The impact of pollutant as selection pressure on conjugative transfer of dioxin-catabolic plasmids harbored by Rhodococcus sp. strain p52

Plasmid-mediated bioaugmentation has potential application in the cleanup of recalcitrant environmental pollutants. In this study, we examined the influence of various contaminants (in different categories or different amounts) as a selection pressure on the spread of catabolic plasmids within an activated sludge bacteria community bioaugmented with Rhodococcus sp. strain p52 harboring pDF01 and pDF02. The distinguishable genera of transconjugants were isolated under the stresses of phenanthrene, dibenzothiophene, and dibenzo-p-dioxin. The three contaminants exerted different degrees of influence on the activated sludge bacteria bearing the catabolic plasmids. The relatively high ratios of transconjugant-bearing catabolic plasmids were detected in the reactor fed with dibenzo-p-dioxin. As dibenzo-p-dioxin from 10 to 80 mg/L was fed into the reactors, the ratios of transconjugant-bearing catabolic plasmids increased. Additionally, levels of ROS and extracellular LDH of activated sludge bacteria in the contaminants-fed reactors increased, comparing with that in the control reactor, indicating that the contaminants exerted toxicity which promoted the cell membrane permeability of the activated sludge bacteria. Our study provides a characterization of the recalcitrant contaminants as a selection pressure that can modulate catabolic plasmid transfer during genetic bioaugmentation for the removal of contaminants.


Introduction
The escalating rate of industrialization has led to the release of various recalcitrant organic compounds into the environment during the past century. Among these compounds, persistent organic pollutants comprise one of the most notorious groups, as they include polychlorinated dibenzodioxins, polychlorinated dibenzofurans, and polychlorinated biphenyls (Kulkarni et al. 2008;Urban et al. 2014). With stable structures and high hydrophobicity, they are resistant to biodegradation in natural environments (Iyer et al. 2016). There are obstacles to the efficient removal of these pollutants by bacteria in treatment engineering facilities or by in situ remediation systems (Herrero and Stuckey 2015). Fortunately, some excellent microbial degraders have been equipped with evolved catabolic pathways (Chakraborty and Das 2016;Kolvenbach et al. 2014;Sharma et al. 2018) and can be exploited to reinforce the indigenous microbiota that facilitate bioaugmentation (Carlos et al. 2017;Herrero and Stuckey 2015). Bioaugmentation is operated in two nuanced approaches: inoculating efficient degrading strains to dominate the degradation of pollutants (cell bioaugmentation) or introducing mobile genetic elements (MGEs) to disseminate catabolic ability to indigenous bacterial populations (genetic bioaugmentation) (Carlos et al. 2017). Due to the frequent failure of the inoculated cells during interactions with biological and non-biological environmental factors, the latter has significant advantages over the former (Carlos et al. 2017;Herrero and Stuckey 2015). In most cases, the key genes responsible for degrading the recalcitrant pollutants are located on MGEs, including plasmids, transposons, integrons, and genomic islands or phages (Top et al. 2002;Tsuda et al. 1999;van der Meer et al. 2001). In particular, conjugative or mobilizable plasmids are the major carriers during the horizontal transfer of genes, and they have played a major role in genetic bioaugmentation (Bathe 2004;Carlos et al. 2017). The selection of donor-plasmid pairs is a priority in the application of genetic bioaugmentation. A highly transferrable plasmid with a broad host range and encoding catabolic pathways for target pollutants are prerequisites for genetic bioaugmentation application. Additionally, robust donor adaptation to the environment (e.g., reactors or contaminated sites) is preferable, although independent of the donor's long-term survival. After introducing the plasmid-harboring donor into an indigenous bacterial community, the proportions of catabolic plasmid (or catabolic pathway) bearers in the bacterial community seem to determine the removal efficiency of the target pollutant (Top and Springael 2003). Environmental conditions, including biotic and abiotic factors, can influence genetic bioaugmentation. These factors include the morphological and physiological characteristics of donor cells, the ratio and phylogenetic relatedness of donors and recipients, temperature, pH, and the features of contaminants (Alderliesten et al. 2020;Carlos et al. 2017;Shintani et al. 2018;Stallwood et al. 2005).
During contaminant bio-treatment, the contaminant itself plays a critical role in plasmid-mediated genetic bioaugmentation. On the one hand, the recalcitrant organic compounds can provide nutrients for the growth of the bacterial cells that have acquired catabolic plasmids. Thus, the plasmid bearers gain a unique niche with competitive advantages over the bacteria without plasmids. On the other hand, the recalcitrant contaminants exert toxic effects on the bacterial cells and trigger a cascade of responses in the cells to ameliorate the situation (Zhao et al. 2019). In this sense, the contaminants serve as a selection pressure to promote the catabolic plasmidbearing population in the bacterial community, like a driver of genetic bioaugmentation. However, the expression and replication of plasmids require energy and resources of the host cells, and thus catabolic plasmids also carry a burden (i.e., fitness cost) to the bearer (San Millan et al. 2014). As a result, the plasmid-bearing cells within a bacterial population are prone to be replaced by the nonbearing cells in the absence of selection pressure (Ikuma and Gunsch 2012). Consequently, selection pressure is required to maintain the ratio of transconjugants in the bacterial community. During contaminant treatment by microorganisms, continuously present selection pressure modulates catabolic plasmid transfer and confers a fitness advantage to the hosts (i.e., transconjugants) (Fan et al. 2019). Although attention has been paid to the influence of some contaminants as selection pressures during catabolic plasmid transfer (Bouma and Lenski 1988;Ikuma and Gunsch 2012;San Millan and Maclean 2017), much remains unclear, especially the action modes of the contaminants. For example, there is little information concerning the impacts of the category and strength of the selection pressure (deriving from different contaminants) on catabolic plasmid transfer compared to the flourishing studies on the influence of antibiotics on horizontal transfer of resistance genes. Thus, it is worthwhile to focus attention on the effects of contaminants as a selection pressure on catabolic plasmid transfer.
We previously isolated Rhodococcus sp. strain p52, which can utilize dibenzofuran as the sole carbon and energy source (Peng et al. 2013). This strain contains two gene clusters (dfdA1A2A3A4 and dbfA1A2) encoding two angular dioxygenases located on two plasmids, pDF01 and pDF02 (Peng et al. 2013), that can initiate the dihydroxylation of a wide range of aromatic compounds such as dibenzofuran, dibenzo-p-dioxin, dibenzothiophene, and phenanthrene (Kasuga et al. 2013;Peng et al. 2013). The plasmids pDF01 and pDF02 contain the catabolic genes for an integrated dioxin degradation pathway. The catabolic plasmids can be conjugatively transferred into a broad range of bacterial strains (Sun et al. 2017). Genetic bioaugmentation of activated sludge with the plasmids harbored by strain p52 enhanced the removal of dibenzofuran in laboratory-scale sequencing batch reactors. The catabolic plasmids pDF01 and pDF02 have shown potential application in genetic bioaugmentation for contaminant remediation (Ren et al. 2018). The purpose of the present study is to examine the effect of contaminants as a selection pressure on genetic bioaugmentation with the catabolic plasmids. To achieve this, we examined the spread of the catabolic plasmids under various selection pressures (different categories or different amounts of contaminant) within an activated sludge bacteria community bioaugmented with strain p52. We investigated the host range and monitored the ratios of transconjugant-bearing catabolic plasmids under the different selection pressures. Additionally, we measured levels of reactive oxygen species (ROS) and extracellular lactate dehydrogenase (LDH) of the activated sludge bacteria under the contaminants stresses. Our study will provide a characterization of the recalcitrant contaminants as a selection pressure to modulate catabolic plasmid transfer during genetic bioaugmentation for the removal of contaminants.

Activated sludge reactor setup
A series of laboratory-scale activated sludge reactors was set up. The working volume of the reactor was 1 L, and the inner diameter and the length of the reactor were 90 mm and 300 mm, respectively. An air diffuser was installed at the bottom of each reactor, and the aeration rate was 0.8 L/min. The activated sludge (taken from the aeration tank of the wastewater treatment plant of Qingdao Campus, Shandong University) was aerated for a week to exhaust carbon sources. The initial mixed liquid-suspended solids (MLSS) concentration of each reactor was 3.6 g/L, and the experimental temperature and pH were controlled at 25°C and 6.0−8.0, respectively. The hydraulic retention time was 24 h, operating in a sequencing batch mode (each cycle was 5 min for influent, 20 min for precipitation, 5 min for effluent, and the remaining time for aeration. The volume exchange ratio was 50%). The synthetic wastewater was fed into the reactors in the composition 0.462 g/L NH 4 Cl, 0.108 g/L KH 2 PO 4 , 0.540 g/L MgSO 4 ·7H 2 O, 0.216 g/L KCl, 0.066 g/L CaCl 2 , 0.006 g/L yeast extract, and 1.8 mL/L trace element solution. The trace element solution consisted of 0.240 g/L MnCl 2 ·4H 2 O, 0.240 g/L ZnCl 2 , 0.300 g/L H 3 BO 3 , 0.060 g/L KI, 0.300 g/L Na 2 MoO 4 ·2H 2 O, 0.116 g/L CoCl 2 · 6H 2 O, and 0.060 g/L CuSO 4 ·5H 2 O. The pH value of the synthetic wastewater was around 7.0 (van Loosdrecht et al. 2016). Dibenzo-p-dioxin, dibenzothiophene, and phenanthrene were dissolved in ethanol as stock solutions with a final concentration of 20 g/L. To test the strength of selection pressure, dibenzo-p-dioxin in a concentration gradient of 10, 20, 40, and 80 mg/L was supplemented to the synthetic wastewater. To test the category effect of selection pressure, dibenzo-p-dioxin, dibenzothiophene, and phenanthrene with concentrations of 40 mg/L were supplemented to the synthetic wastewater. There were two different types of control reactors, including strain p52-noninoculated control for testing the removal of contaminants by activated sludge bacteria, and strain p52inoculated control for testing the influence of selection pressure. In strain p52-inoculated control reactor, sodium acetate trihydrate was supplemented to the synthetic wastewater instead of the contaminant. Each test was repeated at least three times. The COD concentration of influent in the reactor was about 3100 mg/L, and the ratio of COD, N (NH 4 + -N) and P (PO 4 3 -P) was 100:5:1. Except the strain p52-noninoculated control, other reactors were inoculated with 10 mL of strain p52 cell suspension of a cell density of 10 6 −10 7 CFU/mL in the reactor. Strain p52 cell suspension was prepared as follows: strain p52 was pre-cultured in a 250-mL Erlenmeyer flask containing 50 mL LB medium and shaken in an incubator at 30°C and 180 rpm for 24 h. The cells were centrifuged and washed with carbon-free mineral salts medium (MSM) three times, then resuspended in the MSM (OD600=1).
To determine the concentrations of residual contaminants and MLSS in the reactors, 50 mL and 100 mL of the mixed liquor samples were taken from the reactor at the end of each cycle. The sampling for analysis was carried out within the first 2 days in this study. Extraction and analysis of the residual contaminants were conducted as described in a previously study (Ren et al. 2018). The MLSS concentration was determined according to standard methods (American Public Health Association 1998). The specific removal rate for each contaminant was calculated as follows: where I and R are the initial concentration and residual concentration of the contaminant, respectively, and t is the time for removal. The specific removal rate for each contaminant within the first day was calculated in this study. Data are shown as the means with standard deviations of the triplicate tests.

Transconjugant acquisition
Prior to isolating transconjugants, the seed sludge and the mixed liquor samples from the strain p52-noninoculated control reactors fed with contaminants were subjected to DNA extraction and amplification of dfdA and dbfA fragments using primers shown in Table 1. After confirming the absence of dfdA and dbfA harbored by bacteria in non-bioaugmented activated sludge, the mixed liquor samples were taken from the strain p52-bioaugmented reactors fed with contaminants, and spread on the selective plates prepared from dibenzofuransupplemented carbon-free MSM at the end of each operation cycle. After being cultured at 30°C for 4 days, the colonies (mostly white round colonies) with morphological characteristics different from strain p52 (orange-red round colonies) were chosen for further confirmation by colony PCR as described by Ren et al. (2018). The primers used in this study are shown in Table 1. The proper sized PCR products checked by electrophoresis were sequence analyzed for further confirmation. Simultaneously, transconjugants were identified by amplification and sequencing analysis of the 16S rRNA gene for Blast searches against the NCBI nucleotide database. The same operation was performed for strain p52-inoculated control reactor (without contaminant).

DNA extraction and quantitative polymerase chain reaction analysis
At the end of each operation cycle, 10 mL mixed liquor samples were taken from each reactor for total DNA extraction using a soil DNA isolation kit (Omega Biotek, Inc., Norcross, GA) according to the manual. The qPCR was conducted as previously described using the iCycler thermocycler (BIO-RAD), and absolute copy numbers of target DNA fragments were calculated based on the obtained Ct value and the standard curve (Yun et al. 2006). The qPCR primer information is shown in Table 1. Based on the absence of dfdA and dbfA harbored by bacteria in the non-bioaugmented activated sludge, the copy numbers of dfdA and dbfA fragments were determined to evaluate the amount of catabolic plasmids harbored by transconjugants and strain p52. Meanwhile, the copy number of the specific fragment targeting an intergenic spacer region (ISR) between the 16S and 23S rRNA genes of strain p52 was determined to evaluate the amount of strain p52. And the copy number of the 16S rRNA gene of the total bacteria in the mixture was determined. The ratio of transconjugantbearing pDF01 or pDF02 in the bacterial community was assessed based on the absolute copy number of the target DNA detected by qPCR modified from that of Bellanger et al. (2014), and calculated as the ratio of copy numbers of transconjugant-bearing pDF01 or pDF02 (indicated by plasmid amount subtracted the contribution of strain p52) to donor. The calculation formula was as follows: where I t is the copy number of dfdA (located on pDF01) or dbfA (located on pDF02) fragment within t days; R t and R o are the copy numbers of strain p52 ISR within t days and the initial copy number of strain p52 ISR, respectively. A is the copy number of pDF01 or pDF02 per cell of Rhodococcus sp. strain p52; these are 1 and 4, respectively, which were quantified according to San Millan et al. (2014). Data are shown as the means with standard deviations of the triplicate tests.

ROS and extracellular LDH assay for activated sludge bacteria in reactors
Triplicate activated sludge mixed liquor samples were taken from activated sludge reactors for ROS assays according to the instructions of the Reactive Oxygen Species Assay Kit (Beyotime, Jiangsu, China). Briefly, the sample was centrifuged to remove the supernatant and washed with PBS buffer; diluted 2′,7′-dichlorodihydrofluorescein diacetate (DCFH-DA) was added, and the sample was incubated at 37°C in the dark for 20 min. After DCFH-DA enters the cell, it is h y d r o l y z e d b y i n t r a c e l l u l a r e s t e r a s e t o 2 ′ , 7 ′dichlorodihydrofluorescein (DCFH), which cannot penetrate the cell membrane. Intracellular ROS can oxidize DCFH to produce fluorescent 2′,7′-dichlorofluorescein (DCF). The fluorescence intensity of DCF at 488-nm excitation wavelength and 525-nm emission wavelength was measured using a microplate reader (TECAN, USA). Data are reported as the means with standard deviations for triplicate assays.
To assess cell membrane integrity, the activity of LDH released from bacterial cells into the water phase of the mixed liquor was measured. After the end of each cycle, 1.8 mL of the mixed liquor was taken from each reactor for LDH detection according to the instructions of the LDH Release Assay Kit (Beyotime, Jiangsu, China). Briefly, the sample was centrifuged, and the supernatant was mixed with LDH detection working solution (consisting of diaphorase, 2-p-iodophenyl-3-nitrophenyltetrazolium chloride (INT), and lactic acid) and incubated at 25°C for 30 min in the dark. Nicotinamide adenine dinucleotide (NAD + ) was reduced to reduce nicotinamide adenine dinucleotide (NADH) by LDH, and then NADH and INT were catalyzed by diaphorase to generate NAD + and formazan. The absorbance of the produced formazan at a wavelength of 490 nm was measured using a spectrophotometer (HACH, USA). Three assays of extracellular LDH contents under different selection pressures were conducted. Data are reported as the means with standard deviations for triplicate assays.

Statistical analysis
Statistical analysis was performed using SPSS version 26.0. The differences between means compared to controls were analyzed by one-way analysis of variance

Isolation of transconjugants in activated sludge fed with different contaminants
Transconjugants were isolated from activated sludge during contaminant removal, yet failed from the control reactors due to plasmid loss after subculture. These transconjugants were confirmed to harbor two catabolic plasmids, pDF01 and pDF02 (Fig. 1). A total of nine genera-affiliated transconjugants were isolated, and the general features of different genera of transconjugants are listed in Table 2. The distinguishable genera of transconjugants were isolated under the stresses of different contaminants. In particular, transconjugants belonging to all nine genera were isolated during treatment of dibenzo-p-dioxin. Transconjugants belonging to Arthrobacter and Klebsiella were isolated from activated sludge during treatment of each of the three contaminants.
The average G+C contents of the transconjugants were compared to the donor strain according to the GenBank genome database. The average genomic G+C content of most species of transconjugants differed from that of the strain p52 by more than 5.0%, and the difference in genomic G+C content between Acinetobacter and strain p52 was 27.5%. The results indicated a broad host range of the catabolic plasmids (Drønen et al. 1998).
The specific removal rates of the activated sludge for phenanthrene, dibenzothiophene, and dibenzo-p-dioxin are shown in Fig. 2. Comparing with non-bioaugmented control, activated sludge bacteria in the bioaugmented reactors degraded these compounds more efficiently. This indicated that bioaugmentation of the reactors with strain p52 enhanced the degradation of different contaminants. And the enhancement of degradation was contributed by strain p52 and the transconjugants in activated sludge that had acquired pDF01 and pDF02.

Catabolic plasmid spread within activated sludge bacteria community under the stress of different contaminants
Previous studies have demonstrated that dibenzo-p-dioxin, dibenzothiophene, and phenanthrene are all substrates for the dioxygenases encoded on catabolic plasmids (Kasuga et al. 2013;Peng et al. 2013). The three recalcitrant compounds were fed into reactors to test the effect of different contaminants as positive stresses on catabolic plasmid transfer. As activated sludge bacteria could not efficiently utilize these compounds, the MLSS of the reactor slightly decreased (p>0.05) over time ( Fig. SM-1, online resource). The amount of the total activated sludge bacteria was monitored (Fig. 3a). The amount of the total bacteria in all reactors increased within 2 days, and the increases were far greater in the reactors fed with phenanthrene and dibenzothiophene than in the sodium acetate-fed control reactor. Simultaneously, the amount of strain p52 in all bioaugmented reactors increased significantly after 1 day and were higher than in the control (Fig. 3b), indicating the adaptability and robustness of the introduced strain p52 in the reactors with the contaminants. As shown in Fig. 3c and d, the contaminants exerted different degrees of influence on the activated sludge bacteria bearing the catabolic plasmids. The relatively high ratio of transconjugant-bearing catabolic plasmid was detected in the reactor fed with dibenzo-p-dioxin within two days, and the least was observed with phenanthrene. The ratio of transconjugant-bearing catabolic plasmid in all reactors fed with the contaminants increased with time. In addition, different ratios for pDF01 and pDF02 were observed, even if the difference in their copy numbers within a cell was deducted, indicating the distinct dynamics of the two catabolic plasmids in sludge bacterial cells. To examine the strength of selection pressure on catabolic plasmid transfer, dibenzo-p-dioxin of different concentration was fed into the reactors. The amount of total activated sludge bacteria and strain p52 in reactors fed with different concentrations of dibenzo-p-dioxin was monitored by qPCR. As shown in Fig 4a and b, the amount of total activated sludge bacteria and strain p52 increased in the reactors fed with dibenzo-p-dioxin. Specifically, the amount of activated sludge bacteria increased significantly on day 2 compared to that on day 1, except in the reactor containing 10 mg/L dibenzo-pdioxin. A similar trend was observed for the amount of strain p52. The ratios of transconjugant-bearing catabolic plasmids were evaluated ( Fig. 4c and d). As dibenzo-p-dioxin concentrations were supplemented from 10 to 80 mg/L, the ratios of transconjugant-bearing catabolic plasmids gradually increased within the experimental period. The results indicated that the concentration of dibenzo-p-dioxin, in other words, the strength of selection pressure, promoted the transconjugant proportion in the activated sludge bacterial community. Additionally, the ratio of transconjugant-bearing pDF02 was disproportionate to pDF01, with the exclusion of the difference in their copy numbers within a cell, indicating that transconjugants containing only one catabolic plasmid might be present in the sludge bacterial community. Accordingly, only one catabolic plasmid could be detected by PCR amplification of the target genes on the catabolic plasmids in some transconjugants ( Fig. SM-2, online resource).

ROS content and membrane permeability of activated sludge bacteria under the stress of different contaminants
In order to understand the mechanism of selective pressure on catabolic plasmid transfer, the ROS content of activated sludge bacteria in the reactor fed with contaminant was compared to that in the control reactor fed with sodium acetate. As shown in Fig. 5a, the ROS levels of bacteria in the reactors fed with phenanthrene, dibenzothiophene, and dibenzo-p-dioxin were higher than that in the control reactor during 48-h monitoring. Additionally, when dibenzo-p-dioxin was fed into the reactor at different concentration, the ROS level of activated sludge bacteria showed different degree of increase compared to that in the control reactor (Fig. 5b). The results demonstrated that the contaminants could promote ROS level of activated sludge bacteria, indicating that the contaminant exerted cytotoxicity on activated sludge bacteria. Release of LDH is an indicator of cell membrane leakage (Myllyluoma et al. 2008). Cell membrane permeability of the activated sludge bacteria was evaluated by the activity of extracellular LDH of the activated sludge bacteria. The level of extracellular LDH of activated sludge bacteria increased significantly in the contaminant-fed reactor, comparing with that in the sodium acetate-fed control reactor during 48-h monitoring (Fig. 6a). On the other hand, release of LDH by the activated sludge bacteria was stronger in the reactors fed with dibenzo-p-dioxin at different concentrations than that in the control reactor (Fig. 6b). The results were in accord with the ROS level variation of the activated sludge bacteria under the stresses of contaminants, indicating that the contaminants  The red column represents specific removal rate in the reactor bioaugmented with Rhodococcus strain p52, and the grey column represents that in non-bioaugmented control reactor. Double asterisk denotes the group is statistically different from the control group (p < 0.01) exerted toxicity which promoted the cell membrane permeability of the activated sludge bacteria.

Discussion
Plasmid-mediated catabolic gene transfer has played an important role in bacterial evolution, and at present the process contributes to the efficient cleanup of emerging recalcitrant pollutants in the environment (Urban et al. 2014). In the present study, we characterized the influence of contaminants as selection pressures on the spread of catabolic plasmids within the bacterial community. Both the category and amount of the contaminants modulated catabolic plasmid transfer, and thus influenced the efficient removal of the contaminants.
For application to genetic bioaugmentation, we examined the conjugative transfer of catabolic plasmids in the activated sludge reactor, a model system in bio-treatment engineering. The abundant activated sludge bacterial species included an ample recipient pool that facilitated profiling the host range of the catabolic plasmids. In previous studies, pDF01 and pDF02 have shown a relatively broad range of hosts and were often transferred concomitantly (Ren et al. 2018;Sun et al. 2017). We isolated transconjugants belonging to nine genera from activated s l u d g e d u r i n g t r e a t m e n t o f d i b e n z o -p -d i o x i n , dibenzothiophene, and phenanthrene. Furthermore, conjugative transfer of pDF01 and pDF02 occurred between bacteria with distant phylogenetic relationships, as indicated by the significant differences in their genomic G+C content. There are biological obstacles to obtaining genes from donor bacteria with excessively different G+C contents, more specifically, which may lead to repression of plasmid gene or degrade the plasmid by restriction system in transconjugant cell (Alderliesten et al. 2020;Dimitriu et al. 2019;Popa and Dagan 2011). Our results demonstrated the possibility of horizontal transfer of catabolic genes crossing phylogenetic boundaries in the environment. Horizontal transfer of the accessary genes to distantly related microorganisms, such as antibiotic resistance genes, has received considerable attention, since beneficial to bacteria may yet afflict humans (Qiu et al. 2012). As inferred from the diverse catabolic pathways  respectively. Single asterisk denotes the group is statistically different from the control group (p < 0.05) Although selection pressure plays critical roles during genetic bioaugmentation (Digiovanni et al. 1996), it is difficult to distinguish the actual influence on plasmid lateral transfer or on the growth of transconjugants (in other words, vertical transmission of the plasmids). Studies have shown that the presence of selection pressure is not a prerequisite when conjugative transfer of plasmids occurs (Guo et al. 2015). To date, there is fragmentary information on selection pressure modulating the conjugation process. Limited knowledge of the mechanisms of selection pressure acting on the conjugation process follows from the arguments of selection pressure promoting the spread of antibiotic resistance genes (Pu et al. 2021). Selection pressures such as those imposed by antibiotics or heavy metals can cause global cellular responses that indirectly stimulate conjugation or particularly induce an SOS response, resulting in directly up-regulating the expression of conjugation machinery genes (Beaber et al. 2004;Lopatkin et al. 2016;Zhang et al. 2018). Unlike the controversial opinion that selection pressure influences the conjugation process, researchers are prone to regard the selection pressure as an advantage for the competitive growth of transconjugants under positive selection (Andersson and Hughes 2014; Top et al. 2002). It is reasonable to assume that the catabolic genes carried by the plasmids confer the plasmid-bearer the ability to utilize the contaminants for growth. Hence, the presence of selection pressure promotes the propagation of transconjugants.
Regarding the effect of selection pressure on plasmid transfer, the different methods used to detect transconjugants could influence the results ). In the present study, the proportion of transconjugants in the activated sludge bacteria was indicated by the ratio of transconjugant-bearing catabolic plasmid measured by qPCR analysis. When we calculated the ratio of transconjugants, the values according to the results of qPCR were higher than those measured by the spread plate method (Fig. SM-3, online resource). Actually, the results from distinct detection methods reflected different perspectives. In general, the host range of a plasmid is usually narrower than the transfer host range (plasmids can be transferred by conjugation) and wider than the stably maintained (long-term) host range (Shintani et al. 2014;Suzuki et al. 2010). In the present study, selection for the phenotypes of catabolic plasmid-bearing transconjugants proceeded differently under the individual contaminant stresses. On the one hand, the category of contaminants selected distinguishing genera of the plasmid hosts, although permissive recipients belonging to Arthrobacter and Klebsiella were present. On the other hand, the category and amounts of contaminants influenced the ratios of transconjugant-bearing catabolic plasmids in the community. In particular, under the stress of dibenzo-p-dioxin, relatively abundant hosts were observed, and the highest ratio of transconjugant-bearing catabolic plasmid was detected. Since the transconjugants were detected after the donor had been introduced into the activated sludge for at least 24 h, the ratio of transconjugant-bearing catabolic plasmid herein involved contributions from the transfer frequency of the catabolic plasmids and the proliferation of transconjugants. However, when we discuss contaminant selection, we should dissect the dynamics between the conjugation step and transconjugant propagation. As has been reported previously, chemicals can trigger an increase in intracellular ROS along with increases of membrane permeability that facilitate the process of conjugative plasmid transfer (Han et al. 2019;Lu et al. 2020;Wang et al. 2018;Zhang et al. 2019). In the present study, we have detected increases of the ROS levels in activated sludge bacteria during treatment of different contaminants, and the elevation of the membrane permeability was indicated by extracellular LDH level of the activated sludge bacteria under the stress of different contaminants. The rise in ROS level is linked to triggering the SOS response (Han et al. 2019), merging into the global regulation network in cells, which may upregulate genes on plasmids responsible for conjugation (Baharoglu et al. 2010;Banuelos-Vazquez et al. 2017;Butala et al. 2009;Fornelos et al. 2016;Jones and Holland 1985). Therefore, the contaminants as a selection pressure might promote the conjugation process. However, the levels of ROS and extracellular LDH did not indicate the action strength of the individual contaminant stress. Further studies are needed to elucidate the individual contaminant triggered-regulation in the donor and recipient cells during the conjugation process. In the present study, dibenzo-p-dioxin, dibenzothiophene, and phenanthrene were all degradable substrates following the pathways encoded by the catabolic plasmids (Kasuga et al. 2013;Peng et al. 2013). Thus, after acquisition of the plasmids, the transconjugants in the activated sludge bacteria gained the ability to utilize the contaminants. Under the stresses of the different contaminants, the resulting high ratio of the transconjugant-bearing catabolic plasmid occurred with the relatively high removal rate of dibenzo-p-dioxin; this could be attributed to the propagation of the transconjugants by utilizing the contaminants as carbon and energy sources, although this does not exclude the possible distinction of conjugation efficiency under the different stresses. Similarly, the stronger the dibenzo-p-dioxin concentration in the reactor, the greater the amounts of carbon and energy sources available to support the growth of transconjugants, thereby leading to higher proportions of the transconjugants. Herein, the contaminants as selection pressures could provide nutrients for the activated sludge bacteria that carried the catabolic plasmids, thereby conferring benefits to the host sufficient to offset the fitness cost. This study has characterized the recalcitrant contaminants as a selection pressure modulating catabolic plasmid spread in a model system. When multiple carbon sources are available in a contaminated environment, the influence of trace contaminants on catabolic plasmid transfer remains an open question.

Supplementary Information
The online version contains supplementary material available at https://doi.org/10.1007/s11356-021-15682-9. Author contribution Gang Zhao conducted the experiments, prepared the figures, and wrote main manuscript text. Yanan Wu, Xu Wang, and Meng Chen edited the manuscript and proofread the full manuscript. Li Li prepared the main framework of the overall experiments and rearranged the contents. All authors read and approved the final manuscript.
Funding This work was supported by the Natural Science Foundation of China (nos. 21876100 & 22076102).
Data availability The datasets used and/or analysed during the current study are available from the corresponding author on reasonable request.

Declarations
Ethics approval and consent to participate Not applicable.

Consent for publication Not applicable.
Competing interests The authors declare no competing interests.