Environ. Sci. Technol. 1996, 30, 1180-1184
Biodehalogenation: Oxidative and Hydrolytic Pathways in the Transformations of Acetonitrile, Chloroacetonitrile, Chloroacetic Acid, and Chloroacetamide by Methylosinus trichosporium OB-3b CHARLES E. CASTRO,* STEPHEN K. O’SHEA, WEN WANG, AND ELEANOR W. BARTNICKI The Environmental Toxicology Graduate Program, University of California, Riverside, California 92521, and CEC Consulting, 1090 Madison Place, Laguna Beach, California 92651
Resting cell suspensions of the soil methylotroph Methylosinus trichosporium OB-3b rapidly metabolize acetonitrile (ACN) and chloroacetonitrile (CCN) to bicarbonate. Formate, a natural substrate for this organism, is a major metabolite in both cases. The first step in these processes, established with 13Clabeled substrates and NMR analysis, is an oxygen insertion into the C-H bond to yield the corresponding cyanohydrins. With acetonitrile, the loss of HCN from the initially formed cyanohydrin (HOCH2CN) produces formaldehyde. The latter is rapidly converted to formate and CO2. With chloroacetonitrile, the first formed cyanohydrin [HOCHCl(CN)] may lose HCN to produce formyl chloride or lose HCl to yield formyl nitrile. The generation of CO, in addition to formate, implicates formyl chloride and/or formyl nitrile as intermediates. With neither ACN nor CCN is the nitrile moiety directly attacked or hydrolyzed. In particular, the nitrile hydrolysis products from CCN, chloroacetamide (CAM), and chloroacetic acid (CA) are not detected. In addition, hydrolysis of the C-Cl moiety of CCN does not occur. In contrast, incubation of CAM and CA with this organism does result in the corresponding hydroxy amide and acid directly. The slower dehalogenation of the amide and acid entail a direct microbial hydroxylation of the C-Cl bond. Based upon this study and related work with methylene and ethylene halides, it may be concluded that in the series ClCH2X (X ) Cl, Br, CN) oxygen insertion is the initial step in dehalogenation by this organism. However, in the series ClCH2Y (Y ) CH2Cl, CO2H, CONH2), the dehalogenation is the result of a direct hydrolysis of the C-Cl bond. This organism also transforms
1180
9
ENVIRONMENTAL SCIENCE & TECHNOLOGY / VOL. 30, NO. 4, 1996
cyanide ion, but incubations with K13CN resulted in no detectable products.
Introduction Acetonitrile is the simplest member of the alkyl cyanide class. It is a broadly used solvent because of its capacity to dissolve both organic and inorganic substances. It is also poisonous (1). The potential conversion and decontamination of this solvent by soil microorganisms is not well charted, but a general path for the bioconversion of nitriles is thought to entail hydrolysis of the nitrile to the corresponding amide followed by hydrolysis of the latter to the corresponding acid (2) (eq 1): O RCN
RCNH2
O RCOH
(1)
This pathway has been defined for the metabolism of acetonitrile by Pseudomonas sp. (3). The genes responsible for nitrile hydratase from P. chloroaphis have recently been cloned (4). The enzyme hydrolyzes nitriles to the corresponding amides. However, the detection of thiocyanate (SCN-) from dogs treated with acetonitrile led Lang to postulate, a century ago (5), an oxygen insertion into the CH3 moiety to yield formaldehyde cyanohydrin, followed by decomposition of the latter to CN- and metabolism to SCN-. The soil methylotroph Methylosinus trichosporium OB3b possesses a high concentration of the enzyme methane monooxygenase (6). The soluble form of this enzyme s-MMO has a powerful capacity for oxidation by oxygen insertion into carbon-hydrogen bonds and multiple bonds (7). A particulate form of the enzyme (p-MMO) can also be present in this organism, and it is favored (over s-MMO) by high copper content in the growth media. The organism we have used was grown in the absence or in the presence of low copper concentrations. We have recently described the rates and pathways for the rapid metabolism of polyhalomethanes by this organism (8). All of these substances, CH3X, CH2X2, and CHX3 (X ) Cl, Br), are initially transformed by an oxygen insertion into the C-H bond. These oxidatived conversions were completely inhibited by acetylene establishing methane monooxygenase as the responsible enzyme. We wished to extend this work to encompass the simple alkyl cyanides, because this class of substrate contains two potential sites of initial attack: The C-H bond adjacent to the nitrile linkage and the multibonded CtN moiety itself. The two substrates chosen for scrutiny are acetonitrile and chloroacetonitrile. The latter represents a structure similar to the methylene halides (CH2X2), except that one of the halogens is replaced by cyanide. The general hydrolytic pathway in eq 1 would produce chloroacetamide and chloroacetic acid from chloroacetonitrile. Thus, these chlorinated species have also been examined. In addition, all three haloorganics may be used as 13C-NMR reactivity probes for the environment (9). Hence, we wished to compare the rates of their * Address correspondence to this author at CEC Consulting.
0013-936X/96/0930-1180$12.00/0
1996 American Chemical Society
FIGURE 1. Time course for chloride ion release from chloroacetonitrile by Methylosinus trichosporium OB-3b, pH 7.4, 0.1 M phosphate buffer.
transformation and the nature of the pathways with which they are transformed by this soil methylotroph.
Experimental Methods Materials. [1,2-13C]-Chloroacetic acid was prepared by the slow decomposition of sulfuryl chloride in [1,2-13C]-acetic acid in the presence of phosphorous pentachloride. The amide was prepared by aminolysis of the methyl ester. The latter was obtained by converting the chloro acid first to chloroacetyl chloride and then to the ester. [1,2-13C]Chloroacetonitrile was prepared by dehydrating the amide with phosphorus pentoxide. The details of these syntheses have recently been reported (9, 10). [1,2-13C]-Acetonitrile and [13C]potassium cyanide were purchased from Cambridge Isotopes, Andover, MA. They were used without purification. The 13C-NMR of both substances accorded with the literature. The organism M. trichosporium OB-3b was grown at 30 °C on methane in a low copper medium in a 5-L New Brunswick BioFlow II fermentor. A fixed ratio of methane to air 1:4 at 250 mL/min and a stirring speed of 300 rpm were employed. Centrifuged cells were suspended in phosphate buffer (0.1 M, pH 7.4) and recentrifuged thrice before reaction (11). Cells grown in the absence of any copper did not show an enhanced rate of substrate metabolism. Methods. All transformations were conducted with resting cells at a concentration of 0.1 g wet wt /mL in phosphate buffer, pH 7.4, in the manner previously described (8). Substrates were 10-3 M originally. As before, 125-mL Erlenmeyer flasks filled with air and 20 mL of cells were employed. The flasks were sealed with a serum-capped stopcock. The total volume of each flask was ∼160 mL. Approximately 0.5-1-mL samples were removed at the desired time for 13C-NMR analysis. These samples were transferred to (5 mm i.d.) stoppled NMR tubes. The cell suspension in the NMR tubes was immersed in boiling water for 1 min, and the sample was held at 5 °C until the NMR acquisition could begin. For the lock, 100 µL of D2O/mL of cell suspension was added. NMR Analysis. The spectrum of naturally abundant (1.1%) 13C-labeled compounds at 0.001 M is invisible under our conditions because of the low sensitivity of the 13C
FIGURE 2. Time course for chloride ion release from chloroacetamide (CAM) and chloroacetic acid (CA) by Methylosinus trichosporium OB-3b, pH 7.4, 0.1 M phosphate buffer.
nucleus to NMR. At higher concentrations, however, spectra can be observed. The dimethylformamide (DMF) resonances in Figure 4 illustrate this. We have employed 13C-labeled substrates in this work, and this has two main effects upon the spectra. It greatly enhances the detectability of these substances at low concentration. Thus, only labeled compounds, the substrate and products derived from it, are seen. (An exception is DMF when it is used as solvent for the substrate.) Because the substrates each contain two labeled carbons, there is another unique feature to these spectra. The resonance of each carbon is effected (split) by its adjoining 13C-labeled neighbor such that it appears as a doublet. Thus, in our work all doublets represent two carbon species derived from the substrate. The single carbon metabolites (carbon monoxide, bicarbonate, formate, cyanide) appear as singlets. The low concentrations employed still require a relatively high number of acquisitions to increase the signal to noise ratio in the final spectrum. An excellent introduction to 13CNMR and spectral analysis (12) is available. The 13C-NMR spectra were obtained with a General Electric QE-300 NMR spectrometer in the manner previously described. Generally an overnight (16 h) acquisition was employed (8, 11, 13). Kinetics. For the organic halides, chloride ion release rates were monitored by direct potentiometry as previously described (8, 11, 13).
Results Kinetics. The rate of chloride ion release from chloroacetonitrile (CCN) by M. trichosporium OB-3b is shown in Figure 1. The rate of conversion varied by 20% (( 10%) from batch to batch of cells. As with the polyhalomethanes (8), the rate of conversion fell off sharply after the cells were held for 1 week. The half-life at this cell density,
VOL. 30, NO. 4, 1996 / ENVIRONMENTAL SCIENCE & TECHNOLOGY
9
1181
FIGURE 3. (a) 13C-NMR spectrum of the incubation of Methylosinus trichosporium OB-3b (0.1 g/mL) with [1,2-13C]acetonitrile (1 × 10-3 M) at 0.5 h, pH 7.4, 0.1 M phosphate buffer. (b) Sample at 24 h.
estimated from initial slopes, is 11 min. This is very close to that observed for the methyl halides under these conditions (8). In contrast to this fast conversion, the corresponding amide chloroacetamide (CAM) and chloroacetic acid (CA) under identical conditions are slow to react, and reactions often do not go to completion (Figure 2). Half-lives estimated from initial slopes for these latter are ∼3.4 h. We had previously noted that CA was inert or very slow to react with this organism. Indeed it is the end metabolite from the partial conversion of ethylene dichloride by this organism (13). The kinetics for acetonitrile and cyanide transformations were not monitored, but some idea of the conversions can be obtained from the NMR data presented below. Cyanide ion is completely consumed in 24 h. Acetonitrile is more than 62% converted to the cyanohydrin in 0.5 h. Reactions Pathways. The NMR analyte contains the entire incubation mixture. Consequently, all 13C-labeled substrates are observed directly, and there is no doubt they are products derived from the parent substrate. Figure 3a shows the 13C-NMR spectrum of the acetonitrile incubation at 0.5 h. The resonances for starting nitrile are at δ 2 (CH3) and 118 (CN). The peak at 112.5 is a machine center spike. Clearly a large new resonance is observed at δ 47 and a close examination of the nitrile resonance at 118 indicates an overlay of two doublets so that the δ 47 doublet is also coupled to it. At 24 h (Figure 3b), the methyl doublet for acetonitrile is nearly gone. The δ 47 doublet has increased and a new singlet at δ 172 corresponding to formic acid has appeared. The overlap of the nitrile doublets in the CN range (δ ∼118) is clearer in this spectrum. The unknown doublet at δ 47 was deduced to result from the
1182
9
ENVIRONMENTAL SCIENCE & TECHNOLOGY / VOL. 30, NO. 4, 1996
cyanohydrin of formaldehyde. This was established by synthesizing the cyanohydrin in situ via the reaction of excess formalin with potassium cyanide in water (eq 2). H2O
H2CdO + CN- 798 HOCH2CN + OH-
(2)
The only new resonance in this solution is exactly at δ 47. Note: H2CdO, H2C(OH)2, and CN- exhibit resonances at δ 181, 62, and 164, respectively. Because of excess formaldehyde and the generation of cyanohydrin, cyanide was not observed in this solution. Thus, the path of metabolism of acetonitrile by OB-3b entails an initial oxygen insertion into the C-H bond to generate formaldehyde cyanohydrin. The latter loses HCN to yield formaldehyde, which is further oxidized to formate and carbon dioxide (bicarbonate) by this organism (eq 3). The conversion to formaldehyde and CH3CN OB-3b
(HO–CH2–CN)
HCN + H2CO (3)
I HCO3–
HCO2H
formate to CO2 are a part of the natural metabolism of methane by this organism. In this case, the intermediate oxygen insertion product, the cyanohydrin, has a sufficient lifetime such that it is directly observed by NMR. A close examination of Figure 3b also shows small but discernible resonances for HCO3- (δ 161) and CN- (δ 164). Chloroacetonitrile (CCN) is a more complicated substrate and the associated 13C-NMR spectra taken at 5 min, 15 min, and 30 min, and 24 h are shown in Figure 4. For this substance as well as the CAM and CA substrates, the
incubations were begun by injecting the cell suspension with the requisite amount of a 1 M solution of the substrate in dimethyl formamide (DMF). The DMF resonances appear as singlets at δ 165, 38, and 32 in the spectra. The 5-min spectrum, in addition to CCN (ClCH2 at δ 25, d and CN at δ 117, d), exhibits a small but clear doublet at δ 17 along with two singlets corresponding to formate (δ 172) and cyanide (δ 164). At 15 min, the δ 17, d resonance is barely discernible. Cyanide and formate have increased. At 30 min, formate is the dominant product, but cyanide and traces of bicarbonate (δ 161) can be seen. At 24 h, carbon monoxide (δ 182), formate, cyanide, and bicarbonate resonances along with carbon dioxide (δ 123) are visible. In addition, in the DMF methyl region, there are three new singlets at δ 33, 35, and 23. These are unassigned, and we presume that they may represent utilization of the CCN carbon source for the building of other cellular constituents containing 13C-labeled methyl moieties. From these spectra and the results with acetonitrile, chloroacetamide, and chloroacetic acid (shown below), we deduce the following pathway for the metabolism of chloroacetonitrile (eq 4): OH Cl–CH2–CN
OB-3b
O
-HCl
(Cl–CH–CN) -HCN II O (b) (a) (H–C–Cl)
(H–C–CN)
H2O
H2O
HCO2H + HCN
(4)
CO + HCN
CO + HCl
HCO2H + HCl
We know the initial reaction is not a direct microbiological hydroxylation of the C-Cl bond. If this were to occur, the cyanohydrin of formaldehyde should be generated. It has a finite lifetime under these conditions as we have noted above, but no resonances in any of the spectra from CCN exhibits the δ 47 d for HOCH2CN. Moreover, no resonance corresponding to chloroacetamide or chloroacetic acid are discernible. Thus, the CN moiety is not hydrolyzed in this case as it is not with acetonitrile. We assign the initial doublet to the intermediate II (eq 4). This, both halohydrin and cyanohydrin, would be expected to decompose by two pathways (4a and b). Path 4b with the loss of HCN would generate formyl chloride. This substance is unstable and is established to decompose to formate by hydrolysis and to CO by loss of HCl (14). We have inferred the same intermediate in the conversion of the methylene halides to CO and formate by this organism (8). Path a, loss of HCl from II, would produce formyl nitrile. As far as we are aware, this substance is unknown. Like formyl chloride, we presume that the substance would hydrolyze to formate and HCN or lose HCN to produce CO. We cannot eliminate this path for the decomposition of II. With chloroacetic acid (CA) the 5- and 24-h spectra are essentially the same. Both show a large amount of unreacted CA (ClCH2 δ 44 d and CO2H δ 176 d) (Figure 5b). The hydrolysis product glycolic acid is observable (HOCH2 at δ 62 d) as the first main product, and a trace of bicarbonate is observable. Thus, though poorly converted by OB-3b, the organism does directly hydrolyze the CCl bond of CA (eq 5): OB-3b
H2O + ClCH2CO2H 98 HOCH2CO2H + Cl- + H+ (5)
FIGURE 4. (a) 13C-NMR spectrum of the incubation of Methylosinus trichosporium OB-3b (0.1 g/mL) with [1,2-13C]chloroacetonitrile (10-3 M) at 5 min, pH 7.4, 0.1 M phosphate buffer. DMF resonances at 32, 37, and 166 δ. (b-d) Samples at 15 min, 30 min, and 1 day, respectively.
Note: Blank incubations with CA in buffer, but without cells, exhibit no hydrolysis in 2 weeks. Like CA, chloroacetamide (CAM) undergoes a direct microbiological hydrolysis to hydroxyacetamide (HOCH2 at δ 62 d, CONH at δ 180 d) though a greater conversion occurs with CAM than with CA (Figure 5a). A greater conversion to bicarbonate is also apparent. An apparent 20% conversion to the hydroxyl amide was evident in 3 h. Cyanide. In an effort to further understand the fate of cyanide, a few incubations of OB-3b with K13CN were undertaken under the conditions employed for other substrates. After 2 h, the CN- resonance (δ 164 s) had vanished, but no 13C-labeled resonances were observable. We presume that the one carbon products, not NOE enhanced, are too dilute to observe.
Discussion The path of transformation of acetonitrile and chloroacetonitrile by OB-3b observed in this work follows that previously established for the methyl halides and polyhalomethanes (8). A general formulation of these reactions allows a correlation of the response of this organism to a defined set of structural types that are of environmental consequence. Thus, in the sequence, CH3X (X ) Cl, Br, CN) and in the series XCH2Y (Y ) Br, Cl; X ) Cl, Br, CN), the first step is an oxygen insertion into the C-H bond (eqs 6 and 7). The resulting halohydrins or cyanohydrins are
VOL. 30, NO. 4, 1996 / ENVIRONMENTAL SCIENCE & TECHNOLOGY
9
1183
FIGURE 5. (a) 13C-NMR spectrum of the incubation of Methylosinus trichosporium OB-3b (0.1 g/mL) with [1,2-13C]chloroacetamide (10-3 M) at 24 h, pH 7.4, 0.1 M phosphate buffer. (b) Same conditions with [1,2-13C]chloroacetic acid as substrate. CH3X
OB-3b
HO–CH2–X
(6)
OH YCH2X
OB-3b
Y–CH–X
(7)
unstable chemically and decompose in water by loss of HX or HY as outlined above. With the methyl halides and methylene halides, the first produced halohydrins (eqs 6 and 7) could be reasonably inferred but not directly observed. With the nitriles, however, the cyanohydrins can be directly observed by NMR. In addition to the oxygen insertions typical of this organism, another transformation, the hydrolysis of C-Cl bonds, can be a general means of dehalogenation by OB3b. That is, the direct hydrolysis of chloroacetic acid and chloroacetamide follows the same path previously observed with ethylene dichloride. Thus in the series XCH2Z where X ) Cl and Z ) CH2Cl, CO2H, CONH2, the dehalogenation proceeds via hydrolysis (eq 8): OB-3b
XCH2Z 98 HOCH2Z
(8)
The dehalogenation of these substances and the oxidation of acetonitrile by OB-3b emphasizes the important role soil methylotrophs may play in transforming xenobiotics in the environment. The complete transformation of cyanide is particularly astonishing. However, a direct extrapolation of these results with a single organism to the environment is difficult because of the large number of unknown factors in any environmental sample. Other organisms or abiotic processes may dominate. In essence, the nature and pathway of metabolism of these substrates by OB-3b, and perhaps other soil methylotrophs, will parallel our findings. The rates of conversion would be expected to be slower (15) because of a presumably lower cell density and a lower concentration of substrate at any given site. As an example of diversity, the hydrolysis of the
1184
9
ENVIRONMENTAL SCIENCE & TECHNOLOGY / VOL. 30, NO. 4, 1996
nitrile and amide linkages of CCN and CAM, not observed with OB-3b, readily occurs in a variety of environmental samples (10).
Acknowledgments We thank the Kearney Foundation for Soil Sciences for partial support of this work.
Literature Cited (1) Merck Index, 10th ed.; Windholtz, E., Budavari, S., Blumett, R. F., Otterbein, E. S., Eds.; Merck & Co.: Rahway, NJ, 1983. (2) Ahmed, A. E.; Farooqui, M. Y. H.; Trieff, N. M. Nitriles. In Bioactivation of Foreign Compounds; Anders, M. W., Eds.; Academic: New York, 1985; Chapter 17, pp 485-518. (3) Firmin, J. L.; Gray, D. O. Biochem. J. 1976, 158, 223-229. (4) Nishiyama, M.; Horinouchi, S.; Kobyashi, M.; Nagasawa, T.; Kamada, H.; Beppu, T. J. Bacteriol. 1991, 173, 2465-2472. (5) Lang, S. Arch. Exp. Pathol. Pharmakol. 1894, 34, 247-248. (6) Fox, B. G.; Froland, W. A.; Jollie, D. R.; Lipscomb, J. D. Methods Enzymol. 1990, 188, 191-202. (7) Green, J.; Dalton, H. J. Biol. Chem. 1989, 264, 17698-17703. (8) Bartnicki, E. W.; Castro, C. E. Environ. Toxicol. Chem. 1994, 13, 241-245. (9) Castro, C. E.; O’Shea, S. K.; Bartnicki, E. W. Environ. Sci. Technol. 1995, 29, 2154-2156. (10) Castro, C. E.; O’Shea, S. K.; Wang, W.; Bartnicki, E. W. Environ Sci. Technol. 1996, 30, 1185-1191. (11) Riebeth, D. M.; Belser, N. O.; Castro, C. E. Environ. Toxicol. Chem. 1992, 11, 497-501. (12) Silverstein, R M.; Bassler, G. C.; Morrill, T. C. Spectrometric Identification of Organic Compounds; John Wiley & Sons: New York, 1981. (13) Castro, C. E.; Riebeth, D. M.; Belser, N. O. Environ. Toxicol. Chem. 1992, 11, 749-755. (14) Krauskopf, K. B.; Rollefson, G. K. J. Am. Chem. Soc. 1934, 56, 2542-2548. (15) Castro, C. E. Environ. Toxicol. Chem. 1993, 12, 1609-1618.
Received for review June 15, 1995. Revised manuscript received November 2, 1995. Accepted November 27, 1995.X ES950422M X
Abstract published in Advance ACS Abstracts, February 15, 1996.
