Biotransformation of N-Nitrosodimethylamine by Pseudomonas mendocina KR1

Abstract

N-Nitrosodimethylamine (NDMA) is a potent carcinogen and an emerging contaminant in groundwater and drinking water. The metabolism of NDMA in mammalian cells has been widely studied, but little information is available concerning the microbial transformation of this compound. The objective of this study was to elucidate the pathway(s) of NDMA biotransformation by Pseudomonas mendocina KR1, a strain that possesses toluene-4-monooxygenase (T4MO). P. mendocina KR1 was observed to initially oxidize NDMA to N-nitrodimethylamine (NTDMA), a novel metabolite. The use of ¹⁸O₂ and H₂¹⁸O revealed that the oxygen added to NDMA to produce NTDMA was derived from atmospheric O₂. Experiments performed with a pseudomonad expressing cloned T4MO confirmed that T4MO catalyzes this initial reaction. The NTDMA produced by P. mendocina KR1 did not accumulate, but rather it was metabolized further to produce N-nitromethylamine (88 to 94% recovery) and a trace amount of formaldehyde (HCHO). Small quantities of methanol (CH₃OH) were also detected when the strain was incubated with NDMA but not during incubation with either NTDMA or HCHO. The formation of methanol is hypothesized to occur via a second, minor pathway mediated by an initial α-hydroxylation of the nitrosamine. Strain KR1 did not grow on NDMA or mineralize significant quantities of the compound to carbon dioxide, suggesting that the degradation process is cometabolic.

N-Nitrosodimethylamine (NDMA) is present in groundwater and drinking water primarily as a by-product of wastewater and drinking water disinfection and from past military testing and disposal of 1,1-dimethylhydrazine, a component of liquid rocket propellant that contained NDMA as an impurity (9, 20, 21). NDMA is a potent mutagen and a suspected human carcinogen, so its presence in drinking water is of significant concern (1, 22, 34, 36). Although there is no federal drinking water standard for NDMA, the United States Environmental Protection Agency has estimated that concentrations exceeding 0.7 ng/liter may significantly increase cancer risk (34), and the California Office of Environmental Health Hazard Assessment recently set a draft public health goal of 3 ng/liter (24).

The metabolism of NDMA in animals and its toxicological effects have been widely studied (1, 24, 36). The mammalian metabolism of NDMA is initiated by the cytochrome P450-dependent mixed-function oxidase system and follows either a demethylation (α-hydroxylation) or a denitrosation pathway depending on the site of oxidative attack (1, 32, 36) (Fig. 1). The demethylation route (Fig. 1A) results in the formation of the methyldiazonium ion, a strong alkylating agent which is thought to account for much of the carcinogenic activity of NDMA. This ion either binds to macromolecules or spontaneously disassociates to methanol and nitrogen gas (29, 36). The denitrosation pathway (Fig. 1B) results in the formation of nitrite, methylamine, and formaldehyde as metabolites (36).

FIG. 1.

Demethylation and denitrosation pathways of NDMA metabolism in mammalian cells (modified from references 7 and 36).

In contrast to the extensive data from animal studies, there is relatively little information concerning the microbial degradation of NDMA. Previous studies have revealed that the compound can be biodegraded under both aerobic and anaerobic conditions and in numerous environments, including surface and vadose soils, groundwater, and lake water, but the organisms responsible for NDMA metabolism in nature are largely speculative (3, 9, 12, 15, 25, 27, 31, 37). In addition, various pure cultures expressing broad-specificity monooxygenase enzymes have now been observed to degrade NDMA through cometabolism, including Methylosinus trichosporium OB3b (soluble methane monooxygenase) (28, 38), Pseudomonas mendocina KR1 (toluene-4-monooxygenase [T4MO]) (28, 30), Rhodococcus ruber ENV425 (propane monooxygenase) (30), and Mycobacterium vaccae JOB5 (propane monooxygenase) (28). Although NDMA degradation in natural environments and by pure cultures has been reported, few data exist concerning the microbial pathways of NDMA metabolism.

The objective of this work was to elucidate the pathway of NDMA biotransformation by P. mendocina KR1. This strain possesses T4MO, a four-component diiron monooxygenase enzyme that transforms toluene to para-cresol (26, 35). This enzyme has broad substrate specificity and has also been observed to oxidize a number of other compounds of environmental concern, including a variety of halogenated alkanes and alkenes (14, 16-18). P. mendocina KR1 was also recently reported to be capable of metabolizing NDMA (28, 30), but the degradation pathway is unknown. The data presented herein suggest that the bacterium oxidizes the nitrosamine primarily to N-nitrodimethylamine (NTDMA), which is then metabolized further to produce N-nitromethylamine (NTMA) and formaldehyde. This microbial pathway appears to differ significantly from both the demethylation and denitrosation pathways mediated by cytochrome P450 enzymes in eukaryotes (Fig. 1) and which have been previously proposed as likely routes of microbial metabolism (12, 34, 38, 39).

MATERIALS AND METHODS

Chemicals.

NTDMA and NTMA were obtained from Ron Spanggord (SRI International, Menlo Park, CA). [¹⁴C]NDMA (specific activity of 55 mCi/mmol) was purchased from American Radiolabeled Chemical (St. Louis, MO), and ¹⁸O₂ was from Icon Isotopes (Summit, NJ). All other chemicals used in this study, including D₂O and H₂¹⁸O, were from Sigma-Aldrich and were reagent grade or higher purity.

Analytical methods.

NDMA, NTDMA, and NTMA were analyzed using the high-pressure liquid chromatography method previously developed for the quantification of nitro-2,4-diazabutanal, a key metabolite of hexahydro-1,3,5-trinitro-1,3,5-triazine (RDX) biotransformation under aerobic conditions (5, 6). Analyses of nitrate (NO₃⁻), nitrite (NO₂⁻), formate (HCOO⁻), formaldehyde (HCHO), and methanol (CH₃OH) were performed as previously reported (2, 11). Methylamine and dimethylamine were analyzed by the method detailed in the work of Gui et al. (8). The capture and quantification of ¹⁴CO₂ were performed as described previously (4).

An alternative gas chromatography (GC) system (Agilent 6890; Agilent Technologies, Palo Alto, CA) coupled to a 5973 quadrupole mass spectrometer (electron impact) was used in the experiments with ¹⁸O₂, D₂O, and H₂¹⁸O. One microliter of methylene chloride extract was injected in the splitless mode on a 50-m by 0.2-mm by 0.33-μm HP-5MS capillary column (Agilent Technologies, Palo Alto, CA). The column was heated at 40°C for 2 min and then warmed to 200°C at a rate of 15°C/min. Helium was used as carrier gas at an average velocity of 28 cm/s. The injector temperature was set at 150°C, and the detector interface was maintained at 200°C. Data were collected in the scan mode between 35 and 200 atomic mass units.

Growth and assay conditions for metabolic studies.

P. mendocina KR1 was initially cultured in a nephelo culture flask (Bellco Biotechnology) with basal salts medium (BSM) (28). Neat toluene (100 μl per 100-ml culture) was placed in the sidearm tube of the flask. The toluene subsequently volatilizes and dissolves in the culture medium with time to provide an organic growth substrate to the strain. The cells were incubated at 30°C and agitated on a rotary shaker at 175 rpm. After growth to either early or mid-log phase (optical density at 600 nm [OD₆₀₀] of 0.4 and 0.9, respectively), the cells were harvested and washed twice in BSM to remove residual toluene. The cells were then added to 6-ml vials containing BSM with NDMA (325 μM or as specified) in the absence of toluene. The final absorbance (OD₆₀₀) of the culture in the reaction vials was 2.0. Two sets of control reaction mixtures were prepared, the first containing NDMA without bacterial cells and the second containing bacteria without NDMA. The vials were covered in aluminum foil to prevent photolysis of NDMA and incubated at 30°C (175 rpm). The experiments in which ¹⁸O₂ was added were performed similarly, except that an 0.5-ml bacterial suspension in BSM was added to 2-ml vials. Before the addition of NDMA to the cell suspension, ¹⁸O₂ (0.5 ml) was injected into the vial using a gastight syringe.

NDMA degradation by cloned T4MO.

To confirm that toluene-4-monooxygenase was the enzyme responsible for the NDMA degradation by P. mendocina KR1, NDMA assays were conducted with Pseudomonas putida PPO200:AF, a microbial strain expressing cloned T4MO. The T4MO gene was cloned into this organism using the pNM185 expression vector (19). The parent strain (P. putida PPO200), which was cured of the TOL plasmid and is unable to oxidize toluene, was tested as a negative control (13). The parent strain and clone were grown in 250-ml shake flasks containing 100 ml of mineral salts solution (10) plus 0.4% (wt/vol) glutamic acid sodium salt. The strains were grown to late logarithmic phase, washed in mineral salts solution, and then added to duplicate 160-ml serum bottles containing mineral salts solution with ∼225 μM of NDMA. Cell-free control samples were also prepared. Subsamples were collected after 1, 4, and 7 days; extracted as described below; and analyzed for NDMA by Environmental Protection Agency method 8015B.

To extract NDMA, aqueous subamples (0.5 ml) were removed from each serum bottle and placed in clean 2-ml gas chromatography autosampler vials (GC vial) containing 0.5 ml of a solvent mixture (80% methylene chloride-20% acetone). The vials were placed horizontally on an orbital shaker (Lab-Line, Melrose Park, IL) and mixed at ∼300 rpm in the dark for 4 h to promote partitioning of NDMA to the solvent phase. A 200-μl volume of the solvent was then removed from each vial with a glass syringe and placed into a clean GC vial for analysis.

RESULTS

Transformation of NDMA by toluene-grown P. mendocina KR1.

Resting cells of P. mendocina KR1 that were grown on toluene as a sole carbon source and harvested in early log phase (OD₆₀₀ = 0.4) completely transformed NDMA (323 μM) within 27 h. In the control bottles containing BSM but no cells, the concentration of NDMA remained unchanged throughout the incubation period (data not shown). The disappearance of NDMA was concomitant with the rapid and transient production of NTDMA (Fig. 2). The presence of NTDMA was confirmed by GC-mass spectrometry (MS) (m/z 90 Da) and by comparison with a reference standard (Fig. 3A). When the headspace of a vial containing active cells was amended with ¹⁸O₂, the NTDMA peak showed molecular masses at 90 and 92 Da, corresponding to a mixture of (CH₃)₂NN¹⁶O¹⁶O and (CH₃)₂NN¹⁶O¹⁸O, respectively (Fig. 3B). This observation shows that a single oxygen atom originating from air was added to NDMA. The incubation of NDMA with cells and either D₂O or H₂¹⁸O did not lead to any changes in the mass of NTDMA [i.e., all was present as (CH₃)₂NN¹⁶O¹⁶O], confirming that the oxygen in NTDMA was not derived from water.

FIG. 2.

Biotransformation of NDMA by P. mendocina KR1 after growth on toluene and harvest at early log phase (OD₆₀₀ of 0.4). Symbols: ○, NDMA; □, N-nitrodimethylamine; ▵, N-nitromethylamine; +, formaldehyde; ×, methanol. The symbols indicate averages of duplicate experiments, and the error bars indicate 1 standard deviation.

FIG. 3.

GC-MS spectrum of the metabolite NTDMA produced during NDMA transformation using resting KR1 cells in the absence (A) or presence (B) of ¹⁸O₂.

The disappearance of NTDMA was accompanied by the formation of NTMA, formaldehyde, and methanol (Fig. 2). At the end of the 27-h incubation, 284 and 51 μM of NTMA and methanol were accumulated, respectively (Table 1). This represents a nearly stoichiometric accumulation of NTMA (88% of the 323 μM expected). However, the quantity of methanol observed at the end of the study reflects only 16% of that expected based on the total transformation of NDMA to NTMA and methanol. Other potential metabolites, including methylamine and dimethylamine (detection limit of 0.5 mg/liter), nitrite and nitrate (detection limit of 1 mg/liter), nitrous oxide (N₂O; detection limit of 10 μg/liter), and nitric oxide (detection limit of 30 μg/liter), were not detected.

TABLE 1.

Stoichiometry of nitrogen and carbon during NDMA transformation by P. mendocina KR1^a

Substrate (μM)	Growth stage	Transformation rate (μM h−1)	Metabolite (μM)					Stoichiometry (% recovery)
			NTDMA	NTMA	HCHO	HCOO−	CH3OH	N	C
NDMA (323)	Early log (OD600 = 0.4)	39	Transient (max of 72 μM)	284	Transient (max of 3.8 μM)	NDc	51	88	60
NDMA (278)	Mid-log (A600 = 0.9)	132	ND	262	Transient (tr)d	Transient (max of 20 μM)	22	94	55
NTDMA (211)	Mid-log (A600 = 0.9)	193	NAb	210	Transient (tr)	Transient (max of 26 μM)	ND	99	50

^aThe data represent the means from duplicate or triplicate samples, and standard deviations were <5% of the mean.

^bNA, not applicable.

^cND, not detected.

^dtr, trace.

The production of N-nitromethylamine and formaldehyde from NDMA is consistent with hypothetical monooxygenase reactions. The possibility that methanol was generated by enzymatic reduction of formaldehyde (which was observed in small amounts as a transient intermediate) was tested by incubating the strain in BSM amended with 45 μM formaldehyde and collecting samples with time. Methanol was not detected during this study (data not shown). In addition, a new lot of NDMA was purchased from Sigma-Aldrich (≥99.9% pure) to ensure that a putative impurity in the original NDMA stock was not the source of methanol. In experiments with the new stock, P. mendocina KR1 rapidly transformed NDMA into NTMA, yielding a molar ratio of NTMA/NDMA of 91% after approximately 6.5 h. Similar amounts of methanol were detected as in the previous experiment, with a molar ratio (methanol/NDMA) of ∼10%. This suggests that methanol was not present as or derived from an impurity in the NDMA stock.

In a separate study in which P. mendocina KR1 was harvested later in the growth cycle (i.e., mid-log phase rather than early log phase), the NDMA degradation rate was observed to be approximately threefold higher than that for the early-log-phase cells even though the starting culture density was the same (Table 1). In this experiment, 278 μM of NDMA was completely transformed within 3.5 h, whereas in the previous experiment, >30% (∼110 μM) of the added NDMA remained after 6 h of incubation. Interestingly, NTDMA was not observed as an intermediate during this experiment, presumably because it was degraded too rapidly. A total of 262 μM of NTMA was detected at the conclusion of the study, which accounted for ∼94% of that expected based on the initial quantity of NDMA added.

In order to gain more insight into the mechanism of transformation of NDMA by P. mendocina KR1, mid-log-phase cells were harvested and incubated with 211 μM of NTDMA. The NTDMA was stoichiometrically transformed into NTMA (99% conversion) within 4 h (Fig. 4). Approximately 26 μM of formate and a trace amount of formaldehyde were detected, but neither persisted beyond 6 h (Fig. 4; Table 1). Most importantly, methanol was not detected as a product from NTDMA metabolism. This confirms that the alcohol is not formed from NTDMA or its subsequent products but rather is derived from NDMA, probably through a second, minor metabolic path.

FIG. 4.

Transformation of NTDMA by P. mendocina KR1 after growth on toluene. Cells were harvested at mid-log phase (OD₆₀₀ of 0.9). Symbols: ○, NTDMA; ▵, N-nitromethylamine; □, formate. The symbols indicate averages of duplicate experiments, and the error bars indicate 1 standard deviation.

In order to evaluate whether the transient formaldehyde or formate was utilized by the strain as a carbon source (i.e., incorporated into the biomass) or mineralized (liberated as CO₂), a set of experiments was performed with [¹⁴C]NDMA (270 μM solution containing 500,000 dpm). After 5 days of incubation, the radiolabeled carbon was distributed as follows: 1.7% was liberated as CO₂, 0.2% was found within the cells, and 92.7% was in the supernatant. These data indicate that the strain is not utilizing any of the metabolites generated during NDMA transformation as an energy source. A longer incubation time (10 days) did not lead to the degradation of methanol or N-nitromethylamine (data not shown). Slightly higher levels of mineralization were observed when the culture was fed toluene during the incubation period, but extents of mineralization became stationary below 10% after about 27 days with a slight accumulation in the cells (1.2%).

NDMA degradation by cloned T4MO.

The T4MO clone P. putida PPO200:AF degraded NDMA to below detection by GC-flame ionization detection analysis (∼14 μM) within 24 h (data not shown). Conversely, the parent strain (P. putida PPO200) did not metabolize NDMA. The breakdown product NTDMA was observed during NDMA metabolism by the T4MO clone.

DISCUSSION

The data presented support previous reports (28, 30) that P. mendocina KR1 is capable of degrading NDMA and provide the first confirmed pathway for this reaction. Based on metabolite analysis and data from ¹⁸O₂-enriched samples, P. mendocina KR1 adds an oxygen atom (O) from O₂ to convert the nitroso functional group in NDMA into the nitro functional group of NTDMA (Fig. 5). The addition of a single oxygen to NDMA is consistent with other reactions carried out by T4MO, which is known to attack alkanes, alkenes, and aromatics (16-18, 23, 35). The photolytic formation of NTDMA from NDMA has been previously reported (33). However, the addition of an oxygen atom to the nitroso group of NDMA by T4MO, instead of one of the methyl groups, is novel and somewhat unexpected. This reaction differs significantly from both the denitrosation and demethylation pathways of NDMA metabolism catalyzed by P450 enzymes of eukaryotes and proposed for various bacteria (1, 12, 32, 38).

FIG. 5.

Proposed pathway of NDMA metabolism by P. mendocina KR1.

It is likely that subsequent α-hydroxylation of one of the methyl groups of NTDMA leads to the formation of N-nitrohydroxymethylmethylamine, CH₃CH₂OHNNO₂, which then decomposes to N-nitromethylamine and formaldehyde (Fig. 5). Both NDMA and NTDMA were converted into N-nitromethylamine and formaldehyde. Formate was also detected during the oxidation of both compounds, likely from the further oxidation of formaldehyde (both compounds were transient). The oxidation rate of NTDMA (193 μM h⁻¹) was higher than the rate calculated for NDMA (132 μM h⁻¹). A similar reaction has been proposed for the oxidation of NTDMA by cytochrome P450 2E1 in rat liver microsomes, generating N-nitromethylamine and formaldehyde (7). Only trace levels of formaldehyde were detected after NTDMA oxidation by P. mendocina KR1. However, formaldehyde is very reactive and may have either polymerized or otherwise reacted with cells or soluble extracellular compounds.

The methanol detected during NDMA degradation by KR1 may be generated via the demethylation (α-hydroxylation) mechanism shown in Fig. 1. Experiments revealed that the alcohol is not generated from formaldehyde or NTDMA, nor is it present as an impurity. For example, when NTDMA, the first metabolite of NDMA, was incubated with active cells of KR1, stoichiometric quantities of N-nitromethylamine were produced and methanol was not detected (Table 1). In contrast, strain KR1 produced between 22 and 51 μM of methanol (∼8 to 16% of that expected if all NDMA was metabolized to methylamine and methanol), and the methanol concentration was higher (with a concomitant decline in N-nitromethylamine) when the degradation rate of the cells was lower (Table 1). Taken as a whole, the data suggest that P. mendocina KR1 degrades NDMA by two pathways: a major pathway mediated by T4MO, which leads to N-nitromethylamine via the initial oxidation of NDMA to NTDMA, and a second, minor pathway which produces methanol as previously suggested for the degradation of NDMA via demethylation as shown in Fig. 1. Many of the other intermediates generated via the second route are unstable and not readily detected. This pathway represents less than 10% of the NDMA degraded by strain KR1.

The rapid degradation of NDMA by P. putida PPO200:AF, a clone expressing T4MO, but not by the parent strain, which was cured of the TOL plasmid, shows that T4MO is responsible for the initial oxidation of NDMA. Moreover, NTDMA, the initial product detected during oxidation of NDMA by strain KR1, was also observed during NDMA oxidation by the T4MO clone. The NTDMA was initially detected by GC-flame ionization detection and was subsequently identified by GC-MS. These data provide further evidence that T4MO catalyzes the oxidation of NDMA to NTDMA. This is in accordance with the work of Sharp et al. (28), who previously showed that NDMA oxidation by T4MO-expressing strains such as KR1 and Ralstonia pickettii PKO1 was inhibited by acetylene, a specific inhibitor of certain monooxygenases.

In summary, the data from this work suggest that P. mendocina KR1 transforms NDMA to N-nitromethylamine and formaldehyde via the intermediary formation of N-nitrodimethylamine, using an oxidative pathway that is different from those previously described for eukaryotes. Metabolism of NDMA by a second, minor pathway which generates methanol as a metabolite is also indicated. This metabolic route may be the α-hydroxylation pathway described for eukaryotes, in which methanol and nitrogen gas are formed from the reaction of the methyldiazonium ion with water.

P. mendocina KR1 did not grow on NDMA or mineralize significant quantities of the compound to carbon dioxide, suggesting that the reaction is cometabolic rather than a growth-linked process. However, the concentrations of NDMA present in groundwater and soils (i.e., part-per-trillion to low part-per billion levels) are generally too low to support significant microbial growth. Thus, cometabolic reactions involving broad-specificity oxygenases and/or other enzymes may be important in determining the longevity of this carcinogen in groundwater and other environments. An improved understanding of NDMA biodegradation by pure cultures and in natural environments will provide critical information for better predicting the fate of NDMA and hopefully for developing new bioremediation approaches to treat it.