Isolation, and identification of the pretilachlor-degrading strain B2
A pure bacterial culture with gram-positive and nonmotile cells was isolated and named B2. B2 Colonies were convex, opaque and red on LB agar after two days of cultivation. In addition, strain B2 tested positive for urease and catalase but negative for nitrate reduction, oxidase and starch hydrolysis. The 16S rRNA gene sequence showed that strain B2 had 100% similarity with strain R. erythropolis DSM43066T and 99.93% similarity with R. erythropolis NBRC100887T, forming a subclade with these two strains (Fig. S1). Therefore,strain B2 was preliminarily identified as Rhodococcus sp. basing on its characteristics.
Optimization of the cultivation conditions for pretilachlor-degradation
Three factors (pH, temperature, and inoculum size) with significant effects on microbial degradation were selected as the cultivation conditions to optimize using the CCD model. The data in Table S3 were used in multiple regression analysis, and response variable Y can be obtained using the following quadratic polynomial model equation:
Y=84.31+2.00A-0.622B+2.57C+3.3AB-2.39AC+0.76BC-8.81A²-3.26B²-0.76C² (4)
The ANOVA results for the quadratic response surface model are shown in Table 2. The regression model for pretilachlor degradation was statistically significant (P < 0.05) with R2 = 0.9501, and the results showed that A, C, AB, AC, A2, B2 significantly affected the pretilachlor degradation by strain B2. Thus, pH, inoculum size had significant effects on the degradation rate. Based on the P values of AB and AC (0.0075 and 0.037), the pH-temperature and inoculum size-pH interaction effects on the pretilachlor degradation were highly significant. Therefore, a response surface analysis was conducted to determine the impacts of the interaction between pH and temperature on pretilachlor-degradation by strain B2 (Fig. 1). These results revealed the maximum rate of pretilachlor degradation by strain B2 was 86.1% under the optimal conditions of pH 6.98, 30.1°C, and inoculum size of 0.3 g/L.
Kinetics of pretilachlor degradation by strain B2
The effect of the initial pretilachlor concentration on the degradation of pretilachlor was calculated via nonlinear least squares regression analysis using Origin 9.0pro and the results are shown in Fig. 2. The kinetic parameters were as follows: qmax = 3.28 d−1, KS = 53.51 mg/L, and Ki = 25.38 mg/L. The Sm was 36.84 mg/L, indicating that the theoretical efficiency of pretilachlor degradation was highest at this concentration, when the concentration of pretilachlor was more than 36.84 mg/L, the inhibition of strain B2 by the pretilachlor was obvious. This result may be attributed to the toxicity of pretilachlor to the assayed strain. The Andrews model was as follows:
Statistical regression results revealed the parameters of pretilachlor degradation kinetics (Table. S4). The resulting correlation coefficient R2 = 0.8285, indicates that the model was an excellent fit to the experimental data. The ability of strain B2 to degrade pretilachlor degradation increased at low pretilachlor concentrations, but decreases at higher pretilachlor concentrations. And this kinetic model is helpful for predicting the microbial bioremediation of pretilachlor by strain B2.
Identification of the metabolites resulting from pretilachlor degradation by strain B2
The products of pretilachlor degradation by strain B2 were detected by GC-MS. A compound from the control sample had the same RT as the pretilachlor standard (RT = 11.5 min, Fig. 3A). The molecular ion (M+) of peak (RT = 11.5 min) was 311 m/z with characteristic fragment ions mass spectral data showing a 96% match with pretilachlor in the NIST library (Fig. 3C).
In addition, a new metabolite appeared at a retention time of 11.1 min (Fig. 3B). As we could not obtain the standard for the metabolite, and it was identified using GC-MS analyses. The M+ peak of this product was 269 m/z, and the characteristic fragment ions were 237.9 m/z(M+−CH2OH), 220.0 m/z(M+−CH2Cl), 202.0 m/z (M+−CH3ClOH), 192.0 m/z(M+−CH2COCl), 176.0 m/z(M+−(CH2CH3)2Cl), 162.1 m/z(M+−(CH2CH3)2CH2Cl), 147.1 m/z(M+−CH2COCl(CH2)2OH) and 132.0 m/z(M+−CH2COCl(CH2)2OHCH3), (Fig. 3D), These mass spectral data of the metabolite correspond to N-hydroxyethyl-2-chloro-N-(2,6-diethyl-phenyl)-acetamide which has not previously been reported during pretilachlor degradation, indicating it to be a novel product. In addition, LC-MS/MS analysis demonstrated the presence of pretilachlor and the metabolite N-hydroxyethyl-2-chloro-N-(2,6-diethyl-phenyl)-acetamide, results that are consistent with the corresponding GC-MS analysis (Figure S9 in Supporting Information). However, the metabolite could not be further metabolized by strain B2. Therefore, pretilachlor degradation process by strain B2 involves O-dealkylation, representing a new mechanism of initial chloroacetamide herbicide degradation.
Screening of a mutant, TB2, defective in pretilachlor degradation
When grown on LB agar containing 100 mg/L pretilachlor, the colonies of strain B2 could produce a visible transparent halo, and N-hydroxyethyl-2-chloro-N-(2,6-diethyl-phenyl)-acetamide, which is more water-soluble than pretilachlor, was formed. In our present study, we observed that a few cells of strain B2 did not generate a transparent halo after continuous streaking on fresh LB agar plates, and one such isolate was named TB2. Resting cell transformation experiments revealed that TB2 could not metabolize pretilachlor (Fig. S2), suggesting that the related gene responsible for O-dealkylation in pretilachlor degradation was deleted in the mutant TB2.
Genome comparison between strains B2 and TB2
The draft genomes of strains B2 and TB2 were sequenced with the Illumina MiSeq system and were shown to be 6,873,325 bp and 6,728,834 bp in length, respectively. Furthermore, a comparison of the genomes of strain B2 and TB2, resulted in the identification of a fragment from scaffold 51 of strain B2 that was absent in the genome of mutant TB2. The absence of this fragment was then verified by PCR. after which the flanking regions of scaffold 51 were confirmed by SEFA-PCR. Finally, a 115,851-bp fragment was acquired. And sequence comparison combined with PCR demonstrated that a 32,147-bp region of this fragment was absent in mutant TB2 (Fig. 4A).
ORF analysis of the fragment absent in strain TB2
A gene cluster consisting of five genes, EthRB2, EthAB2, EthBB2, EthCB2 and EthDB2, was identified by an analysis of all ORFs, and the encoded amino acid sequence of the genes in the missing fragment were identified in NCBI (Table S6). EthRB2, encoding the AraC/XylS family regulator (37 kDa), shares the highest similarity with the EthR from Rhodococcus sp. T3-1 (100%). EthAB2, encoding a ferredoxin reductase (43 kDa), shares the highest similarity with EthA from Rhodococcus sp. T3-1 (100%). EthBB2, encoding a cytochrome P450 oxidase (44 kDa), shares the highest similarity with EthB from Rhodococcus sp. T3-1 (97.5%). EthCB2, encoding a 2Fe-2S ferredoxin (11 kDa), shares the highest similarity with EthC from Mycobacterium sp. CH28 (99.06%). EthDB2, encoding a protein of unknown function (10 kDa), shares the highest similarity with EthD from Mycobacterium sp. CH28 (90.29%) (Fig. 5). The gene EthB was termed EthBB2 (EthB from strain B2), and its inferred amino acid sequence was aligned to that of EthBT3-1 from Rhodococcus sp. strain T3-1, as shown in Fig. 4B. Ten amino acid differences were observed between the two proteins, which may confer different physical properties to EthBB2. The high similarity of the two proteins indicated the occurrence of horizontal gene transfer event (Fig. 4A). The upstream eth gene cluster contains two gene fragments (tnpA1, and tnpA2) encoding the proteins displaying >99% sequence identity with the Tn3 family transposase (TnpA) and one fragment (tnp1) belonging to the IS30 family transposase (Table S6). However, a transposase was not identified downstream of the eth gene cluster. In contrast, two transposons, IS3-type class II, flanked the EthRABCD gene cluster of R. ruber IFP 2001 [37].
Functionally complementation of the Eth gene cluster in strain TB2
To determine the function of the gene cluster EthRABCDB2, the recombinant plasmid pQeth1 containing EthRABCDB2 was introduced into strain E. coli DH5α and strain TB2. The recombinant strain TB2 (pQeth1) acquired the ability to degrade pretilachlor and generate a visible transparent halo in LB agar supplemented with 100 mg/L pretilachlor, which was similar to that observed for strain B2. HPLC results showed that TB2 (pQeth1) could degrade pretilachlor, and exhibited the O-dealkylation activity (Fig. S2). A similar phenomenon was found in strain Rhodococcus sp. R-XP(pQeth1). However, E. coli DH5α(pQeth1 or pQeth2) and strain TB2(pQeth2) failed to degrade pretilachlor, indicating that EthRABCDB2 was expressed at a low level or the original promoter could not promote transcription of the cluster in E. coli, and EthDB2 was essential for degradation. In order to verify the above hypothesis, the EthABCDB2 and EthABCB2 fragments under the control of the T7 promoter in vector pET-29a(+) were introduced into E. coli BL21(DE3). Whole-cell transformation assay results showed that the IPTG-induced suspension of E. coli BL21(DE3) harboring EthABCD (but not EthABC) was able to degrade pretilachlor , although with low activity (data not shown), These results indicated that EthRB2, EthCB2 gene are not essential and that the original promoter is important for the Eth gene cluster. Therefore, EthABCDB2 and EthABDB2 were reconstituted with original promoters to analyze the degradation of chloroacetanilide herbicides. The HPLC results showed that strain TB2 (pQeth3, and pQeth4) could convert pretilachlor, butachlor, alachlor, acetochlor and propisochlor to the corresponding metabolites, indicating that EthD was likely a ferredoxin gene.
Expression of the gene cluster EthABCDB2 and reconstitution of the chloroacetanilide herbicide degradation enzyme in vitro
The components EthAB2, EthBB2, EthCB2 and EthDB2 were expressed in E. coli BL21(DE3)pLysS separately, and each recombinant protein was purified using Ni-affinity chromatography. The Mw value of the four proteins were consistent with the theoretically calculated values (Fig. S3). Purified EthCB2-His6 and EthDB2-His6, were mixed with EthAB2-His6 and EthBB2-His6 in vitro. The results showed that the EthABDB2-His6 and EthABCDB2-His6 (not EthABCB2-His6) mixture showed pretilachlor-degrading activity indicating that EthD is a ferredoxin. The catalytic activity of EthABDB2 toward pretilachlor was 3.41 ± 0.4 μmol/min/mg. EthABDB2 performed N-dealkoxymethylation to acetochlor, alachlor, propisochlor butachlor, O-dealkylation to pretilachlor, but it was unable to degrade metolachlor. The GC-MS results showed that strain EthABDB2 could convert butachlor, alachlor, acetochlor and propisochlor to the corresponding metabolites CDEPA (for butachlor and alachlor) or CMEPA (for propisochlor and acetochlor) (Fig. S4-S7). These metabolites are generated from C-N bond cleavage by N-dealkoxymethylation, and based on the comparison of the chemical structures of chloroacetanilide herbicides, the number of C-atoms between the N and O in the side chain affects the degradation. According to these results, the mechanism of pretilachlor degradation (O-dealkylation) was different from that of the other four chloroacetanilide herbicides(N-dealkylation), and the degradation pathway of chloroacetanilide herbicides by EthABDB2 was proposed (Fig. 4C). The recombinant strain TB2 (pQeth4) could not degrade metolachlor, suggesting that steric hindrance blocks enzyme-substrate interactions. EthBB2, a cytochrome P450 monooxygenase of the multicomponent system, plays a key role in chloroacetanilide herbicide degradation, and EthABDB2 from strain B2 has a broader substrate spectrum than that of the corresponding enzymes from strain T3-1. Thus, EthABDB2 is a better enzyme for practical bioremediation of chloroacetanilide herbicides.
Characteristics of EthABDB2
The effects of different environmental factors on the enzymatic activity of EthABDB2 were determined (Fig. S8). The enzyme activity was assessed at 10-65 °C, and was shown to function optimally at 30 °C (Fig. S8B). In addition, EthABDB2 showed the enzyme showed high activity at pH 7.0-8.5, with an optimum pH of 7.5 in Tris-HCl buffer (Fig. S8A), and decreased activity was lost observed at pH values below 4.0 or above 10.0. Metal ions play an important role in the enzyme activity. As shown in Table S5, Fe2+ and Mg2+ could strongly enhance EthABDB2 activity, while Ca2+ could also slightly increase enzyme activity. However, the divalent cations Ba2+, Co2+ and Zn2+ slightly decreased enzyme activity, while Ag+, Cu2+, Hg2+, Ni2+ Cr2+ and Mn2+ significantly inhibited its enzyme activity. Chemical agents EDTA severely inhibited the enzyme activity, indicating that metal ions are required for its enzymatic reaction.