A Synergistic Consortium Involved in rac-Dichlorprop Degradation as Revealed by DNA Stable Isotope Probing and Metagenomic Analysis

ABSTRACT

rac-Dichlorprop, a commonly used phenoxyalkanoic acid herbicide, is frequently detected in environments and poses threats to environmental safety and human health. Microbial consortia are thought to play key roles in rac-dichlorprop degradation. However, the compositions of the microbial consortia involved in rac-dichlorprop degradation remain largely unknown. In this study, DNA stable isotope probing (SIP) and metagenomic analysis were integrated to reveal the key microbial consortium responsible for rac-dichlorprop degradation in a rac-dichlorprop-degrading enrichment. OTU340 (Sphingobium sp.) and OTU348 (Sphingopyxis sp.) were significantly enriched in the rac-[¹³C]dichlorprop-labeled heavy DNA fractions. A rac-dichlorprop degrader, Sphingobium sp. strain L3, was isolated from the enrichment by a traditional enrichment method but with additional supplementation of the antibiotic ciprofloxacin, which was instructed by metagenomic analysis of the associations between rac-dichlorprop degraders and antibiotic resistance genes. As revealed by functional profiling of the metagenomes of the heavy DNA, the genes rdpA and sdpA, involved in the initial degradation of the (R)- and (S)-enantiomers of dichlorprop, respectively, were mostly taxonomically assigned to Sphingobium species, indicating that Sphingopyxis species might harbor novel dichlorprop-degrading genes. In addition, taxonomically diverse bacterial genera such as Dyella, Sphingomonas, Pseudomonas, and Achromobacter were presumed to synergistically cooperate with the key degraders Sphingobium/Sphingopyxis for enhanced degradation of rac-dichlorprop.

IMPORTANCE Understanding of the key microbial consortium involved in the degradation of the phenoxyalkanoic acid herbicide rac-dichlorprop is pivotal for design of synergistic consortia used for enhanced bioremediation of herbicide-contaminated sites. However, the composition of the microbial consortium and the interactions between community members during the biodegradation of rac-dichlorprop are unclear. In this study, DNA-SIP and metagenomic analysis were integrated to reveal that the metabolite 2,4-dichlorophenol degraders Dyella, Sphingomonas, Pseudomonas, and Achromobacter synergistically cooperated with the key degraders Sphingobium/Sphingopyxis for enhanced degradation of rac-dichlorprop. Our study provides new insights into the synergistic degradation of rac-dichlorprop at the community level and implies the existence of novel degrading genes for rac-dichlorprop in nature.

INTRODUCTION

The raceme of the dichlorprop, rac-dichlorprop [(R,S)-2-(2,4-dichlorophenoxy)propanoic acid], a phenoxyalkanoic acid herbicide, is widely used worldwide to control broad-leaf weeds in crop farms and nonagricultural fields (1, 2). Due to its extensive use and high solubility in water, rac-dichlorprop was persistently residual in diverse environments, including groundwater and soil (3, 4). rac-Dichlorprop has moderate to low toxicity for most aquatic organisms (48-h 50% effective concentration [EC₅₀] of >100 mg liter⁻¹) and is a potential risk to human health (5, 6).

The microorganism-mediated environmental fate of rac-dichlorprop is of interest. rac-Dichlorprop degraders, including Sphingobium herbicidovorans MH, Delftia acidovorans MC1, Alcaligenes sp. strain CS1, Ralstonia sp. strain CS2, Stenotrophomonas maltophilia PM, Rhodoferax sp. strain P230, and Sphingopyxis sp. strain DBS4, have been isolated by traditional cultivation-dependent approaches in recent decades (7–13). Microbial degradation of phenoxyalkanoic acid herbicides usually initiates with the cleavage of the side chain to produce corresponding phenols, which is catalyzed by Fe(II)/α-ketoglutarate-dependent dioxygenases. For nonchiral herbicides such as 2,4-dichlorophenoxyacetate (2,4-D) and 4-chloro-2-methylphenoxyacetate (MCPA), the dioxygenase involved is typically TfdA, which is encoded by the tfdA gene (14). Regarding the chiral herbicide rac-dichlorprop, the (R)- and (S)-enantiomers of dichlorprop are initially transformed to 2,4-dichlorophenol and pyruvate by RdpA and SdpA, respectively. The corresponding genes rdpA and sdpA encoding these two enantiospecific dioxygenases were characterized in Sphingobium herbicidovorans MH and Delftia acidovorans MC1, respectively (15, 16). 2,4-Dichlorophenol was further catabolized via a classic pathway (15, 16). A microbial consortium rather than a single species is well recognized to play more important roles in the enhanced degradation of organic pollutants (17). However, the key microbial consortium and the interactions between community members during the synergistic degradation of rac-dichlorprop remain largely unclear (18).

Stable isotope probing (SIP) is a cultivation-independent technique that can be used to effectively investigate the functional microorganisms involved in biogeochemical processes (19–25). In addition, metagenomics is the study of the entire genomes of all microbes from an environmental sample and provides useful taxonomic, functional, and evolutionary information (26). In our previous study, an enrichment culture capable of degrading rac-dichlorprop was obtained from a rac-dichlorprop-contaminated soil in an abandoned pesticide factory in Nanjing, Jiangsu, China (31°57′9.58″N, 118°50′24.07″E), and a bacterial strain, Sphingopyxis sp. DBS4, capable of utilizing rac-dichlorprop as sole carbon source for growth was isolated from the enrichment (13). Knowing the importance of the microbial consortium in rac-dichlorprop degradation, DNA-SIP and metagenomic analysis were integrated to reveal the key microbial consortium involved in rac-dichlorprop degradation in the enrichment in this study. The interactions between the key degraders and the metabolite degraders were also investigated, and a new rac-dichlorprop degrader was isolated with additional supplementation of the antibiotic ciprofloxacin, which was instructed by the metagenomic analysis of the antibiotic resistance genes (ARGs) of ¹³C-labeled DNA. In addition, novel genes involved in rac-dichlorprop degradation were indicated to exist in nature.

RESULTS

Key bacteria involved in rac-dichlorprop degradation as revealed by DNA-SIP.

Biodegradation of rac-[¹²C]dichlorprop ([¹²C]dichlorprop) and rac-[¹³C]dichlorprop ([¹³C]dichlorprop) by the enrichment was carried out in microcosms. No significant differences between the degradation of [¹²C]dichlorprop and [¹³C]dichlorprop were found (see Fig. S1 in the supplemental material). rac-Dichlorprop (30 mg liter⁻¹) was completely degraded within 6 days by the enrichment, while nearly no degradation of rac-dichlorprop was found in the control.

Genomic DNA of the [¹²C]dichlorprop- and [¹³C]dichlorprop-labeled samples was extracted at 4 days (approximately 75% of rac-dichlorprop was degraded) and then separated by isopycnic centrifugation. The relationships between buoyant density and the fraction number are listed in Fig. S2. Quantification of functional degrading genes in each fraction by quantitative real-time PCR (qPCR) showed that the abundances of rdpA and sdpA genes were higher in fractions 5 to 7 (buoyant density, 1.734 to 1.725 g ml⁻¹) of the [¹³C]dichlorprop-labeled samples (highest in fraction 6) than in fractions 8 to 9 (buoyant density, 1.720 to 1.716 g ml⁻¹) of the [¹²C]dichlorprop-labeled samples (Fig. S3). The DNA in fractions 5 to 7 of the [¹³C]dichlorprop-labeled samples was referred to as the “heavy DNA.”

The 16S rRNA genes of each fraction (from 5 to 9) of [¹²C]dichlorprop- and [¹³C]dichlorprop-labeled samples were analyzed by high-throughput sequencing. In total, 1,773,096 sequences were obtained from the 30 fractions (five fractions of two samples in triplicate), and 474 different operational taxonomic units (OTUs) (at 97% identity) were detected in all samples. Different abundances of taxonomically affiliated OTUs in equivalent fractions (fractions 5 to 9) between [¹³C]dichlorprop- and [¹²C]dichlorprop-labeled samples were found (Fig. S4). The genomic DNA of those microorganisms that have assimilated [¹³C]dichlorprop will shift from fractions 8 to 9 (with normal buoyant density of 1.720 to 1.716 g ml⁻¹) to fractions 5 to 7 (with higher buoyant density of 1.734 to 1.725 g ml⁻¹). So, bacteria (represented by OTUs) showing relatively higher abundances in fractions 5 to 7 of the [¹³C]dichlorprop-labeled samples compared to those in [¹²C]dichlorprop-labeled samples were defined as key species involved in rac-dichlorprop degradation. The relative abundances of OTU340 and OTU348 in fractions 5 to 9 of [¹²C]dichlorprop-labeled samples were 0.30%, 0.24%, 0.42%, 0.31% and 0.10% and 0.30%, 0.07%, 0.32%, 0.11%, and 0.01%, respectively. However, in equivalent fractions 5 to 9 of [¹³C]dichlorprop-labeled samples, the relative abundances of OTU340 and OTU348 were 2.18%, 12.63%, 4.11%, 0.50%, and 0.01% and 29.34%, 45.20%, 2.94%, 0.11%, and 0.02%, respectively. The results showed that OTU340 and OTU348 were enriched at fractions 5 to 7 (with a higher buoyant density of >1.725 g ml⁻¹) in [¹³C]dichlorprop-labeled samples (Fig. 1). By phylogenetic analysis, OTU340 and OTU348 were assigned to the genera Sphingobium and Sphingopyxis, respectively (Fig. S5). The results indicated that Sphingobium and Sphingopyxis species have assimilated [¹³C]carbon from rac-dichlorprop and are the key species involved in rac-dichlorprop degradation.

FIG 1.

Relative abundances of OTU340 (Sphingobium) and OTU348 (Sphingopyxis) at different buoyant densities of [¹³C]- and [¹²C]dichlorprop-labeled samples.

Functional profiling of metagenomes.

Metagenome sequencing was employed to reveal the potential metabolic pathways of rac-dichlorprop degraders and the associations between rac-dichlorprop degraders and antibiotic resistance genes (ARGs). The metagenomes of fraction 6 of [¹³C]- and [¹²C]dichlorprop-labeled samples were sequenced. As a result, a total of 213,576,692 and 214,582,596 clean reads were generated, and 57,683 and 36,285 contigs were obtained, respectively. In the metagenome of the [¹²C]dichlorprop-labeled fraction 6, Dyella (26.8%), Dokdonella (15.5%), and Methylobacillus (10.4%) were the three most dominant genera, followed by Pseudomonas (7.2%), Achromobacter (4.9%), Rhodanobacter (4.4%), Rudaea (4.3%), and Mizugakiibacter (3.0%). In contrast, in the metagenome of the [¹³C]dichlorprop-labeled fraction 6, Sphingopyxis (23.4%), Dyella (16.0%), and Sphingobium (12.1%) were the three most dominant genera, followed by Dokdonella (9.8%), Pseudomonas (4.1%), Sphingomonas (3.4%), Rudaea (2.7%), Achromobacter (2.7%), Rhodanobacter (2.6%), and Methylobacillus (2.3%) (Fig. S6).

The potential metabolic pathways of the metagenomes were predicted by KEGG database (Fig. 2a and b). At KEGG pathway level 1, “metabolism” was the most abundant functional category in both metagenomes (Fig. 2a). Interestingly, the relative abundance of “xenobiotics biodegradation and metabolism” of the [¹³C]dichlorprop-labeled metagenome (4.7%) was higher than that of the [¹²C]dichlorprop-labeled metagenome (3.3%) at KEGG pathway level 2 (Fig. 2b). These results indicated that genomic DNA of functional microorganisms capable of degradation of xenobiotics (including rac-dichlorprop) was enriched in fraction 6 of the heavy DNA sample.

FIG 2.

KEGG functional abundance analysis. (a) Relative abundances of [¹²C]- and [¹³C]dichlorprop-labeled samples at KEGG pathway level 1; (b) relative abundances of [¹²C]- and [¹³C]dichlorprop-labeled samples at KEGG pathway level 2.

To explore the associations between rac-dichlorprop-degrading bacteria and antibiotic resistance genes (ARGs), the [¹³C]dichlorprop-labeled metagenome was subjected to a BLAST search against the Comprehensive Antibiotic Resistance Database (CARD). At the class level, Dyella, Sphingopyxis, and Sphingobium contributed 21.5%, 16.4%, and 10% of the antibiotic resistance (AR), respectively (Fig. S7a). Because Sphingobium and Sphingopyxis were involved in rac-dichlorprop degradation as revealed by DNA-SIP, the CARD was further applied to analyze the relationships between ARGs and the genera Sphingobium and Sphingopyxis. As shown in Fig. S7b, Sphingobium and Sphingopyxis harbored most ARGs jointly at the antibiotic resistance ontology (ARO) level. Sphingobium contributed 22.9% and 9.6% and Sphingopyxis contributed 15.8% and 24.5% of the fluoroquinolone ARGs mfd and patA, respectively. These results suggested that fluoroquinolone antibiotics could be used to enrich and isolate rac-dichlorprop-degrading strains (potentially belonging to the genus Sphingobium or Sphingopyxis) from enrichment.

The contributions of active taxa (genus level) for rac-dichlorprop degradation in the [¹³C]dichlorprop-labeled metagenome were further analyzed. Subsequently, a stacked chart was generated by calculating the relative abundances of microbial taxa in the metabolic pathways in the [¹³C]dichlorprop-labeled metagenome (Fig. S8a). The contributions of active taxa (genus level) for metabolic pathways were calculated. The abundances of six genera (Sphingobium, Sphingopyxis, Dyella, Sphingomonas, Pseudomonas, and Achromobacter) were all associated with the “chlorocyclohexane and chlorobenzene degradation pathways” (KEGG pathway level 3, ko00361) and contributed 46.1%, 16.3%, 12.4%, 9.4%, 3.0%, and 1.63% to the ko00361 pathway, respectively. Notably, Sphingobium and Sphingopyxis appeared to associate directly with metabolic activities, and their contributions to degradation seemed to be better defined. The taxonomic affiliations of enzymes responsible for 2,4-dichlorophenol degradation were mostly assigned to the four genera Dyella, Sphingomonas, Pseudomonas, and Achromobacter based on the analysis of the [¹³C]dichlorprop-labeled metagenome (Fig. S8b). These findings indicated that Dyella, Sphingomonas, Pseudomonas, and Achromobacter species (most likely to be involved in the degradation of the metabolite 2,4-dichlorophenol) might interact with the key degraders Sphingobium/Sphingopyxis for the synergistic degradation of rac-dichlorprop.

Isolation of new degraders for rac-dichlorprop and its metabolite from the enrichment.

Due to the close associations between rac-dichlorprop degraders and fluoroquinolone antibiotic resistance genes, a traditional continuous enrichment method with additional supplementation of 5 mg liter⁻¹ ciprofloxacin, a fluoroquinolone antibiotic, was used to isolate rac-dichlorprop degraders from the enrichment. A bacterial strain of Sphingobium, L3, capable of degrading rac-dichlorprop, was successfully isolated from the enrichment. The 16S rRNA gene sequence of Sphingobium sp. strain L3 shared 99.7% similarity with that of the DNA-SIP-identified OTU340 (Fig. S5), further confirming that OTU340 (Sphingobium) was truly involved in rac-dichlorprop degradation. rac-Dichlorprop, (R)-dichlorprop, or (S)-dichlorprop (30 mg liter⁻¹) was completely degraded by strain L3 (inoculated at an initial optical density at 600 nm [OD₆₀₀] of 0.05) within 12 h in mineral salt medium (MSM) (Fig. S9) and the OD₆₀₀ of strain L3 cells increased to the final value of 0.09, indicating that rac-dichlorprop could be utilized by strain L3 as the sole carbon and energy source for growth. Strain L3 preferred to degrade (R)-dichlorprop over (S)-dichlorprop (Fig. S9). Strain L3 harbored both rdpA and sdpA genes, which showed 100% similarities to rdpA (AF516752) and sdpA (AJ628860) genes from Sphingobium herbicidovorans MH, respectively (data not shown).

In addition, a bacterial strain of Achromobacter sp., D1, was isolated from the enrichment using 2,4-dichlorophenol as the enriching substrate, and the 16S rRNA gene sequence of Achromobacter sp. D1 shared 100% similarity with OTU265 (Fig. S10a). It was found that 30 mg liter⁻¹ 2,4-dichlorophenol was completely degraded by strain D1 within 12 h in MSM (Fig. S10b). The OD₆₀₀ of strain D1 cells increased from the initial 0.05 to a final 0.08 (Fig. S10b), indicating that 2,4-dichlorophenol could be utilized by strain D1 as the sole carbon and energy source for growth. However, strain D1 did not degrade rac-dichlorprop.

Enhanced degradation of rac-dichlorprop by a synergistic consortium.

The degradation efficiencies for rac-dichlorprop and cell growths of different artificial consortia were compared. The 2,4-dichlorophenol degrader Achromobacter sp. D1 instead of the nondegrader Escherichia coli significantly promoted the degradation efficiencies of both Sphingobium sp. L3 (the rac-dichlorprop degrader isolated from the enrichment in this study) and Sphingopyxis sp. DBS4 (the rac-dichlorprop degrader isolated from the enrichment previously in our lab [13]) (Fig. 3, left). In addition, the biomass of the consortia strain L3 plus strain D1 and strain DBS4 plus strain D1 increased more than that of consortia strain L3 plus E. coli and strain DBS4 plus E. coli (Fig. 3, right). All these data showed that Achromobacter sp. D1 can synergistically cooperate with Sphingobium sp. L3/Sphingopyxis sp. DBS4 in the enrichment to promote rac-dichlorprop degradation and cell growth.

FIG 3.

Degradation of rac-dichlorprop by and cell growth of different artificial consortia. Strains used in the artificial consortia were strain DBS4 (rac-dichlorprop degrader isolated previously in our lab), strain L3 (rac-dichlorprop degrader isolated in this study), strain D1 (2,4-dichlorophenol degrader isolated in this study), and E. coli (the nondegrader). Each strain was inoculated at an initial OD₆₀₀ of 0.02, so the consortia containing two strains were inoculated at an initial OD₆₀₀ of 0.04.

Taxonomical assignment of known rac-dichlorprop-degrading genes to bacterial genera.

To explore the associations between rac-dichlorprop degraders and known degrading genes, the assigning between rac-dichlorprop-degrading bacteria (Sphingobium and Sphingopyxis) and functional genes (rdpA and sdpA) was further analyzed. A total of 1,348 and 65,908 reads matching rdpA and 1,320 and 68,896 reads matching sdpA were found in the metagenomes of [¹²C]dichlorprop- and [¹³C]dichlorprop-labeled fraction 6, respectively (Fig. 4a), showing that rdpA and sdpA genes were enriched in the [¹³C]dichlorprop-labeled metagenome.

FIG 4.

Taxonomical assignment of rdpA and sdpA genes to bacterial genera. (a) rdpA and sdpA gene read numbers in [¹²C]- and [¹³C]dichlorprop-labeled metagenomes. (b) rdpA and sdpA genes taxonomically assigned to bacterial genera based on metagenomic analysis of the [¹²C]- and [¹³C]dichlorprop-labeled metagenomes.

Phylogenetic analysis of rdpA and sdpA genes found in the metagenomes showed that these initial degrading genes were closely related to the rdpA and sdpA genes reported in Sphingobium herbicidovorans MH (Fig. S11). As shown in Fig. 4b, both rdpA and sdpA genes were taxonomically assigned to the genus Sphingobium, while neither rdpA nor sdpA was assigned to the genus Sphingopyxis. This result indicated that Sphingopyxis species in our enrichment might harbor novel genes for rac-dichlorprop degradation, which is consistent with our previous finding that neither the rdpA nor the sdpA gene was found in our isolated strain Sphingopyxis sp. DBS4 (13).

DISCUSSION

Microorganisms usually live in complex communities. Community members interact with each other, shaping the structure and function of the community. rac-Dichlorprop degradation in environments is commonly carried out by a microbial consortium rather than a single species. In this study, DNA-SIP and metagenomic analysis were integrated to investigate the key microbial consortium involved in the degradation of rac-dichlorprop. Microorganisms represented by OTU340 and OTU348 were shown to play key roles in rac-dichlorprop degradation by the culture-independent method DNA-SIP. Phylogenetic analysis of OTU340 showed its close relationship to Sphingobium species, a Sphingomonadaceae member, which has been extensively confirmed for its versatile capacities to degrade various organic pollutants (27–30). Interestingly, a bacterial strain of Sphingobium sp. L3, capable of utilizing rac-dichlorprop as the sole carbon and energy source for growth, was also isolated from the enrichment, further confirming the role of Sphingobium species. In addition, the SIP-identified OTU348 shared 100% similarity to the 16S rRNA gene sequence of Sphingopyxis sp. DBS4 (MH553106) (see Fig. S5 in the supplemental material), a rac-dichlorprop-utilizing strain previously isolated from the same enrichment in our laboratory (13). Our results confirmed that Sphingobium and Sphingopyxis species were directly responsible for rac-dichlorprop degradation in the enrichment.

Analysis of the initial degrading genes for (R)- and (S)-dichlorprop in the [¹²C]- and [¹³C]dichlorprop-labeled metagenomes revealed a higher abundance of rdpA and sdpA genes in the [¹³C]dichlorprop-labeled metagenome, showing that these functional genes were enriched in the heavy DNA fraction. Both rdpA and sdpA genes were taxonomically assigned to the genus Sphingobium, while neither the rdpA nor the sdpA gene was assigned to the genus Sphingopyxis. Interestingly, nearly identical rdpA (AF516752) and sdpA (AJ628860) genes were found in the genome of Sphingobium sp. strain L3 (data not shown), while neither the rdpA nor the sdpA gene was detected in Sphingopyxis sp. DBS4 (13). These findings indicated that a potential novel dioxygenase gene involved in the initial degradation of rac-dichlorprop exists in the enrichment, especially in Sphingopyxis species.

At the genus level, Sphingobium, Sphingopyxis, Dyella, Sphingomonas, Pseudomonas, and Achromobacter showed 46.1%, 16.3%, 12.4%, 9.4%, 3.0%, and 1.63% relative contributions to the “chlorocyclohexane and chlorobenzene degradation pathways” (KEGG pathway level 3, ko00361), respectively. Sphingobium and Sphingopyxis species, as confirmed by DNA-SIP and degrading-strain isolation experiments, utilized rdpA, sdpA, and even novel genes to enantioselectively transform (R)-dichlorprop and (S)-dichlorprop to pyruvate and the nonchiral metabolite 2,4-dichlorophenol. Then, rac-dichlorprop-degrading strains have assimilated [¹³C]carbon from 2,4-dichlorophenol. A part of 2,4-dichlorophenol was assimilated by 2,4-dichlorophenol-degrading strains. Based on metagenomic analysis, taxonomically diverse bacterial species of genera such as Dyella, Sphingomonas, Pseudomonas, and Achromobacter were presumed to synergistically cooperate with Sphingobium/Sphingopyxis species for enhanced degradation of rac-dichlorprop (Fig. 5).

FIG 5.

A synergistic consortium involved in rac-dichlorprop degradation. Specific strains including Sphingobium, Sphingopyxis, Dyella, Sphingomonas, Pseudomonas, and Achromobacter species and the initial dioxygenases involved in the degradation of (R)- and (S)-dichlorprop are shown. TCA cycle, tricarboxylic acid cycle.

The metabolite 2,4-dichlorophenol was reported to be hydroxylated to 3,5-dichlorocatechol by a phenol hydroxylase, and 3,5-dichlorocatechol was finally mineralized via 2,4-dichloro-cis,cis-muconate, trans-2-chlorodienelactone, cis-2-chlorodienelactone, 2-chloromaleylacetate, and maleylacetate (15). Sphingomonas agrestis 58-1, Pseudomonas sp. strain GT241-1, and Achromobacter sp. strain LZ35 were reported to use 2,4-dichlorophenol as the sole carbon source for growth (31–36). More interestingly, the Achromobacter sp. D1, capable of utilizing 2,4-dichlorophenol as its sole carbon source for growth, was also isolated from the enrichment. In natural environments, degradation of organic compounds is often carried out by a microbial consortium rather than a single species. Many experiments have shown that cell growth and substrate utilization are frequently higher in a mixed culture (consortium) than in a single species (17, 37, 38). The enhanced degradation of pollutants by a consortium compared to a pure culture is potentially due to the complementarity of degradation pathways of individual members in the consortium (39, 40). On the other hand, cross-feeding (metabolites produced by one strain being utilized by others) between the community members also plays important roles in the better performance of the consortium (41). In our study, Dyella, Sphingomonas, Pseudomonas, and Achromobacter species (e.g., Achromobacter sp. D1) in the enrichment are most likely involved in the degradation of 2,4-dichlorophenol or its metabolites, which reduces the inhibition effect of 2,4-dichlorophenol on the key degraders such as Sphingobium sp. L3 and Sphingopyxis sp. DBS4 (see Fig. S12 in the supplemental material). In addition, the cross-feeding between Sphingobium/Sphingopyxis and Dyella, Sphingomonas, Pseudomonas, and Achromobacter members also promotes the cell growth of functional microorganisms and increases the stability of the community, thereby enhancing the degradation of rac-dichlorprop.

MATERIALS AND METHODS

Chemicals and medium.

rac-Dichlorprop (CAS RN 120-36-5) and (R)-dichlorprop (CAS RN 15165-67-0) were obtained from Aladdin Reagent Co., Ltd. (Shanghai, China). (S)-Dichlorprop was obtained by separation of the rac-dichlorprop at Chiralway Biotech Co., Ltd. (Shanghai, China). ¹³C-labeled rac-dichlorprop ([¹³C₆]dichlorprop, racemic form [see Fig. S13 in the supplemental material]) was obtained from Cambridge Isotope Laboratories, Inc. (Tewksbury, MA). Except where otherwise stated, rac-dichlorprop was used in this study. All other chemicals and solvents were of highest analytical reagent grade and were purchased from Sigma-Aldrich (Germany). Mineral salt medium [MSM; 1.0 g NaCl, 1.0 g (NH₄)₂SO₄, 1.5 g K₂HPO₄, 0.5 g KH₂PO₄, and 0.2 g MgSO₄·7H₂O per liter of water, pH 7.0] was used in this study.

Enrichment culture capable of degrading rac-dichlorprop.

An enrichment culture capable of degrading rac-dichlorprop that was derived from a rac-dichlorprop-contaminated soil in an abandoned pesticide factory in Nanjing, Jiangsu, China (31°57′9.58″N, 118°50′24.07″E), was obtained previously (13). The culture was incubated on a rotary shaker at 30°C and 180 rpm, and then 5 ml of subculture was inoculated into 100 ml of fresh MSM containing 30 mg liter⁻¹ of rac-dichlorprop and transferred every 7 days. The rac-dichlorprop-degrading capacity of the enrichment was detected via high-performance liquid chromatography (HPLC) as described below.

SIP experiment.

Two milliliters of the enrichment cultures was added to 18 ml of MSM supplemented with 30 mg liter⁻¹ of unlabeled rac-dichlorprop ([¹²C]dichlorprop) or ¹³C-labeled rac-dichlorprop ([¹³C]dichlorprop) and then incubated at 30°C. MSM (20 ml) supplemented with 30 mg liter⁻¹ of [¹²C]dichlorprop or [¹³C]dichlorprop was used as the control. At 4 days, approximately 75% of rac-dichlorprop (30 mg liter⁻¹) was degraded in both samples (Fig. S1), and then the DNA was extracted, respectively, using a FastDNA Spin kit (MP Biomedicals, Santa Ana, CA, USA) according to the manufacturer’s instructions. DNA content was determined using an ND-2000 UV-visible (UV-Vis) spectrophotometer (NanoDrop Technologies, Wilmington, DE). The experiments were carried out in triplicate.

Approximately 5 μg DNA of each sample was dissolved in Tris-EDTA (pH 8.0)–CsCl solution at a final buoyant density of ∼1.72 g ml⁻¹ and then poured into a Quick-Seal centrifuge tube (13 by 51 mm, 5.1 ml; Beckman Coulter, Pasadena, CA). Buoyant density was obtained by a digital refractometer (model AR200; Leica Microsystems Inc., Buffalo Grove, IL). The centrifuge tube was ultracentrifuged (Optima L-100XP; Beckman Coulter) at 45,000 rpm (20°C) for 44 h after heat sealing. Subsequently, DNA was fractionated in the tube, and then targeted fractions were collected by a fraction recovery system (Beckman Coulter). Finally, 14 fractions (400 μl each) were collected and the DNA in each fraction was purified using the method described by Thomas et al. (20).

The primers 338F (5′-ACTCCTACGGGAGGCAGCA-3′) and 806R (5′-GGACTACHVGGGTWTCTAAT-3′) (42, 43) were used to amplify the V3-V4 variable regions of the bacterial 16S rRNA gene derived from fractions 5 to 9, and then the amplification products were sequenced at Shanghai Majorbio Bio-pharm Biotechnology Co., Ltd. (Shanghai, China), using the Illumina HiSeq 2500 platform. Raw Fastq files were demultiplexed and quality filtered by Trimmomatic and merged using FLASH. Trimmed sequences were clustered to operational taxonomic units (OTUs) with a 97% similarity cutoff using UPARSE, and chimeric sequences were removed using UCHIME. The taxonomical assignment of OTUs was performed by the RDP Classifier algorithm against the Silva database using a confidence threshold of 70%.

Quantification of the rdpA and sdpA genes by qPCR.

The rdpA [the initial gene for (R)-dichlorprop degradation] or sdpA [the initial gene for (S)-dichlorprop degradation] gene copy numbers in each collected fraction of [¹²C]dichlorprop and [¹³C]dichlorprop samples were analyzed by quantitative real-time PCR in an Applied Biosystems 7300 real-time PCR system using a SYBR Premix Ex Taq RT-PCR kit (TaKaRa). The rdpA and sdpA genes were amplified using the primer pairs rdpA-f/rdpA-r and sdpA-f/sdpA-r1/sdpA-r2, respectively (44). Calibration curves of qPCR were obtained using serial dilutions of the plasmid pCETA/Blunt-Zero/rdpA carrying the rdpA gene or the plasmid pCETA/Blunt-Zero/sdpA carrying the sdpA gene. An equation was used to calculate the number of gene copies as previously described (45).

Metagenome sequencing.

As revealed by qPCR, the abundances of the functional genes rdpA and sdpA were highest in fraction 6 (buoyant density, 1.729 g ml⁻¹) of the [¹³C]dichlorprop sample. The triplicate metagenomes of fraction 6 from [¹³C]dichlorprop samples were combined as a metagenome and then randomly fragmented into ∼400-bp fragments using a Covaris M220 focused ultrasonicator (Covaris, Inc., Woburn, MA, USA). Then, a library was prepared using the NEXTflex rapid DNA-Seq kit (Bioo Scientific Corp., Austin, TX, USA) according to the manufacturer’s instructions. The metagenome of fraction 6 of the [¹²C]dichlorprop sample was also sequenced and used as the control. The two metagenomes were sequenced by Shanghai Majorbio Bio-pharm Biotechnology Co., Ltd. (Shanghai, China), using an Illumina NovaSeq 6000 sequencing platform (Illumina, San Diego, CA, USA). Adaptor removal and quality filtering were performed on raw reads using FASTP (https://github.com/OpenGene/fastp), and quality-trimmed reads that contained N (ambiguous bases) or were shorter than 50 bp were discarded by FASTP. To acquire contigs, MEGAHIT was used to assemble high-quality reads (46). The open reading frames (ORFs) of the assembled contigs were predicted by MetaGene (47). ORFs (length of >100 bp) were selected to establish the Non-Redundant (NR) gene catalog. Representative sequences of the NR gene catalog were annotated based on the NCBI NR database using blastp with an E value cutoff of 1e⁻⁵ using Diamond (48). The KEGG annotation was conducted using Diamond against the KEGG database (version 94.2) with an E value cutoff of 1e⁻⁵. To explore the distribution patterns between rac-dichlorprop degraders and ARGs, annotation of ARGs was conducted using Diamond against the CARD (https://card.mcmaster.ca/home, version 3.0.9) (E value of ≤10⁻⁵). Prediction of genes (ARGs, rdpA, sdpA, and 2,4-dichlorophenol catabolic genes) was carried out using blastp against the NCBI NR database with an E value of ≤10⁻⁵. The blastp results were annotated using MEGAN for assignment of taxonomic genus (49). The relative function/gene contribution of a taxon was represented by the abundance of the taxon relative to the total abundance of all taxa involved in the function/gene, which was calculated by summing the abundance of the taxon participating in the functions/genes assigned to that total abundance of all taxa involved in the functions/genes (50). The sum of abundances of all taxa/functions/genes detected in each subject was 1. Top 10 relative function/gene contribution of a taxon was selected for analysis, while other taxa were merged into others.

Isolation of degrading strains from the enrichment culture.

To isolate rac-dichlorprop-degrading strains from the enrichment culture, 5 mg liter⁻¹ ciprofloxacin was added because strains involved in rac-dichlorprop degradation might be resistant to fluoroquinolone antibiotics (indicated by metagenomic analysis). After continuative enrichment with ciprofloxacin for three rounds, the enrichment culture was diluted and plated onto LB agar plates containing 5 mg liter⁻¹ ciprofloxacin. Individual colonies grown on the plates were checked for their capabilities to degrade rac-dichlorprop. Finally, a rac-dichlorprop-degrading strain, designated L3, was isolated.

To isolate 2,4-dichlorophenol-degrading strains from the enrichment culture, the enrichment culture was diluted with sterile water, diluted on LB agar plates containing 30 mg liter⁻¹ 2,4-dichlorophenol, and then incubated at 30°C for 1 week. Individual colonies on the plates were further tested for their 2,4-dichlorophenol-degrading capability. Finally, a 2,4-dichlorophenol-degrading strain, designated D1, was isolated. A taxonomy analysis of the isolated strains was carried out as described previously (51), and strains L3 and D1 were identified as Sphingobium and Achromobacter species, respectively.

Degradation of rac-dichlorprop by artificial consortia.

Degradations of rac-dichlorprop by different artificial consortia with or without the 2,4-dichlorophenol degrader were compared. Strains used in the artificial consortia were Sphingopyxis sp. strain DBS4 (rac-dichlorprop degrader isolated previously in our lab [13]), Sphingobium sp. strain L3 (rac-dichlorprop degrader isolated in this study), Achromobacter sp. strain D1 (2,4-dichlorophenol degrader isolated in this study), and E. coli (the nondegrader). Artificial consortia were set as follows: (i) strain DBS4 and strain D1 (OD₆₀₀, 0.02:0.02), (ii) strain L3 and strain D1 (OD₆₀₀, 0.02:0.02), (iii) strain DBS4 and E. coli (OD₆₀₀, 0.02:0.02), and (iv) strain L3 and E. coli (OD₆₀₀, 0.02:0.02). Strain DBS4 alone (OD₆₀₀, 0.02) and strain L3 alone (OD₆₀₀, 0.02) were set for comparison. The concentration of rac-dichlorprop and the cell density of each consortium were measured every 4 h by HPLC and UV spectrophotometry, respectively. All treatments were carried out in triplicate.

HPLC analysis of rac-dichlorprop.

rac-Dichlorprop in liquid medium was extracted with an equal volume of dichloromethane-acetic acid (99:1, vol/vol), and the extract was evaporated and redissolved in methanol. The resulting solution was filtered through a 0.22-mm-pore Millipore membrane. rac-Dichlorprop was determined by an HPLC system (UltiMate 3000 Titanium) coupled with a C₁₈ reversed-phase column (5 μm, 250 mm by 4.6 mm). The mobile phase, consisting of 80% methanol (methanol-water, vol/vol) and 1.0% glacial acetic acid, was delivered at a flow rate of 1.0 ml min⁻¹ at 30°C. The injection volume was 20 μl, and the detection wavelength of rac-dichlorprop was 230 nm.

Data availability.

Sequences of 16S rRNA gene amplicons and metagenomics are available at NCBI under BioProject ID PRJNA730391.