Selective Removal of Nitrogen from Quinoline and Petroleum by ...

11 downloads 0 Views 96KB Size Report
0099-2240/00/$04.000. Feb. 2000, p. 688–693. Vol. 66, No. 2. Copyright © 2000, American Society for Microbiology. All Rights ..... emulsion phase at 2,000.
APPLIED AND ENVIRONMENTAL MICROBIOLOGY, Feb. 2000, p. 688–693 0099-2240/00/$04.00⫹0 Copyright © 2000, American Society for Microbiology. All Rights Reserved.

Vol. 66, No. 2

Selective Removal of Nitrogen from Quinoline and Petroleum by Pseudomonas ayucida IGTN9m JOHN J. KILBANE II,1* RAJARAM RANGANATHAN,1 LISA CLEVELAND,1 KEVIN J. KAYSER,1 CLAUDIA RIBIERO,2 AND MONICA M. LINHARES2 Institute of Gas Technology, Des Plaines, Illinois 60018,1 and Petrobras Research & Development Center Cidade Universitaria, 21949-900 Rio de Janiero, Brazil 2 Received 31 August 1999/Accepted 23 November 1999

Enrichment culture experiments employing soil and water samples obtained from petroleum-contaminated environments succeeded in the isolation of a pure culture possessing the ability to utilize quinoline as a sole nitrogen source but did not utilize quinoline as a carbon source. This culture was identified as Pseudomonas ayucida based on a partial 16S rRNA gene sequence, and the strain was given the designation IGTN9m. Examination of metabolites using thin-layer chromatography and gas chromatography-mass spectrometry suggests that P. ayucida IGTN9m converts quinoline to 2-quinolinone and subsequently to 8-hydroxycoumarin. Resting cells of P. ayucida IGTN9m were shown to be capable of selectively removing about 68% of quinoline from shale oil in a 16-h treatment time. These results suggest that P. ayucida IGTN9m may be useful in petroleum biorefining for the selective removal of organically bound nitrogen from petroleum. sonably well studied, there is very little information concerning the use of quinoline-degrading microorganisms to remove nitrogen from petroleum. Several quinoline-degrading Pseudomonas cultures were found to have no ability to remove significant levels of nitrogen from crude oil or asphaltene fractions of petroleum (2). The removal of 20 to 45% of nitrogen from crude oil by unspecified mixed cultures has been claimed, but the ability of these cultures to metabolize quinoline is unknown (12). The purpose of this investigation was to isolate aerobic bacterial cultures capable of utilizing quinoline as a nitrogen source, but incapable of utilizing quinoline as a carbon source, and subsequently to examine the metabolic pathway of quinoline degradation and the ability of such cultures to selectively remove nitrogen from petroleum.

The quality of petroleum is progressively deteriorating as the highest-quality petroleum deposits are preferentially produced (10, 17). Consequently, concern about the concentration of compounds and contaminants such as sulfur, nitrogen, and metals in petroleum will intensify. These contaminants not only contribute to environmental pollution resulting from the combustion of petroleum but also interfere with the processing of petroleum by poisoning catalysts and contributing to corrosion (8, 17). The selective removal of contaminants from petroleum while retaining the fuel value is a difficult technical challenge. New processes are needed (10). The selective removal of sulfur from dibenzothiophene and from petroleum by various microbial cultures has been demonstrated, and this technology is now being commercialized for the biorefining of petroleum (13). Biorefining can also potentially be used to remove nitrogen and metals from petroleum, but so far this area of research has received very little attention (12). Quinoline is perhaps the most widely studied organonitrogen compound as regards biodegradation, and quinoline is considered to be representative of many organonitrogen compounds typically found in petroleum (6). Many aerobic and anaerobic microbial cultures that can degrade quinoline have been found (9). The majority, if not the entirety, of microbial cultures described in the literature that metabolize quinoline do so by fully degrading it and can therefore utilize quinoline as a sole source of carbon, energy, and nitrogen (2, 5, 7, 9, 14, 18–22). However, for use in a petroleum biorefining application, it would be preferable if nitrogen was selectively removed from quinoline, leaving the carbon and the calorific value of the molecule intact. The metabolic pathways utilized by various aerobic quinoline-degrading microorganisms have been investigated, and even though the cultures were capable of fully degrading or mineralizing quinoline, some metabolic pathways were shown to initiate the degradation of quinoline by selectively oxidizing and removing nitrogen from quinoline (9, 19). While the biodegradation of quinoline has been rea-

MATERIALS AND METHODS Bacterial cultures and growth conditions. Environmental samples of soil and/or water were obtained from petroleum and coal processing sites, compost, and other sites where contamination with petroleum hydrocarbons exist. The environmental samples were used to inoculate nutristat and shake flask enrichment culture experiments to obtain cultures capable of using quinoline as a sole source of nitrogen. All experiments employed a defined mineral salts medium containing, per liter, 0.37 g of KH2PO4, 0.25 g of MgSO4 䡠 7H2O, 0.07 g of CaCl2 䡠 2H2O, 0.02 g of FeCl3, and 20 g of glucose, glycerol, and succinate (6.66 g of each). This medium was adjusted to a pH of 6.5 to 7, and nitrogen was supplied in the form of quinoline which was present at concentrations ranging from 1 to 20 mM. For control experiments, (NH4)2SO4 was used at a concentration of 1.3 g/liter. Agar plates were prepared by adding 15 g of agar per liter of mineral salts medium and adding ammonia or quinoline as a nitrogen source at 5 mM. Nutristats and shake flasks were operated at room temperature (approximately 26°C) or 30°C. The working volume of nutristats was 1 liter, and shake flask experiments generally used 50 or 100 ml of liquid medium. Flow rates of the nutristats were adjusted to achieve hydraulic retention times ranging from 2 days to as long as a month, and the flow rates as well as the organonitrogen test compound were altered as needed to ensure that the nutristats created an environment suitable for the selection of cultures with improved abilities to selectively cleave C-N bonds. This means that the bacterial cell density in the nutristats and shake flasks ranged from 102 to 108 cells/ml, but generally cell densities of 104 to 105 cells/ml were maintained. The bacteria isolated from the effluent of nutristats and/or from shake flasks were subjected to chemical mutagenesis and/or physical mutagenesis using nitrosoguanidine (NTG) and shortwave UV irradiation, respectively. For NTG mutagenesis, cell suspensions containing approximately 109 cells/ml were incubated for 30 min at 30°C with shaking with 10 ␮g of NTG per ml (1). The cells were recovered by centrifugation at 10,000 ⫻ g for 15 min and used to inoculate nutristats or enrichment culture

* Corresponding author. Mailing address: Institute of Gas Technology, 1700 S. Mt. Prospect Rd., Des Plaines, IL 60018. Phone: (847) 768-0723. Fax: (847) 768-0546. E-mail: [email protected]. 688

VOL. 66, 2000

SELECTIVE REMOVAL OF NITROGEN FROM QUINOLINE

experiments. UV mutagenesis was performed by placing 5 ml of bacterial cell suspension (approximately 108 cells/ml) in plastic petri dish bottoms which were exposed to a germicidal UV lamp (General Electric) at a distance of 10 cm for 90 s. The cell suspension was agitated or swirled continuously during UV exposure. These conditions of NTG and UV mutagenesis were chosen to result in the death of about 99% of the cell population as determined by quantifying CFU on nutrient agar plates (Difco, Detroit, Mich.) before and after mutagenesis. The mutagenized cells were then used to reinoculate nutristats, start additional shake flask experiments, and streak onto agar plates containing organonitrogen test compounds. Care was taken to ensure that the amount of biomass that was added back to nutristats in the form of inocula was insufficient to provide a significant amount of nitrogen in the form of dead biomass. Cells from the effluents of nutristats, shake flasks, and agar plates were routinely tested using the nitrogen bioavailability assay described below. Pure cultures were obtained by streaking samples obtained from nutistats or from positive nitrogen bioavailability assays. Single colonies were obtained from nutrient agar and from mineral salts agar plates. Pure cultures that repeatedly yielded positive results in nitrogen bioavailability assays were selected. The determination of 16S rRNA gene sequences for the determination of the species of bacterial isolates was performed by MIDI Labs (Newark, Del.). Nitrogen bioavailability assay. The nitrogen bioavailability assay used defined mineral salts medium in growth tests in which organonitrogen model compounds such as quinoline, pyridine, carbazole, and porphyrin served as sources of carbon and/or nitrogen. Growth tests were performed using six conditions: 1, test compound as sole source of carbon and nitrogen; 2, test compound as sole source of carbon (alternative nitrogen source, ammonia, was available); 3, test compound as sole source of nitrogen (alternative carbon source, glucoseglycerol-succinate, was available). 4, test compound present as well as alternative sources of carbon and nitrogen; 5, only alternative nitrogen (ammonia) and carbon (glucose-glycerol-succinate) sources were available (the test compound was not present); 6, nitrogen compounds were not present, but alternative carbon (glucose-glycerol-succinate) sources were available. These six growth conditions constitute a bioassay for the ability of a culture to metabolize organonitrogen compounds. When carbon and nitrogen sources other than the test compounds were needed, they were supplied in the form of a glucose-glycerol-succinate mixture (20 g/liter) and as ammonia (20 mM), respectively. The nitrogen bioavailability assay described above can be performed with any organonitrogen test compound which is ordinarily used at a concentration of from 1 to 20 mM. The various cultures to be tested were inoculated into test tubes or shake flasks containing medium components appropriate for the six test conditions. The cultures were then incubated aerobically with agitation for 2 to 5 days, at room temperature (approximately 26°C) or at 30°C. The growth of the cultures was monitored by measuring the turbidity (optical density) of the cultures in the various test conditions or by determining CFU. Turbidity of cell suspensions in test tubes was determined with a Klett-Summerson colorimeter equipped with a green filter such that 100 Klett units corresponds to a cell density of about 5 ⫻ 108 cells/ml. CFU were determined by plating dilution series of bacterial cells onto nutrient agar (Difco, Detroit, Mich.) plates. The nitrogenfree sample (test condition 6) served as a negative control, while the samples amended with both carbon and nitrogen sources (test conditions 4 and 5) served as positive controls and should produce healthy microbial growth unless the test compound is toxic to the culture being tested. In this event, only condition 5 should result in healthy growth. The amount of bacterial growth observed in test conditions 1, 2, and 3 in comparison with the amount of growth observed in test conditions 4, 5, and 6 indicates the ability of cultures to use the organonitrogen test compound as a source of carbon and/or nitrogen. Those cultures which showed better growth in test condition 3 than condition 1 or 2 may have been preferentially utilizing the organonitrogen compound as a nitrogen source only and were examined more thoroughly in further experiments. Spectrophotometric scans to detect metabolites. UV/visible spectrophotometric scans were performed at 240 to 900 nm on a Beckman DU-65 spectrophotometer. Supernatants from cultures grown using ammonia or quinoline as sole nitrogen sources were obtained by centrifugation at 10,000 ⫻ g for 15 min and then compared spectrophotometrically to identify new peaks formed due to accumulation of metabolites specific to the metabolism of quinoline. Substrate range tests. Microbiological tests were performed to determine the range of organonitrogen compounds that could serve as sole sources of nitrogen for growth for the various pure cultures obtained in the project. Tests were also performed to determine if various organonitrogen compounds are inhibitory to the growth of these microbial cultures. Growth tests were performed according to the nitrogen bioavailability assay procedures previously described. The organonitrogen compounds and control compounds used included 2-methyl-␤naphthothiazole, 2-methyl benzothiazole, 2(methylmercapto) benzimidazole, 1,1-methylene bis(3-methyl piperidine), thiazole, 1-butylpyrrolidine, 2-methylene-1,3,3-trimethyl indoline, 2-methyl-3-propylpyrazine, 2-phenylbenzothiazole, 2-methyl quinoxaline, 2-methyl indoline, carbazole, quinoline, 8-hydroxyquinoline, 5-hydroxyisoquinoline, 3,4-dihydro-2(1H)-quinoline, quinazoline, 2,4-quinolinediol, isoquinoline, 3-methyl isoquinoline, isocarbostyril, protoporphyrin, pyridine, phenyl benzothiazole, nicotinic acid, imidazole, indole, HEPES buffer, urea, guanine, lysine, tryptophan, 4-hydroxycoumarin, 7-hydroxycoumarin, and ammonium sulfate. These chemicals were chosen to determine the range of organonitrogen compounds that could be used as nitrogen sources by bacterial

689

cultures. All chemicals were obtained from Sigma (St. Louis, Mo.) or Aldrich Chemical Company (Milwaukee, Wis.) and were of the highest purity obtainable. TLC for identification of metabolites. Thin-layer chromatography (TLC) was performed on Whatman silica C18 plates by the method described by Watson and Cain (22). The running-phase solvents used were hexane, acetic acid, and xylene in the ratio of 5:1:2. Supernatants from bacterial cultures grown with quinoline as the sole source of nitrogen were obtained after centrifugation at 10,000 ⫻ g for 15 min and were then acidified to pH 2 with HCl. Typically 100 ml of aqueous supernatant was extracted with 0.5 to 1.0 volume of ethyl acetate which was recovered from the aqueous phase using a separatory funnel. The ethyl acetate extract was then evaporated in a hood, resulting in concentration of the sample from 20- to 1,000-fold prior to the analysis of the extracts by TLC. Typically 10 to 50 ␮l of ethyl acetate sample that had been concentrated 100-fold relative to the volume of aqueous supernatant extracted was spotted onto TLC plates. The TLC plates typically were run for about 20 min and then observed under normal lighting and under short-wave (245 nm) and long-wave (366 nm) UV light. Some experiments also utilized resting cells which were prepared by centrifuging from 100 to 500 ml of log-phase cultures grown with either quinoline or ammonia as a nitrogen source. Then the washed cell pellets were resuspended in 5 to 50 ml of mineral salts medium to final cell densities of 109 to 1010 cells/ml. These cell suspensions were incubated with 20 mM quinoline for periods ranging from 15 min to 16 h at 30°C with shaking. Extraction of the supernatants from resting cells as well as growing cells was carried out either by ethyl acetate solvent extraction or with C18 solid-phase extraction cartridges (Waters Associates, Milford, Mass.). The cartridges for solid-phase extraction were rinsed with 2 ml of distilled water, and then various volumes of supernatant sample were filtered through the C18 solid-phase extraction cartridges. The cartridge was extracted with 1-ml aliquots of ethyl acetate. The ethyl acetate layer was drawn from the aqueous layer and dried with sodium sulfate. The extract was stored in amber vials at 4°C until they were analyzed by TLC and/or gas chromatography (GC)-mass spectrometry (MS). Semicarbazide-HCl and 2,4-dinitrophenyl hydrazine (2,4-DNPH), for inhibiting cell respiration and derivatizing metabolites, respectively, were added to some experiments followed by subsequent extraction and TLC analysis as described by Watson and Cain (22). These experiments used resting cell suspensions prepared as described above. A typical incubation mixture which utilized metabolite derivitization consisted of 200 ml of mineral salts medium which contained 2 g (dry weight) of cells, 2 mM semicarbazide-HCl, and 3 to 20 mM quinoline as the nitrogen source. The mixture was incubated for 2 h at 30°C with shaking, and the cells were centrifuged at 10,000 ⫻ g for 15 min. To the supernatant, 5 ml of 0.2% 2,4-DNPH in 2 M HCl was added. This mixture was left overnight at room temperature and extracted twice with 0.5 volume of ethyl acetate. The combined extracts were shaken for 1 min with Na2CO3 and then separated on a TLC plate either as is or after concentration of the ethyl acetate 20- to 1,000-fold by evaporation. The 2,4-DNPH-derivatized spots were identified as being yellow on a white background under normal lighting. These plates were also viewed under UV light to detect any spots not visible as a 2,4-DNPH derivative. Appropriate controls such as cultures grown with NH4Cl, with and without semicarbazide and with and without 2,4-DNPH in the reaction mixture, were also examined. GC-MS. GC-MS analysis was performed on extracts derived from growing and resting cell cultures exposed to quinoline and on compounds eluted from spots observed on TLC plates. Extraction of the supernatants from resting cells as well as growing cells was carried out either by ethyl acetate solvent extraction or with C18 solid-phase extraction cartridges as described above for the preparation of samples for TLC analysis. Additionally, TLC spots of possible metabolites were scraped from the TLC plates and eluted with ethyl acetate and concentrated for analysis by GC-MS. For analysis of the extracts, we used a Hewlett-Packard (HP) 5971 mass selective detector and 5890 series II GC with HP 7673 auto sampler tower and a 30-meter Rezteck XTI-5 column. The final oven temperature was maintained at 300°C. The detection limit was 1 ng or 1 ␮g/ml with a 1-␮l injection. Mass spectrographs were compared with various libraries of mass spectrograph data prepared from known standard compounds. Several chromatograph libraries were consulted to determine the identity of metabolites of organonitrogen compounds. The presence or absence of nitrogen in various compounds was also determined by GC-atomic emission detection using the nitrogen-specific wavelength of 174.2 nm for detection. Tests with petroleum. Pseudomonas ayucida IGTN9m was grown in mineral salts medium using quinoline as the sole nitrogen source to produce 1 liter of culture at an optical density at 600 nm of 1.67, which was harvested by centrifugation at 10,000 ⫻ g for 15 min; the cell pellet was resuspended in 100 ml of mineral salts medium. The culture was divided into two 50-ml portions, and 3 ml of shale oil (1.7% nitrogen) was added to each. The cultures were incubated in 250-ml Erlenmeyer flasks shaking at room temperature overnight (16 h). Then the oil was recovered using a separatory funnel and the centrifugation of any emulsion phase at 2,000 ⫻ g for 10 min. The water-free oil phase was removed by syringe or pipette. Nitrogen analyses were performed by a modified ASTM D-3177 method. The allowed replicate tolerances and reproducibility of these analyses were 0.05%, and the detection limit was 0.002% N. Carbon-hydrogen

690

KILBANE ET AL.

APPL. ENVIRON. MICROBIOL.

FIG. 1. Comparison of MS data of a metabolite of quinoline produced by P. ayucida IGTN9m with 2-quinolinone.

analyses were performed by a modified ASTM D-3177 method. The allowed replicate analysis tolerances and reproducibility of these analyses were 0.5% C and 0.1% H, and the detection limits were 0.3% C and 0.1% H. Sulfur concentrations in oil samples were determined using a LECO SC-132 analyzer, which has a working range of 0.002 to 100%, with an accuracy of ⫾61% (relative). The detection limit for sulfur is 0.002%. The amount of quinoline present in oil samples was determined by GC using an HP 5890 series II GC equipped with a 5921A atomic emission detector and calculating the area under the peak corresponding to a retention time for quinoline.

RESULTS AND DISCUSSION Environmental samples obtained from petroleum-contaminated locations were used to inoculate nutristats and shake flask enrichment culture experiments in which 3 to 20 mM quinoline was supplied as the sole source of nitrogen. Quinoline was initially toxic to the nutristat cultures, making it difficult to establish and maintain bacterial growth. However, after several weeks bacterial growth was obtained. Initially the flow rates of the nutristats were adjusted to achieve hydraulic retention times of about 30 days. As the ability of the microbial culture to tolerate and utilize quinoline improved, the flow rates of the nutristats were increased to be as fast as possible without causing washout of all cells. Eventually, hydraulic retention times of 48 h were achieved. Mixed and pure cultures obtained from the quinoline nutristats were routinely tested using the nitrogen bioavailability assay to detect cultures capable of using quinoline as a nitrogen source but not as a carbon source. All of the quinoline-utilizing cultures initially obtained from the nutristats were found to fully degrade quinoline, utilizing it as a carbon as well as a nitrogen source. Subsequently, the flow rates of the nutristats were increased so that hydraulic retention times decreased from 96 to 4 h or less

over a period of 4 months. During this time cells were routinely obtained from the nutristat effluent, mutagenized, and returned to the nutristat. Mutagenesis was employed because repeated enrichment culture experiments performed without mutagenesis failed to isolate any cultures that utilized quinoline as a nitrogen source, without also using it as a carbon source. Eventually, we obtained a pure culture that yielded nitrogen bioavailability assay results indicating that quinoline was used as a nitrogen but not a carbon source. A partial (500-bp) sequence of the 16S rRNA gene of this gram-negative, rod-shaped bacteria was determined, identifying it as P. ayucida. This culture was designated P. ayucida IGTN9m. P. monteilii, which shows 99% homology, and P. nitroreducens and P. pseudoalcaligenes, which both show 98.3% homology, are closely related but not identical to P. ayucida IGTN9m. P. ayucida IGTN9m was determined to have a cell doubling time of 4.25 h when grown in defined salts medium at 30°C with quinoline as the sole nitrogen source. Substrate range and specificity tests were also carried out for P. ayucida IGTN9m, using prolonged incubations of 2 to 3 weeks. These substrate range tests were repeated several times using inocula derived from all of those organonitrogen compounds that yielded growth in a previous nitrogen bioavailability assay. This was done to address the possible need for adaptation to a given compound before growth could be obtained and to accurately determine the widest range of organonitrogen structures that could be utilized by P. ayucida IGTN9m. Agar plates were streaked after growth was observed on any of the test compounds to check culture purity. Repeated substrate range tests confirmed that the results obtained were representative of the

VOL. 66, 2000

SELECTIVE REMOVAL OF NITROGEN FROM QUINOLINE

691

FIG. 2. Comparison of MS data of a metabolite of quinoline produced by P. ayucida IGTN9m with 7-hydroxy-2H-1-benzopyran-2-one (7-hydroxycoumarin).

abilities of P. ayucida IGTN9m and not a mutant derivative of the culture. The full range of compounds used in these tests is listed in Materials and Methods. Growth was obtained on urea, tryptophan, lysine, guanine, nicotinic acid, quinoline, 3,4-dihydro-2(1H)-quinolinone, 2,4-quinolinediol, 8-hydroxyquinoline, and quinoxaline. Growth with urea, tryptophan, lysine, guanine, and nicotinic acid as nitrogen sources is an ability possessed by many if not most aerobic bacteria and most likely has no relationship to metabolic pathways relevant to the utilization of quinoline. The culture grew in the presence of all test compounds except 1-butylpyrrolidine and 2-methylene-1,3,3trimethyl indoline when quinoline was simultaneously present as an alternate nitrogen source. These results indicate that with the exception of the two compounds mentioned above, the organonitrogen compounds do not have a toxic effect on the culture, and so failure to grow on the other test compounds provides a true reflection of the substrate range of P. ayucida IGTN9m for the metabolism of organonitrogen compounds. Further tests are needed to determine if 1-butylpyrrolidine and/or 2-methylene-1,3,3-trimethyl indoline are intrinsically toxic or if they are metabolized by IGTN9m into inhibitory compounds. Spectrophotometric scans (UV/visible) were carried out on the supernatants of P. ayucida IGTN9m cultures that were grown with quinoline as the sole nitrogen source. This was done to determine if we could detect spectrophotometric peaks other than quinoline that might correspond to the accu-

mulation of metabolites. Control supernatants from the cultures grown with ammonium sulfate were included to rule out artifacts arising from normal metabolism of the culture. Quinoline has an absorption maximum peak at 310 nm. This peak can be observed to decrease with time during the growth of P. ayucida IGTN9m at the expense of quinoline; however, we observed no new peak that might correspond to a metabolite of quinoline (data not shown). TLC was performed on extracts derived from the culture supernatants of P. ayucida IGTN9m grown with quinoline as the sole nitrogen source, as well as with resting cells incubated for various times in the presence of quinoline. Controls consisting of P. ayucida IGTN9m grown using ammonium sulfate rather than quinoline as a nitrogen source were included in all experiments. Additionally, pure chemicals that are possible metabolites of quinoline such as protocatechuate, catechol, pyruvic acid, p-hydroxybenzoic acid, formamide, 8-hydroxyquinoline, and succinic acid dimethyl ester were included on the TLC plates to determine if any of these compounds were formed during the microbial degradation of quinoline. Two spots which have Rf values of 0.73 and 0.88 were identified as possible metabolites of quinoline by P. ayucida IGTN9m, as these compounds were found only in samples derived from the incubation of P. ayucida IGTN9m cells with quinoline. P. ayucida IGTN9m cells incubated with ammonia did not produce these compounds (data not shown). These two compounds could not be accurately identified by TLC analysis alone, and

692

KILBANE ET AL.

APPL. ENVIRON. MICROBIOL.

TABLE 1. Time course of accumulation of quinoline metabolites produced by IGTN9m as measured by GC-MS Metabolite

2-Quinolinone 8-Hydroxycoumarin

Content (%)a 15 min

60 min

4h

47.8 8.4

37.3 17.3

17.2 16.7

a Percentage of the total area of all peaks present in the chromatograph (with the exception of quinoline).

so these spots were scraped from TLC plates, eluted with ethyl acetate, and subjected to GC-MS analysis. GC-MS analysis was performed not only on compounds obtained from TLC spots but also on extracts of culture supernatants from growing cell as well as resting cells of P. ayucida IGTN9m. The GC-MS analysis of extracts of culture supernatants allowed for the possible detection of metabolites that did not yield detectable spots in TLC. The results from the GC-MS analyses provided probable structures for two compounds as metabolites of quinoline produced by P. ayucida IGTN9m: 2-quinolinone and 8-hydroxycoumarin. MS data comparing these metabolites with authentic 2-quinolinone and 7-hydroxy-2H-1-benzopyran-2-one are shown in Fig. 1 and 2, respectively. 8-Hydroxycoumarin is not commercially available, but 7-hydroxycoumarin (7-hydroxy-2H-1-benzopyran-2-one) is, so 7-hydroxycoumarin was used as a standard in these experiments. The metabolite of quinoline produced by P. ayucida IGTN9m had a somewhat shorter retention time in GC analysis than 7-hydroxycoumarin, but it yielded a mass spectrum apparently identical to that of 7-hydroxycoumarin and to published spectra of 8-hydroxycoumarin (18). To further analyze the metabolites of quinoline produced by P. ayucida IGTN9m, the relative abundance of these two metabolites was quantified with resting cells exposed to quinoline for various times. The results (Table 1) strongly indicate that quinoline is converted first to 2-quinolinone and then to 8-hydroxycoumarin. There are undoubtedly other metabolites, as the direct conversion of 2-quinolinone to 8-hydroxycoumarin is unlikely. However, no other intermediates relevant to this conversion were observed in these experiments. The structures of 2-quinolinone and 8-hydroxycoumarin (Fig. 3) suggest a partial pathway for the biotransformation of quinoline by P. ayucida IGTN9m. The oxygenation of the carbon atom adjacent to the nitrogen atom in quinoline to form 2-quinolinone is consistent with the selective cleavage of C-N bonds in quinoline by P. ayucida IGTN9m. Moreover, the results of the substrate range tests which indicate that P. ayucida IGTN9m can utilize quinoline, 3,4-dihydro-2(1H)-quinolinone, 2,4-quinolinediol, 8-hydroxyquinoline, and quinoxaline are also consistent with the partial pathway depicted in Fig. 3. The identification of 2-quinolinone and 8-hydroxycoumarin as metabolites of quin-

oline produced by P. ayucida IGTN9m is also consistent with published results of other investigations of the bacterial metabolism of quinoline (9, 19). Nonetheless, the partial pathway for the metabolism of quinoline by P. ayucida IGTN9m pictured in Fig. 3 must be considered tentative, as the structures of the metabolites have not been identified with absolute certainty. Since 8-hydroxycoumarin does not contain nitrogen, these data demonstrate that P. ayucida IGTN9m is capable of selective removal of nitrogen from quinoline. The subsequent decreased abundance of 8-hydroxycoumarin with longer incubation times (Table 1) indicates that this compound is in turn converted to other, as yet unidentified metabolites. However, it is important to keep in mind that this culture does not use quinoline as a carbon source and is unable to completely degrade it efficiently. Also, many quinoline-degrading microorganisms have been reported to produce pink (18, 19), green (14, 18), and brown (19) metabolites, but P. ayucida IGTN9m produces no colored metabolites derived from quinoline. The metabolism of quinoline by microbial cultures is complicated by the use of different chemical names by various researchers. The initial aerobic metabolites of quinoline are sometimes called 2-quinolinone (14), 2(1H)-quinolinone (15), 2-hydroxyquinoline (7, 18, 19), or 2-oxo-1,2-dihydroquinoline (4, 16). However, these various names refer to the same metabolite, which tautomerizes between different structures as shown in Fig. 3. The oxygen involved in the conversion of quinoline to 2-quinolinone/2-hydroxyquinoline/2-oxo-1,2-dihydroquinoline has been shown to be derived from water (6, 11, 15). Quinoline 2-oxidoreductase has been shown to be responsible for this initial metabolism of quinoline (4, 16). The subsequent conversion of 2-quinolinone/2-hydroxyquinoline/2-oxo-1,2-dihydroquinoline to 2,6-dihydroxyquinoline or 2,8-dihydroxyquinoline has been reported for various microbial cultures (7, 18, 19). The further metabolism of 2,6-dihydroxyquinoline has been reported to result in the degradation of the benzene ring portion of the molecule (7), while the further metabolism of 2,8-dihydroxyquinoline has been reported to result in the formation of 8-hydroxycoumarin with concomitant release of nitrogen and the subsequent degradation of the molecule to 2,3-dihydroxyphenyl-propionic acid (19), cis,cis-muconic acid, and catechol (6). The production of 8-hydroxycoumarin by P. ayucida IGTN9m is similar to the metabolism of quinoline reported for an unidentified Pseudomonas culture (19); however, the inability to use quinoline as a carbon source and the lack of production of colored metabolites make the metabolism of quinoline by P. ayucida IGTN9m unique. Results with petroleum. A test for the removal of organic nitrogen from petroleum was performed to determine the ability of P. ayucida IGTN9m to remove nitrogen from shale oil. Duplicate washed, concentrated P. ayucida IGTN9m cell sus-

FIG. 3. Suggested partial pathway for the degradation of quinoline by P. ayucida IGTN9m.

VOL. 66, 2000

SELECTIVE REMOVAL OF NITROGEN FROM QUINOLINE

TABLE 2. Biotreatment of shale oil with P. ayucida IGTN9m Sample

Contenta Carbon

Hydrogen

Nitrogen

Sulfur

Quinoline

Control

85.01

9.69

1.71

1.41

177,495

Biotreated 1 2

85.21 84.87

9.81 9.85

1.63 1.62

1.34 1.36

58,502 57,415

a Results are reported as a percentage of the neat oil, except for the amount of quinoline, which is reported as the area under the peak corresponding to the retention time of quinoline in gas chromatographs.

pensions were incubated with shale oil samples for 16 h at 30°C. The control sample consisted of shale oil added to sterile mineral salts medium which was incubated for 16 h at 30°C. After incubation, the petroleum samples were recovered and analyzed. The results (Table 2) indicate that resting cells of P. ayucida IGTN9m are capable of removing about 5% of the total organic nitrogen and about 68% of quinoline from shale oil during an overnight (16-h) incubation. In summary, the selective cleavage of carbon-nitrogen bonds by P. ayucida IGTN9m was demonstrated by determining that quinoline is metabolized to compounds tentatively identified as 2-quinolinone and 8-hydroxycoumarin. Initial results indicate only a 5% reduction of total nitrogen from petroleum, but about 68% of the quinoline present in petroleum was selectively removed as a consequence of biotreatment. These results indicate that biorefining processes for the selective removal of nitrogen from petroleum are technically feasible. However, while these results are encouraging, they must be extended with further research and the isolation of cultures which metabolize a broader range of organonitrogen substrates if a practical petroleum biorefining process for the removal of nitrogen is to be developed. ACKNOWLEDGMENTS This research was supported by a grant from Petrobras Research and Development Center and by U.S. Department of Energy contract DE-AC26-99BC15219. REFERENCES 1. Adelberg, E. A., M. Mandel, and C. C. C. Chen. 1965. Optimal conditions for mutagenesis by N-methyl-N⬘-nitro-N-nitrosoguanidine in Escherichia coli K12. Biochem. Biophys. Res. Commun. 18:788–795. 2. Aislabie, J., A. K. Bej, H. Hurst, S. Rothenburger, and R. M. Atlas. 1990. Microbial degradation of quinoline and methylquinolines. Appl. Environ. Microbiol. 56:345–351. 3. Asturias, J. A., and K. N. Timmis. 1993. Three different 2,3-dihydroxyy-1,2dioxygenase genes in the gram-positive polychlorobiphenyl-degrading bac-

693

terium Rhodococcus globerulus P6. J. Bacteriol. 175:4631–4640. 4. Blase´, M., C. Brutner, B. Tshisuaka, S. Fetzner, and F. Lingens. 1996. Cloning, expression and sequence analysis of the three genes encoding quinoline 2-oxidoreductase, a molybdenum-containing hydroxylase from Pseudomonas putida 86. J. Biol. Chem. 271:23068–23079. 5. Brockman, F. J., B. A. Denovan, R. J. Hicks, and J. K. Fredrickson. 1989. Isolation and characterization of quinoline-degrading bacteria from subsurface sediments. Appl. Environ. Microbiol. 55:1029–1032. 6. Fetzer, S. 1998. Bacterial degradation of pyridine, indole, quinoline, and their derivatives under different redox conditions. Appl. Microbiol. Biotechnol. 49:237–250. 7. Grant, D. J. W., and T. R. Al-Najjar. 1976. Degradation of quinoline by a soil bacterium. Microbios 15:177–189. 8. Hegedus, L. L., and R. W. McCabe. 1981. Catalyst poisoning. Catal. Rev. 23:377–476. 9. Kaiser, J. P., Y. Feng, and J. M. Bollag. 1996. Microbial metabolism of pyridine, quinoline, acridine, and their derivatives under aerobic and anaerobic conditions. Microbiol. Rev. 60:483–498. 10. Kassler, P. 1996. World energy demand outlook, p. 229–242. In G. Jenkins (ed.), Energy exploration & exploitation, vol. 14. Multi-Science Publishing Co. Ltd., Berkshire, United Kingdom. 11. Kayser, K. J., B. A. Bielaga-Jones, K. Jackowski, O. Odusan, and J. J. Kilbane. 1993. Utilization of organosulfur compounds by axenic and mixed cultures of Rhodococcus rhodochrous IGTS8. J. Gen. Microbiol. 139:3123– 3129. 12. Lin, M. S., E. T. Premuzic, J. H. Yablon, and W. M. Zhou. 1996. Biochemical processing of heavy oils and residuum. Appl. Biochem. Biotechnol. 57/58: 659–664. 13. Monticello, D. J., S. Murphy, and S. Johnson. 1996. Biorefining and microbial desulfurization: the upgrading of crude oil and bitumen, p. 133–145. In L. Lortie, W. D. Gould, and M. Stichbury (ed.), Proceedings of the 12th Annual General Meeting of BIOMINET. Natural Resources Canada, Ottawa 14. O’Loughlin, E. J., S. R. Kehrmeyer, and G. K. Sims. 1996. Isolation, characterization, and substrate utilization of a quinoline-degrading bacterium. Int. Biodeterior. Biodegrad. 38:107–118. 15. Pereira, W. E., C. E. Rostad, T. J. Leiker, D. M. Updegraff, and J. L. Bennett. 1988. Microbial hydroxylation of quinoline in contaminated groundwater: evidence for incorporation of the oxygen atom of water. Appl. Environ. Microbiol. 54:827–829. 16. Peschke, B., and F. Lingens. 1991. Microbial metabolism of quinoline and related compounds. XII. Rhodococcus spec. B1 compared with the quinoline oxidoreductase from Pseudomonas putida 86. Biol. Chem. Hoppe-Seyler 372:1081–1088. 17. Reeson, S. 1996. Heavy fuel oil: acceptable? Available? Affordable? Energy World 235:9–11. 18. Schwarz, G., E. Senghas, A. Erben, B. Schafer, F. Lingens, and H. Hoke. 1988. Microbial metabolism of quinoline and related compounds. 1. Isolation and characterization of quinoline-degrading bacteria. Syst. Appl. Microbiol. 10:185–190. 19. Shukla, O. P. 1986. Microbial transformation of quinoline by a Pseudomonas sp. Appl. Environ. Microbiol. 51:1332–1342. 20. Truex, M. J., F. J. Brochman, D. L. Johnstone, and J. K. Fredrickson. 1992. Effect of starvation on induction of quinoline degradation for a subsurface bacterium in a continuous flow column. Appl. Environ. Microbiol. 58:2386– 2392. 21. Ulonska, A., W. D. Deckwer, and V. Hecht. 1995. Degradation of quinoline by immobilized Comamonas acidovorans in a three-phase airlift reactor. Biotechnol. Bioeng. 46:80–87. 22. Watson, G. K., and R. B. Cain. 1975. Microbial metabolism of the pyridine ring. Biochem. J. 146:157–172.