Degradation of fluoranthene by Pasteurella sp. IFA and Mycobacterium sp. PYR-1: isolation and identification of metabolites

Authors


 : H. Leskovs¦ek, Department of Environmental Sciences, 39 Jamova Street, Ljubljana, 1000, Slovenia (e-mail : hermina.leskovsek@ijs.si).

Abstract

The findings from a biodegradability study of fluoranthene using two pure bacterial strains, Pasteurella sp. IFA (B-2) and Mycobacterium sp. PYR-1 (AM) are reported. Of total fluoranthene, 24% (B-2) and 46% (AM) was biodegraded in an aqueous medium during 14 d of incubation at room temperature. During this period the bacteria were capable of mineralizing approximately two-thirds (B-2) and four-fifths (AM) of biodegraded fluoranthene to CO2, while one-third (B-2) and one-fifth (AM) of the original fluoranthene remained as stable metabolic products. These metabolites were isolated using liquid–liquid extraction and identified using gas chromatography – mass spectrometry (GC–MS) and derivatization techniques. Two metabolites (9-fluorenone-1-carboxylic acid and 9-fluorenone) were identified by GC–MS directly, while the metabolites 9-fluorenone-1-carboxylic acid, 9-hydroxyfluorene, 9-hydroxy-1-fluorene-carboxylic acid, 2-carboxybenzaldehyde, benzoic acid and phenylacetic acid were determined in their derivatized forms. From the identified metabolites, a fluoranthene biodegradation pathway was proposed for Pasteurella sp. IFA.

Polycyclic aromatic compounds (PAH) are widespread environmental pollutants commonly found in soil, surface waters and sediments. Most of the higher molecular weight PAH and even some low molecular weight PAH are toxic and have potential carcinogenic properties (Richardson & Gangolli 1992). Microbial degradation is the major route through which PAH are removed from the environment (Cho 1997). While lower molecular weight PAH have been shown to be degraded easily (Cerniglia 1984 ; Gibson & Subramanian 1984 ; Heitkamp et al. 1987 ; Weissenfels et al. 1990 ; Cerniglia 1992 ; Cullen et al. 1994 ; Ghoshal et al. 1996), the metabolism of four or more ring PAH has been less extensively studied and is also less well understood (Heitkamp et al. 1987 ; Kelley et al. 1991 ; Walter et al. 1991 ; Weissenfels et al. 1991 ; Cerniglia 1993 ; Kelley et al. 1993 ; Li et al. 1996a,b). It is known that biodegradation of more complex PAH is limited by their low bioavailability, resulting from extremely low water solubility and strong adsorption to soil particles (Weissenfels et al. 1992).

Various micro-organisms that degrade PAH containing fewer than four rings have been isolated, but less is known about the ability of micro-organisms to metabolize the more recalcitrant molecules such as fluoranthene (Barnsley 1975 ; Hietkamp et al. 1988 ; Mueller et al. 1989 ; Kelley & Cerniglia 1991 ; Boldrin et al. 1993). In several laboratories, the degradation of high molecular weight polycyclic aromatic hydrocarbons has been demonstrated using bacterial strains isolated from mixed bacterial culture (Heitkamp & Cerniglia 1988 ; Weissenfels et al. 1990 ; Kelley et al. 1991 ; Boldrin et al. 1993 ; Li et al. 1996b), but Weissenfels et al. (1990) were the first to report that a pure strain of Alcaligenes denitrificans isolated from a mixed culture was able to mineralize fluoranthene completely.

More recently, degradation of PAH by Mycobacteria was also demonstrated by Boldrin et al. (1993) who showed that Mycobacterium sp. strain BB1 was able to utilize phenanthrene, pyrene and fluoranthene as a sole source of carbon and energy and to degrade fluorene co-metabolically. A pure culture of Mycobacterium sp. PYR-1 isolated from oil-contaminated estuarine sediment was reported by Kelley & Cerniglia (1991) to be able to mineralize fluoranthene significantly following enzymatic induction by pyrene. The same authors also reported the isolation and identification of the metabolite 9-fluorenone-1-carboxylic acid by Mycobacterium sp. strain PYR-1 (Kelley et al. 1991). This was the first report of a fluoranthene metabolite in which significant degradation of one of the aromatic rings occurs. They were later able to identify nine more metabolites (Kelley et al. 1993) and reported the possible simultaneous modes of biodegradation by Mycobacterium sp. strain PYR-1.

In this study, the degradability of fluoranthene using a pure Pasteurella sp. IFA strain was investigated and compared with the results from the degradation of fluoranthene by Mycobacterium sp. PYR-1 (known to be able to degrade fluoranthene) under the same experimental conditions.

Materials and methods

Growth medium

The bacterial strains Pasteurella sp. IFA (B-2) and Mycobacterium sp. PYR (AM) were chosen for fluoranthene biodegradation studies in aqueous media. Both strains were previously isolated from an oil-contaminated soil and an estuarine sediment. Mycobacterium sp. PYR-1 was kindly donated by the National Centre for Toxicological Research, Jefferson, Arkansas, USA and Pasteurella sp. IFA by the Institute for Agrobiotechnology, Tulln, Austria. The preparation of the substrate, mineral media with trace elements and the inoculum are described in full elsewhere (Šepičet al. 1996, Šepičet al. 1997).

Biodegradation of fluoranthene in aqueous media

All experiments were set up in aqueous media at room temperature (22–26 °C). In the first experiment using Pasteurella sp. IFA, four cotton wool-stoppered Erlenmeyer flasks (1 litre) were filled with mineral media containing 20 mg l−1 fluoranthene and incubated for 10 d. A similar experiment was set up using Mycobacterium sp. PYR-1 with 15 flasks using the same concentration of fluoranthene (20 mg l−1). Ten flasks were incubated for 8 d and five flasks for 14 d, all at room temperature. The 10 d incubation period for Pasteurella sp. IFA and the 8 d incubation period for Mycobacterium sp. PYR were found to be the optimal incubation times from previous studies (Šepičet al. 1997). A 14 d incubation period for Mycobacterium sp. PYR-1 was chosen to determine whether different metabolic products were accumulated after prolonged incubation.

Analytical procedure

At set intervals, the cultures were extracted three times with equal volumes of ethylacetate (EtAc, 50 ml) and combined to give the neutral extract. The cultures were then acidified to pH 2·5 with H2SO4 and extracted once more with another three volumes of ethylacetate. These were combined to give the acidic extract. Both extracts were dried (anhydrous Na2SO4) and reduced in volume to 1 ml (N2). The extracts were analysed using gas chromatography with mass selective detection (GC–MSD, Hewlett Packard 6890-5972; Waldbronn, Germany). The chromatograph was equipped with an HP–MS5 crosslinked 5% phenylmethyl silicone capillary column (30 m length, 0·25 mm diameter, 0·25 mm film thickness).

Commercially available standard compounds (Aldrich Chemical Co., Milwaukee, WI, USA) matching the published fluoranthene metabolites (Kelley et al. 1993) were dissolved in organic solvents and screened using GC–MS. The structure of these compounds is presented in Fig. 1. The elucidation of the structure of metabolites E, F, H and I was only possible after derivatization when characteristic mass spectra were obtained.

Figure 1&.

emsp;The structure of fluoranthene metabolites 9-fluorenone-1-carboxylic acid (a), 9-fluorenone (b), 9-hydroxyfluorene (c), 9-hydroxy-1-fluorene-carboxylic acid (d), adipic acid (e), phthalic acid (f), 2-carboxybenzaldehyde (g), benzoic acid (h), phenylacetic acid (i)

Derivatization

Derivatization is used to improve either the volatility or the thermal stability of compounds to make them more accessible to gas chromatography. The various fluoranthene metabolites differ in their chemical structure and functional groups, e.g. R-OH, R-COOH, R-C = O, R-CHO and no one derivatizing agent is suitable for all the metabolites ; fortunately there are many derivatizing agents available. The most common derivatization method is the formation of trimethylsilyl (TMS) ethers and esters (Poole 1977) which can be applied to compounds containing an active hydrogen group, thus converting polar-reactive compounds into unpolar-inert compounds. Other functional groups, e.g. ketones may also be transformed to enol-TMS ethers, although oxime formation is a better alternative derivatization technique for these types of compounds.

Silylation.

For those metabolites containing R-OH and R-COOH groups, silylation with N-methyl-N-(trimethylsilyl) trifluoracetamide (MSTFA ; Aldrich) was applied (Sloan et al. 1971 ; Ende & Luftmann 1984). The method involves adding 50 μl MSTFA derivatization agent to 200 μl standard metabolite solution (0·2 mg ml−1) in a small sealed reaction vial. The mixture is then left for 12 h at room temperature with constant agitation to allow the TMS-ethers and esters to form.

Oxime formation.

Derivatization with O-(2,3,4,5,6-pentafluorobenzyl)-hydroxylamine hydrochloride (PFBHA.HCl ; Fluka Chemie AG, Buchs, Switzerland) was used for those metabolites containing carbonyl groups (R-C = O and R-CHO) (Kovác¦ & Anderle 1977). This involved adjusting the pH of 1 mg of metabolite extract to 7·4 using 1·5 ml 0·1 mol l−1 Tris buffer (adjusted using dilute HCl) followed by the addition of 3 ml 0·15 mol l−1 KCl and 1 ml 0·05 mol l−1 PFBHA.HCl. The sample was then stirred for 3 h at room temperature (22–26 °C). The pentafluorobenzyloximes (PFBO) were extracted from reaction solution using 3 volumes of cyclohexane : diethylether solution (4 : 1). The solvent (Merck KGaA, Darmstadt, Germany) was removed (N2) and the PFBO products dissolved in 100 μl toluene.

Each sample extract was derivatized with PFBHA.HCl and divided into two halves. One half of each derivatized sample was analysed by GC–MS, and the remainder derivatized with MSTFA so that all the samples had been derivatized with both derivatizing agents.

Oxygen consumption

The Biochemical Oxygen Demand (BOD) is a standard test used to measure the amount of oxygen required by bacteria to oxidize the organic matter in a water sample. In this case, the test was conducted over a 14 and 28 d period at 20 °C within a controlled environment and the consumed oxygen measured manometrically (10 bottles, WTW Model BSB-Controler Model 1020T; Weilheim, Germany). For each bacterial strain, five samples similar to those used in the biodegradation studies, and five blanks containing DCM to account for oxygen consumption due to the DCM present in the samples, were set up. Carbon dioxide was removed from the samples by placing NaOH pellets in each bottle cap. The BOD was calculated for each set of samples and blank corrected.

Carbon dioxide evolution

Micro-organisms can degrade organic compounds to different extents by either breaking only one bond in the molecule, or by completely degrading the compound to CO2 and H2O e.g. mineralization (Heitkamp et al. 1987 ; Venosa et al. 1992 ; Hanstveit 1992). Therefore, by measuring the evolved CO2, the amount of totally degraded fluoranthene can be estimated. In this study, the amount of CO2 formed over a 1 month period (concentration of CO2 at days 7, 14, 21 and 28) was determined according to the method reported by Thomas (1995). The apparatus consisted of three dreschel bottles connected in series. The first bottle contained the sample (5 mg of fluoranthene in 250 ml of mineral media and the inoculum) ; the second contained a standardized BaOH solution and the third, pure water to act as an airlock. At the end of the experiment, the flasks were flushed using pure nitrogen. The amount of evolved CO2 could then be calculated from the residual amount of BaOH measured by titrating with 0·05 mol l−1 HCl according to the standard procedure. All samples were blank corrected.

Results

Isolation and characterization of metabolites

Neutral and acidic extracts of each biodegraded sample from both bacteria were first analysed using GC–MS operating in the SCAN mode (m/z = 50–500). Fluoranthene metabolic products were identified on the basis of the molecular and fragment ions in the mass spectra and the chromatographic retention time (Rt) of authentic compounds. Unfortunately, the concentration of the metabolites was insufficient for SCAN detection so the sample extracts from the same incubation times were combined and re-analysed. In the chromatograph of the combined neutral extract, the undegraded fluoranthene peak was the major component. The identification of these metabolites was further improved by using GC–MS in the SIM mode. In the non-derivatized neutral and acidic extract of both bacterial samples, only two metabolites were confirmed : 9-fluorenone-1-carboxylic acid and 9-fluorenone.

As direct analysis of the sample proved insufficient for detecting all but two of the known metabolites, even when using SIM, derivatization was applied. Initially, the authentic metabolites of fluoranthene (except metabolite B) were derivatized using MSTFA. The resultant TMS-ethers and esters produced a typical mass fragmentation pattern with an intense ion fragment at M-15 corresponding to the loss of a –CH3 group from the molecular ion. The ions at m/z 73 [(CH3)3Si]+, m/s 75 [HO = Si(CH3)2]+ and m/z 147 [(CH3)2Si = O-Si(CH3)3]+ were characteristic for MSTFA derivatization. The ion at m/z 73 was usually the base ion. Also obvious in the spectra was the [M-90]+ ion formed by the cleavage of the TMS-OH group. GC–MS of the PFBO derivatives revealed intense ions at m/z 181 (F5(C6H5CH2)+), [M-181]+ and [M-197]+. In the last fragment ion, besides (F5(C6H5CH2)+, oxygen was eliminated. The fragment m/z 181 was usually the base ion in PFBO spectra.

The advantages of derivatization are shown in Fig. 2. Metabolite E (adipic acid) before derivatization shows a broad chromatographic peak (Rt 10·2 to 12·2 min) and a non-characteristic mass spectrum with no obvious molecular ion and without any specific fragment ions (Fig. 2a). On the other hand, the derivatized form of metabolite E using MSTFA produces a sharp chromatographic peak (Rt 13·2 to 13·3 min) and a characteristic mass spectrum (Fig. 2b). Even if the molecular ion is hardly noticeable (m/z 290), the [M+-CH3] fragment at m/z 275 and the fragment ions at m/z 73, m/z 75, m/z 111, m/z 129 and m/z 147 are sufficient for identification purposes. The formation of the m/z 129 and m/z 147 fragments is described by Capella & Zorzut (1968).

Figure 2&.

emsp;MS spectrum of metabolic adipic acid. (a) Direct analysis, (b) its TMS-ester. (a) Scan 077 (12·31 min) : METABE.D(−) ; (b) Scan 624 (13·239 min) : METEMI.D(−)

Among the metabolites tested only metabolite B did not contain carboxylic groups and was derivatized only with PFBHA.HCl. The metabolites A, C, D, E, F, G, H and I were derivatized with MSTFA. Metabolites A and G contain both types of functional groups (R-C = O, R-CHO) and were derivatized using both PFBHA.HCl and MSTFA, resulting in PFBO-TMS-ethers and esters.

GC-MS/SIM analysis was able to detect the presence of minor amounts of fluoranthene metabolite derivatives. No PFBO-metabolite derivatives were found in the samples containing Pasteurella sp. IFA after 10 d of incubation. In the neutral and acidic extracts the following metabolites (as TMS-esters) were detected : A, D, G and H. In the acidic extract, metabolite I was found. In both extracts, metabolite A also occurred as PFBO-TMS-ester and metabolite G was shown to be present ; only the derivatized form (PFBO oxime or PFBO-TMS-ester) was not clearly defined because of its similar structure and therefore similar Rt and fragmentation pattern at low sample concentration. Although this problem was repeated throughout the work, the main task was to determine the presence of metabolites and not the derivatized forms, so therefore it was not investigated further. Table 1a shows the percentage of identified fluoranthene metabolites after derivatization in both extracts as a result of degradation by Pasteurella sp. IFA (10 d of incubation).

Table 1.  Metabolites identified after the degradation of fluoranthene by Pasteurella sp. IFA (B-2) after 10 d (a) and by Mycobacterium sp. PYR-1 (AM) following 8 (b) and 14 d (c) of incubation, identified by GC-MS/SIM after derivatization
   Neutral (%)Acidic (%) 
MetaboliteSymbolMol.
wt.
PMPMPMPMTotal
(%)
  • P = PFBO oximes ; M = TMS esters and ethers after derivatization with MSTFA ; PM = PFBO-TMS-esters and ethers after derivatization with both agents.

  • *

    Not clearly defined.

(a) Pasteurella sp. IFA (10 d of incubation)
9-Fluorenone-1-carboxylic acidA224n.d.0,1140,019n.d.0,1590,0530.345
9-FluorenoneB180n.d.n.d.n.d.n.d.n.d.n.d.n.d.
9-HydroxyfluoreneC182n.d.n.d.n.d.n.d.n.d.n.d.n.d.
9-Hydroxy-1-fluorene-carboxylic acidD226n.d.0,008n.d.n.d.0,032n.d.0.040
Adipic acidE146n.d.n.d.n.d.n.d.n.d.n.d.n.d.
Phthalic acidF166n.d.n.d.n.d.n.d.n.d.n.d.n.d.
2-CarboxybenzaldehydeG150n.d.0,1070,006*n.d.0,2370,006*0.356
Benzoic acidH122n.d.5.158n.d.n.d.2.767n.d.7.925
Phenylacetic acidI136n.d.n.d.n.d.n.d.0,013n.d.0.013
(b) Mycobacterium sp. PYR-1 (8 d of incubation)
9-Fluorenone-1-carboxylic acidA224n.d.0,1090,030n.d.0,6750,3951.209
9-FluorenoneB180n.d.n.d.n.d.n.d.n.d.n.d.n.d.
9-HydroxyfluoreneC182n.d.0,004n.d.n.d.n.d.n.d.0.004
9-Hydroxy-1-fluorene-carboxylic acidD226n.d.0,065n.d.n.d.0,515n.d.0.006
Adipic acidE146n.d.n.d.n.d.n.d.n.d.n.d.n.d.
Phthalic acidF166n.d.n.d.n.d.n.d.n.d.n.d.n.d.
2-CarboxybenzaldehydeG1500,0050,0020,003*n.d.0,1010,013*0.124
Benzoic acidH122n.d.3,894n.d.n.d.3,334n.d.7.227
Phenylacetic acidI136n.d.0,007n.d.n.d.0,022n.d.0.029
(c) Mycobacterium sp. PYR-1 (14 d of incubation)
9-Fluorenone-1-carboxylic acidA224n.d.0.1900.092n.d.1.1951.5393.016
9-FluorenoneB180n.d.n.d.n.d.n.d.n.d.n.d.n.d.
9-HydroxyfluoreneC182n.d.n.d.n.d.n.d.n.d.n.d.n.d.
9-Hydroxy-1-fluorene-carboxylic acidD226n.d.0.073n.d.n.d.0.265n.d.0.338
Adipic acidE146n.d.n.d.n.d.n.d.n.d.n.d.n.d.
Phthalic acidF166n.d.n.d.n.d.n.d.n.d.n.d.n.d.
2-CarboxybenzaldehydeG150n.d.0.0500.010*n.d.0.1710.005*0.236
Benzoic acidH122n.d.5.242n.d.n.d.3.924n.d.9.166
Phenylacetic acidI136n.d.0.017n.d.n.d.0.013n.d.0.030

In neutral sample extracts from the experiments with bacterial strain Mycobacterium sp. PYR-1 after 8 d of incubation, metabolite G in PFBO form and metabolites A, C, D, G, H and I as TMS-ethers and esters were present. Again, metabolites A and G were identified in the neutral and acidic extracts as PFBO-TMS-ester. In the acidic extract, the metabolites G-PFBO and C-TMS-ether could not be detected. The remaining metabolites were determined as TMS-esters (A, D, G, H, I) and as PFBO-TMS esters A and G were proved to be present. Table 1b shows percentages of metabolites identified in neutral and acidic extracts and the total percentage of identified metabolites for fluoranthene degradation by Mycobacterium sp. PYR-1 after 8 d of incubation.

Similarly, in the samples containing Mycobacterium sp. PYR-1, no PFBO derivatives were present after 14 d of incubation. However, in the neutral and acidic extracts, the following metabolites (as TMS-esters) were identified : A, D, G, H and I. Metabolites A and G were also identified as PFBO-TMS-esters. Table 1c shows percentages of identified fluoranthene metabolites by Mycobacterium sp. PYR-1 after 14 d of incubation.

Oxygen consumption

For Pasteurella sp. IFA, the BOD for 14 and 28 d of incubation was 3·8 mg l−1 and 5·5 mg l−1, with approximately 4% within the first 24 h, another 3% between 2 and 14 d of incubation, and 3% of total BOD being oxidized between days 14 and 28. For Mycobacterium sp. PYR-1, the BOD was 21·9 mg l−1 and 27·2 mg l−1 over the same period, respectively. The results revealed that most of the oxygen is consumed within the first 48 h of incubation (approximately 30% of total BOD), followed by a much reduced consumption with time (8% between days 2 and 14 and 8% between days 14 and 28).

Carbon dioxide evolution

The amount of CO2 evolved during 28 d of incubation (days 7, 14, 21 and 28) is given in Table 2. The percentage of CO2 evolved was 14% for Pasteurella sp. IFA and 40% for Mycobacterium sp. PYR-1 after 28 d of incubation.

Table 2.  Fluoranthene biodegradation mass balance
BacteriaAM
(8 d)
AM
(14 d)
B-2
(10 d)
Remaining undegraded  fluoranthene (%)595475
Released CO2 (%)414014
Identified metabolites (%) 9139
Total10910798

Discussion

The presence of the same metabolites in the degradation products of each bacterium suggests that Pasteurella sp. IFA degrades fluoranthene by a similar route to Mycobacterium sp. PYR-1. In the case of Pasteurella sp. IFA, five metabolites (9-fluorenone-1-carboxylic acid, 9-hydroxy-1-fluorene-carboxylic acid, 2-carboxybenzaldehyde, benzoic acid and phenylacetic acid) were identified and with Mycobacterium sp. PYR-1 (AM), six metabolites (9-fluorenone-1-carboxylic acid, 9-hydroxyfluorene, 9-hydroxy-1-fluorene-carboxylic acid, 2-carboxybenzaldehyde, benzoic acid and phenylacetic acid). For Pasteurella sp., metabolite 9-hydroxyfluorene (C) was not identified. This is thought to be because of the lower sample volume.

From Table 1 it can be seen that benzoic acid (H) was the main metabolite identified (91·3%) after degradation by Pasteurella sp., followed by (G, 4·1%) 2-carboxybenzaldehyde and (A, 4·0%) 9-fluorenone-1-carboxylic acid. The metabolites phenylacetic acid (I) and 9-hydroxy-1-fluorene-carboxylic acid (D) were both under 0·5%. Even though the incubation times for the two bacteria are not exactly comparable, there appear to be differences in their metabolite patterns, e.g. Pasteurella sp. accumulated a higher amount of benzoic acid and 2-carboxybenzaldehyde but smaller amounts of the other metabolites compared with Mycobacterium sp. PYR-1. Benzoic acid (H) was the most abundant metabolite in all the experiments involving Mycobacterium sp. PYR-1, accounting for 78·8% (AM, 8 d of incubation) and 71·8% (AM, 14 d) of the total. The metabolite with the second highest concentration was 9-fluorenone-1-carboxylic acid (A, 13·1%, 8 d) and 23·4% (14 d), followed by 9-hydroxy-1-fluorene-carboxylic acid (D, 6·3% and 2·7%). The metabolites G, I and C were all under 0·5%. This shows that although Pasteurella sp. is slower in degrading fluoranthene, it is more efficient in degrading initial metabolic products (e.g. A, 9-fluorenone-1-carboxylic acid) and results in higher amounts of accumulated one ring metabolites (e.g. G, 2-carboxybenzaldehyde and H, benzoic acid).

The difference in metabolic pattern between Pasteurella sp. IFA and Mycobacterium sp. PYR-1 can be observed by their different BOD. The metabolism of Pasteurella sp. IFA showed a constant but slow rise in the BOD curve compared with the BOD curve of Mycobacterium sp. PYR-1 ; Mycobacterium sp. PYR-1 showed a sharper rise but then suddenly showed reduction in BOD over time. These results suggest a longer lag phase and a slower but continuous degradation for Pasteurella sp. IFA, while Mycobacterium sp. PYR-1 degrades most of the metabolites in the initial degradation phase.

Of the two bacterial strains investigated, Mycobacterium sp. PYR-1 was more successful in completely mineralizing fluoranthene in aqueous media than Pasteurella sp. IFA. Pasteurella sp. IFA degraded only 25% of total fluoranthene, while 14% of fluoranthene was mineralized to CO2 and 9% remained as stable metabolic products (Table 1). In the case of Mycobacterium sp. PYR-1, 46% of total fluoranthene was biodegraded, 40% of biodegraded fluoranthene was mineralized to CO2, and 13% of biodegraded fluoranthene was identified as stable metabolic products. The CO2 evolution experiment confirmed that Mycobacterium sp. PYR-1 was far more active and more efficient than Pasteurella sp. IFA.

Fluoranthene biodegradation pathway

According to the metabolites identified in this study, the same fluoranthene biodegradation pathway was used by both bacterial strains. The presence of metabolites 9-fluorenone-1-carboxylic acid, 9-hydroxy-1-fluorene-carboxylic acid, 2-carboxybenzaldehyde, benzoic acid and phenylacetic acid observed at all the incubation times for both bacteria suggests two possible fluoranthene biodegradation pathways (Fig. 3).

Figure 3&.

emsp;Proposed fluoranthene biodegradation pathway

In a previous paper (Šepičet al. 1997), a part of the fluoranthene biodegradation pathway for Mycobacterium sp. PYR-1 was proposed. In this study, only the metabolite 9-fluorenone-1-carboxylic acid was identified and it was thought that decarboxylation would lead to the formation of 9-fluorenone. However, even after extensive investigations, the presence of 9-fluorenone could not be confirmed, suggesting that the biodegradation of fluoranthene does not follow this route, although it might be that 9-fluorenone is simply an unstable intermediate which quickly undergoes further degradation, or that the previous intermediate, 9-fluorenone-1-carboxylic acid, is very stable. This would explain why it is found in the second highest concentration. It is more likely that protonation of metabolite 9-fluorenone-1-carboxylic acid (1·209%) leads to the formation of metabolite 9-hydroxy-1-fluorene-carboxylic acid (0·580%). Its further decarboxylation is questionable but it would produce 9-hydroxyfluorene which was detected, although only in one experiment (after 8 d of incubation with Mycobacterium sp. PYR-1). This, again, does not prove the fluoranthene biodegradation pathway. Further oxidation of these metabolites could lead to the formation of the benzoic derivatives which were identified, 2-carboxybenzaldehyde (0·124%), benzoic acid (7·228%) and phenylacetic acid (0·029%).

The presence of a significant amount of 9-fluorenone-1-carboxylic acid confirmed the pathways of fluoranthene degradation by Mycobacterium sp. PYR-1 proposed in the literature by Kelley et al. (1993). The formation of 9-hydroxy-1-fluorene-carboxylic acid represents an initial attack on the fused aromatic ring portion of the fluoranthene molecule. The presence of 9-hydroxy-1-fluorene-carboxylic acid indicates an alternative pathway which occurs simultaneously with the 9-fluorenone-1-carboxylic acid.

In summary, Pasteurella sp. IFA is only able to degrade approximately half the amount of fluoranthene that Mycobacterium sp. PYR-1 degraded over a similar period, i.e. 24% and 46%, respectively. The degradation pattern of the two bacterial strains differs in that Pasteurella sp. IFA shows a steady degradation with time compared with Mycobacterium sp. PYR-1 which shows an initial rapid degradation of fluoranthene followed by a much reduced degradation rate. Also, the amount of total degraded fluoranthene completely mineralized to CO2 is smaller in Pasteurella sp. IFA than in Mycobacterium sp. PYR-1. Although there is an obvious difference in the rate of degradation between both bacterial strains, the identification of the same metabolites in the degradation products of each bacterium confirms that fluoranthene is degraded via similar biodegradation pathways. Finally, even if Pasteurella sp. IFA cannot compete with Mycobacterium sp. PYR-1, it is nonetheless an efficient degrader of fluroanthene and one of only a few identified pure strains of bacteria that are capable of degrading four ring PAH. This is the first reported biodegradation pathway of fluoranthene by Pasteurella sp. IFA.

Acknowledgements

The financial support of the Slovenian Ministry of Science and Technology is gratefully acknowledged. The authors wish to thank Prof. J. Marsel from the University of Ljubljana, Slovenia and Dr David J. Heath from the National Institute of Chemistry, Ljubljana, Slovenia for collaboration and helpful discussions, Dr Andreas P Loibner and Werner Fuchs from the Institute for Agrobiotechnology, Tulln, Austria for activated sludge isolates, and Prof. Carl E. Cerniglia from the National Centre for Toxicological Research, Jefferson, Arkansas, USA for the donation of Mycobacterium sp. PYR-1. They are also grateful to Mrs Karmen Stanic¦ from the National Institute of Biology for cultivation of bacterial cultures.

Ancillary