Microbial Community Analysis Provides Insights into the Effects of Tetrahydrofuran on 1,4-Dioxane Biodegradation

Widespread in groundwater and carcinogenic to humans, 1,4-dioxane (dioxane) is attracting significant attention in recent years. Advanced oxidation processes can effectively remove dioxane but require high energy consumption and operation costs. Biological removal of dioxane is of particular interest due to the ability of some bacteria to mineralize dioxane at a low energy cost. Although dioxane is generally considered recalcitrant to biodegradation, more than 20 types of bacteria can degrade dioxane as the sole electron donor substrate or the secondary electron donor substrate. In the latter case, tetrahydrofuran (THF) is commonly studied as the primary electron donor substrate. Previous work has shown that THF promotes dioxane degradation at a low THF concentration but inhibits dioxane degradation at a high THF concentration. Our work expanded on the previous work by mechanically examining the effects of THF on dioxane degradation in a microbial community context.

ABSTRACT

Tetrahydrofuran (THF) is known to induce the biodegradation of 1,4-dioxane (dioxane), an emerging contaminant, but the mechanisms by which THF affects dioxane biodegradation in microbial communities are not well understood. To fill this knowledge gap, changes in the microbial community structure in microcosm experiments with synthetic medium and landfill leachate were examined over time using 16S rRNA gene amplicon sequencing and functional gene quantitative PCR assays. The overarching hypothesis being tested was that THF promoted dioxane biodegradation by increasing the abundance of dioxane-degrading bacteria in the consortium. The data revealed that in experiments with synthetic medium, the addition of THF significantly increased the abundance of Pseudonocardia, a genus with several representatives that can grow on both dioxane and THF, and of Rhodococcus ruber, a species that can use THF as the primary growth substrate while cometabolizing dioxane. However, in similar experiments with landfill leachate, only R. ruber was significantly enriched. When the THF concentration was higher than the dioxane concentration, THF competitively inhibited dioxane degradation since dioxane degradation was negligible, while the dioxane-degrading bacteria and the corresponding THF/dioxane monooxygenase gene copies increased by a few orders of magnitude.

IMPORTANCE Widespread in groundwater and carcinogenic to humans, 1,4-dioxane (dioxane) is attracting significant attention in recent years. Advanced oxidation processes can effectively remove dioxane but require high energy consumption and operation costs. Biological removal of dioxane is of particular interest due to the ability of some bacteria to mineralize dioxane at a low energy cost. Although dioxane is generally considered recalcitrant to biodegradation, more than 20 types of bacteria can degrade dioxane as the sole electron donor substrate or the secondary electron donor substrate. In the latter case, tetrahydrofuran (THF) is commonly studied as the primary electron donor substrate. Previous work has shown that THF promotes dioxane degradation at a low THF concentration but inhibits dioxane degradation at a high THF concentration. Our work expanded on the previous work by mechanically examining the effects of THF on dioxane degradation in a microbial community context.

INTRODUCTION

The organic compound 1,4-dioxane (dioxane) has been used widely as a chlorinated solvent stabilizer throughout the United States (1). As early as 1999, dioxane was classified as a group B2 carcinogen (a probable human carcinogen) (2). A recent animal study showed that oral exposure to dioxane led to liver tumors in rodents (3). According to a survey completed in 2015, this emerging environmental contaminant is detected in 21% of the U.S. public water systems with concentrations exceeding the 0.35-µg/liter reference level (corresponding to a 10⁻⁶ cancer risk) at 6.9% of these surveyed sites (4). Although no federal maximum contaminant level for dioxane has been set, more than 30 states in the United States have established guidelines and standards for drinking water and groundwater (4–6).

Conventional treatment technologies for volatile organic compounds such as activated carbon adsorption and air stripping are not effective for dioxane (7, 8). Advanced oxidation processes (AOPs) are proven treatment technologies for dioxane, but their use is limited by the high energy cost (9, 10). Biological treatment is a promising method given the capability of some microorganisms to metabolize dioxane or cometabolize dioxane by using tetrahydrofuran (THF), toluene, methane, or propane as the primary growth substrate (11–16). To date, at least 11 bacterial strains capable of metabolizing dioxane and 13 strains capable of cometabolizing dioxane have been identified (17). However, despite that, using microbial consortia is more practical than pure cultures for two reasons. First, almost all biological reactors utilized for water treatment are based on microbial consortia since keeping pure cultures from contamination is difficult in practice. Second, some bacteria, such as Pseudonocardia sp. strain ENV478, cannot mineralize dioxane and produce daughter products such as 2-hydroxyethoxyacetic acid (16, 18). The intermediates can be further mineralized by other bacteria in the consortia.

Dioxane-metabolizing bacteria yet isolated have large half-maximum-rate constant for dioxane (14–16, 18). Therefore, biological reactors are not able to remove dioxane to below the water quality guidelines and standards under common operating conditions. Adding a primary substrate to increase the number of bacteria may address this problem. Among the primary growth substrates, THF is attracting much research interest because (i) dioxane and THF are analogous in structure (8), (ii) almost all reported dioxane-degrading bacteria can also degrade THF (17), and (iii) the degradation of dioxane and THF by these bacteria is initiated by the same THF/dioxane monooxygenase enzyme (19). Among the known dioxane-degrading bacterial strains, some, such as Pseudonocardia dioxanivorans CB1190, Mycobacterium sp. strain PH-06, and Pseudonocardia sp. strain N23, can use both dioxane and THF as the sole carbon and energy source (13, 15, 20–22), and others, such as Rhodococcus ruber T1 and Rhodococcus ruber T5, can only cometabolize dioxane when using THF as the primary growth substrate (17, 23). As a result, THF is well known for promoting dioxane biodegradation. On the other hand, THF is inherently biodegradable and not as environmentally persistent as dioxane (24). In addition, THF has low toxicity and bioaccumulation potential (24). There is no evidence of THF being carcinogenic to humans so far. Therefore, THF might be used in some applications, such as the treatment of industrial wastewater and landfill leachate and the enrichment of dioxane-degrading bacteria. A few authors have studied the effects of THF on dioxane degradation in pure cultures (23, 25, 26) and mixed cultures (27) and concluded that high THF concentrations inhibited dioxane degradation. However, these effects have not been investigated from the perspective of microbial community, which could provide insight into the microbial interactions during dioxane degradation.

The overall objective of the present study was to fill in this knowledge gap by running a set of microcosm studies in which dioxane and THF spiking tests were performed in glass bottles containing a microbial consortium and synthetic mineral medium or an environmental sample (i.e., a landfill leachate). The changes in microbial community structure were tracked, and key functional genes were quantified in response to dioxane and THF throughout the microcosm studies.

RESULTS AND DISCUSSION

Effects of THF on dioxane degradation: three observations.

Table 1 summarizes the data for four sets of microcosm bottles (SD, SDT, LD, and LDT) used in the first spiking test (as well as the second and third spiking tests). Each set consisted of two duplicate bottles (identified as “A” and “B” in each bottle set, e.g., LDT_A and LDT_B). Leachate from a landfill was used as an example of environmental samples (7). The comparison between the SD, SDT, LD, and LDT bottles shows the effects of environmental settings on dioxane degradation, while the comparison between the SD and SDT bottles and between the LD and LDT bottles shows the effects of THF on dioxane degradation. The change in dioxane and THF concentrations in the four sets of bottles is shown in Fig. 1. The four panels in Fig. 1 correspond to the four sets of highly consistent duplicate bottles. The first observation in Fig. 1 is that the addition of THF shortened the time needed to completely degrade dioxane. With THF addition, dioxane was completely degraded in 24 days in the SDT bottles, which was 30% faster than that in the SD bottles. The difference between LD and LDT bottles was more significant, with dioxane degraded 65% faster in the LDT bottles than in the LD bottles. THF promoted dioxane removal because it increased dioxane-degrading bacteria, which is discussed in detail below. The second observation is that THF inhibited dioxane degradation when THF was >50 mg/liter. This observation was supported by the second and third spiking tests (see Fig. S1 in the supplemental material). Further microbial community analysis combined with quantitative PCR (qPCR) analysis, described in the following section, showed that the inhibition was competitive inhibition. Within the first 20 days, ∼73% dioxane was removed in the SD bottles (the average of the two duplicates is reported here unless noted otherwise), but the dioxane removal in the LD bottles was negligible. Thus, we made our third observation, i.e., that leachate significantly slowed down the dioxane degradation. This could be explained by the composition of the leachate. The change of chemical oxidation demand (COD; excluding THF and dioxane) in the leachate bottles was insignificant in the first 20 days (Fig. 2c and d), suggesting that the majority of the compounds in the leachate were not readily biodegradable. About 20 common volatile organic compounds in the leachate were measured (see Table S2 in the supplemental material). Some of them, such as cis-1,2-dichloroethene, were known to inhibit dioxane degradation (28). Interestingly, when THF was present, the negative effect of leachate on dioxane degradation disappeared, as shown by the comparison of Fig. 1b and d: almost all dioxane was removed in the SDT and LDT bottles within the first 20 days. This observation was evaluated in more detail by characterizing the changes in the microbial community composition over time.

TABLE 1.

Test matrix^a

Bottle set	Medium	First spike
SD	Synthetic medium	50 mg/liter dioxane
SDT	Synthetic medium	50 mg/liter dioxane and 150 mg/liter THF
LD	Landfill leachate	50 mg/liter dioxane
LDT	Landfill leachate	50 mg/liter dioxane and 150 mg/liter THF

^aFor all conditions, the second spike was 50 mg/liter dioxane and the third spike was 50 mg/liter dioxane plus 150 mg/liter THF.

FIG 1.

Dioxane and THF removal during the first spiking test. (a) SD; (b) SDT; (c) LD; (d) LDT. S, synthetic medium; L, landfill leachate; D, dioxane; T, THF. A and B represent duplicate bottles.

FIG 2.

sCOD (soluble chemical oxygen demand) change in the first spiking test in the “B” bottles. (a) SD_B; (b) SDT_B; (c) LD_B; (d) LDT_B. Dioxane and THF concentrations were converted to sCOD by multiplying 1.82 and 2.44 (obtained by calculating the theoretical oxygen demand), respectively. The results for the duplicate “A” bottles are shown in Fig. S2 and are similar to the results shown in this figure. The landfill was >30 years old; thus, most of the COD was composed of nonbiodegradable substances.

Microbial community analysis of the four sets of bottles: an overview.

Microbial communities in microcosm samples were compared using nonmetric multidimensional scaling ordination of relativized 16S rRNA sequence data (Fig. 3). This analysis revealed clear sample clusters depending on the type of medium (synthetic medium or leachate) and the addition of THF, suggesting that these factors contributed to disparate microbial communities (Fig. 3). Further, all eight bottles had similar microbial structures at day 0, suggesting a high similarity to the identical inoculum, but by day 10 samples clustered in the leachate group or the synthetic medium group as the SD and SDT bottles were grouped together and far away from the LD and LDT bottles. This trajectory in microbial community composition continued, and by day 20 each group was further separated into two clusters depending on whether THF was added. The ordination results are consistent with the dioxane, THF, and COD degradation data presented in Fig. 1 and 2. For example, the LD bottles had the least change of dioxane, THF, and COD in the first 20 days; correspondingly, the sample distance among the LD bottle samples is the smallest in Fig. 3.

FIG 3.

Nonmetric multidimensional scaling ordination of 16S rRNA sequence data. The suffix numbers 0, 10, and 20 indicate the sampling days.

The relative abundance of Actinobacteria increased by at least 2-fold in all bottles (Fig. 4). This trend was stronger in the SDT and LDT bottles (i.e., >5.8-fold), suggesting a correlation of dioxane degradation with Actinobacteria and a stronger correlation of THF degradation with this class. This observation is consistent with literature showing that many Actinobacteria can degrade dioxane and THF (13, 20, 29).

FIG 4.

Relative abundances of the 16S rRNA gene sequence data at the class level. The suffix numbers 0, 10, and 20 indicate the sampling days.

To determine the key groups of bacteria responsible for the dioxane and THF degradation, 2,985 operational taxonomic units (OTUs) in the samples were filtered using the following two criteria. First, in at least one sample, the OTU relative abundance was >1%. Second, in at least one set of bottles, the relative abundance of the OTUs increased by >2-fold within the first 20 days. This screening resulted in the 10 OTUs summarized in Fig. S3 in the supplemental material. The last four OTUs in Fig. S3, including Rheinheimera, Prosthecobacter debontii, “Exiguobacteraceae” (a taxon proposed by the Greengenes Database curators), and Saprospiraceae, were enriched only in the leachate bottles when the dioxane degradation in the LD bottles was negligible. This indicated that the enrichment of the four OTUs was probably due to the small fraction of biodegradable material present in the leachate (∼200 mg COD/liter, Fig. 2c). The other six OTUs (Fig. 5) were likely relevant to dioxane and THF degradation and are discussed in detail below.

FIG 5.

Bar graph showing the relative abundance of the six OTUs likely responsible for dioxane and/or THF degradation. The suffix numbers 0, 10, and 20 indicate the sampling days.

Since the change of volatile suspended solids (VSS) was less than 27% in all the bottles (see Table S3 in the supplemental material), the total biomass concentration did not significantly change in the spiking tests. Therefore, the relative abundance trend of each OTU is a good approximation of the mass concentration trend of microbes corresponding to the OTUs.

Further analysis of the six OTUs identified as potentially responsible for dioxane and THF degradation. (i) SD bottles: collaborative dioxane degradation by bacteria corresponding to three OTUs.

As described above, ∼73% of the dioxane was removed in the SD bottles during the first 20 days. Of the 73%, the majority (∼61%) was converted to organic intermediates, and only a small fraction (∼12%) was mineralized (see Fig. 2a). However, the dioxane mineralization rate increased to ∼69% at day 30, suggesting that the intermediates from dioxane degradation could be further mineralized. The additional spiking test results (see Fig. S1a in the supplemental material) from days 20 to 55 showed that dioxane was completely removed when it was the sole electron donor, suggesting that some bacteria in the consortium were able to metabolize dioxane (i.e., grow on dioxane).

Corresponding to the conversion of dioxane, three OTUs significantly increased: Pseudonocardia (OTU 85863), Methylocystaceae (OTU 593232), and Xanthomonadaceae (OTU 4477150) (Fig. 5). First, an unclassified Pseudonocardia increased from ∼0.1 to 1.5% (Fig. 5). Pseudonocardia is a genus in the Actinobacteria class and largely contributed to the observed increase in Actinobacteria from 1.2 to 2.6% (see Fig. 4). Several Pseudonocardia strains (15, 20, 22, 30) were reported to be capable of degrading dioxane via metabolic and cometabolic transformation. The increased relative abundance in SD bottles in the absence of THF (Fig. 1a) suggested that this species metabolized dioxane as a carbon and energy source. This species will be further discussed below regarding its ability to degrade THF and the effects of THF on its dioxane degradation.

Second, Methylocystaceae increased from 0 to 3.2% in SD bottles. Methylocystaceae (i.e., type II methanotrophs) is a family of bacteria that assimilate C₁ carbon compounds such as methane and methanol via the serine pathway (31). These bacteria were enriched, probably by using C₁ carbon from biomass decay and dioxane degradation. For example, formate can be formed in microbial metabolism of oxalate and glyoxylate, which are intermediates of dioxane biodegradation (32–34). Methylocystaceae may then grow on formate as the primary substrate (35).

Third, Xanthomonadaceae increased from 0.01 to 1.1%. Similar to the enrichment of Methylocystaceae, Xanthomonadaceae might grow on the carbon from biomass decay and dioxane degradation.

(ii) SDT bottles: enrichment of three additional microbial groups due to THF addition.

Within the first 20 days, ∼92% of the dioxane was removed in the SDT bottles (Fig. 1b). Figure 2b also shows that most THF was mineralized. Compared to the SD bottles, the addition of THF increased the dioxane removal rate. The relative abundance of the three OTUs that increased in the SD bottles (OTU 85863, OTU 593232, and OTU 4477150) also increased in the SDT bottles, but at different rates (Fig. 5). The increase of Pseudonocardia in the SDT bottles was more than two times higher than that in the SD bottles; this might be explained by reports showing that most dioxane-degrading strains in Pseudonocardia can grow on both dioxane and THF (15, 20, 22). On the other hand, the increase of Methylocystaceae and Xanthomonadaceae in the SDT bottles was less than that in the SD bottles, suggesting that Methylocystaceae and Xanthomonadaceae were suppressed. Methylocystaceae and Xanthomonadaceae might be outcompeted by dioxane- and THF-degrading bacteria (Pseudonocardia and Rhodococcus ruber) since all of them likely used and competed for common carbon sources, such as daughter products of dioxane and THF degradation.

More importantly, three additional microbial groups showed substantial increase: Rhodococcus ruber (OTU 4435984, which increased from nearly 0 to as high as 4.2%), Pandoraea (OTU 107349, a genus that increased from 0.001 to 1.1%), and Cupriavidus (OTU 3437276, a genus that increased from nearly 0 to 0.9%). Rhodococcus ruber strains (e.g., T1, T5, ENV425, and 219) are well known for their ability to cometabolize dioxane with THF as the primary substrate (14, 16, 23, 29), consistent with our findings of Rhodococcus ruber in this study. The Cupriavidus and Pandoraea genera are known versatile hydrocarbon degraders and were possibly involved in the metabolism of intermediates in the dioxane and THF biodegradation process (13, 36–39).

In summary, the addition of THF further increased the abundance of Pseudonocardia, which is capable of growth on both dioxane and THF. Rhodococcus ruber, which can cometabolize dioxane by using THF as the primary substrate, was also significantly enriched. These two species together contributed to the promoted dioxane degradation in the SDT bottles (compared to the SD bottles). Cupriavidus and Pandoraea, two possible intermediate utilizers, also proliferated. These plausible interpretations of changes in the microbial community structure explained the higher dioxane removal rate in the SDT bottles than in the SD bottles.

(iii) LDT bottles: enrichment of Rhodococcus ruber and competitive inhibition of dioxane degradation by THF.

In LDT bottles, dioxane and THF were completely degraded within the first 20 days. Of the six OTUs in Fig. 5, only Rhodococcus ruber significantly increased, from 0.001% at day 0 to 8.6% at day 20. As discussed previously, Rhodococcus ruber is well known for its ability to metabolize THF and cometabolize dioxane by using THF, which was ∼46% of the total COD, as the primary substrate. In contrast, none of the six OTUs significantly increased in the LD bottles, corresponding well with the insignificant dioxane degradation in these bottles (Fig. 1). Rhodococcus ruber is also well known to use a variety of organic compounds, such as ethyl tert-butyl ether, polyethylene, alkane, and polycyclic aromatic hydrocarbons (40–43), as growth substrates. The broad substrate specificity made Rhodococcus ruber outcompete other dioxane-degrading bacteria and become the solely enriched dioxane-degrading species in the LDT bottles, while both Rhodococcus ruber and Pseudonocardia were significantly enriched in the SDT bottles (Fig. 5). Despite the differences in the microbial communities of the SDT and LDT bottles, the dioxane degradation profiles for the two sets of bottles were similar. This could be explained by similar numbers of dioxane-degrading bacteria in the two sets of bottles: the sum of the relative abundance of Rhodococcus ruber and Pseudonocardia was 8.0% at day 20 in the SDT bottles, while the relative abundance of Rhodococcus ruber was 8.6% at day 20 in the LDT bottles (Fig. 5).

At the beginning of the second spiking test (i.e., first 10 days of the second spiking test), the dioxane degradation rate was lower in the LDT bottles (Fig. S1d) than in other bottles (Fig. S1a, S1b, and S1c). This can be explained by the microbial community data in Fig. 5. Figure 5 shows that in the bottles other than LDT bottles, Pseudonocardia spp. (dioxane-metabolizing bacteria) were already enriched at the end of the first spiking test and were ready to degrade dioxane at the beginning of the second spiking test. However, in the LDT bottles, Pseudonocardia spp. were negligible, whereas Rhodococcus ruber (dioxane-cometabolizing bacteria) was enriched at the end of the first spiking test. These Rhodococcus ruber strains might be similar to the dioxane-cometabolizing Rhodococcus ruber strains T1 and T5 in Sei et al. (23), which do not degrade dioxane without the presence of the primary substrate THF. After about 10 days in the second spiking test, Pseudonocardia spp. should have been enriched in the LDT bottles to degrade dioxane. In LDT bottles, the dioxane degradation rate was much higher in the third spiking test than in the first two spiking tests because of the availability of THF as the primary substrate in the third spiking test and the already accumulated Rhodococcus ruber (Fig. S1). The cometabolism also explains the abrupt dioxane removal between days 10 and 13 in the first spiking test in the LDT bottles.

To understand how THF inhibited dioxane degradation by Rhodococcus ruber at high concentrations (days 0 to 10) but promoted dioxane degradation at low concentrations (days 10 to 20) in the LDT bottles, the changes associated with THF/dioxane monooxygenase (DXMO) and aldehyde dehydrogenase (ALDH) in the LDT bottles and LD bottles (for comparison purposes) were measured. DXMO encodes an enzyme to oxidize dioxane to 2-hydroxyethoxyacetaldehyde, and ALDH encodes an enzyme to degrade the intermediates, including 2-hydroxyethoxyacetaldehyde, glyoxal, and glycolaldehyde (18, 34). The numbers of gene copies of DXMO and ALDH only slightly increased in the LD bottles (Fig. 6). In the LDT bottle, however, the fold increase in DXMO and ALDH gene copies is consistent with the fold increase in Rhodococcus ruber. Using bottle LDT_A as an example, the relative abundance of Rhodococcus ruber increased by 24-fold from days 10 to 20 (Fig. 5), while the DXMO gene copies increased by 31-fold and the ALDH gene copies increased by 22-fold (Fig. 6). Previous research has shown that the dioxane-transforming strains Rhodococcus ruber T1 and Rhodococcus ruber T5 possess thm/dxm genes, which may be the case in our study; yet definitively identifying these genes as originating from Rhodococcus ruber requires additional analyses, such as strain isolation and genomic sequencing.

FIG 6.

Gene copies per liter of ALDH and DXMO (log concentration) in the leachate medium bottles. The suffix numbers 0, 10, and 20 indicate the sampling days.

Most of the dioxane-degrading bacteria discussed here have been reported in the literature, but we studied them in a microbial community context in two different environmental settings (i.e., synthetic groundwater and a real-world landfill leachate). This led to two findings. First, while THF promoted dioxane degradation in both environmental settings through enriching dioxane-degrading bacteria, the enriched bacteria were different: dioxane-metabolizing bacteria (Pseudonocardia) and dioxane-cometabolizing bacteria (Rhodococcus ruber) in the synthetic groundwater, but only cometabolizing bacteria (Rhodococcus ruber) in the real-world leachate. This observation was repeatable in the duplicate experiments. Second, we provided three lines of evidence to support that THF competitively inhibited dioxane degradation when the THF concentration was high. This is discussed further in the following paragraph.

We conclude that THF promoted dioxane degradation at a low concentration by promoting the growth of dioxane-metabolizing bacteria and dioxane-cometabolizing bacteria but competitively inhibited dioxane degradation at a high concentration for the following three reasons. First, all three spiking tests in the four sets of bottles yielded the same observation that THF inhibited dioxane degradation when the THF concentration was >50 mg/liter but promoted dioxane degradation when the THF concentration was <50 mg/liter (see Fig. S1 in the supplemental material). This observation is typical for competitive inhibition, in which the degradation of multiple compounds is catalyzed by the same enzyme (44). If THF did not inhibit dioxane degradation, dioxane removal would have started at the beginning of the third spiking test in all of the bottles (see Fig. S1). Second, our conclusion is consistent with previous studies on the effect of THF on dioxane with mixed cultures and pure strains such as Rhodococcus ruber T1 and T5 (23, 25, 27). Third, a well-known dioxane-cometabolizing species (Rhodococcus ruber) and two well-known genes involved in dioxane and THF degradation (DXMO and ALDH) (37) increased by a few orders of magnitude in the first 10 days of the first spiking test; they removed >70% of the THF but removed negligible amounts of dioxane. It is worth mentioning that a propane monooxygenase from Mycobacterium dioxanotrophicus PH-06 was recently reported to be capable of degrading dioxane (45). Since we did not find enrichment of Mycobacterium in our experiments, this enzyme was probably not involved in the dioxane degradation in our experiments.

MATERIALS AND METHODS

Overview of the microcosm experiments.

In this study, four sets of glass bottles (SD, SDT, LD, and LDT) were used, and each set consisted of two duplicate 500-ml bottles with 5 ml of activated sludge from a full-scale bioreactor for landfill leachate treatment and 295 ml of sterile synthetic medium or landfill leachate. To minimize contamination and maintain oxygen supply, all bottles were sealed with sterile cotton at the bottleneck. Bottles were placed on a shaker (model SHKE 2000; Thermo Scientific, Dubuque, IA) at 120 rpm. Three spiking tests were carried out, with the first spiking test matrix summarized in Table 1 and the complete test matrix summarized in Table 1. In addition, two sterile control bottles were also prepared: one contained synthetic medium, and the other contained landfill leachate. Other bottle setup details, including the synthetic medium composition and the landfill characterization, are described in sections S3 and S4 in the supplemental material.

Sampling and measurement.

During the experiment, each bottle was sampled about every 2 days using sterile syringes. The samples were measured for dioxane and THF using solid-phase microextraction coupled with a Hewlett-Packard gas chromatography/mass spectrometry system (GC/MS; GC 5890 series II; MSD 5971A) using the selected ion monitoring mode, a method modified from the APHA method 6040E (46). Briefly, 1 ml of standard or sample was transferred to a 20-ml vial that contained 9 ml of deionized water and dioxane-d8 as an internal standard. The vial was then sealed and magnetically stirred at 400 rpm and 65°C for 20 min, with a 100-µm polydimethylsiloxane fiber (Supelco; Bellefonte) immerged to extract THF and dioxane from the samples and standards. After extraction, the fiber was thermally desorbed for 5 min using a GC/MS injector (225°C). The initial oven temperature (40°C) was maintained for 2 min, then increased (15°C/min) to 225°C, and then held for 3 min. The detector temperature was 280°C. The limit of detection is 0.01 mg/liter.

In addition, mixed liquor in the bottles was sampled every 10 days (on days 0, 10, and 20 for the first spiking test) until the dioxane and THF were completely removed in most bottles for analysis of the volatile suspended solids (VSS), the COD, and the microbial community and for qPCR analyses. The VSS were measured according to APHA method 2540E (46), and the COD was measured using the USEPA-approved Hach Laboratory Method 8000 (47).

To determine the microbial community structure in the aforementioned samples, iTag sequencing (Illumina MiSeq) of 16S rRNA genes was carried out. In total, the microbial communities in 24 samples were analyzed. 16S rRNA genes were amplified from 10 ng of purified genomic DNA in duplicate using the modified archaeal and bacterial primers 515FB and 806RB (48, 49) in accordance with the protocol described by Caporaso et al. (50, 51) and used by the Earth Microbiome Project (52), with a slight modification: the annealing temperature was modified to 60°C. Amplicons were sequenced using an Illumina MiSeq in 250 × 250-bp mode. A total of 709,558 sequences resulted from the MiSeq runs. Raw sequences were joined, demultiplexed, and then quality filtered using the default parameters in QIIME version 1.9.0 (53).

Sequences were then clustered into operational taxonomic units (OTUs), which were defined as ≥97% 16S rRNA gene sequence similarity using SUMACLUST (54) and SortMeRNA (55) using Greengenes v13.5 (56) for taxonomy. The open and closed reference tables resulting from more than one sequencing run were converted to closed references only to generate one OTU table. The resulting OTU table was filtered to keep only OTUs that had 10 sequences or more (resulting in 2,985 OTUs) and normalized using cumulative sum scaling (57).

The functional genes for aldehyde dehydrogenase (ALDH) and dioxane monooxygenase (DXMO) were quantified in duplicate using qPCR assays. Specifically, the ALDH primers sad/aldehyde dehydrogenase (ALDH-For and ALDH-Rev), the DXMO primers dxmB/dioxane monooxygenase (DXMO-For and DXMO-Rev), and qPCR thermal cycler protocol described in Gedalanga et al. (37) were used. These functional genes were first amplified from Pseudonocardia dioxanivorans CB1190 DNA extractions (37). Functional gene amplicons were then cloned and sequenced to verify that the correct DNA fragments were amplified during PCR. These amplicons were linearized, purified, and quantified by fluorometry and used for the standard curve generation in subsequent qPCR assays.

Data availability.

Sequences are available in the NCBI Sequence Read Archive (BioProject no. PRJNA530571) and at http://mason.eoas.fsu.edu.