Delftia sp. LCW, a strain isolated from a constructed wetland shows novel properties for dimethylphenol isomers degradation

Background Dimethylphenols (DMP) are toxic compounds with high environmental mobility in water and one of the main constituents of effluents from petro- and carbochemical industry. Over the last few decades, the use of constructed wetlands (CW) has been extended from domestic to industrial wastewater treatments, including petro-carbochemical effluents. In these systems, the main role during the transformation and mineralization of organic pollutants is played by microorganisms. Therefore, understanding the bacterial degradation processes of isolated strains from CWs is an important approach to further improvements of biodegradation processes in these treatment systems. Results In this study, bacterial isolation from a pilot scale constructed wetland fed with phenols led to the identification of Delftia sp. LCW as a DMP degrading strain. The strain was able to use the o-xylenols 3,4-DMP and 2,3-DMP as sole carbon and energy sources. In addition, 3,4-DMP provided as a co-substrate had an effect on the transformation of other four DMP isomers. Based on the detection of the genes, proteins, and the inferred phylogenetic relationships of the detected genes with other reported functional proteins, we found that the phenol hydroxylase of Delftia sp. LCW is induced by 3,4-DMP and it is responsible for the first oxidation of the aromatic ring of 3,4-, 2,3-, 2,4-, 2,5- and 3,5-DMP. The enzyme may also catalyze both monooxygenation reactions during the degradation of benzene. Proteome data led to the identification of catechol meta cleavage pathway enzymes during the growth on ortho DMP, and validated that cleavage of the aromatic rings of 2,5- and 3,5-DMPs does not result in mineralization. In addition, the tolerance of the strain to high concentrations of DMP, especially to 3,4-DMP was higher than that of other reported microorganisms from activated sludge treating phenols. Conclusions LCW strain was able to degraded complex aromatics compounds. DMPs and benzene are reported for the first time to be degraded by a member of Delftia genus. In addition, LCW degraded DMPs with a first oxidation of the aromatic rings by a phenol hydroxylase, followed by a further meta cleavage pathway. The higher resistance to DMP toxicity, the ability to degrade and transform DMP isomers and the origin as a rhizosphere bacterium from wastewater systems, make LCW a suitable candidate to be used in bioremediation of complex DMP mixtures in CWs systems. Electronic supplementary material The online version of this article (10.1186/s12866-018-1255-z) contains supplementary material, which is available to authorized users.


Background
It is well known that the wastewater of coal conversion and coke operation processes contains high concentrations of phenolic compounds, constituting up to 80% of the total chemical oxygen demand (COD) [1]. The phenolic fraction of the this wastewater is mainly represented by phenol, cresols, resorcinols and dimethylphenols (DMP) [2], substances which are considered toxic, carcinogenic, mutagenic and teratogenic [3]. The DMPs, have shown to be even more persistent and toxic than other phenolic compounds [4] and they are often released to the environment in higher concentration than the permissible limits established [1,5].
Biological treatment of phenols and their derivatives has been the main approach for their removal [6,7]. Biodegradation of the six DMP isomers, in pure or mixed cultures, often involves different Pseudomonas species [8][9][10]. Degradation of 2,6-DMP has only been reported by Mycobacterium sp. strain DM1 [11]. Studies aiming at the removal of DMP in a mixture with other substituted phenols in laboratory-scale sludge units have also been performed [6,12]. Recently, constructed wetlands (CWs) have been investigated for their potential to remove coke oven wastewater [13]. Promisingly, Schultze-Nobre et al., reported efficient removal of a mixture of three DMP-isomers (2,6-, 3,4-and 3,5-DMP) in a laboratory scale CW [14].
Complex physicochemical and biological processes in CWs include the interaction of plants, microorganism and pollutants. The main role in the transformation and mineralization of organic pollutants is played by microorganisms present in the rhizosphere, an environment suitable for highly efficient transformations of complex contaminants [15,16]. There is, however, little knowledge about the bacterial removal of DMP in CWs and detailed microbiological and molecular biological studies are needed to reveal the biological activities towards DMP as a basis for better management of CWs.
In this work, we aimed to isolate strains from a constructed wetland treating phenolic wastewater in order to characterize the ability to degrade DMP isomers, the toxicity tolerance toward the isomers and to reveal the genomic and proteomic mechanisms supporting the degradation. The first results indicated that Delftia sp. LCW is a versatile bacterium with novel properties to degrade DMP isomers and it could be used to enhance phenolic compounds removal in CWs.

DMP-degrading bacterium, growth and co-metabolism assays
From the isolation trials, LCW was the strain recovered with capabilities to use DMP as sole carbon and energy source. According to its 16S rRNA gene sequence, the isolate is a member of the genus Delftia. The 16S rRNA gene is identical to that of Delftia sp. strain SM-1 (JN001163) and differs by only 1-2 bases from that of various Delftia acidovorans isolates (e.g. D. acidovorans SPH-1, CP000884 and D. acidovorans NBRC 14950, AB680719; 99.9% sequence identity) but by 7-8 bases from that of Delftia tsuruhatensis and Delftia lacustris strains, including the respective type strains (D. lacustris 332 EU888308; D. tsuruhatensis T7, AB075017).
The strain was able to use 2,3-DMP and 3,4-DMP as sole carbon and energy source ( Fig. 1a and b). Complete depletion of the isomers was observed and both substrates yielded similar bacterial biomass (Table 1).

Proteomic analysis
The label-free shotgun proteomics approach allowed the identification of 1814 proteins (Additional file 2). First, an unsupervised clustering approach (Principal Component Analysis, PCA) based on the different treatments, revealed clear segregation by sample treatment. The variances among the treatment groups showed that replicate analyses are clustering together. The PCA analysis also showed a greater segregation between the proteome profile of 3,4-DMP and 2,3-DMP among the four treatments (Additional file 3). The proteins that were significantly different between 3,4-and 2,3-DMP, were functionally assigned according to KO (KEEG Orthology), which characterizes gene functions in order to infer high level protein functions of the organism.. The main functional categories of the proteins that differed between 3,4-and 2,3-DMP treatments can be found in Additional file 4. In total, five proteins related with aromatic degradation pathways could be identified ( Table 2). Phenol hydroxylase, catechol 2,3-dioxygenase, 2-hydroxymuconic-semialdehyde dehydrogenase and 4-oxalocronate tautomerase were identified in all treatments. Protocatechuate 4,5-dioxygenase was identified only in 2,3-DMP grown cells. In Fig. 4, the protein abundance of related DMP proteins present in 2,3and 3,4-DMP as singles isomers are indicated. There were not significant differences of the abundances of the proteins between for 2,3-and 3,4-DMP treatments, for any of the proteins related with aromatic degradation (phenol hydroxylase, p = 0.160; catechol 2,3-dioxygenase, p = 0.227, 2-hydroxymuconic semialdehyde dehydrogenase, p = 0.488 and 4-oxalocronate tautomerase p = 0.227) (Additional file 1: Table B).
In general, DMP biodegradation have been reported for all isomers, mainly through four different pathways, 1) monooxygenation of the aromatic ring to form dimethylcatechols followed by extradiol ring-cleavage as reported for 2,3-and 3,4-DMP degradation [28,31], 2) oxidation of the methylsubstituent to form methyl-substituted gentisates followed by degradation through a gentisate degradative pathway as reported for 2,5-, and 3,5-DMP degradation [32,33]; 3) oxidation of both methylsubstituents to form 4-hydroxyisophthalate followed by hydroxylation and degradation via a protocatechuate pathway as reported for 2,4-DMP degradation [34,35] or 4) two successive monooxygenation of the aromatic ring 2,6-dimethylhydroquinone and 2,6-dimethyl-3-hydroxyhydroquinone as reported for 2,6-DMP degradation [11]. Proteome analysis showed the expression of a catechol meta cleavage pathway by Delftia sp. LCW in the presence of 3,4-DMP and 2,3-DMP including a catechol 2,3-dioxygenase, a 2-hydroxymuconic semialdehyde dehydrogenase and a 4-oxalocronate tautomerase. Catechols will be formed by a phenol monooxygenase, where the gene encoding the alpha-subunit (P3 subunit) was also detected in the genome through PCR, in    Although catechol meta cleavage enzymes were detected in the presence of 2,3-DMP, an additional enzyme was identified when the isomer was provided as sole carbon source. This enzyme was the protocatechuate 4,5-dioxygenase (A9C0K2). Other proteins belonging to the protocatechuate pathway were not detected in the proteomic analysis. The protocatechuate pathway has been described only in 2,4-DMP degradation by Pseudomonas putida NCBIMB 9866 where it is formed through the subsequent oxidations both methyl group of 2,4-xylenol to form 4-hydroxyisophthalate and hydroxylation to protocatechuate [35]. Our results indicate that Delftia sp. LCW produced enzymes of the protocatechuate pathway in the presence of 2,3-DMP provided as a sole  carbon source. However, this is probably due to fortuitous induction. Since proteins were identified by searching against UNIPROT database for genera related to Delftia or reported DMP-degrading bacteria, only the anticipated proteins listed could be identified. When a mixture of 3,4-and 2,3-DMP was provided, there was an evident shorter lag phase for the degradation of 2,3-DMP. Fostered degradation of DMP in mixtures with other aromatic compounds was previously observed in a sequencing batch reactor, where degradation kinetics of 3,4-DMP were improved in the presence of 4-nitrophenol (4-NP) [12]. Likewise, experiments with C. testosteroni had shown that the addition of phenol increased cell growth and shortened the time necessary for 4-Clorophenol (CP) degradation [27]. Similar experiments with Pseudomonas putida XQ23 showed acceleration of 2,3-DMP degradation in the presence of DMP mixtures [37]. Those finding, in addition to higher abundance of meta pathway proteins in the mixture of the ortho DMP, suggested that the presence of 3,4-DMP in the mixture induced the expression of the catechol meta cleavage pathway and fostered the degradation 2,3-DMP.
Screening of catabolic genes using degenerated primers and PCR led to the detection of the fragment of the phenol hydroxylase large subunit [38,39]. The phenol hydroxylases comprise related family of enzymes capable to hydroxylate mainly phenols and their methyl-substituted derivatives to the corresponding catechols [40,41]. Some phenol hydroxylases families have showed also the ability to transform toluene or benzene [42]. Phylogenetical relationships of phenol hydroxylase from LCW, showed that the closest protein is phenol hydroxylases from Delftia AZ1-13 (ACZ44761), which phenotype (L6) was a representative in a lab-scale reactor treating phenols [43]. Phylogenetical clustering of the phenol hydroxylase of LCW revealed similarities with the phenol hydroxylase Comamonas testosteroni TA441 (BAA34172) [44]. It was also similar to other strains with reported genes for catechol meta cleavage pathway e.g. Cupriavidus pinatubonensis [28,45]. Another related phenol hydroxylase of LCW, belonging to Pseudomonas sp. M1 (ABM96259) with high similarities to the dmpKLMNOP hydroxylase gen of CF600 [46] .
Delftia sp. LCW was also able to use benzene as growth substrate. One known pathway for the degradation of benzene involves two successive monooxygenations that are catalyzed by different soluble diiron monooxygenases [45]. Nevertheless, some phenol monooxygenases may catalyze both successive monooxygenations [40]. This characteristic was found in Pseudomonas sp. M1, where the multicomponent hydroxylase phc gene was responsible for the initial oxidation of phenol or benzene [46], this subunit is indeed similar to the one found in LCW. In addition, the phenol hydroxylase gene large subunit sequence found in LCW showed similarities with the toluene/benzene/chlorobenzene-monooxygenase (tbc1D) from Burkholderia sp. strain JS150 (AAG40791). Considering the detection of this gene sequence, the corresponding protein expressed and the phylogenetic similarities with other well-characterized functional proteins, we can propose that this phenol hydroxylase detected in Delftia sp. LCW may be responsible of catalyzing both monooxygenation reactions during the degradation of benzene.
Considering that only 3,4-DMP contributed to 8.5% for the phenol toxicity in ash heap water from an oil shale industry [52] and given the fact that this isomer can impair the performance of biological treatment processes of phenolic wastewater [4], the tolerance and degradation of 3,4-DMP by Delftia sp. LCW, together with its role for co-metabolic transformation of other DMP isomers support the proposal of LCW strain as a promising microbial tool component on further developments of wastewater treatment technologies of phenolic compounds.

Conclusions
LCW strain was able to degraded complex aromatics compounds. DMPs and benzene are reported for the first time to be degraded by a member of Delftia genus. In this strain, the isomer 3,4-DMP acted as an inducer of phenol hydroxylase enzyme, that is responsible for a first oxidation of the DMPs, and it is followed by a further meta cleavage pathway for ortho-DMPs isomers. Such induction also led to the transformation of 2,4-, 2,5-and 3,5-DMPs. The higher resistance to DMP toxicity, and the understanding of degradation pathways for DMPs by LCW strain, as well as its origin as a rhizosphere bacterium from CWs, make of Delftia sp. LCW a suitable candidate to be used in bioremediation of coke-coal contaminated sites and is important to establish technological alternatives for wastewater treatment for coal-coke industry effluents.

Isolation and identification of DMP degrading bacteria
Bacterial strains were isolated by selective enrichment from a pilot-scale horizontal sub-surface CW fed with groundwater with benzene, phenols and m-cresols. In order to select strains with potential ability to degrade DMP for subsequent assays, bacterial growth on each of the six single isomers of DMP as a sole carbon source and energy was determined. The 16S rRNA gene from the only isolate showing growth on DMP isomers and called strain LCW, was amplified by PCR using universal bacterial primers 27F and 1492R [53]. Sanger sequencing of the purified PCR product was obtained (GATC Biotech AG, Cologne, Germany) and the 16S rRNA gene sequence was compared to public databases using RDP [54] to identify closest reference sequences The 16S rRNA gene sequence from strain LCW is available in KY643688k with accession number KY643688 [55].

DMP degradation and co-metabolism assays
In order to evaluate degradation and co-metabolic capabilities of the isolated, pure cultures of the strain were transferred from plates to liquid LB medium and incubated overnight at 30°C with agitation. Then, cells were harvested by centrifugation, washed three times with phosphate buffer (50 mM of NaH 2 PO 4 , pH 7.0) and re-suspended in 50 mL of liquid M9 [56] (turbidity ofṼ 0.05 at 620 nm) with the corresponding carbon source (as described below). First, the selected isolate was grown on 3,4-DMP and 2,3-DMP separately with a concentration of 70 mg L − 1 each. In addition, co-metabolic transformation by the strain was tested using each isomer (2,3-, 2,4-, 2,5-, 2,6-and 3,5-DMP) mixed with 3,4-DMP in concentrations of 35 mg L − 1 per isomer (i.e. 70 mg L − 1 DMP in total). Each treatment was set up in triplicates. Samples were taken from each culture every 4 and 12 h to determine cell number and DMP concentration, until the stationary phase was reached. Growth was followed by cell counting using a Coulter Counter® (Beckman Coulter Inc) and the software Multisizer 3 Version 3.51® (Beckman Coulter Inc.). The concentration of DMP in the cultures was measured in cell free supernatants by HPLC (Promicence, Shimadzu) equipped with a UV detector and a 2.7 μm by 3.0 × 150 mm poroshell 120 EC-C18 column (Agilent Technologies, USA). For the separation of the DMP isomers, the mobile phase A consisted of formic acid (0.1%) and phase B consisted of acetonitrile (100%), using the following gradients over a total run time of 45 min: 20% B to 40% B with a flow rate of 0.2 mL min − 1 [14]. For each treatment, a flask prepared under the same conditions without bacterial inoculum was set up in order to determine abiotic losses. In order to compare differences in degradation, the approximate bacterial yield for each mixture was calculated according to the following formula: where; C tf ¼ number of cells per mL at the end of the expontial phase C t0 ¼ number of cells per mL at time cero * A theoretical value of 3 mg mL − 1 of dry mass equivalent to 1 × 10 7 cell mL − 1 for Gram negative rod bacteria was calculated [57] C from DMP consumed = (DMP t0 mg. mL −1 − DMP tf mg. mL −1 ) × 0.787* where,  [38,39,58], intradiol catechol dioxygenases [59] and extradiol catechol dioxygenases [39,60] in the genome of strain LCW was performed by PCR using previously described degenerate primers (Additional file 5). The obtained PCR products were purified and sequenced (GATC Biotech AG, Cologne, Germany). The obtained nucleic acid sequences were translated and compared to entries in public databases [61]. The only detected gene sequence from this PCR specific target survey is available in GeneBank with accession number MF804845. Protein translation was obtained through EMBL-EBI translation tool service [62], where, frame 3 from the positive strand was chosen. PATRIC genome database [63] was used in order to check the detected gene in other related genus members.

Proteomics analysis
Protein extraction, LC-MS measurements and protein identification were performed as described previously by Lünsmann et al., [64]. Briefly, for protein extraction, cells were grown on 3,4-DMP and 2,3-DMP individually and on mixtures of 3,4-with 2,3-and 3,4-with 3,5-DMP as described before. Bacterial cells were harvested in the middle of the exponential phase. Lysis was achieved by ultrasonication and proteins were recovered by denaturation and solubilization with urea buffer. The protein lysates were applied on a polyacrylamide gel electrophoresis. Gel pieces were cut and the proteins were lysed using trypsin overnight. The resulting peptide lysates were separated by liquid chromatography hyphenated with mass spectrophotometry (LC-MS). For the identification of proteins, the acquired LC-MS data were searched against the public database Uni-protKB using the closest related genera, i.e. Delftia, Pseudomonas, Ralstonia, Comamonas and bacterium AZ1-13, as inferred from the sequenced Phenol hydroxylase fragment. LC-MS spectra were searched using the Proteome Discoverer (Thermo Fisher Scientific, v1.4, San Jose, CA, USA). Search settings were: Sequest HT search engine, trypsin (full specific), MS tolerance 10 ppm, MS/MS tolerance 0.02 Da, two missed cleavage sides, dynamic modifications oxidation (Met), static modifications carbamidomethylation (Cys). Only peptides that passed the FDR thresholds of < 1% FDR q-value (Percolator) and rank 1 peptides were considered for further analysis. Label-free quantification was done using the Top-3 peptide area for approach. These linear area values were log 10 -transformed, median normalized and furthermore missing value imputation were performed. Principal component analysis (PCA) was performed using InfernoRDN (version 1.6044.35184 July 19, 2016). Hypothesis significance tests were performed with Prostar [65] using the threshold of the log fold change (FC) of 1 and the threshold of -log (p-value) of 2 for protein abundances. The proteins significantly differing between incubations with the 3,4-and 2,3-DMP were functionally assigned with the annotation tool KEGG BlastKOALA [66].

Toxicity test
Delftia sp. LCW was grown in mineral medium M9 with 4 g L − 1 of disodium succinate as carbon and energy source in 50-mL flasks. Once the cultures reached the exponential phase of growth, DMP isomers were added separately at different concentrations (one flask per concentration) ranging from 45 to 350 mg L − 1 . No DMP isomers were added to the control culture. For each flask, the OD 620mn was measured every hour until the growth stopped in one of the cultures flasks. To determine the toxicity of each compound, specific growth rate μ (h − 1 ), was calculated as follows [67]: where, μ [h −1 ] = Growth rate x t1 = OD 620 at time t 1 x t2 = OD 620 at time t 2 t 1 = one hour after toxin was added [h] t 2 = final time when the growth stopped [h] Inhibited growth (%) was defined as the percentages of the growth rates μ (h −1 ) of toxified cultures divided by the growth rate of the control culture. EC 50 was calculated for each isomer by interpolation of the concentration resulting in 50% growth inhibition. Log P ow values were obtained from Toxnet Database [68].

Statistical analyses
In order to indicate significant differences among values of maximal bacterial biomass in the cometabolic assay and protein abundances of selected proteins in proteomics data, normal distributions were tested with Shapiro Wilk and further parametric or not parametric test were performed accordingly. In addition, Pearson correlation was performed to correlate EC 50 and Log P ow values for all isomers. Analyses were performed using SigmaPlot Version 13.0 (®Systat Sofware, Inc.). P-values and specific statistical tests are described in Additional file Additional files Additional file 1: Table A. p-values for the cometabolic assay, and Table B. p-values of paired t-tests for the comparison between 2,3-and 3,4-DMP. (DOCX 16 kb) Additional file 2: Table S2. Proteome raw data for DMP treatments (XLSX 22980 kb) Additional file 3: Figure S1. Principal component Analysis of the proteomic profile of strain LCW for the four DMP treatments. (PNG 37 kb) Additional file 4: Figure S2. Proteins functional category of Delftia sp. LCW for the proteins with significant differences between 3,4-and 2,3-DMP isomers (proteins with p-value < 0.001 for FC were selected for the analysis). (TIF 100 kb) Additional file 5: Table S1. Primers used for targeting catabolic genes in the genome of