8.66 ± 0.72
a b c d e
IC, incubation control; BS, biostimulation; TV, Trametes versicolor; LT, Lentinus tigrinus; LS, lignocellulosic substrate; SO, soybean oil; Br30, Brij 30; Mn2+ , manganese ions. Samples that underwent MPN–DGGE and/or pyrosequencing processing are displayed with a+ symbol. All concentrations are expressed as mg kg−1 of dry soil and data are the means of three independent experiments. Cultivable PAHs-degrading specialized bacteria (CHDB), expressed as Log MPN g−1 soil and CHDB/CHB percent ratios; data are the means of three independent experiments. 16SrRNA and ITS region gene copies quantified by qPCR, expressed as Log gene copies g−1 ; data are the means of three independent experiments.
S. Lladó et al. / Journal of Hazardous Materials 283 (2015) 35–43
amended by mixing it with an additional 10% (w/w) of noninoculated sterilized (121 ◦ C for 45 min) LS. All microcosms were incubated at 28 ◦ C for 60-d in the dark and their moisture content kept constant by the periodic addition of sterile deionized water. All experiments were carried out in three parallel replicates under non-axenic conditions. Non-amended controls were prepared and incubated as above and will henceforth be referred to as incubation controls (IC). Subsequent analyses were carried out on the upper soil layer and the results were normalized by taking into account the dilution effect of the LS mixed with the polluted soil. All treatments are detailed in Table 1. Procedures for analyzing the organic contaminants, biochemical determinations, CHB and CHDB most probable number (MPN) counts, and quantitative PCR assay of both the bacterial and fungal communities are previously described in Ref. [7] and detailed in Table 1. 2.3. DGGE molecular profiling Cultivable hydrocarbon-degrading bacteria (CHDB) were studied by means of DGGE and followed different approaches. Samples were selected as follows: (i) initial soil and controls (IC, IC + SO, IC + Br30); (ii) assessment of the native CHBD communities after soil biostimulation (BS-LS treatments); (iii) Effect of bioaugmentation over native CHBD populations. TV-LS + SO + Mn2+ , LT-LS + SO + Mn2+ were selected for sake of comparison with the biostimulated counterparts where PAH degradation was remarkable (BS-LS + SO + Mn2+ ). Samples for DNA extraction were collected from the highest positive dilutions from most-probable-number assays on microtiter plates grown in a PAH mixture (dibenzothiophene, fluorene, anthracene and phenanthrene [11]), at time 0 and after 60 d, and placed in sterile Eppendorf tubes stored at −20 ◦ C prior to analysis. To obtain DNA from microtiter plates, a composite sample of 1.6 mL containing 200 L of each replicate (n = 8) from the last positive dilution (n = 8) was centrifuged to harvest the microbial biomass. DNA from pellets was extracted with a bead-beating protocol using the Power Soil DNA extraction kit (MoBio Laboratories, Solano Beach, CA, USA), in accordance with the manufacturer’s instructions. Extracted DNA belonging to the same soil triplicate was pooled before further analyses. Primers and procedures are described in Ref. [12]. The nucleotide sequences identified in this study were deposited in the GenBank database under accession numbers JQ894738–JQ894777. 2.4. Multiplexed bacterial/fungal tag-encoded FLX amplicon pyrosequencing Treatments selected for barcoded 16S rRNA and ITS gene pyrosequencing analysis are shown in Table 1. In order to obtain deeper sequencing, only the most representative samples were selected based on the following criteria: (i) initial soil and controls (IC, IC + SO, IC + Br30); (ii) assessment of the native communities related to the presence of MAs and LS, in those samples with highest PAHs degradation (BS-LS + SO + Mn2+ , BS-LS + Br30 + Mn2+ ); (iii) effect of bioaugmentation over native populations (TV-LS + SO + Mn2+ , PT-LS + SO + Mn2+ ) as already mentioned in the previous section for MPN–DGGE samples. Samples for DNA extraction were collected from previously homogenized soil test tubes. Thus, a sample of 250 mg of each microcosm was extracted with a bead-beating protocol using the Power Soil DNA extraction kit (MoBio Laboratories, Solano Beach, CA, USA), in accordance with the manufacturer’s instructions. Extracted DNA was quantified using Quant-iT Picogreen dsDNA Kit (Invitrogen; Carlsbad, CA, USA). Extracts belonging to the same soil triplicate
37
were pooled prior to subsequent analyses [4]. A diluted DNA extract (1:10) in ultra-pure water was used as a template for PCR. Each DNA composite sample from triplicate microcosms was amplified separately with both 16S rRNA (eubacteria) and ITS (fungi) gene primers containing unique multiplex identifier (MID) tags. MID1 through MID6, MID25 and MID26 from the extended MID set recommended by Roche Diagnostics [13] were used for ITS gene amplification (fungi). In addition, MID9 through MID14, MID7 and MID16 were used for 16S rRNA gene amplification (eubacteria). Each forward primer began at the 5 end with the primer adaptor A (5 -CCATCTCATCCCTGCGTGTCTCCGAC3 ), followed by the library key sequence (5 -TCAG-3 ), the selected MID sequence, and the template sequence. Reverse primers were designed by replacing adaptor A with adaptor B (5 -CCTATCCCCTGTGTGCCTTGGCAGTC-3 ), without key and MID sequences. The template specific to the 16S rRNA gene amplification were the forward primer 341F (5 -CCTACGGGAGGCAGCAG-3 ) [14] and the reverse primer 802R (5 -TACCAGGGTATCTAATCC-3 ) [15], whilst primer pair ITS1F (5 -CTTGGTCATTTAGAGGAAGTAA3 ) and ITS2 (5 -GCTGCGTTCTTCATCGATGC-3 ) [16] were used to perform the fungal amplicon libraries. Reactions were carried out following the same PCR program used for the DGGE analyses, as described above and purified with a PCR clean-up system (Promega, WI, USA) and eluted in 50 L of ultrapure water. The DNA concentration of pooled amplicons was then measured using Quant-iT Picogreen dsDNA Kit (Invitrogen, Carlsbad, CA, USA) prior to combining them all (16S and ITS libraries) into a single sample at a concentration suitable for the pyrosequencing protocol. The sample was finally submitted to the Genomic Department from Parc Científic de Barcelona (University of Barcelona) for sequencing, using the 454 Life Sciences Titanium Platform (Roche Diagnostics, Branford, CT, USA).
2.5. 454-Pyrosequencing data analysis Trimming of the 16S barcoded sequences into libraries was carried out using QIIME software version 1.7.0 [17]. This resulted in eight distinct datasets. The following steps were performed using QIIME: Denoising using Denoiser [18]; Operational taxonomic unit (OTU) picking using uclust [19]; sequence alignment using PyNAST [20]; Chimera detection using ChimeraSlayer [21]; Taxonomy assignment of individual datasets using the RDP Classifier 2.2 with a bootstrap cutoff of 80% [22]; Alpha-diversity analyses using QIIME 1.7.0; Beta-diversity analyses using UniFrac [23], including dendograms constructed using thetayc distances among samples. Singleton OTUs were removed prior to Alpha and Beta-diversity analyses. To eliminate the effect of the sampling effort, sub-sampling was performed normalizing the number of sequences of all samples to those of the sample with lowest number of sequences. Segregation of the ITS barcoded sequences into libraries was carried out using QIIME software version 1.7.0. This again resulted in eight distinct datasets. Denoising and OTU picking was performed as aforementioned but with ITS sequences the alignment is not possible due to the high variance existing between them. Chimera check, removing of singleton OTUs, sub-sampling and alpha and beta-diversity analyses were also performed as described above. UNITE database was used in order to taxonomically classify the ITS sequences. Analysis of UNITE output files was performed using MEGAN software version 4.0 [24]. Data from pyrosequencing datasets were submitted to the Sequence Read Archive of the National Center for Biotechnology Information (NCBI) under the study accession number SRA051395.
38
S. Lladó et al. / Journal of Hazardous Materials 283 (2015) 35–43
3. Results 3.1. Bacterial diversity The bacterial sequences obtained by pyrosequencing clustered into 8987 OTUs at a 97% similarity threshold. 16S rRNA gene barcoded pyrosequencing analysis revealed high bacterial diversity in the initial soil (Table 2), where Proteobacteria and Actinobacteria were the dominant phyla (Fig. 1A) making up 62% and 16% of all bacterial OTUs, respectively. Among Proteobacteria phylum, Alpha and Gammaproteobacteria were identified as the most important classes in the initial soil (Fig. 1B), accounting to 68% and 19% of all the Proteobacteria sequences. In addition, Sphingomonadaceae (Alphaproteobacteria) was the most abundant bacterial family detected in the initial soil through pyrosequencing, making up 19% of the total sequence diversity (Supplementary information Table S1). After 60 days of treatment, the Shannon diversity index in IC microcosms did not show significant variation in terms of bacterial diversity with respect to the initial soil (Table 2). Nevertheless, the addition of both MAs led to a remarkable shift in the total bacterial population; in particular, a dramatic decline in bacterial biodiversity and a substantial variation with respect to the initial soil condition were observed upon Brij 30 supplementation (IC + Br30), as reported in Table 2 and in the cluster tree in Fig. 2A, respectively. In particular, no OTUs belonging to Actinobacteria were detected (accounting below 0.0001% of relative abundance) in Brij 30-supplemented IC microcosms (Fig. 1A). Conversely, Proteobacteria appeared to thrive in soil amended with both MA (IC + SO and IC + Br30), this being more evident in the presence of Brij 30 where their relative abundance reached 83% of the whole bacterial sequences. Soybean oil had a positive impact on Sphingomonadaceae-based phylotypes, which accounted for 34% of all 16S rDNA sequences (Supplementary information Table S2). Pyrosequencing analysis also showed that the addition of LS boosted bacterial biodiversity up with respect to the IC + SO and IC + Br30 treatments. A higher relative importance of the Xanthomonadaceae (16%) and Burkholderiaceae (11%) families was detected in the BS-LS + SO + Mn2+ and BS-LS + Br30 + Mn2+ treatments, respectively, although members of the Sphingomonadaceae
family became less abundant (Supplementary information Table S1). Consistently with what reported above for IC microcosms, Brij 30 had also a negative impact on Actinobacteria abundance even in the presence of LS. 16SrDNA pyrosequencing libraries showed substantially similar biodiversity profiles either in the presence or in the absence of the bioaugmented WRF (Fig. 1A). 3.2. Potential PAHs-degrading bacteria MPN–DGGE studies (Fig. 3) of CHBD showed that potential PAHs degraders, such as Alphaproteobacteria (Rhizobiaceae and Bradyrhizobiaceae) and Actinobacteria (Mycobacterium), were present in the initial soil (Table 3). New bands belonging to Beta and Gammaproteobacteria classes (i.e. Comamonadaceae and Xanthomonadaceae, respectively) appeared in the IC treatment, where no MA was added to the soil. A visible shift in bacterial diversity was detected in IC + Br30 when compared to the IC and IC + SO MPN–DGGE profiles. Furthermore, the addition of SO promoted the occurrence of band 9 (B9 in Fig. 3), a phylotype closely related to the genus Agrobacterium (Alphaproteobacteria). B9 coincided with band 29 (B29) in the BS-LS + SO + Mn2+ treatment. On the other hand, LS addition combined with the MA Brij 30 (BS-LS + Br30 + Mn2+ ) apparently stimulated the growth of Cupriavidus (Betaproteobacteria) (B34 in Fig. 3). Moreover, the addition of LS promoted the appearance of a phylotype belonging to Alcaligenaceae family (Betaproteobacteria), closely related to Pigmentiphaga genus. 3.3. Fungal diversity The fungal sequences obtained by pyrosequencing clustered into 3309 OTUs at a 97% similarity threshold. ITS fungal region barcoded pyrosequencing analysis revealed a high fungal diversity in the initial soil (Table 2). Fusarium and Scedosporium were the two predominant fungal genera in the soil at the end of the biopiling step, representing 23% and 25% of the total ITS sequences, respectively. Regardless of the treatment applied, Fusarium became predominant in soil after 60 d, this being particularly evident in IC + Brij30 and IC + SO where its relative abundance was >90%, while
Table 2 Estimated richness, diversity and sample coverage for 16S rRNA gene and ITS1 libraries of creosote polluted soil microcosms. Treatment 16S Initial soil IC IC + SO IC + Br30 BS-LS + SO + Mn BS-LS + Br30 + Mn TV-LS + SO + Mn LT-LS + SO + Mn ITS InitialSoil IC IC + SO IC + Br30 BS-LS + SO + Mn BS-LS + Br30 + Mn TV-LS + SO + Mn LT-LS + SO + Mn a b c d e
NSa
OTUsb
Chao1c
Shannond
ESCe
18,137 11,390 5304 2557 1984 1591 2030 27,657
3339 2630 1017 554 728 546 746 4613
8156 (7580; 8810) 6857 (6297; 7503) 2547 (2230; 2946) 1515 (1249; 1881) 2122 (1795; 2550) 1885 (1515; 2395) 2034 (1743; 2412) 11,312 (10,611; 12,093)
5.52 (5.47; 5.55) 5.64 (5.59; 5.69) 4.82 (4.75; 4.88) 4.62 (4.54; 4.71) 5.24 (5.14; 5.34) 4.89 (4.78; 5.00) 5.10 (5.00; 5.20) 5.67 (5.63; 5.70)
0.88 0.85 0.88 0.86 0.74 0.75 0.73 0.90
11,712 – 2067 1055 225 173 2195 1213
2167 – 507 314 117 93 608 367
4.76 (4.71; 4.81) – 4.23 (4.14; 4.34) 4.16 (4.03; 4.29) 4.09 (3.90; 4.28) 3.94 (3.73; 4.14) 4.83 (4.74; 4.92) 4.55 (4.44; 4.66)
0.86 0.83 0.77 0.58 0.55 0.80 0.78
7729 (6891;8715) – 1443 (1189; 1792) 1223 (906; 1711) 563 (339; 1013) 458 (263; 879) 2044 (1665; 2560) 1388 (1044; 1905)
Number of sequences for each library. Calculated with QIIME at the 3% distance level. Chao1 richness index calculated using QIIME at the 3% distance level (values in brackets are 95% confidence intervals). Shannon diversity index calculated using QIIME at the 3% distance level (values in brackets are 95% confidence intervals). Estimated sample coverage: Cx = 1 − (Nx/n), where Nx is the number of unique sequences and n is the total number of sequences.
S. Lladó et al. / Journal of Hazardous Materials 283 (2015) 35–43
39
Fig. 1. Biodiversity of eubacterial phyla (A) and Proteobacteria classes (B) expressed as relative OTUs abundance (%), in the initial soil and in different 60-d-old microcosms based on the classification of partial 16S rRNA sequences, using RDP-classifier. (TV, Trametes versicolor; LT, Lentinus tigrinus; BS, biostimulation experiments; LS, lignocellulosic substrate; SO, soybean oil; Br30, Brij 30; Mn2+ , manganese ions).
Scedosporium dropped below 5% of the total relative abundance (Fig. 4). Generally, fungal community was less affected by MAs when LS was added as soil amendant (Fig. 4) and Fusarium was still the predominant genus, representing 54–86% of all sequences. The relative abundance of the genus Scedosporium (9–36%) in the presence of LS was invariably higher than in IC (Fig. 4). Other fungal genera were present in soil at a much lesser extent (Fig. 4). In WRF-bioaugmented microcosms, T. versicolor ITS phylotypes were detected at low percentages while no L. tigrinus sequences were found, thus implying that the exogenously added fungi were unable to thrive in soil (Fig. 4). Fusarium (52–65%) and Scedosporium (28–38%) were the most predominant genera also in the bioaugmentation treatments.
purpose, we used 454-pyrosequencing to study the most predominant microbial taxa, as well as their shifts, in a PAH contaminated soil which underwent biostimulation and bioaugmentation treatments. Furthermore, in order to find bacterial key players that may be involved in the degradation of highly recalcitrant PAHs, a strategy involving DGGE analysis of viable hydrocarbonoclastic (CHDB) bacteria from MPN plates was selected. We chose this approach due to low CHDB/CHB ratio in DNA directly extracted from soil, which hampered the DGGE visualization and successful band excising/sequencing of the DNA of CHDB. Aware of the possible underestimation of hydrocarbonoclastic bacteria diversity, we deliberately chose this approach in an attempt to enhance the probability to obtain CHDB sequences. Indeed, the low complexity of the MPN–DGGE profiles allowed 42 bands to be excised and successfully sequenced.
4. Discussion 4.1. Bacterial community structure in pristine soil The main objective of the present work was to reach a more in-depth understanding of the biological processes such as potential fungal–bacterial interactions over different bioremediation approaches in a real industrial contaminated soil [7]. With this
Proteobacteria and Actinobacteria were the predominant phyla in the initial soil. There is substantial variability in the abundance of members of different phyla in different soils, but Proteobacteria
40
S. Lladó et al. / Journal of Hazardous Materials 283 (2015) 35–43
Fig. 2. Thetayc cluster tree showing the relationship of bacterial (A) and fungal (B) communities in the different microcosms to one another based on pyrosequencing libraries. The scale bar represents the distance between clusters in Thetayc units. (TV, Trametes versicolor; LT, Lentinus tigrinus; BS, biostimulation experiments; LS, lignocellulosic substrate; SO, soybean oil; Br30, Brij 30; Mn2+ , manganese ions).
and Actinobacteria are commonly present at very high levels [25]. Moreover, the high relative abundances of Alpha and Gammaproteobacteria in the initial soil of the present work are consistent with the results reported in a previous work targeting the same creosote-polluted soil [5], where the same phyla were the most abundant at the end of (diverse) biostimulation treatments. Concerning Alphaproteobacteria, we also reported high abundance of the Sphingomonadaceae family in the initial polluted soil. Due to their broad presence in polluted sites and their wide range of metabolic pathways, members of the Sphingomonadaceae family are considered to be powerful PAH degraders of the soil environment [26,27]. Although Sphingomonas genus made up 51% of the
Fig. 3. Denaturing Gradient Gel Electrophoresis profiles of PCR-amplified 16S rRNA gene fragments (V3–V5 regions) of last positive dilution of MPN plates. The progressive numbers disposed at the left side of each lane correspond to the codes assigned to the excised and sequenced. Gel was carried out at a denaturing concentration from 40% to 60%. (TV, Trametes versicolor; LT, Lentinus tigrinus; BS, biostimulation experiments; LS, lignocellulosic substrate; SO, soybean oil; Br30, Brij 30; Mn2+ , manganese ions).
total family (Supplementary information Table S2), no members of this genus were detected in the MPN–DGGE analyses of the initial soil, highlighting the inconsistency between cultivation-dependent approaches and molecular methods. However, the presence of potential PAHs degraders in the initial soil was also confirmed by the MPN–DGGE profiles which disclosed the presence of known potential hydrocarbonoclastic genera such as Agrobacterium, Bradyrhizobium, Rhizobium and Mycobacterium. This result suggests that a processes of acclimatization to the polluted environment occurred. In this regards, it is well known that acclimatization of microbial communities to the soil pollutants is a key factor in increasing degradation [28]. Therefore, a significant part of the bioremediation treatments
Aspergillus
Inial Soil
Chaetomium Cosmopora IC + SO Fusarium Hansfordia IC + Br30
Hebeloma Hypocraceae Malasezzia
BS-LS+SO+Mn
Morerella Muscoda
BS-LS+Br30+Mn
Nectriaceae Peziza TV-LS+SO+Mn
Scedosporium Sordariales Trametes
LT-LS+SO+Mn
Others 0%
10%
20%
30%
40%
50%
60%
70%
80%
90%
100%
Fig. 4. Biodiversity of fungal genera expressed as relative OTUs abundance (%), in the initial soil and in different 60-d-old microcosms based on the classification of partial ITS1 sequences, using UNITE database. (TV, Trametes versicolor; LT, Lentinus tigrinus; IC, incubation controls; BS, biostimulation experiments; LS, lignocellulosic substrate; SO, soybean oil; Br30, Brij 30; Mn2+ , manganese ions).
e
c
d
b
Alcaligenaceae () 98% 503 X – – – – – – – – 16S B41
Band detection (+) above 1% of relative intensity. Sequences were aligned against the GenBank database with the BLAST search alignment tool. Phylogenetic groups were defined by using the Ribosomal Data Project (RDP) Naive Bayesian Classifier [22]. Family is represented. ␣, , ␥ represent ␣-proteobacteria, -proteobacteria and ␥-proteobacteria, respectively. B1 = B9 = B29; B3 = B10 = B19 = B20 = B22 = B31;B5 = B38 = B42;B7 = B14; B8 = B11 = B17 = B21 = B25 = B28 = B32 = B37 = B40; B12 = B13 = B15 = B18 = B23 = B24 = B26 = B30 = B35; B16 = B27. Band detected by means of gel migration. Band not sequenced.
480 455 459 495 352 350 496 481 468 486 501 507 – – – X – – – – – – – – – – X X – – X – – – – – – – – – – – X X – – X X X – X – – – X X – X – – – – – – – – X X X – – – – – – – – X Xe X – – – – X – X – – – X – X – – – – – – – X X X – Xe – – – X X X X – – – – – – – –
LT SO TV SO BS Br30 BS SO BS IC Br30 IC SO IC Initial soil
16S B1d 16S B2 16S B3 16S B5 16S B6 16S B7 16S B8 16S B12 16S B16 16S B33 16S B34 16S B36
a
99% 99% 98% 100% 99% 99% 98% 98% 97% 100% 100% 100%
Agrobacterium tumefaciens (NR 041396.1) Bradyrhizobium liaoningense (NR 041785.1) Rhizobium oryzae (NR 044393.1) Mycobacterium monacense (NR 041723.1) Hydrogenophaga intermedia (NR 024856.1) Achromobacter denitrificans (NR 042021.1) Pseudoxanthomonas Mexicana (NR 025105.1) Pigmentiphaga kullae (NR 025112.1) Azoarcus indigens (NR 024851.1) Mycobacterium rutilim (NR 043761.1) Cupriavidus campinensis (NR 025137.1) Uncultured soil bacterium clone from a PAHs polluted soil (DQ907006) Bordetella himzii (NR 027537.1)
Closest organism in GenBank database (accession no.) Length (bp) Band detectiona Band
Table 3 Properties of DGGE bands: designations and accession numbers for the band sequences and levels of similarity to related organisms.
% similarityb
Phylogenetic groupc
Rhizobiaceae (␣) Bradyrhizobiaceae (␣) Rhizobiaceae (␣) Mycobacteriaceae (Actinobacteria) Comamonadaceae () Alcaligenaceae () Xanthomonadaceae (␥) Alcaligenaceae () Rhodocyclaceae () Mycobacteriaceae (Actinobacteria) Burkholderiaceae () –
S. Lladó et al. / Journal of Hazardous Materials 283 (2015) 35–43
41
assayed was aimed at exploiting the hydrocarbon-depleting ability of the resident microbiota. 4.2. Effect of surfactant amendment on bacterial community composition It is noteworthy that surfactants can increase the biodegradation rates of highly recalcitrant PAHs and are widely used in bioremediation studies [29]. In the present work, we found remarkable shifts in the soil bacterial community when MAs were added to the soil, especially related to the non-ionic surfactant Brij 30 (Fig. 2A). In addition, a toxic effect of the Brij 30 on the CHDB populations, as inferred by MPN counts, was previously observed [7]. Such a decrease in the size of the PAH-degrading population, as well as the shift in the diversity structure of the CHDB population (MPN–DGGE profiles), could explain the lower degradation rate of 4-ring PAHs detected when Brij 30 was added to the soil in the absence of LS [7]. Coincidently, pyrosequencing data reveals that both Actinobacteria (Mycobacteriaceae) and Bacteroidetes (Chitinophagaceae) populations are highly depleted after surfactant Brij 30 supplementation. Therefore, the obtained results from this diversified approach would indicate that both families may be directly related on 4-ring PAH biodegradation in the polluted soil. On the other hand, when SO was used as MA in non-amended soil (IC + SO), the community shifts observed by means of pyrosequencing were concomitant with higher CHB–MPN counts but lower HMW–PAH depletion levels suggesting that soybean oil was being used as a carbon source by the bacteria. However, higher rate of benzo(a)pyrene depletion were observed in IC + SO treatment and this coincided with the presence of the genus Agrobacterium (Rhizobiaceae family) according to MPN–DGGE. Bacteria in the Rhizobiaceae family are commonly found in polluted environments [30], but no evidence of benzo(a)pyrene degradation ability has been found in the literature. Moreover, the class of Rhizobiaceae was present at low percentages (
Contents lists available at ScienceDirect
Journal of Hazardous Materials journal homepage: www.elsevier.com/locate/jhazmat
Pyrosequencing reveals the effect of mobilizing agents and lignocellulosic substrate amendment on microbial community composition in a real industrial PAH-polluted soil ˜ d,∗ S. Lladó a,b , S. Covino b , A.M. Solanas a , M. Petruccioli c , A. D’annibale c , M. Vinas a
Department of Microbiology, University of Barcelona, Diagonal 645, E-08028 Barcelona, Spain Institute of Microbiology, Academy of Sciences of the Czech Republic, Vídenská 1083, 142 20 Prague 4, Czech Republic c Department for Innovation in Biological, Agro-Food and Forest Systems [DIBAF], University of Tuscia, Via S. Camillo de Lellis snc, 01100 Viterbo, Italy d GIRO Joint Research Unit IRTA-UPC, Institute of Research and Technology Food and Agriculture [IRTA], Torre Marimon, E-08140 Caldes de Montbui, Spain b
h i g h l i g h t s • • • • •
Soil microbial community assessment through classical (MPN) and molecular tools (DGGE and pyrosequencing) is provided. A failure of exogenous white rot fungi to colonize the polluted soil is shown by DGGE and pyrosequencing. Surfactant Brij 30 hampers 4-ring PAHs degradation due to toxicity over Actinobacteria and Bacteroidetes populations. A high prevalence of Fusarium and Scedosporium populations is revealed during soil bioremediation. Cupriavidus, Mycobacterium and Chithinophagaceae are potential HMW–PAH degraders in the soil.
a r t i c l e
i n f o
Article history: Received 3 June 2014 Received in revised form 19 August 2014 Accepted 21 August 2014 Available online 16 September 2014 Keywords: Soil bioremediation White rot fungi (WRF) Creosote HMW–PAH Non-ionic surfactant Next generation sequencing (NGS)
a b s t r a c t Bacterial and fungal biodiversity throughout different biostimulation and bioaugmentation treatments applied to an industrial creosote-polluted soil were analyzed by means of polyphasic approach in order to gain insight into the microbial community structure and dynamics. Pyrosequencing data obtained from initial creosote polluted soil (after a biopiling step) revealed that Alpha and Gammaproteobacteria were the most abundant bacterial groups, whereas Fusarium and Scedosporium were the main fungal genera in the contaminated soil. At the end of 60-days laboratory scale bioremediation assays, pyrosequencing and DGGE data showed that (i) major bacterial community shifts were caused by the type of mobilizing agent added to the soil and, to a lesser extent, by the addition of lignocellulosic substrate; and (ii) the presence of the non-ionic surfactant (Brij 30) hampered the proliferation of Actinobacteria (Mycobacteriaceae) and Bacteroidetes (Chitinophagaceae) and, in the absence of lignocellulosic substrate, also impeded polycyclic aromatic hydrocarbons (PAHs) degradation. The results show the importance of implementing bioremediation experiments combined with microbiome assessment to gain insight on the effect of crucial parameters (e.g. use of additives) over the potential functions of complex microbial communities harbored in polluted soils, essential for bioremediation success. © 2014 Elsevier B.V. All rights reserved.
1. Introduction Microbial biodegradation is the main process occurring in natural decontamination processes and bioremediation treatments
∗ Corresponding author. Tel.: +34 934674040; fax: +34 934674042. E-mail addresses: [email protected] (S. Lladó), [email protected] (S. Covino), [email protected] (A.M. Solanas), [email protected] (M. Petruccioli), ˜ [email protected] (A. D’annibale), [email protected] (M. Vinas). http://dx.doi.org/10.1016/j.jhazmat.2014.08.065 0304-3894/© 2014 Elsevier B.V. All rights reserved.
exploiting such microbial activities represent a valuable alternative for the clean-up of matrices contaminated by recalcitrant pollutants, such as polycyclic aromatic hydrocarbons (PAHs) [1]. Furthermore, bacteria, yeasts and filamentous fungi showing remarkable PAH-degrading capabilities are ubiquitous in the terrestrial environment. In spite of the importance of resident microbial communities for success of bioremediation treatments [2], little is known about the shifts occurring in bacterial and fungal populations during reclamation of real historically polluted soils [3–5]. Moreover, when allochthonous microorganisms such
36
S. Lladó et al. / Journal of Hazardous Materials 283 (2015) 35–43
as white rot fungi (WRF) are inoculated in the polluted matrix for mycoaugmentation treatments, an in-depth analysis of the cooperative or antagonistic relationships existing between exogenously added fungi and the native microflora is essential [6]. In our previous study [7], a creosote-polluted soil, previously biotreated by dynamic biopiling and characterized by the presence of a highly recalcitrant PAH fraction, was further treated by biostimulation of the indigenous soil microbiota and mycoaugmentation with two WRF. The impact of single or combined soil amendments on both PAH depletion and resident microbial community evolution in the proposed bioremediation approaches were comparatively evaluated [7]. The low WRF colonization rates as well as the effects of soil supplementation with (i) mobilizing agents (MAs) and (ii) a lignocellulosic substrate (LS) on the autochthonous microbial communities were also discussed [7]. On an overall basis, biostimulating the soil by adding sterile LS promoted a significantly higher total petroleum hydrocarbon (TPH) and high molecular weight-PAHs (HMW–PAHs) removal than that achieved by mycoaugmentation with the two WRF. This prompted us to investigate the overall microbial processes occurring in soil during different bioremediation approaches, in order to improve our knowledge about native microbial communities and their role in polluted sites. As far as we know, the present work report for the first time the use of pyrosequencing to study both fungal and eubacterial native populations throughout different bioremediation strategies in a real PAH-polluted soil. 2. Materials and methods 2.1. Soil, materials and microorganisms A composite sample of an aged creosote-contaminated soil was collected after four months of biopiling at pilot-scale [8]
and henceforth referred to as initial soil. Its main properties were as follows [7]: TPH content, 2815 mg kg−1 ; pH 7.5; waterholding capacity, 33.7%; cultivable heterotrophic bacteria (CHB) and hydrocarbon-degrading bacteria (CHDB) were 1.3 × 107 and 2.3 × 106 MPN g−1 , respectively with a CHDB vs CHB ratio of 21%. Brij 30 (dodecyl tetraethylene glycol ether), soybean oil and manganese (II) sulfate (MnSO4 ) were purchased from SigmaAldrich (Saint Louis, MO, USA). Two different WRF strains, Trametes versicolor ATCC 42530 and Lentinus tigrinus CBS 577.79, were cultivated to produce inocula as described previously [7]. 2.2. Microcosm set-up Sterile soybean oil (SO) and Brij 30 were used at a final concentration of 4.5% (w/w) for soil pre-treatment as reported elsewhere [9]. The lignocellulosic substrate (LS) composed of a wheat straw/wheat bran mixture (80:20, w/w) was added to 16 cm × 3.5 cm test tubes and sterilized (121 ◦ C for 45 min) [10]. Hence, LS moisture was adjusted to 70% (w/w) with sterile deionized water, while MnSO4 (20 mg kg−1 ) was added as a manganese source when the treatment required it. Pre-colonized wheat (Triticum aestivum) seeds were used as a fungal inoculum for the LS at 5% (w/w) in those treatments where T. versicolor and L. tigrinus were assayed. Both inoculated and non-inoculated test tubes, the latter being referred to as biostimulation experiments (BS), were incubated for 7 d at 28 ◦ C. Regardless of the treatment, the LS mass amounted to 10% of the total dry weight for each microcosm. Subsequently, a layer of either mobilizing agent (MA) supplemented or bulk soil (25 g), the moisture content of which had previously been adjusted to 60% of its water-holding capacity (w/w), was added to the test tubes. The soil had been previously
Table 1 Treatment description, molecular analyses carried out and summary of chemical and microbiological results [7]. Treatmenta
DGGE 16S MPNb
Pyrosequencing 16S/ITSb
TPHc
4-Ring PAHs
5-Ring PAHs
CHDBd
qPCRe 16S
qPCR ITS
Initial soil
+
+
2815 ± 233
272 ± 10
117 ± 4
6.37 ± 0.07 (21%)
6.45 ± 0.25
6.56
IC IC + SO IC + Br30
+ + +
+ + +
1439 ± 51 1395 ± 30 1515 ± 179
96 ± 10.2 157 ± 8.1 214 ± 8.4
80 ± 3 71 ± 2 77 ± 3
6.21 ± 0.01 (6.9%) 6.32 ± 0.02 (0.9%) 3.07 ± 0.06 (0.004%)
7.34 ± 0.23 6.83 6.80 ± 0.1
6.92 6.78 6.74 ± 0.12
BS-LS BS-LS + SO BS-LS + Br30
+ + +
− − −
1077 ± 242 1098 ± 207 1255 ± 68
76 ± 3.3 77 ± 12 115 ± 12
62 ± 5 55 ± 6 62 ± 3
6.76 ± 0.16 (2.2%) 6.07 ± 0.05 (0.2%) 5.36 ± 0.07 (0.08%)
9.12 ± 0.1 9.46 ± 0.07 9.19 ± 0.14
8.79 ± 0.27 9.04 ± 0.05 8.67 ± 0.5
BS-LS + Mn2+ BS-LS + SO + Mn2+ BS-LS + Br30 + Mn2+
+ + +
− + +
1106 ± 26 810 ± 27 766 ± 27
63 ± 3.9 56 ± 7.8 58 ± 3.3
54 ± 1 37 ± 2 33 ± 2
6.74 ± 0.26 (2.6%) 6.03 ± 0.01 (0.15%) 5.20 ± 0.22 (0.05%)
9.31 ± 0.1 9.32 ± 0.34 8.86 ± 0.22
8.71 ± 0.25 9.24 8.51 ± 0.57
TV-LS TV-LS + SO TV − LS + Br30
− − −
− − −
1545 ± 153 1338 ± 204 1552 ± 29
77 ± 5.6 81 ± 9 127 ± 10.7
78 ± 6 68 ± 0.1 70 ± 3
6.11 ± 0.02 (0.62%) 6.98 ± 0.18 (3.8%) 5.65 ± 0.09 (0.14%)
10.2 ± 0.21 9.70 ± 0.1 8.92 ± 0.52
8.64 ± 0.19 9.62 ± 0.64 9.41 ± 0.21
TV-LS + Mn2+ TV-LS + SO + Mn2+ TV-LS + Br30 + Mn2+
− + −
− + −
1417 ± 155 1449 ± 65 1436 ± 60
78 ± 12 94 ± 11 114 ± 10
70 ± 8 62 ± 3 67 ± 3
6.06 ± 0.21 (0.24%) 7.71 ± 0.06 (16.8%) 6.07 ± 0.05 (0.65%)
9.44 ± 0.35 9.44 ± 0.38 9.27 ± 0.29
8.88 ± 0.34 9.23 ± 0.59 9.00 ± 0.35
LT-LS LT-LS + SO LT-LS + Br30
− − −
− − −
1396 ± 15 1467 ± 170 1578 ± 42
79 ± 6.5 91 ± 14 131 ± 14
70 ± 2 64 ± 9 70 ± 3
6.19 ± 0.37 (1.05%) 7.66 ± 0.17 (12%) 5.72 ± 0.20 (0.14%)
9.38 ± 0.06 9.62 ± 0.19 9.45 ± 0.06
8.94 ± 0.08 9.51 ± 0.16 8.92 ± 0.28
LT-LS + Mn2+ LT-LS + SO + Mn2+ LT-LS + Br30 + Mn2+
− + −
− + −
1147 ± 53 1093 ± 71 1260 ± 2
61 ± 6.7 72 ± 13 112 ± 3.7
58 ± 2 51 ± 2 69 ± 0.2
5.60 ± 0.03 (0.16%) 7.62 ± 0.24 (27.6%) 6.11 ± 0.01 (0.19%)
9.38 ± 0.16 9.64 ± 0.24 9.21 ± 0.7
8.96 ± 0.06 9.55 ± 0.35 8.66 ± 0.72
a b c d e
IC, incubation control; BS, biostimulation; TV, Trametes versicolor; LT, Lentinus tigrinus; LS, lignocellulosic substrate; SO, soybean oil; Br30, Brij 30; Mn2+ , manganese ions. Samples that underwent MPN–DGGE and/or pyrosequencing processing are displayed with a+ symbol. All concentrations are expressed as mg kg−1 of dry soil and data are the means of three independent experiments. Cultivable PAHs-degrading specialized bacteria (CHDB), expressed as Log MPN g−1 soil and CHDB/CHB percent ratios; data are the means of three independent experiments. 16SrRNA and ITS region gene copies quantified by qPCR, expressed as Log gene copies g−1 ; data are the means of three independent experiments.
S. Lladó et al. / Journal of Hazardous Materials 283 (2015) 35–43
amended by mixing it with an additional 10% (w/w) of noninoculated sterilized (121 ◦ C for 45 min) LS. All microcosms were incubated at 28 ◦ C for 60-d in the dark and their moisture content kept constant by the periodic addition of sterile deionized water. All experiments were carried out in three parallel replicates under non-axenic conditions. Non-amended controls were prepared and incubated as above and will henceforth be referred to as incubation controls (IC). Subsequent analyses were carried out on the upper soil layer and the results were normalized by taking into account the dilution effect of the LS mixed with the polluted soil. All treatments are detailed in Table 1. Procedures for analyzing the organic contaminants, biochemical determinations, CHB and CHDB most probable number (MPN) counts, and quantitative PCR assay of both the bacterial and fungal communities are previously described in Ref. [7] and detailed in Table 1. 2.3. DGGE molecular profiling Cultivable hydrocarbon-degrading bacteria (CHDB) were studied by means of DGGE and followed different approaches. Samples were selected as follows: (i) initial soil and controls (IC, IC + SO, IC + Br30); (ii) assessment of the native CHBD communities after soil biostimulation (BS-LS treatments); (iii) Effect of bioaugmentation over native CHBD populations. TV-LS + SO + Mn2+ , LT-LS + SO + Mn2+ were selected for sake of comparison with the biostimulated counterparts where PAH degradation was remarkable (BS-LS + SO + Mn2+ ). Samples for DNA extraction were collected from the highest positive dilutions from most-probable-number assays on microtiter plates grown in a PAH mixture (dibenzothiophene, fluorene, anthracene and phenanthrene [11]), at time 0 and after 60 d, and placed in sterile Eppendorf tubes stored at −20 ◦ C prior to analysis. To obtain DNA from microtiter plates, a composite sample of 1.6 mL containing 200 L of each replicate (n = 8) from the last positive dilution (n = 8) was centrifuged to harvest the microbial biomass. DNA from pellets was extracted with a bead-beating protocol using the Power Soil DNA extraction kit (MoBio Laboratories, Solano Beach, CA, USA), in accordance with the manufacturer’s instructions. Extracted DNA belonging to the same soil triplicate was pooled before further analyses. Primers and procedures are described in Ref. [12]. The nucleotide sequences identified in this study were deposited in the GenBank database under accession numbers JQ894738–JQ894777. 2.4. Multiplexed bacterial/fungal tag-encoded FLX amplicon pyrosequencing Treatments selected for barcoded 16S rRNA and ITS gene pyrosequencing analysis are shown in Table 1. In order to obtain deeper sequencing, only the most representative samples were selected based on the following criteria: (i) initial soil and controls (IC, IC + SO, IC + Br30); (ii) assessment of the native communities related to the presence of MAs and LS, in those samples with highest PAHs degradation (BS-LS + SO + Mn2+ , BS-LS + Br30 + Mn2+ ); (iii) effect of bioaugmentation over native populations (TV-LS + SO + Mn2+ , PT-LS + SO + Mn2+ ) as already mentioned in the previous section for MPN–DGGE samples. Samples for DNA extraction were collected from previously homogenized soil test tubes. Thus, a sample of 250 mg of each microcosm was extracted with a bead-beating protocol using the Power Soil DNA extraction kit (MoBio Laboratories, Solano Beach, CA, USA), in accordance with the manufacturer’s instructions. Extracted DNA was quantified using Quant-iT Picogreen dsDNA Kit (Invitrogen; Carlsbad, CA, USA). Extracts belonging to the same soil triplicate
37
were pooled prior to subsequent analyses [4]. A diluted DNA extract (1:10) in ultra-pure water was used as a template for PCR. Each DNA composite sample from triplicate microcosms was amplified separately with both 16S rRNA (eubacteria) and ITS (fungi) gene primers containing unique multiplex identifier (MID) tags. MID1 through MID6, MID25 and MID26 from the extended MID set recommended by Roche Diagnostics [13] were used for ITS gene amplification (fungi). In addition, MID9 through MID14, MID7 and MID16 were used for 16S rRNA gene amplification (eubacteria). Each forward primer began at the 5 end with the primer adaptor A (5 -CCATCTCATCCCTGCGTGTCTCCGAC3 ), followed by the library key sequence (5 -TCAG-3 ), the selected MID sequence, and the template sequence. Reverse primers were designed by replacing adaptor A with adaptor B (5 -CCTATCCCCTGTGTGCCTTGGCAGTC-3 ), without key and MID sequences. The template specific to the 16S rRNA gene amplification were the forward primer 341F (5 -CCTACGGGAGGCAGCAG-3 ) [14] and the reverse primer 802R (5 -TACCAGGGTATCTAATCC-3 ) [15], whilst primer pair ITS1F (5 -CTTGGTCATTTAGAGGAAGTAA3 ) and ITS2 (5 -GCTGCGTTCTTCATCGATGC-3 ) [16] were used to perform the fungal amplicon libraries. Reactions were carried out following the same PCR program used for the DGGE analyses, as described above and purified with a PCR clean-up system (Promega, WI, USA) and eluted in 50 L of ultrapure water. The DNA concentration of pooled amplicons was then measured using Quant-iT Picogreen dsDNA Kit (Invitrogen, Carlsbad, CA, USA) prior to combining them all (16S and ITS libraries) into a single sample at a concentration suitable for the pyrosequencing protocol. The sample was finally submitted to the Genomic Department from Parc Científic de Barcelona (University of Barcelona) for sequencing, using the 454 Life Sciences Titanium Platform (Roche Diagnostics, Branford, CT, USA).
2.5. 454-Pyrosequencing data analysis Trimming of the 16S barcoded sequences into libraries was carried out using QIIME software version 1.7.0 [17]. This resulted in eight distinct datasets. The following steps were performed using QIIME: Denoising using Denoiser [18]; Operational taxonomic unit (OTU) picking using uclust [19]; sequence alignment using PyNAST [20]; Chimera detection using ChimeraSlayer [21]; Taxonomy assignment of individual datasets using the RDP Classifier 2.2 with a bootstrap cutoff of 80% [22]; Alpha-diversity analyses using QIIME 1.7.0; Beta-diversity analyses using UniFrac [23], including dendograms constructed using thetayc distances among samples. Singleton OTUs were removed prior to Alpha and Beta-diversity analyses. To eliminate the effect of the sampling effort, sub-sampling was performed normalizing the number of sequences of all samples to those of the sample with lowest number of sequences. Segregation of the ITS barcoded sequences into libraries was carried out using QIIME software version 1.7.0. This again resulted in eight distinct datasets. Denoising and OTU picking was performed as aforementioned but with ITS sequences the alignment is not possible due to the high variance existing between them. Chimera check, removing of singleton OTUs, sub-sampling and alpha and beta-diversity analyses were also performed as described above. UNITE database was used in order to taxonomically classify the ITS sequences. Analysis of UNITE output files was performed using MEGAN software version 4.0 [24]. Data from pyrosequencing datasets were submitted to the Sequence Read Archive of the National Center for Biotechnology Information (NCBI) under the study accession number SRA051395.
38
S. Lladó et al. / Journal of Hazardous Materials 283 (2015) 35–43
3. Results 3.1. Bacterial diversity The bacterial sequences obtained by pyrosequencing clustered into 8987 OTUs at a 97% similarity threshold. 16S rRNA gene barcoded pyrosequencing analysis revealed high bacterial diversity in the initial soil (Table 2), where Proteobacteria and Actinobacteria were the dominant phyla (Fig. 1A) making up 62% and 16% of all bacterial OTUs, respectively. Among Proteobacteria phylum, Alpha and Gammaproteobacteria were identified as the most important classes in the initial soil (Fig. 1B), accounting to 68% and 19% of all the Proteobacteria sequences. In addition, Sphingomonadaceae (Alphaproteobacteria) was the most abundant bacterial family detected in the initial soil through pyrosequencing, making up 19% of the total sequence diversity (Supplementary information Table S1). After 60 days of treatment, the Shannon diversity index in IC microcosms did not show significant variation in terms of bacterial diversity with respect to the initial soil (Table 2). Nevertheless, the addition of both MAs led to a remarkable shift in the total bacterial population; in particular, a dramatic decline in bacterial biodiversity and a substantial variation with respect to the initial soil condition were observed upon Brij 30 supplementation (IC + Br30), as reported in Table 2 and in the cluster tree in Fig. 2A, respectively. In particular, no OTUs belonging to Actinobacteria were detected (accounting below 0.0001% of relative abundance) in Brij 30-supplemented IC microcosms (Fig. 1A). Conversely, Proteobacteria appeared to thrive in soil amended with both MA (IC + SO and IC + Br30), this being more evident in the presence of Brij 30 where their relative abundance reached 83% of the whole bacterial sequences. Soybean oil had a positive impact on Sphingomonadaceae-based phylotypes, which accounted for 34% of all 16S rDNA sequences (Supplementary information Table S2). Pyrosequencing analysis also showed that the addition of LS boosted bacterial biodiversity up with respect to the IC + SO and IC + Br30 treatments. A higher relative importance of the Xanthomonadaceae (16%) and Burkholderiaceae (11%) families was detected in the BS-LS + SO + Mn2+ and BS-LS + Br30 + Mn2+ treatments, respectively, although members of the Sphingomonadaceae
family became less abundant (Supplementary information Table S1). Consistently with what reported above for IC microcosms, Brij 30 had also a negative impact on Actinobacteria abundance even in the presence of LS. 16SrDNA pyrosequencing libraries showed substantially similar biodiversity profiles either in the presence or in the absence of the bioaugmented WRF (Fig. 1A). 3.2. Potential PAHs-degrading bacteria MPN–DGGE studies (Fig. 3) of CHBD showed that potential PAHs degraders, such as Alphaproteobacteria (Rhizobiaceae and Bradyrhizobiaceae) and Actinobacteria (Mycobacterium), were present in the initial soil (Table 3). New bands belonging to Beta and Gammaproteobacteria classes (i.e. Comamonadaceae and Xanthomonadaceae, respectively) appeared in the IC treatment, where no MA was added to the soil. A visible shift in bacterial diversity was detected in IC + Br30 when compared to the IC and IC + SO MPN–DGGE profiles. Furthermore, the addition of SO promoted the occurrence of band 9 (B9 in Fig. 3), a phylotype closely related to the genus Agrobacterium (Alphaproteobacteria). B9 coincided with band 29 (B29) in the BS-LS + SO + Mn2+ treatment. On the other hand, LS addition combined with the MA Brij 30 (BS-LS + Br30 + Mn2+ ) apparently stimulated the growth of Cupriavidus (Betaproteobacteria) (B34 in Fig. 3). Moreover, the addition of LS promoted the appearance of a phylotype belonging to Alcaligenaceae family (Betaproteobacteria), closely related to Pigmentiphaga genus. 3.3. Fungal diversity The fungal sequences obtained by pyrosequencing clustered into 3309 OTUs at a 97% similarity threshold. ITS fungal region barcoded pyrosequencing analysis revealed a high fungal diversity in the initial soil (Table 2). Fusarium and Scedosporium were the two predominant fungal genera in the soil at the end of the biopiling step, representing 23% and 25% of the total ITS sequences, respectively. Regardless of the treatment applied, Fusarium became predominant in soil after 60 d, this being particularly evident in IC + Brij30 and IC + SO where its relative abundance was >90%, while
Table 2 Estimated richness, diversity and sample coverage for 16S rRNA gene and ITS1 libraries of creosote polluted soil microcosms. Treatment 16S Initial soil IC IC + SO IC + Br30 BS-LS + SO + Mn BS-LS + Br30 + Mn TV-LS + SO + Mn LT-LS + SO + Mn ITS InitialSoil IC IC + SO IC + Br30 BS-LS + SO + Mn BS-LS + Br30 + Mn TV-LS + SO + Mn LT-LS + SO + Mn a b c d e
NSa
OTUsb
Chao1c
Shannond
ESCe
18,137 11,390 5304 2557 1984 1591 2030 27,657
3339 2630 1017 554 728 546 746 4613
8156 (7580; 8810) 6857 (6297; 7503) 2547 (2230; 2946) 1515 (1249; 1881) 2122 (1795; 2550) 1885 (1515; 2395) 2034 (1743; 2412) 11,312 (10,611; 12,093)
5.52 (5.47; 5.55) 5.64 (5.59; 5.69) 4.82 (4.75; 4.88) 4.62 (4.54; 4.71) 5.24 (5.14; 5.34) 4.89 (4.78; 5.00) 5.10 (5.00; 5.20) 5.67 (5.63; 5.70)
0.88 0.85 0.88 0.86 0.74 0.75 0.73 0.90
11,712 – 2067 1055 225 173 2195 1213
2167 – 507 314 117 93 608 367
4.76 (4.71; 4.81) – 4.23 (4.14; 4.34) 4.16 (4.03; 4.29) 4.09 (3.90; 4.28) 3.94 (3.73; 4.14) 4.83 (4.74; 4.92) 4.55 (4.44; 4.66)
0.86 0.83 0.77 0.58 0.55 0.80 0.78
7729 (6891;8715) – 1443 (1189; 1792) 1223 (906; 1711) 563 (339; 1013) 458 (263; 879) 2044 (1665; 2560) 1388 (1044; 1905)
Number of sequences for each library. Calculated with QIIME at the 3% distance level. Chao1 richness index calculated using QIIME at the 3% distance level (values in brackets are 95% confidence intervals). Shannon diversity index calculated using QIIME at the 3% distance level (values in brackets are 95% confidence intervals). Estimated sample coverage: Cx = 1 − (Nx/n), where Nx is the number of unique sequences and n is the total number of sequences.
S. Lladó et al. / Journal of Hazardous Materials 283 (2015) 35–43
39
Fig. 1. Biodiversity of eubacterial phyla (A) and Proteobacteria classes (B) expressed as relative OTUs abundance (%), in the initial soil and in different 60-d-old microcosms based on the classification of partial 16S rRNA sequences, using RDP-classifier. (TV, Trametes versicolor; LT, Lentinus tigrinus; BS, biostimulation experiments; LS, lignocellulosic substrate; SO, soybean oil; Br30, Brij 30; Mn2+ , manganese ions).
Scedosporium dropped below 5% of the total relative abundance (Fig. 4). Generally, fungal community was less affected by MAs when LS was added as soil amendant (Fig. 4) and Fusarium was still the predominant genus, representing 54–86% of all sequences. The relative abundance of the genus Scedosporium (9–36%) in the presence of LS was invariably higher than in IC (Fig. 4). Other fungal genera were present in soil at a much lesser extent (Fig. 4). In WRF-bioaugmented microcosms, T. versicolor ITS phylotypes were detected at low percentages while no L. tigrinus sequences were found, thus implying that the exogenously added fungi were unable to thrive in soil (Fig. 4). Fusarium (52–65%) and Scedosporium (28–38%) were the most predominant genera also in the bioaugmentation treatments.
purpose, we used 454-pyrosequencing to study the most predominant microbial taxa, as well as their shifts, in a PAH contaminated soil which underwent biostimulation and bioaugmentation treatments. Furthermore, in order to find bacterial key players that may be involved in the degradation of highly recalcitrant PAHs, a strategy involving DGGE analysis of viable hydrocarbonoclastic (CHDB) bacteria from MPN plates was selected. We chose this approach due to low CHDB/CHB ratio in DNA directly extracted from soil, which hampered the DGGE visualization and successful band excising/sequencing of the DNA of CHDB. Aware of the possible underestimation of hydrocarbonoclastic bacteria diversity, we deliberately chose this approach in an attempt to enhance the probability to obtain CHDB sequences. Indeed, the low complexity of the MPN–DGGE profiles allowed 42 bands to be excised and successfully sequenced.
4. Discussion 4.1. Bacterial community structure in pristine soil The main objective of the present work was to reach a more in-depth understanding of the biological processes such as potential fungal–bacterial interactions over different bioremediation approaches in a real industrial contaminated soil [7]. With this
Proteobacteria and Actinobacteria were the predominant phyla in the initial soil. There is substantial variability in the abundance of members of different phyla in different soils, but Proteobacteria
40
S. Lladó et al. / Journal of Hazardous Materials 283 (2015) 35–43
Fig. 2. Thetayc cluster tree showing the relationship of bacterial (A) and fungal (B) communities in the different microcosms to one another based on pyrosequencing libraries. The scale bar represents the distance between clusters in Thetayc units. (TV, Trametes versicolor; LT, Lentinus tigrinus; BS, biostimulation experiments; LS, lignocellulosic substrate; SO, soybean oil; Br30, Brij 30; Mn2+ , manganese ions).
and Actinobacteria are commonly present at very high levels [25]. Moreover, the high relative abundances of Alpha and Gammaproteobacteria in the initial soil of the present work are consistent with the results reported in a previous work targeting the same creosote-polluted soil [5], where the same phyla were the most abundant at the end of (diverse) biostimulation treatments. Concerning Alphaproteobacteria, we also reported high abundance of the Sphingomonadaceae family in the initial polluted soil. Due to their broad presence in polluted sites and their wide range of metabolic pathways, members of the Sphingomonadaceae family are considered to be powerful PAH degraders of the soil environment [26,27]. Although Sphingomonas genus made up 51% of the
Fig. 3. Denaturing Gradient Gel Electrophoresis profiles of PCR-amplified 16S rRNA gene fragments (V3–V5 regions) of last positive dilution of MPN plates. The progressive numbers disposed at the left side of each lane correspond to the codes assigned to the excised and sequenced. Gel was carried out at a denaturing concentration from 40% to 60%. (TV, Trametes versicolor; LT, Lentinus tigrinus; BS, biostimulation experiments; LS, lignocellulosic substrate; SO, soybean oil; Br30, Brij 30; Mn2+ , manganese ions).
total family (Supplementary information Table S2), no members of this genus were detected in the MPN–DGGE analyses of the initial soil, highlighting the inconsistency between cultivation-dependent approaches and molecular methods. However, the presence of potential PAHs degraders in the initial soil was also confirmed by the MPN–DGGE profiles which disclosed the presence of known potential hydrocarbonoclastic genera such as Agrobacterium, Bradyrhizobium, Rhizobium and Mycobacterium. This result suggests that a processes of acclimatization to the polluted environment occurred. In this regards, it is well known that acclimatization of microbial communities to the soil pollutants is a key factor in increasing degradation [28]. Therefore, a significant part of the bioremediation treatments
Aspergillus
Inial Soil
Chaetomium Cosmopora IC + SO Fusarium Hansfordia IC + Br30
Hebeloma Hypocraceae Malasezzia
BS-LS+SO+Mn
Morerella Muscoda
BS-LS+Br30+Mn
Nectriaceae Peziza TV-LS+SO+Mn
Scedosporium Sordariales Trametes
LT-LS+SO+Mn
Others 0%
10%
20%
30%
40%
50%
60%
70%
80%
90%
100%
Fig. 4. Biodiversity of fungal genera expressed as relative OTUs abundance (%), in the initial soil and in different 60-d-old microcosms based on the classification of partial ITS1 sequences, using UNITE database. (TV, Trametes versicolor; LT, Lentinus tigrinus; IC, incubation controls; BS, biostimulation experiments; LS, lignocellulosic substrate; SO, soybean oil; Br30, Brij 30; Mn2+ , manganese ions).
e
c
d
b
Alcaligenaceae () 98% 503 X – – – – – – – – 16S B41
Band detection (+) above 1% of relative intensity. Sequences were aligned against the GenBank database with the BLAST search alignment tool. Phylogenetic groups were defined by using the Ribosomal Data Project (RDP) Naive Bayesian Classifier [22]. Family is represented. ␣, , ␥ represent ␣-proteobacteria, -proteobacteria and ␥-proteobacteria, respectively. B1 = B9 = B29; B3 = B10 = B19 = B20 = B22 = B31;B5 = B38 = B42;B7 = B14; B8 = B11 = B17 = B21 = B25 = B28 = B32 = B37 = B40; B12 = B13 = B15 = B18 = B23 = B24 = B26 = B30 = B35; B16 = B27. Band detected by means of gel migration. Band not sequenced.
480 455 459 495 352 350 496 481 468 486 501 507 – – – X – – – – – – – – – – X X – – X – – – – – – – – – – – X X – – X X X – X – – – X X – X – – – – – – – – X X X – – – – – – – – X Xe X – – – – X – X – – – X – X – – – – – – – X X X – Xe – – – X X X X – – – – – – – –
LT SO TV SO BS Br30 BS SO BS IC Br30 IC SO IC Initial soil
16S B1d 16S B2 16S B3 16S B5 16S B6 16S B7 16S B8 16S B12 16S B16 16S B33 16S B34 16S B36
a
99% 99% 98% 100% 99% 99% 98% 98% 97% 100% 100% 100%
Agrobacterium tumefaciens (NR 041396.1) Bradyrhizobium liaoningense (NR 041785.1) Rhizobium oryzae (NR 044393.1) Mycobacterium monacense (NR 041723.1) Hydrogenophaga intermedia (NR 024856.1) Achromobacter denitrificans (NR 042021.1) Pseudoxanthomonas Mexicana (NR 025105.1) Pigmentiphaga kullae (NR 025112.1) Azoarcus indigens (NR 024851.1) Mycobacterium rutilim (NR 043761.1) Cupriavidus campinensis (NR 025137.1) Uncultured soil bacterium clone from a PAHs polluted soil (DQ907006) Bordetella himzii (NR 027537.1)
Closest organism in GenBank database (accession no.) Length (bp) Band detectiona Band
Table 3 Properties of DGGE bands: designations and accession numbers for the band sequences and levels of similarity to related organisms.
% similarityb
Phylogenetic groupc
Rhizobiaceae (␣) Bradyrhizobiaceae (␣) Rhizobiaceae (␣) Mycobacteriaceae (Actinobacteria) Comamonadaceae () Alcaligenaceae () Xanthomonadaceae (␥) Alcaligenaceae () Rhodocyclaceae () Mycobacteriaceae (Actinobacteria) Burkholderiaceae () –
S. Lladó et al. / Journal of Hazardous Materials 283 (2015) 35–43
41
assayed was aimed at exploiting the hydrocarbon-depleting ability of the resident microbiota. 4.2. Effect of surfactant amendment on bacterial community composition It is noteworthy that surfactants can increase the biodegradation rates of highly recalcitrant PAHs and are widely used in bioremediation studies [29]. In the present work, we found remarkable shifts in the soil bacterial community when MAs were added to the soil, especially related to the non-ionic surfactant Brij 30 (Fig. 2A). In addition, a toxic effect of the Brij 30 on the CHDB populations, as inferred by MPN counts, was previously observed [7]. Such a decrease in the size of the PAH-degrading population, as well as the shift in the diversity structure of the CHDB population (MPN–DGGE profiles), could explain the lower degradation rate of 4-ring PAHs detected when Brij 30 was added to the soil in the absence of LS [7]. Coincidently, pyrosequencing data reveals that both Actinobacteria (Mycobacteriaceae) and Bacteroidetes (Chitinophagaceae) populations are highly depleted after surfactant Brij 30 supplementation. Therefore, the obtained results from this diversified approach would indicate that both families may be directly related on 4-ring PAH biodegradation in the polluted soil. On the other hand, when SO was used as MA in non-amended soil (IC + SO), the community shifts observed by means of pyrosequencing were concomitant with higher CHB–MPN counts but lower HMW–PAH depletion levels suggesting that soybean oil was being used as a carbon source by the bacteria. However, higher rate of benzo(a)pyrene depletion were observed in IC + SO treatment and this coincided with the presence of the genus Agrobacterium (Rhizobiaceae family) according to MPN–DGGE. Bacteria in the Rhizobiaceae family are commonly found in polluted environments [30], but no evidence of benzo(a)pyrene degradation ability has been found in the literature. Moreover, the class of Rhizobiaceae was present at low percentages (
