Nitrite Elimination and Hydrolytic Ring Cleavage in 2,4,6-Trinitrophenol (Picric Acid) Degradation

Abstract

Two hydrogenation reactions in the initial steps of degradation of 2,4,6-trinitrophenol produce the dihydride Meisenheimer complex of 2,4,6-trinitrophenol. The npdH gene (contained in the npd gene cluster of the 2,4,6-trinitrophenol-degrading strain Rhodococcus opacus HL PM-1) was shown here to encode a tautomerase, catalyzing a proton shift between the aci-nitro and the nitro forms of the dihydride Meisenheimer complex of 2,4,6-trinitrophenol. An enzyme (which eliminated nitrite from the aci-nitro form but not the nitro form of the dihydride complex of 2,4,6-trinitrophenol) was purified from the 2,4,6-trinitrophenol-degrading strain Nocardioides simplex FJ2-1A. The product of nitrite release was the hydride Meisenheimer complex of 2,4-dinitrophenol, which was hydrogenated to the dihydride Meisenheimer complex of 2,4-dinitrophenol by the hydride transferase I and the NADPH-dependent F₄₂₀ reductase from strain HL PM-1. At pH 7.5, the dihydride complex of 2,4-dinitrophenol is protonated to 2,4-dinitrocyclohexanone. A hydrolase was purified from strain FJ2-1A and shown to cleave 2,4-dinitrocyclohexanone hydrolytically to 4,6-dinitrohexanoate.

2,4,6-Trinitrophenol (TNP) and 2,4-dinitrophenol (DNP) are nitroaromatic compounds of versatile use in chemical synthesis. They occur as off-stream chemicals during the production of aniline, which is one of the most important starting materials in chemical synthesis. Furthermore, TNP and its salts have been used as explosives. Large quantities of TNP in waste streams of aniline production necessitate remediation. Several bacteria of the Actinomycetales family (notably of the genera Rhodococcus and Nocardioides) grow aerobically on TNP and/or DNP and utilize the compounds as sole nitrogen, carbon, and energy sources (3, 4, 10, 19). This capacity can be harnessed for bioremediation once we understand the underlying mechanisms of these processes.

It was previously established that two hydrogenations take place in the initial attack on TNP (6, 7, 8, 11) (Fig. 1A, panel 1). In Rhodococcus opacus HL PM-1, TNP is hydrogenated at the aromatic nucleus by the hydride transferase II (HTII) encoded by npdI (Fig. 1B) and the NADPH-dependent F₄₂₀ reductase (NDFR) encoded by npdG. The hydride Meisenheimer complex of TNP (H⁻-TNP) (Fig. 1A, panel 2) thereby formed is further hydrogenated by the hydride transferase I (HTI) encoded by npdC and the NDFR, producing the dihydride Meisenheimer complex of TNP (2H⁻-TNP) (panel 3a). In Nocardioides simplex FJ2-1A, the same reactions take place except that a single hydride transferase performs both hydrogenations (7).

FIG. 1.

(A) Upper degradation pathway of TNP. Panel 1, TNP; panel 2, H⁻-TNP; panel 3a, aci-nitro form of 2H⁻-TNP; panel 3b, nitro form of 2H⁻-TNP; panel 4, DNP; panel 5, H⁻-DNP; panel 6, 2,4-DNCH; panel 7, 4,6-DNH; HTII, accession number AAK38104; NDFR, AAK38102; HTI, AAK38097; tautomerase, AAK38103. (B) npd gene cluster of R. opacus HL PM-1 showing the proteins which have been functionally identified (see also reference 8).

More than a decade ago, 4,6-dinitrohexanoate was identified as a dead-end metabolite of TNP degradation resulting from two hydride transfers to TNP. This observation coincided with the hypothesis that 2H⁻-TNP was a dead-end metabolite of TNP degradation (11). Much later it was suggested that nitrite is eliminated from 2H⁻-TNP to produce the hydride Meisenheimer complex of DNP (H⁻-DNP) in N. simplex FJ2-1A (7). Hence, 2H⁻-TNP is a metabolite of productive TNP degradation. Detection of H⁻-DNP suggests that the pathways for TNP and DNP degradation converge, although indisputable identification of H⁻-DNP is still required to confirm the hypothesis. NpdH, a gene product of the npd gene cluster of R. opacus HL PM-1, was recently shown to convert 2H⁻-TNP to an unknown product X (Fig. 1A, panel 3b), which was preliminarily suggested to be a tautomer of protonated 2H⁻-TNP (8); this evoked the issue of which form of 2H⁻-TNP is the substrate for nitrite release.

To address these issues, three enzymes in the TNP degradation pathway were identified and the metabolites were confirmed. Hence, evidence is supplied for the convergent TNP and DNP (Fig. 1A, panel 4) degradation pathways, with H⁻-DNP (panel 5) as the first common metabolite. We have shown that four unusual catabolic reactions (three ring hydrogenations and a hydrolytic ring fission) take place in the upper TNP degradation pathway.

MATERIALS AND METHODS

Bacterial strains and growth conditions.

N. simplex FJ2-1A or R. opacus HL PM-1 was grown in cultures in conical flasks at 30°C in 50 mM phosphate buffer (pH 7.5) containing 0.7 mM picrate, 20 mM sodium acetate, 0.5 g of yeast extract liter⁻¹, 0.5 g of proteose peptone liter⁻¹, 0.5 g of Casamino Acids liter⁻¹, and mineral salts. Mineral salts without nitrogen contained 20 mg of Fe(III)-citrate liter⁻¹, 1 g of MgSO₄^·7 H₂O liter⁻¹, 50 mg of CaCl₂^·2 H₂O liter⁻¹, and 1 ml of SL 6 trace element solution (18). After consumption of 0.7 mM picrate, 0.35 mM picrate was added and the cells were harvested after growth for a further hour. Cultures were harvested by centrifugation immediately after decolorization of the medium. N. simplex FJ2-1A and R. opacus HL PM-1 were also grown in medium as described above with the addition of 1.8% (wt/vol) agar with 0.7 mM 2-nitrocyclohexanone (2-NCH) as the sole nitrogen source. Cells were frozen in liquid nitrogen and stored at −30°C.

Escherichia coli BL21(DE3) (pNTG11) expressing npdH, E. coli JM109 (pNTG6) expressing npdC, and E. coli TOP10 (pDMW10) expressing npdG (8) were grown at 37°C in Luria-Bertani medium containing 100 μg of ampicillin ml⁻¹. Overnight cultures were inoculated into Luria-Bertani medium and grown at 37°C to an optical density at 600 nm of 0.4. Cultures were induced for 4 h at 30°C with 1 mM IPTG (isopropyl-β-d-thiogalactopyranoside).

Preparation of cell extracts.

About 20 g of wet cells of N. simplex FJ2-1A or 2 g of wet cells of the E. coli strains was suspended in 50 mM Tris-HCl (pH 7.5) and lysed with a French press as described before (8). The protein concentration was determined with a dye reagent concentrate (protein assay; Bio-Rad, Munich, Germany) by the method of Bradford (5).

Enzyme purification.

npdH, npdC, and npdG from R. opacus HL PM-1 were expressed in E. coli BL21(DE3) (pNTG11), E. coli JM109 (pNTG6), and E. coli TOP10 (pDMW10), respectively, and the proteins were purified as His-tag fusion proteins by Ni-nitrilotriacetic acid affinity chromatography as described before (8). Imidazole was removed from the enzyme-containing fractions with pD10 desalting columns (Amersham Pharmacia, Freiburg, Germany). Samples were concentrated by ultrafiltration (Vivaspin 2; Vivascience AG, Hanover, Germany).

The nitrite-eliminating enzyme was purified from N. simplex FJ2-1A at 4°C by fast-performance liquid chromatography (LC) (LCC 500 controller, 500 pump, UV-1 monitor, REC-482 recorder, and FRAC autosampler; Pharmacia, Uppsala, Sweden). The cell extract (340 mg of protein) was passed through a Q Sepharose column (HP HR 16/10; Pharmacia) preequilibrated with basic buffer (50 mM Tris-HCl [pH 7.5]) at a flow rate of 1 ml min⁻¹. The activity was eluted from the column with a linear gradient of 0 to 1 M NaCl (200 ml) in basic buffer at 0.25 M NaCl. Ammonium sulfate (1 M) was added to the active fractions and applied to a Phenyl Superose column (HR10/10; Pharmacia) preequilibrated with the same buffer. Enzyme was eluted with a linear gradient (65 ml) of 1 to 0 M ammonium sulfate in basic buffer at 0.41 M (NH₄)₂SO₄ and a flow rate of 0.5 ml min⁻¹. Active fractions were applied to a gel filtration column (Superdex 200 Prep-Grade; Pharmacia) (1 by 30 cm), and the enzyme was eluted with basic buffer at a flow rate of 1 ml min⁻¹. The hydrolase was purified from N. simplex FJ2-1A as described above except that elution from the Q Sepharose column was at 0.33 M NaCl and from the Phenyl Superose HR column was at 0.54 M (NH₄)₂SO₄.

The molecular mass of native proteins was determined using a gel filtration calibration kit (Amersham Pharmacia). Purity and molecular mass of protein subunits were determined by matrix-assisted laser desorption ionization-time of flight (MALDI-TOF) with a G1025A Hewlett-Packard LD-TOF system (GSG Mess- und Analysegeräte Vertriebsgesellschaft mbH, Karlsruhe, Germany) and by sodium dodecyl sulfate-polyacrylamide gel electrophoresis (SDS-PAGE) with a 10% (vol/vol) polyacrylamide gel by the method of Laemmli (9) on a Mini-PROTEAN 3 electrophoresis cell (Bio-Rad).

Enzyme assays.

Enzyme assays were performed with a Cary 50 biospectrophotometer controlled by Cary WinUV Biopackage software (Varian, Mulgrave, Australia). Reactions with the tautomerase were performed as previously described for NpdH (8).

The activity of the nitrite-eliminating enzyme was assayed by measuring the increase in absorbance of H⁻-DNP at 450 nm or repeated recording of UV-visible spectra between 280 and 600 nm in 1-min cycles. The test was conducted with 50 mM Tris-HCl (pH 8.0) containing 0.1 mM 2H⁻-TNP and 6 μg of nitrite-eliminating enzyme. Specific activities were calculated by using the experimentally determined extinction coefficient of 944 M⁻¹ cm⁻¹ (H⁻-DNP at 450 nm; 20°C; pH 8.0).

The activity of HTI with respect to H⁻-DNP was measured as described before (8) except that 100 μM H⁻-DNP was the substrate instead of H⁻-TNP. The decrease in absorbance of H⁻-DNP was monitored at 450 nm. Specific activities were calculated by using the experimentally determined extinction coefficient as described above.

Enzyme activity of the hydrolase was detected by repeated recording of UV-visible spectra between 280 and 600 nm in 1-min intervals. The reaction was performed with 50 mM Tris-HCl (pH 8.0) or 50 mM phosphate buffer (pH 8.0) containing 4 μg of hydrolase and 100 μM 2-NCH (commercially available substrate) or 2,4-DNCH (catabolic substrate obtained by the turnover of H⁻-DNP by HTI and NDFR). Specific activities with 2-NCH as substrate were determined at 340 nm with an extinction coefficient of 3,100 M⁻¹ cm⁻¹ in 50 mM phosphate buffer (pH 8.0).

One unit of enzyme activity was defined as the amount of enzyme that converts 1 μmol of substrate per min.

Analytical methods.

Metabolites were detected and quantified at 210, 340, 390, or 420 nm by high-performance LC (HPLC) analysis (Chromeleon Chromatography Data Systems 4.38 [equipped with a UVD 170S/340S UV/Vis detector, a P 580 pump, and a Gina 50 autosampler; Dionex, Idstein, Germany]) with a Lichrospher 100 RP-18 column (Merck, Darmstadt, Germany) (250 by 4 mm; particle size, 5 μm). The mobile phase consisted of 20% (vol/vol) methanol and 80% (vol/vol) 50 mM phosphate buffer (pH 8.0). Proteins were removed from metabolites by filtration with pD10 desalting columns (Amersham Pharmacia) or with Amicon filters (Centricon YM-10; Millipore, Bedford, Mass.).

LC mass spectra were obtained by negative-mode electrospray ionization (ESI⁻) on an HP1100 HPLC system (Hewlett-Packard, Waldbronn, Germany) coupled to a VG Platform II Quadrupole mass spectrometer (Micromass, Manchester, United Kingdom). Samples were resolved on a C8 reversed-phase column (Gromsil 100; Grom, Herrenberg, Germany) (125 by 4.6 mm; particle size, 5 μm) or on a Lichrosorb 100 RP-18 column (Merck) (250 by 4 mm; particle size, 5 μm). The mobile phase consisted of 5 mM ammonium formate buffer (pH 8.0) or 20% MeOH plus 5 mM ammonium formate buffer (pH 8.0).

Nuclear magnetic resonance (NMR) spectra of the aci-nitro and nitro forms of 2H⁻-TNP (10 mM each) were recorded with an ARX 500 spectrometer (Bruker, Rheinstetten, Germany) at room temperature at a nominal frequency of 500.14 MHz (¹H) and 125.76 MHz (¹³C) for 15 s. Samples were dissolved in D₂O, and the aci-nitro and nitro forms were detected at pH 8.0. Chemical shifts (δ) are given in parts per million relative to tetramethylsilane as the internal standard.

Nitrite concentrations were determined spectrophotometrically by the method of Montgomery (14).

Amino acid sequencing and sequence analysis.

The amino-terminal end of the nitrite-eliminating enzyme or the hydrolase was automatically sequenced with a 476 protein sequencing system (Applied Biosystems, Foster City, Calif.). Database searches were performed with BLAST (1) and FASTA (17) software.

Chemicals.

2H⁻-TNP and H⁻-DNP were prepared according to the method of Severin et al. (22, 23) and stored at −20°C. All chemicals used were of the highest available purity and were purchased from Fluka (Taufkirchen, Germany), Merck, Roth (Karlsruhe, Germany), and Sigma-Aldrich (Taufkirchen, Germany).

RESULTS

Enzymatic proton-shift tautomerization of 2H⁻-TNP.

The aim was to identify the two tautomeric forms of 2H⁻-TNP and to show whether or not NpdH was involved in the tautomeric reaction. Under alkaline conditions (pH ≥ 8), 2H⁻-TNP exists as a double-charged anion (the so-called aci-nitro form) (Fig. 1, panel 3a). HPLC analysis also revealed the tautomeric nitro form (panel 3b): the aci-nitro form was detected with a retention time of 1.86 min, whereas the nitro form was eluted after 2.28 min.

The structures of the aci-nitro and nitro forms of 2H⁻-TNP were identified by ¹H NMR and ¹³C NMR spectroscopy. The NMR data of the aci-nitro form corresponded to the data of authentic 2H⁻-TNP (as described before by Ebert et al.) (7). The ¹H NMR spectrum (500 MHz, D₂O) of the nitro form showed a heptuplet at δ 3.74 ppm and two doublets of doublets at δ 3.61 and 3.52 ppm (Table 1). This resonance pattern was similar to that of the nitro form of protonated 2H⁻-TNT (24) and can be analyzed as an (AB)₂X five-spin system. The signals at δ 3.61 and 3.52 ppm correspond to the double set of diastereotopic methylene protons H^A and H^B at C-3 and C-5. The geminal-coupling constant ²J (H^A, H^B) of 11.73 Hz demonstrated the presence of a doublet of doublets due to the symmetric equivalence of the diastereotopic methylene protons. The vicinal-coupling constants (4.38 and 6.53 Hz) are assigned to the methylene spectrum between H^A and H^B on one side and the H^X proton on the other side. The H^X resonance displayed a heptuplet at δ = 3.74 ppm.

TABLE 1.

¹H NMR (500 MHz, D₂O) data of the nitro form of 2H⁻-TNP

The tautomers were in equilibrium, with an estimated peak area ratio of 20:80 (aci-nitro-nitro form; pH 8) calculated from the HPLC peak areas at A₃₉₀ (the UV maximum for either form). The extinction coefficient for the aci-nitro form at pH 8.0 was 11,864 M⁻¹ cm⁻¹. We could not calculate the extinction coefficient of the nitro form, since we did not produce it in a pure form. Therefore, we estimated the relative ratios from the peak areas for both forms detected by HPLC.

As the trisodium salt of chemically synthesized 2H⁻-TNP is dissolved in buffer, the complex is instantaneously protonated to the aci-nitro form, which then slowly isomerizes to the nitro form (22, 23). When the solution was allowed to stand at room temperature, the time needed to reach the equilibration ratio of 20:80 (aci-nitro-nitro form) was approximately 1 h at pH 7 whereas in the presence of NpdH it took approximately 0.1 min. At pH 8, the time required to reach the equilibration ratio was approximately 5 h whereas in the presence of NpdH it was approximately 1 min. Hence, the time period required for reaching the equilibrium ratio was dependent on the pH. The tautomerase catalyzed the proton-shift tautomerization, accelerating the rate 300- to 600-fold. The tautomerase is thus responsible for the proton-shift tautomerism between the two tautomeric forms.

Purification and molecular characterization of the nitrite-eliminating enzyme.

Nitrite-eliminating activity was detected in the crude extracts of both R. opacus HL PM-1 and N. simplex FJ2-1A, releasing stoichiometric amounts of nitrite from 2H⁻-TNP. The enzyme was purified (as described in Materials and Methods and summarized in Table 2) from N. simplex FJ2-1A due to the easier lysis of this strain compared to the lysis of strain HL PM-1. The specific activity of the purified protein was 57 U mg⁻¹. SDS-PAGE results indicated a purity of >95% and an estimated molecular mass of 35.3 kDa. MALDI-TOF measurements gave a signal at m/z 30.50 kDa. The molecular mass of the purified nitrite-eliminating enzyme (as determined by gel filtration) was estimated to be 42 kDa. Hence, the protein is a monomer. N-terminal amino acid sequencing of the purified enzyme revealed the sequence M K N L E L A Y V G L (G) V H E X L V Y Y A A (Q/T) (H) L D (L) Y R (uncertain amino acids are in parentheses). Comparisons to sequences in databases (GenBank and National Center for Biotechnology Information) identified no similarity to any known protein sequences.

TABLE 2.

Purification of the nitrite-eliminating enzyme from N. simplex FJ2-IA^a

Purification step	Vol (ml)	Total protein (mg)	Sp act (U/mg)	Total activity (U)	Recovery (%)	Purification (severalfold)
Crude extract	35.0	340.0	1.3	432.7	100.0	1.0
Q Sepharose	9.0	37.0	4.1	151.0	35.0	3.2
Phenyl Superose	1.0	1.6	40.4	65.0	15.0	31.8
Gel filtration	0.4	0.2	57.0	14.3	3.3	44.9

^aCells were harvested in the exponential-growth phase. The wet-cell mass was 12 g.

Conversion of the aci-nitro form of 2H⁻-TNP to H⁻-DNP by the nitrite-eliminating enzyme.

To unambiguously identify the substrate and the product of nitrite elimination, the nitrite-eliminating enzyme was incubated with 2H⁻-TNP and the reaction was recorded spectrophotometrically. The enzymatic turnover of 2H⁻-TNP was similar to that reported before (7). Isosbestic points at 338 and 413 nm indicated that the initial reaction mixture contained the aci-nitro form of 2H⁻-TNP (pH 8.0) and that the reaction product was H⁻-DNP. During the turnover, stoichiometric amounts of nitrite were released. To identify H⁻-DNP, an authentic standard was prepared and identified by ¹H NMR (500 MHz, D₂O) [δ (H3, H3′) = 3.83 ppm (s); d (H5) = 7.49 ppm (d); d (H6) = 5.89 ppm; J (H5, H6) = 10.2 Hz]. The results corresponded to the data presented by Behrend and Heesche-Wagner (3). HPLC analysis of H⁻-DNP indicated a retention time of 1.9 min and the same UV-visible spectrum for both the standard and the product of the enzymatic conversion. Further evidence was given by HPLC-mass spectrometry (MS) analysis, revealing a single peak at an ion mass of m/z 185 which corresponded to the molecular ion [M^·]⁻ of H⁻-DNP.

Since 2H⁻-TNP exists as two tautomeric structures, we investigated whether the nitrite-eliminating enzyme possessed selectivity towards the aci-nitro or the nitro form. To show this, the final equilibrium mixture of 20:80 (aci-nitro-nitro form) served as a substrate to begin the experiment (Fig. 2A). Addition of the nitrite-eliminating enzyme showed that only the aci-nitro form (and not the nitro form) was converted to H⁻-DNP (Fig. 2B). After removal of the nitrite-eliminating enzyme and addition of the tautomerase to solution B, the aci-nitro form developed rapidly (Fig. 2C). Removal of the tautomerase followed by addition of the nitrite-eliminating enzyme to solution C demonstrated once more that only the aci-nitro form of 2H⁻-TNP was converted to H⁻-DNP (Fig. 2D). Hence, the nitrite-eliminating enzyme uses the aci-nitro form of 2H⁻-TNP as the only substrate.

FIG. 2.

Conversion of 2H⁻-TNP (Fig. 1, panel 3a) to H⁻-DNP (Fig. 1, structure 5) by the tautomerase and the nitrite-eliminating enzyme. The aci-nitro and nitro forms of 2H⁻-TNP and H⁻-DNP were detected at 390 nm. The HPLC peak areas were integrated for estimations of the percentages of each compound. For each step, the enzyme of the previous reaction was removed prior to HPLC analysis by filtration. Reactions were analyzed after 1 min. 2H⁻-TNP at pH 8 after equilibration, showing the equilibration ratio of 20:80 (aci-nitro-nitro forms) (A) after addition of the nitrite-eliminating enzyme to solution A (B), after addition of the tautomerase to solution B (C), and after renewed addition of the nitrite-eliminating enzyme to solution C (D). Black columns, aci-nitro form of 2H⁻-TNP; grey-shaded columns, nitro form of 2H⁻-TNP; white columns, H⁻-DNP.

HTI converts H⁻-DNP to 2,4-DNCH.

The HTI was previously shown to hydrogenate H⁻-TNP to 2H⁻-TNP (8). To show whether the analogous substance H⁻-DNP additionally served as a substrate, enzyme assays were performed with H⁻-DNP, F₄₂₀, NADPH, and HTI. Repetitive spectroscopic recording during the turnover of H⁻-DNP by the HTI of R. opacus HL PM-1 showed a decrease in absorbance at 441 and 306 nm. A concomitant increase in absorbance at 340 nm demonstrated the generation of a new product, and HPLC analysis revealed a metabolite with a retention time of 3.2 min. The corresponding UV-visible spectrum displayed absorbance maxima at 232 and 340 nm. For confirmatory identification, the new product was analyzed by coupled HPLC-ESI⁻-MS and the intense signal at m/z 187 was assigned to the molecular anion [M^·]⁻ of 2H⁻-DNP. At pH 7.5, 2H⁻-DNP is protonated to 2,4-DNCH (Fig. 1A, panel 6). The specific activity of the HTI for H⁻-DNP was 29.6 U mg⁻¹.

Purification and characterization of the hydrolase converting 2,4-DNCH to 4,6-DNH.

To identify the next product and enzyme in the pathway, the activity converting the protonated form of 2H⁻-DNP (i.e., 2,4-DNCH) was assayed for in crude extracts for subsequent enzyme purification. Since substrate amounts of 2,4-DNCH were unavailable by chemical or biochemical synthesis and the analogous compound 2-NCH was commercially obtainable, 2-NCH was used for further experiments. Both N. simplex FJ2-1A and R. opacus HL PM-1 grew on 2-NCH as the sole source of nitrogen. Crude extracts of both strains cultured with DNP or 2-NCH showed hydrolase activity for 2-NCH and biologically generated 2,4-DNCH. For identification, the enzyme was purified from N. simplex FJ2-1A, with 2-NCH used as the test substrate (see Materials and Methods and Table 3). SDS-PAGE showed a single polypeptide band at 15.3 kDa with a purity of > 98%. The specific activity for 2-NCH was 24.4 U mg⁻¹.

TABLE 3.

Purification of the hydrolase from N. simplex FJ2-IA^a

Purification step	Vol (ml)	Total protein (mg)	Sp act (U/mg)	Total activity (U)	Recovery (%)	Purification (severalfold)
Crude extract	35.0	340.2	0.4	129.3	100.0	1.0
Q Sepharose	9.0	65.0	1.9	125.0	96.0	5.1
Phenyl Superose	3.0	4.3	9.5	28.5	22.0	25.0
Gel filtration	0.8	0.5	24.4	12.2	9.0	64.1

^aThe hydrolase was purified from the same cell extract as the nitrite-eliminating enzyme (see Table 2).

The molecular mass of the hydrolase was determined by MALDI-TOF measurements, giving a signal at m/z 16.989 kDa. Since the molecular mass of the purified enzyme estimated from gel filtration was 63 kDa, it can be assumed that the enzyme consists of four identical subunits. N-terminal amino acid sequencing of the purified protein by automated Edman degradation revealed the sequence M R K F W (H) V G I N V T D M D K S I E F Y R K V G F D V (S) Q (S) K (uncertain amino acids are in parentheses). A FASTA (17) search assigned a sequence identity of 78% to the product of orfF, which encodes a putative lyase in R. opacus HL PM-1 (accession number AAK38100) (21, 25). Hence, we propose to rename orfF as npdF.

The hydrolase was specific for 2,4-DNCH and 2-NCH and showed no activity for other metabolites (such as TNP, H⁻-TNP, 2H⁻-TNP, and H⁻-DNP) of the catabolic pathway. Repeated UV-visible spectroscopic recording displayed a decrease in absorbance at 340 nm, indicating the disappearance of 2,4-DNCH. HPLC analysis revealed a new product with a retention time of 3.4 min. The corresponding UV-visible spectrum displayed maximum absorbance at 203 nm and a band of low intensity at 255 nm. The data were in agreement with the properties of 4,6-dinitrohexanoate (4,6-DNH) described by Lenke and Knackmuss (11). Further confirmation of the structure was obtained by coupled HPLC-ESI⁻-MS analysis. A signal at m/z 205 corresponded to the molecular anion [M^·]⁻ of 4,6-DNH.

Chemical and enzymatic hydrolysis of 2,4-DNCH.

To investigate chemical versus enzymatic hydrolysis of 2,4-DNCH (Fig. 1A, panel 6), UV-visible spectra of the reaction were compared under basic and acidic conditions. Hydrolysis of 2,4-DNCH (panel 6) to 4,6-DNH (panel 7) was shown to be reversible, and the equilibrium of the two forms was pH dependent. In alkaline solution (pH > 8) the reaction mixture contained only 2,4-DNCH, whereas 4,6-DNH predominated under acidic conditions (pH < 5). Figure 3 shows the gradual spontaneous formation of 4,6-DNH from 2,4-DNCH at pH 7.5. The extinction coefficient of 4,6-DNH was not determined, since the compound could not be prepared in a pure form; therefore, the HPLC peak areas at A₂₁₀ were calculated. In the presence of 2 μg of hydrolase, ring cleavage accelerated, showing a 15-fold increase in the peak areas corresponding to 4,6-DNH after 10 min. This shows that the hydrolase converts 2,4-DNCH to 4,6-DNH. When 4,6-DNH was incubated at pH 7.5, HPLC analysis after 30 min revealed its chemical instability as described previously (11).

FIG. 3.

Increases in levels of 4,6-DNH during chemical or enzymatic ring fission of 2,4-DNCH at pH 7.5. Peak areas were integrated from the HPLC retention peaks at 210 nm at the given times.

DISCUSSION

It has been known for more than 10 years that the polynitroaromatic chemical TNP undergoes ring reduction in R. opacus HL PM-1 and that H⁻-TNP is the first metabolite (11, 20). More recently, it became evident that related bacteria like Nocardioides spp. have the same capacity and that they use the ring reduction mechanism a second time; the resulting product is 2H⁻-TNP, which serves as the substrate for nitrite release to form H⁻-DNP (3, 7). In this study, we have identified the enzymes and metabolites of TNP degradation up to ring cleavage, completing depiction of the upper TNP degradation pathway.

The present results confirm the previous suggestion that NpdH from R. opacus HL PM-1 is a tautomerase catalyzing equilibration between the nitro and the aci-nitro form of 2H⁻-TNP (8). Since nitrite elimination by the nitrite-eliminating enzyme was observed with the aci-nitro form only, this suggests that the tautomerase might be responsible for circumventing accumulation of the metabolically inert nitro form. Because tautomerization is also spontaneous, we believe that the rate of TNP degradation can be reduced in an npdH deletion mutant but not halted.

A similar pH-dependent tautomerization of the protonated dihydride Meisenheimer complex of TNT (2H⁻-TNT) has been observed in 2,4,6-trinitrotoluene (TNT) dead-end metabolism in R. opacus HL PM-1 (24). Since the strain accumulates 2H⁻-TNT and is assumed to be a dead-end product (24), the tautomerase may function in TNP catabolism only. Alternatively, the nitrite-eliminating enzyme may not be able to convert 2H⁻-TNT, suggesting that TNT is not utilized as a nitrogen source. Pak et al. also described the formation of tautomers of 2H⁻-TNT, although they suggested that this would lead to denitration and productive degradation (16).

The stoichiometric release and use of nitrite in TNP or DNP degradation has been demonstrated before (10, 11), and the product of nitrite release from 2H⁻-TNP was shown to be H⁻-DNP (3, 7). A second source of H⁻-DNP is hydrogenation of DNP, which is the first attack on DNP in DNP degradation: the HTII of R. opacus HL PM-1 and the hydride transferase of N. simplex FJ2-1A hydrogenate not only picric acid but also DNP to H⁻-DNP (6, 8). Thus, H⁻-DNP is a common metabolite of the two converging pathways of TNP and DNP. It seems reasonable that these two pathways might have coevolved, since TNP and DNP are structurally very similar and occur in combination in waste streams from aniline production.

Until recently the fate of H⁻-DNP was unclear. Behrend and Heesche-Wagner showed that NADPH is required for further degradation by Nocardioides sp. CB 22-2 (3) and suggested that a monooxygenolytic hydroxylation at the para position might take place, forming 2-nitrohydroquinone with release of nitrite. It was also hypothesized that H⁻-DNP could be converted to 2H⁻-DNP, finally giving rise to 4,6-DNH (12). The present results confirm this: the HTI of R. opacus HL PM-1 hydrogenated H⁻-DNP to 2H⁻-DNP, which is protonated to form 2,4-DNCH (Fig. 1, panel 6). The mechanism is analogous to the reduction of H⁻-TNP (Fig. 1). In the presence of NDFR and the coenzymes F₄₂₀ and NADPH, hydride is transferred to the C5 carbon atom of H⁻-DNP.

Chemical hydrolysis of 2-NCH under alkaline conditions has been described by several groups as a reverse Claisen condensation (2, 13). We showed that the hydrolytic ring opening is part of a pH-dependent equilibrium between 2,4-DNCH and 4,6-DNH. Under neutral conditions the reaction was very slow, however, such that a hydrolase should be required for rapid conversion to 4,6-DNH. In fact, crude extracts of R. opacus HL PM-1 or N. simplex FJ2-1A caused 4,6-DNH to disappear rapidly with the release of nitrite, although a product could not be identified (data not shown). This indicates that 4,6-DNH is a true metabolite in the biodegradation of TNP and DNP; this differs from previous findings in accordance with which 4,6-DNH was suggested to be chemically formed as a minor dead-end product by spontaneous hydrolysis (11).

In parallel with the proposed mechanism for DNP degradation described above and previously (3), Blasco et al. (4) suggested an alternative pathway of DNP degradation on the basis of detection of 3-nitroadipate in the supernatant of resting cells of Rhodococcus sp. strain RB1. Our observations support none of these suggested mechanisms. We have provided evidence that in the upper TNP degradation pathway, three hydrogenations and ring fission take place. The bacteria appear to have evolved enzymes to cope with the highly electron-deficient aromatic ring, which obviously needs to be reduced for subsequent hydrolytic ring cleavage. This is in contrast to the general strategy of aerobic bacteria catabolizing nitroaromatic structures, i.e., oxygenolytic elimination of nitro groups and ring cleavage (15).

We suspect that the remaining two nitro groups are cleaved from 4,6-DNH, forming carboxylic acid(s) which are funneled into the tricarboxylic acid cycle. To show this would demand radioactive labeling for detection of aliphatic compounds. Further, molecular approaches such as the creation of deletion mutants should aid in revealing the metabolites and enzymes of the lower TNP degradation pathway.