Isomer-selective biodegradation of high-molecular-weight azaarenes in PAH-contaminated environmental samples

Abstract

Polycyclic aromatic nitrogen heterocycles, or azaarenes, normally co-occur with polycyclic aromatic hydrocarbons (PAHs) in contaminated soils. We recently reported that nontarget analysis using high resolution mass spectrometry of samples from four PAH-contaminated sites revealed a previously unrecognized diversity and abundance of azaarene isomers and their methylated derivatives. Here we evaluated their biodegradability by natural microbial communities from each site in aerobic microcosm incubations under biostimulated conditions. The removal of total quantifiable azaarenes ranged from 15–85%, and was related to the initial degree of weathering for each sample. While three-ring azaarenes were readily biodegradable, the five-ring congeners were the most recalcitrant. Microbial-mediated removal of four-ring congeners varied for different isomers, which might be attributed to the position of the nitrogen atom that also influences the physicochemical properties of azaarenes and possibly the susceptibility to transformation by relevant microbial enzymes. The presence of methyl groups also influenced azaarene biodegradability, which decreased with increasing degree of methylation. Several oxidation products of azaarenes were detected, including ketones and dioxygenated derivatives of three- and four-ring compounds. Our results indicate the susceptibility of some azaarenes to bioremediation, while suggesting the potential implications for risk from the persistence of less-biodegradable isomers and the formation of oxidized-azaarene derivatives.

GRAPHICAL ABSTRACT

1. Introduction

Polycyclic aromatic nitrogen heterocycles (PANHs), or azaarenes, are known to be common co-contaminants at sites contaminated with polycyclic aromatic hydrocarbons (PAHs). However, due to the lack of regulatory requirements and analytical challenges (e.g. lack of standards and reference materials), PANHs had received far less attention than homocyclic PAHs, but interest in their environmental distribution has increased in recent years. PANHs have been detected in urban soils (Bandowe et al., 2014), atmospheric particulate material (Alves et al., 2017) or in paved road dust (Wei et al., 2015) as a result of fuel or biomass combustion among other sources; nevertheless, they reach particularly high concentrations in industrial soils historically contaminated by coal tar, creosote or oil derivatives (Lundstedt et al., 2014). PANHs tend to be more polar than their PAH counterparts, and thus show higher leachability (Larsson et al., 2018) and environmental mobility (Arp et al., 2014; Enell et al., 2016). As a consequence, PANHs have been found at significant concentrations in groundwater at sites associated with contamination by creosote (Pereira et al., 1983) or coal tar (Blum et al., 2011), and in water and sediment samples from riverine environments (Machala et al., 2001; Siemers et al., 2017). However, the range of azaarene congeners that have been studied in environmental systems has been limited to a few analytical targets (Anyanwu and Semple, 2015a). Although the toxicological data for PANHs is still fragmentary, the toxicity of several azaarene isomers and their metabolites is well-known (Bleeker et al., 1999; Wiegman et al., 2001). Some of the high-molecular-weight (HMW) compounds (containing 4 or more rings) are mutagenic and carcinogenic (Bleeker et al., 2002), and are major contributors of AhR-mediated activities (Machala et al., 2001). In addition to their carcinogenic potential, some azaarenes present developmental toxicity (Chlebowski et al., 2017; Hawliczek et al., 2012) and endocrine disrupting activities (Brinkmann et al., 2014; Hawliczek et al., 2012). Therefore, their fate in the environment is important to consider.

Relatively limited research has been conducted on the biodegradation of PANHs in soil and sediment systems, particularly for alkyl-substituted PANHs or those of higher molecular weight (HMW, four or more rings). A few studies have reported that bacteria isolated from soil can degrade or transform some low-molecular-weight (LMW) PANHs containing two rings (quinoline, isoquinoline) and three rings (phenanthridine, acridine) (Parshikov et al., 2012; Pereira et al., 1988; Sutherland et al., 2009; Van Herwijnen et al., 2004). However, little is still known about their fate in the environment, and to our knowledge only two reports focused on azaarene degradation in soil. Anyanwu and Semple analyzed the fate of PANH analogues of phenanthrene spiked into a pristine agricultural soil (Anyanwu and Semple, 2015b), and recently Biache and colleagues demonstrated the removal of four target PANHs during the slurry incubation of three PAH-contaminated soil samples inoculated with a suspension of soil from a gasification plant (Biache et al., 2017). These works have been limited to a small number of target low molecular weight compounds, such as quinoline, carbazole, acridine or N-containing analogues of phenanthrene, and do not consider the fate of higher molecular weight PANHs or methylated derivatives. The capabilities of autochthonous microbial communities or the effects of the N position, degree of methylation or aromatic structure on the susceptibility of azaarenes to microbial transformation remain to be elucidated.

In a previous paper (Tian et al., 2017b), we reported the detection of a wide range of azaarenes and methylated azaarenes in soil and sediment samples from four different PAH-contaminated sites, using high-resolution mass spectrometry (HRMS) combined with Kendrick mass defect filtering. Based on the observation of a large number of azaarene congeners in all four samples using these methods, we set up microcosm incubations under biostimulated conditions to evaluate their propensity to biodegradation by indigenous microbial communities. We hypothesized that biodegradability of PANH isomers could be influenced by their molecular weight, degree of methylation and/or structural features (such as position of the N atom). Organic extracts from microcosms were analyzed by HRMS to obtain insight into congener biodegradation patterns, isomer-selective degradation, and potential formation of transformation products.

2. Materials and methods

2.1. Chemicals.

Anhydrous sodium sulfate, monobasic and dibasic sodium phosphates, and high-performance liquid chromatography (HPLC)-grade solvents, including dichloromethane (DCM), acetone, and methanol, were purchased from Fisher Scientific (Pittsburgh, PA, U.S.A.). Acridine, phenanthridine, benzo[h]quinoline, benzo[c]acridine, and the PAH standard solution (EPA 610 PAH mixture) were obtained from Sigma-Aldrich (St. Louis, MO, U.S.A.). Benzo[a]acridine was purchased from LGC standards (Teddington, Middlesex, UK). Acridine-d₉ and dibenzo[a,j]acridine were from Cambridge Isotope Laboratories (Tewksbury, MA, U.S.A).

2.2. Contaminated samples.

PAH-contaminated samples (three soils and one sediment) were collected from three sites located in North Carolina (NC), U.S.A (Hu et al., 2012; Tian et al., 2017a), and a site in Andalucía, Spain (Tejeda-Agredano et al., 2013). Details on the origin and characteristics of these four samples were provided in our previous work (Tian et al., 2017b), reproduced here as Table S1 in Supporting Information. The samples are referred to below as FS (coal tar-contaminated soil from a former manufactured-gas plant in NC that had been used to feed a lab-scale bioreactor); HS (creosote-contaminated soil from the Holcomb Superfund site wood-treatment facility in NC); KM (creosote-contaminated sediment from the Kerr McGee Superfund site wood-treatment facility in NC); and SC (creosote-contaminated soil from a wood-treatment facility in Andalucía). The KM and SC samples were the most highly contaminated, and the HS and KM samples demonstrated the highest degree of weathering as defined by the ratio of HMW PAHs to low-molecular-weight PAHs (Table S1). The SC sample was both the most highly contaminated and the least weathered of the four samples.

2.3. Microcosm incubations.

The biodegradation patterns of azaarene congener groups were evaluated during aerobic incubations of the four samples. Aliquots of 1 g of air-dried, sieved (2 mm) and homogenized samples were placed in sterile 125-mL Erlenmeyer flasks with PTFE-coated screw caps, containing 30 mL of 10 mM phosphate buffer (pH 7.5) supplemented with urea as nitrogen source. A higher concentration of urea (1 mM) was used for the more highly contaminated samples (SC and KM), while a lower concentration (0.2 mM) was applied to the FS and HS samples, to maintain an approximate C:N molar ratio of 10:1 in accordance with their respective total PAH concentrations (Leys et al., 2005). This C:N molar ratio has been described as optimal to maintain maximal PAH degradation rates and extends during bioremediation of contaminated soils (Leys et al., 2005). Triplicate flasks were sacrificed initially (time 0) and after 7, 21, and 42 days of incubation. To evaluate the effect of the addition of an external nitrogen source on azaarene removal, identical triplicate microcosms except for the addition of urea were set up for the FS and SC samples; these microcosms were sacrificed after 42 days for azaarene analysis. For each sample, triplicate inhibited controls were prepared in the same manner as for biostimulated microcosms, but with the addition of phosphoric acid to pH 2, and sacrificed after 42 days of incubation. Before extraction, inhibited controls were neutralized with a concentrated NaOH solution to facilitate azaarene extraction. All flasks were incubated at 25 °C and under agitation at 150 rpm. Every two days, the caps of flasks were loosened for 5 minutes to avoid oxygen limitations.

At the end of incubation, the entire contents from each flask were transferred to a 30-mL centrifuge tube with PTFE-lined septum. The slurry was then centrifuged at 3900 rpm, and the pellets were stored at −20 °C until extraction. Solvent extractions on the pelleted soils were conducted as described previously (Tian et al., 2017b). In brief, each sample was submitted to two successive overnight extractions using 20 mL of a DCM:acetone mixture (1:1, v/v) with agitation on a wrist-action shaker. Prior to extraction, acridine-d9 (100 μg) and anthracene-d10 (200 μg) were spiked into the samples in acetone solution as recovery standards. Sodium sulfate (10 g) and glass beads were added for better extraction efficiency. The collected extracts were filtered through a 0.2 μm pore-size nylon membrane, brought up to 50 mL with DCM and acetone (1:1, v/v), and stored at 4 °C until further analysis.

2.4. Instrumental analysis.

Nontarget analysis and quantitative target analysis on selected PANHs were conducted as described in Tian et al and its Supporting Information (Tian et al., 2017b). Briefly, nontarget analysis was conducted by HPLC-HRMS using electrospray ionization in positive (ESI+) mode on an Agilent 6520 accurate mass quadrupole time-of-flight MS (qTOFMS, Agilent Technologies, Santa Clara, CA). Data processing and analysis were conducted with XCMS Online (https://xcmsonline.scripps.edu), and the resulting feature tables were imported and analyzed with an in-house script written in R language (R 3.3.3). Features corresponding to PANH isomers were distinguished by transforming the exact mass according to the Kendrick scale based on CH₂, and then calculating the mass defect (Tian et al., 2017b). Target analysis was performed on a TSQ Quantum Ultra triple quadrupole MS (QqQ MS, Thermo Fisher Scientific, Waltham, MA) in ESI+ mode. Quantification of nonalkylated azaarenes was based on the MS/MS transition from [M+H]⁺ to [M+H-28]⁺, corresponding to the loss of a CH₂N group, and further confirming the identities of azaarene isomers, as described previously (Tian et al., 2017b). The compounds for which we had standards (acridine, phenanthridine, benzo[h]quinoline, benzo[c]acridine, benzo[a]acridine, and dibenzo[a,j]acridine) were used to semi-quantify all isomers of the standard itself, assuming that the response factors were equivalent for each isomer. Semi-quantification of azapyrene and azabenzo[a]pyrene isomers was based on the relative response factor of benzo[a]acridine and dibenzo[a,j]acridine, respectively. Concentrations of USEPA-regulated PAHs were determined by HPLC with fluorescence detection as described elsewhere (Tian et al., 2017a).

Features corresponding to candidate azaarene metabolites were selected from the output of the nontarget analysis data processing using XCMS online. All the prioritized features fulfilled the following criteria: i) were significantly “upregulated” in incubated samples with respect to time 0 (fold change >3); ii) contained one nitrogen atom (derived from PANH); iii) presented double bond-equivalents (DBE) > 9 (indicative of a polyaromatic structure, containing at least two aromatic rings); and iv) contained at least one oxygen atom.

2.5. Statistical analysis.

The susceptibility to degradation of azaarene isomer groups was analyzed by hierarchical cluster analysis using Ward’s minimum variance method, and the distances between clusters were measured by Euclidean distance. The number of groups was set at 2 (“readily biodegradable” and “relatively recalcitrant”). The differences between clustered groups were assessed by Mann-Whitney-Wilcoxon test. All data analyses were conducted in R (version 3.3.3). Statistically significant differences (p < 0.05) between pairs of samples were evaluated by t-tests using Microsoft Excel 2010.

3. Results and discussion

In our previous publication (Tian et al., 2017b), we assessed the diversity and abundance of azaarenes containing a single N hetero-atom in four PAH-contaminated samples. Collectively across the four samples, the detected azaarenes represented 8 homologous series of congeners ranging from three rings to five rings. Each series contained congener groups corresponding to the unsubstituted azaarenes and their monomethyl-, dimethyl- and/or trimethyl-derivatives; the greatest intensities (which we infer as abundance, although we recognize that this is not always accurate (Huba et al., 2016)) were observed for the unsubstituted and monomethyl derivatives. Overall, 232 congeners were detected, as summarized in Table 1. There were five major congener groups, designated A, B, C, E, and F. Groups A, B and C are cata-condensed compounds with molecular formulas for the non-alkylated congeners C₁₃H₉N (3-ring), C₁₇H₁₁N (4-ring) and C₂₁H₁₃N (5-ring), respectively; groups E and F are peri-condensed compounds with molecular formulas C₁₅H₉N (4-ring) and C₁₉H₁₁N (5-ring), respectively. Each congener within a group was given a label (A1, A2, etc.).

Table 1.

Detailed information on azaarene congeners detected in the four PAH-contaminated samples (adapted from Tian et al. 2017b).

			Exact mass (Da)
Example chemical (group)a	Formula	DBE	M	[M+H]+	Confidence levelb	# of isomers
azafluorene	C12H9N	9	167.0735	168.0806	3	4
C1-azafluorene	C13H11N	9	181.8910	182.0962	3	4
C2-azafluorene	C14H13N	9	195.1048	196.1118	3	6
acridine (A)	C13H9N	10	179.0735	180.0806	1, 2	6
C1-acridine	C14H11N	10	193.0891	194.0965	3	11
C2-acridine	C15H13N	10	207.1048	208.1120	3	19
C3-acridine	C16H15N	10	221.1204	222.1275	3	3
phenylquinoline	C15H11N	11	205.0891	206.0965	3	7
C1-phenylquinoline	C16H13N	11	219.1048	220.1117	3	7
C2-phenylquinoline	C17H15N	11	233.1204	234.1277	3	8
azapyrene (E)	C15H9N	12	203.0735	204.0807	2	7
C1-azapyrene	C16H11N	12	217.0891	218.0965	3	18
C2-azapyrene	C17H13N	12	231.1048	232.1120	3	15
C3-azapyrene	C18H15N	12	245.1204	246.1275	3	4
benzo[a]acridine (B)	C17H11N	13	229.0891	230.0965	1, 2	13
C1-benzo[a]acridine	C18H13N	13	243.1048	244.1120	3	16
C2-benzo[a]acridine	C19H15N	13	257.1204	258.1278	3	13
C3-benzo[a]acridine	C20H17N	13	271.1361	272.1433	3	1
phenylacridine	C19H13N	14	255.1048	256.1118	3	5
C1-phenylacridine	C20H15N	14	269.1204	270.1278	3	1
azabenzo[a]pyrene (F)	C19H11N	15	253.0891	254.0965	3	17
C1-azabenzo[a]pyrene	C20H13N	15	267.1048	268.1122	3	22
C2-azabenzo[a]pyrene	C21H15N	15	281.1204	282.1276	3	3
dibenzo[a,j]acridine (C)	C21H13N	16	279.1048	280.1120	1, 2	10
C1-dibenzo[a,j]acridine	C22H15N	16	293.1204	294.1277	3	7
C2-dibenzo[a,j]acridine	C23H17N	16	307.1361	308.1433	3	5
Total						232

^aOf the 8 homologous series of isomers, there were 5 major groups as indicated in the text (A, B, C, E, and F). Representative chemicals in each group are indicated in parentheses.

^bConfidence levels of the identification were assigned according to the criteria established by Schymanski et al. (2014). Level 1, structure confirmed using a reference standard with identical retention time and MS. Level 2, probable structure proposed based on unambiguous MS match compared with data in literature or databases; information on HPLC elution profiles for specific groups of isomers were also obtained from literature data (Švábenský et al., 2009; Lintelmann et al., 2010). Level 3, tentative candidates proposed based on MS spectra and molecular formula, but authentic standard was not available for any representative of the homologue group; positions of the N substituent and arrangement of the aromatic ring could not be assigned. In some groups, different compounds have different levels of confidence, so both levels are shown.

3.1. Biodegradability of native azaarenes.

Each of the four PAH-contaminated samples was incubated for six weeks under biostimulated conditions to evaluate the biodegradability of the native azaarenes. The concentrations of the 34 quantifiable non-methylated congeners (belonging to the five major groups A, B, C, E and F) were monitored throughout the incubation. The data are summarized in SI; the time-course evolution of each congener in each sample is shown in Figure S1, and initial (T0) and final (T42) concentrations in microcosm incubations are provided in Tables S2 and S3. Their concentrations in inhibited controls after six weeks of incubation did not show any significant differences with respect to initial (T0) concentrations, indicating that potential abiotic losses during incubation were negligible (Tables S2 and S3). The extent of total azaarene degradation was moderate for FS and HS samples (36% and 43%, respectively) (Table S2). However, there were big differences (p=0.001) between the more-highly contaminated KM and SC samples (Table S3); in agreement with its higher degree of weathering, in the KM sample the five quantified azaarene groups were degraded only 15% after six weeks. In contrast, the much less-weathered SC sample had the highest extent of degradation, with a total azaarene removal of 85%.

The extent of degradation after six weeks was also compared across the different azaarene congener groups (Figure 1). In general, the LMW congener group A (3-ring PANHs) was the most extensively degraded (40–97%) except for the HS sample (27%), which might be attributed to the low initial concentrations in this sample (3.6 μg/g total group A compounds). The higher extent of degradation for congener group A is in agreement with the general trend of PAH biodegradation observed in many studies, for which LMW compounds are generally more easily degraded than 4- or 5-ring PAHs. Indeed, across samples the 5-ring congener groups F (15–46% removal) and C (0–26% removal) were the most recalcitrant.

Figure 1.

Concentrations of the five major azaarene congener groups over time during aerobic microcosm incubations in each of the samples. Note the differences in the y-axis scales. FS, coal tar-contaminated soil from a manufactured-gas plant site (North Carolina, USA); HS, creosote-contaminated soil from the Holcomb Superfund site (North Carolina, USA); KM, creosote-contaminated sediment from the Kerr-McGee Superfund site (North Carolina, USA); SC, creosote-contaminated soil from Andalucía, Spain.

For the more-weathered samples HS and KM, the azaarene analogues of fluoranthene and pyrene (group E) were more extensively degraded (49% and 45%) than the benzo[a]anthracene and chrysene analogues (group B, 45% and 10% removal, respectively); these differences are statistically significant for the more-contaminated KM sample (p= 0.0008). In contrast, for the less-weathered samples FS and SC, congener group B was more substantially degraded (44% and 86%, respectively) than group E (39% and 50%, respectively); these differences were significant for both samples (p=0.05 and p=0.005, respectively). It is not possible to conclude definitively from this evidence that the degree of weathering (ratio of HMW to LMW PAHs) was the primary factor influencing differences in the extent of biodegradation among the samples. However, it is plausible that weathering would decrease the bioavailability of residual HMW contaminants and also would remove the more-readily biodegradable fractions of PANHs and unsubstituted PAHs if cometabolism is important for azaarene biodegradation. It is also of interest that for the less-weathered SC sample, the order of azaarene removal was A>B>E>F>C (Figure 1). This result is inconsistent with the typical pattern for PAH biodegradation in which fluoranthene and pyrene (azaarene analogues in group E) are more biodegradable than benzo[a]anthracene and chrysene (azaarene analogues in group B) (Bossert and Bartha, 1986).

The biodegradability of the methylated congener groups was also evaluated qualitatively. Figure 2 shows the time course evolution of Kendrick mass-defect (KMD) plots for the SC sample; KMD plots are an efficient and convenient way of summarizing the congener groups as explained in Tian et al (2017b). In general, the removal of the non-methylated compounds was greater than the removal of the mono- and di-methylated homologs, while tri- and tetra-methylated families were not significantly removed (Figure 2). The lower susceptibility to biodegradation with increasing degree of methylation could probably be related with steric hindrance for microbial oxidation, resulting from the presence and distribution of methyl groups, limiting the accessibility of the active site of oxygenases to accommodate the substrates. Similar to unsubstituted PANHs, methylated derivatives of 5-ring compounds (groups C and F) were generally resistant to microbial attack. Greater removal of the mono- and dimethylated congeners than tri- and tetramethyl derivatives is in agreement with the degradation trends generally observed during microbial degradation of alkylated PAHs within petrogenic mixtures or in crude oil-polluted samples (Vila and Grifoll, 2009; Wang et al., 1998).

Figure 2.

Kendrick mass defect (KMD) plots over the time series from aerobic microcosm incubations of the SC sample under biostimulated conditions. The KMD plot describes the abundance and diversity of the detected azaarenes in eight homologous series of congener groups; the five groups of greatest interest (most abundant congeners) are labeled as follows: A (C₁₃H₉N), E (C₁₅H₉N), B (C₁₇H₁₁N), F (C₁₉H₁₁N), and C (C₂₁H₁₃N). Overlapping bubbles are isomers with an identical molecular formula and position on the KMD plot, while each horizontal series of bubbles corresponds to homologous series of isomers with increasing methylation from left to right. The color darkened by overlapping suggests the emergence of multiple isomers, and the areas of the bubbles are proportional to the intensities. SC, creosote-contaminated soil from Andalucía, Spain.

3.2. Isomer-selective biodegradation.

Within a given congener group we found significant differences in the extent of biodegradation among isomers, suggesting a selective microbial transformation of certain isomers. Such isomer-selective biodegradation has previously been observed in structurally diverse pollutants such as nonylphenols (Lu and Gan, 2014) and alkylated PAHs (Bayona et al., 1986; Lamberts et al., 2008). To quantitatively assess the relative biodegradability of individual non-methylated isomers over the course of the incubations, the concentration data from the time series for all samples were normalized by averaging the ratio of concentrations between neighboring time points for each isomer; this approach accounts for both the rate and extent of biodegradation over the six-week incubations. Isomers were then categorized by cluster analysis into two significantly different (p < 0.01) groups: relatively recalcitrant compounds and readily biodegradable compounds (Figure 3). It is important to note that “relatively recalcitrant” does not imply non-biodegradable. Here it means a statistically significant, lesser removal of a compound relative to one categorized as readily biodegradable when considering all of the time-series data for all four samples. Absolute biodegradability of a particular congener (significantly lower concentration at 42 days compared to the initial concentration) in a given sample is illustrated in Table S2.

Figure 3.

Hierarchical cluster analysis of the quantifiable non-methylated azaarene congeners according to their biodegradation behavior (using a time-series measure that combines the rate and extent of removal) during microcosm incubations. Ward’s minimum variance method was used, and the congeners were categorized into the two clearly separated groups indicated as “readily biodegradable” and “relatively recalcitrant”. B[h]Q, benzo[h]quinoline; B[c]Ac, benzo[c]acridine; B[a]Ac, benzo[a]acridine; DB[a,j]Ac, dibenzo[a,j]acridine.

As expected from the analysis of congener groups described above, all the isomers in groups C and F (5-ring compounds) clustered in the relatively recalcitrant group, while isomers in group A (3-ring compounds) fell into the readily biodegradable group (Figure 3). However, for the 4-ring congener groups B and E, significant differences were observed in the biodegradability among isomers. While most of the B isomers clustered in the readily biodegradable group, isomer B4 clustered in the relatively recalcitrant group. Two E group isomers (E3 and E5) were classified among the readily biodegradable compounds, whereas three isomers (E1, E2 and E4) were classified as relatively recalcitrant.

Examples of differences in biodegradation between isomers are shown in Figure 4. Of particular interest is the difference between isomers B1 and B6 in the SC sample (Figure 4c), which differ only in the position of the N atom in the structure. The great difference between E4 (azapyrene) and E5 (azafluoranthene) is striking (Figures 3, 4a and 4b), as their PAH counterparts are both relatively biodegradable. The differences in environmental persistence between these groups of isomers may have significant implications for risk. Despite the abovementioned scarcity in toxicological data, the relatively recalcitrant benzo[a]acridine (B1) has been reported as significantly more genotoxic than benzo[c]acridine (B6) (Bleeker et al., 1999), whereas the latter showed higher acute toxicity for marine flagellates (Wiegman et al., 2001). For azapyrene (E4), its mutagenic and carcinogenic potential were suggested using the Ames and AISA assays, respectively (Tanga et al., 1986), and more-recent work demonstrated strong developmental toxicity in zebrafish compared to the PAH 11H-benzo[b]fluorene and the alkylated-PAH retene (Hawliczek et al., 2012). Among the B- and E-group isomers, the relationship between structure and biodegradation potential is still unclear due to the lack of authentic standards. However, the HPLC elution order of isomers (Figure 3 in Tian et al., 2017b) generally corresponded to isomer biodegradability (Figure 3), such that later-eluting isomers were generally more biodegradable (illustrated for two of the samples in Figure 4). This observation suggests that one of the parameters controlling the elution (polarity or basicity) could be inversely related to biodegradability. However, given that homocyclic PAHs tend to be less biodegradable as hydrophobicity increases, a relationship governed by polarity seems less likely. Indeed, it has been demonstrated that sorption mechanisms other than partitioning play an important role in the sorption of N-heterocyclic compounds to soils and minerals (Bi et al., 2006), the overall process being dominated by the cation exchange capacity of protonated azaarenes. Considering that, basicity could be a relevant factor influencing PANH biodegradability in soils. Alternatively, it is also possible that the apparent trend is coincidental, and that the position of the N atom or ring structure (e.g., azapyrene vs azafluoranthene) determines biodegradability based on enzyme specificity.

Figure 4.

Overlayed extracted ion chromatograms (EICs) for C₁₅H₉N (m/z 204.0800) from KM and SC (b) samples at day 0 and after 42 days of microcosm incubation, and for C₁₇H₁₁N (m/z 230.0800) from the SC sample (c) at day 0 and after 21 days of incubation. Each panel depicts the isomer-selective extent of degradation within an azaarene congener group. Data in panel c are shown for 21 days of incubation because several of the analytes were completely removed by Day 42. Structures E4 (azapyrene) and E5 (azafluoranthene) were inferred from literature data (Lintelmann et al., 2010; Švábenský et al., 2009), while those of B1 (benzo[a]acridine) and B6 (benzo[c]acridine) were identified using authentic standards.

3.3. Effect of exogenous nitrogen addition on azaarene biodegradation.

The presence of a nitrogen atom in the azaarene molecule suggested its potential to be used as a nitrogen source for microbial growth, facilitating the assimilation of the carbon skeleton, and thus promoting azaarene removal under nitrogen limited conditions. To evaluate whether the addition of an external nitrogen source influenced azaarene biodegradation, we conducted 42-day incubations of the FS and SC samples in the presence and absence of added nitrogen (Figure S2). In general, in the absence of added nitrogen there was significantly lower degradation of all congener groups, and in no case did the omission of the external nitrogen source lead to improved degradation of any isomer. This finding does not exclude the possibility that the azaarenes could serve as a nitrogen source for the microbial community, but neither does it provide support for such a possibility (we do not know the background concentrations of other possible nitrogen sources in these samples). Conversely, the addition of nitrogen clearly led to significantly greater removal of a number of the isomers in both samples, particularly of some of those concluded to be recalcitrant to biodegradation (e.g., E1, E2, E4, F5 or F9 for the FS sample, and E1, E2, F5 or F9 for the SC sample). Overall, these results underscore the benefit of adding an external nitrogen source to maximize biostimulation.

3.4. Biodegradability of azaarenes relative to their PAH analogues.

We compared the concentrations and biodegradation behaviors of azaarene congener groups with respect to their corresponding homocyclic counterparts. It has been estimated that NSO-heterocyclic polyaromatics are on the order of 1% to 10% of the total mass of homocyclic PAHs in environmental mixtures (Bleeker et al., 2002; Neilson, 1998). For two of the samples in our case (HS and SC), ratios of azaarene congeners to their respective PAH analogues were substantially higher than that (Figure S3). During the microcosm incubations, the proportions of azaarenes to analogue PAHs remained relatively constant for the weathered samples HS and KM. However, they generally increased over time for the HMW congeners in the highly contaminated and much less-weathered SC sample (Figure S3), suggesting the greater recalcitrance of HMW azaarene congeners relative to their homocyclic counterparts. Considering the higher water solubility of azaarenes, this increased recalcitrance could be a factor of the abovementioned sorption to soil materials due to cation exchange of protonated azaarenes and/or steric hindrance for microbial attack derived from the presence of the N atom in the molecule. Depending on the toxicity of azaarenes relative to their PAH counterparts, these observations may have implications for residual risk following bioremediation. In this sense, Machala and colleagues (Machala et al., 2001) identified that the AhR-inducing properties of the five-ring azaarene dibenz[a,j]acridine were similar to those observed for the prototypical carcinogenic HMW-PAH benzo[a]pyrene, and this potential was even higher for its isomer dibenz[a,h]acridine, which showed an AhR-inducing potency 50-fold higher than that of benzo[a]pyrene.

3.5. Transformation products from azaarenes.

The nontarget analysis approach was applied to detect the formation or fate of putative transformation products from azaarenes during the microcosm incubations. Considering that incubations were performed under aerobic conditions, primary attention was paid to oxygenated derivatives of azaarenes that included one or two oxygen atoms. Four major features (groups of related analytes) were detected, whose occurrence and degradation were specific to samples and time points (Figure 5). The first group corresponded to the formula C₁₃H₉NO (m/z 196.0760), which includes acridone, phenanthridinone, and their isomers (Figure 5a). These compounds have been detected in estuarine environments (de Voogt and Laane, 2009), and similar ketones derived from quinolines were suggested as indicators of natural attenuation in groundwater at coal tar-contaminated sites (Reineke et al., 2007). Their concentrations decreased during the incubation process, suggesting that they were already present in the initial samples and were readily biodegradable. Other groups of compounds possessed the formulas C₁₃H₉NO₂ (m/z 212.0706) (data not shown), C₁₄H₁₁NO₂ (m/z 226.0864) (Figure 5b), and C₁₇H₁₁NO₂ (m/z 262.0862) (Figure 5c), corresponding to dioxygenated azaarenes. Although these formulas are consistent with dihydroxy-compounds, analogous HMW dihydroxy-PAHs are highly unstable and easily autoxidized to quinones (Jouanneau and Meyer, 2006; Penning et al., 1996); therefore, we do not believe such dihydroxy compounds would have been present long enough to have been detected in the samples. Other possibilities include the monohydroxy derivatives of acridone-like structures (Figures 5b and 5c), which are more-stable tautomers of dihydroxy compounds (Fetzner, 1998). Because of the lack of reference standards and further information, these derivatives are recognized as level 4 suspects (Schymanski et al., 2014). In the SC samples the concentrations of these analytes peaked at 7 d (for C₁₄H₁₁NO₂) or 21 d (for C₁₇H₁₁NO₂) and decreased thereafter (Figure 5b and c), suggesting their transient formation and subsequent removal during the biodegradation process. In addition, their emergence was concomitant with the removal of the corresponding parent azaarene congener groups, further suggesting that they were oxidized products from azaarenes. Except for the acridine derivatives (C₁₃H₉NO) (Bobeldijk et al., 2002; Brinkmann et al., 2014), the other oxidized azaarenes with formulas C₁₃H₉NO₂, C₁₄H₁₁NO₂, and C₁₇H₁₁NO₂ have not been reported in the literature.

Figure 5.

EICs of azaarene derivatives in the SC sample. (a) C₁₃H₉NO (m/z 196.0760); (b) C₁₄H₁₁NO₂ (m/z 226.0864); (c) C₁₇H₁₁NO₂ (m/z 262.0862). Structures are hypothetical, and positions of the N atom shown in the structures are illustrative only.

4. Conclusions

Current risk assessment approaches during remediation of PAH-contaminated soils generally consider reference levels for only 16 regulated PAHs. However, we have demonstrated the co-occurrence of a wide range of PANHs, including HMW compounds and methylated derivatives. Considering the known toxicological properties of several azaarenes, here we assessed their potential environmental fate during bioremediation. Like their PAH counterparts, there are substantial differences in their susceptibility to biodegradation as a function of molecular weight (number of rings) and degree of methylation. However, the position of the N atom and other parameters associated with their physicochemical properties (such as basicity or polarity) affect the susceptibility to biodegradation between isomers of the same congener group. We have also demonstrated the potential for accumulation of azaarene transformation products under aerobic conditions. These findings on the biodegradation of azaarenes will lend support to future research in environmental monitoring as well as bioremediation for this overlooked class of contaminants.

Highlights.

• Fate of polycyclic aromatic N-heterocycles assessed in four PAH-contaminated soils

• 8 series of congeners analyzed by nontarget screening and mass defect filtering

• Biodegradability varied as a function of molecular weight and degree of methylation

• Selective degradation was observed between isomers of the same congener group

• Nontarget analysis identified the transient accumulation of PANH-oxidation products