Hydride-Meisenheimer Complex Formation and Protonation as Key Reactions of 2,4,6-Trinitrophenol Biodegradation by Rhodococcus erythropolis

Abstract

Biodegradation of 2,4,6-trinitrophenol (picric acid) by Rhodococcus erythropolis HLPM-1 proceeds via initial hydrogenation of the aromatic ring system. Here we present evidence for the formation of a hydride-Meisenheimer complex (anionic ς-complex) of picric acid and its protonated form under physiological conditions. These complexes are key intermediates of denitration and productive microbial degradation of picric acid. For comparative spectroscopic identification of the hydride complex, it was necessary to synthesize this complex for the first time. Spectroscopic data revealed the initial addition of a hydride ion at position 3 of picric acid. This hydride complex readily picks up a proton at position 2, thus forming a reactive species for the elimination of nitrite. Cell extracts of R. erythropolis HLPM-1 transform the chemically synthesized hydride complex into 2,4-dinitrophenol. Picric acid is used as the sole carbon, nitrogen, and energy source by R. erythropolis HLPM-1.

The widespread use of nitrophenols as intermediate chemicals of large-scale syntheses of N-substituted aromatic compounds has led and still leads to considerable environmental problems (10, 16, 20). 2,4,6-Trinitrophenol (picric acid) was used as an explosive and is consequently found as a contaminant in ground water at certain military sites and former production facilities (23). It is also a major byproduct of large-scale nitration of benzene and is therefore found in waste streams (11, 14). Thus, biodegradation of picric acid is of great industrial concern. This is documented by a recent patent on the degradation of picric acid and other nitrophenols by single bacterial isolates (12).

Due to its three nitro groups, picric acid is an aromatic compound with a high positive redox potential and is thus readily susceptible to initial reductive transformations rather than oxidative processes. Attack by microbial oxygenases, as shown for a number of mono- and dinitroaromatic compounds (20), is therefore unknown for this substance. Early reports assuming oxygenolytic elimination of the first nitro group of picric acid (5, 6) have not been verified.

In the early nineties, some new observations of the biodegradation of picric acid and 2,4-dinitrophenol were made, particularly through the enrichment of an organism capable of utilizing picric acid as the sole nitrogen source (8). A temporary orange-red color of the culture medium was observed when Rhodococcus erythropolis HLPM-1 was grown with picric acid as a nitrogen source. This change in color was not due to the reduction of a nitro group as described in an earlier report (22). Based on the identification of metabolites like 2,4-dinitrophenol and 4,6-dinitrohexanoate, a new mechanism for the transformation of picric acid was proposed. For the productive part of the degradation pathway, this mechanism involved addition of one hydride ion to the aromatic system followed by rearomatization and elimination of nitrite. Attempts to isolate or synthesize this novel Meisenheimer complex failed and did not allow its unequivocal identification (8).

Mineralization and utilization of picric acid as the sole carbon and energy source were described by different laboratories. Information on the overall catabolic mechanism, however, is still missing (12, 14, 15, 19).

This communication provides unequivocal evidence for the structural and chemical properties of the hydride-Meisenheimer complex and its protonated species on the basis of the chemical synthesis and spectroscopic characterization. Obviously, the formation of this complex is a key step in the utilization of picric acid as the sole carbon, nitrogen, and energy source.

MATERIALS AND METHODS

Growth of bacteria.

Experiments with polynitrophenols as the sole carbon, nitrogen, and energy source for R. erythropolis HLPM-1 were performed in mineral medium containing (per liter) 25 mg of CaCl₂ · H₂O, 20 mg of Fe(III) citrate · H₂O, 1.0 g of MgSO₄ · 7H₂O, and 1 ml of an inorganic trace element solution (without EDTA and iron, following reference 13) in phosphate buffer (pH 7.4). The phosphate buffer contained 1.36 g of KH₂PO₄/liter and 7.1 g of Na₂HPO₄ · 2H₂O/liter in ultrapure water. This medium was supplemented with picric acid from a stock solution (50 mM, adjusted to pH 7.4) or with 2,4-dinitrophenol to the concentrations described in the text. The cells were incubated in baffled Erlenmeyer flasks on a rotary shaker at 100 rpm and 30°C. Growth was monitored by the optical density at 546 nm versus the colored culture supernatant. To achieve a homogeneous suspension, cells were sonified in an ultrasonic water bath (35 kHz, 200 W) for 30 to 60 s.

Preparation of cell extracts.

Cells of R. erythropolis HLPM-1 grown with 0.5 mM picric acid and 10 mM succinate were harvested by centrifugation and washed twice with phosphate buffer (50 mM, pH 7.4). After resuspension in phosphate buffer, cells were disrupted by three passages through a French pressure cell at 130 MPa. The crude extract was centrifuged at 100,000 × g for 35 min at 4°C. The cytosolic supernatant (cell extract) was stored on ice until use.

Enzyme assay.

Transformation of the hydride complex of picric acid (I) was monitored spectrophotometrically by repeated recording of the UV visible spectrum in the wavelength range between 250 and 600 nm of a solution containing the chemically synthesized hydride complex (I) and cell extract (400 μg of protein/ml) from R. erythropolis HLPM-1 in 50 mM phosphate buffer at pH 7.4 and 25°C.

Analytical methods.

Reversed-phase high-performance liquid chromatography (HPLC) analyses of picric acid (t_R = 7.1 min), 2,4-dinitrophenol (t_R = 12.5 min), nitrite (t_R = 2.4 min), and other metabolites were performed on a 125- by 4.6-mm RP 8 column (particle size, 5 μm) equipped with a precolumn (20 by 4.6 mm) by using a mobile phase of 20% (vol/vol) acetonitrile and 0.26% H₃PO₄ in water. Separation of the hydride complex of picric acid (t_R = 3.1 min) was performed by ion pair chromatography on a column of the same size and material as described above with an isocratic eluent consisting of 30% methanol–water and 5 mM tetrabutylammonium hydrogen sulfate (PicA; Waters, Milford, Mass.). The compounds were detected by UV absorption at 210 nm. Metabolites were identified through in situ recording of the absorption spectra (200 to 600 nm).

¹H and ¹³C nuclear magnetic resonance (NMR) spectra were recorded on a WM 400 (Bruker GmbH, Karlsruhe, Germany) with tetramethylsilane as an internal standard at 400.13 MHz (¹H) and 100.61 MHz (¹³C).

Synthesis of the hydride-Meisenheimer complex of picric acid (I).

Acetonitrile (HPLC-quality) was dried through the addition of molecular sieve (0.3 nm; Merck, Darmstadt, Germany), which was heated in a vacuum prior to use. For the synthesis, 2.8 mmol of dry picric acid was dissolved in 6.6 ml of absolute acetonitrile at −20°C. About 700 mg of molecular sieve was added and slightly shaken for 30 min. The molecular sieve was removed and washed with 1 ml of acetonitrile. To this water-free solution of picric acid a cold (−20°C) solution of 2 mmol of (H₃C)₄NB₃H₈ in 6.6 ml of acetonitrile was added within 7 min. The temperature was maintained at −20°C. Orange-red crystals immediately precipitated and were quickly removed in small samples from the reaction mixture by a glass pipette and dropped on a suction filter. This procedure allowed instant removal of the reaction mixture from the product and prevented further transformation. The crystals were quickly washed with a few drops of cold acetonitrile. Residual solvent was evaporated in a freeze-dryer under high-vacuum conditions within 7 h. Subsequent darker-colored precipitates were discarded.

Chemicals.

All chemicals used were of the highest purity commercially available. Picric acid and 2,4-dinitrophenol were obtained from Fluka (Neu-Ulm, Germany) as moistened preparation. The water was removed through high-vacuum evaporation. 2-Amino-4,6-dinitrophenol (picramic acid) from Tokyo Casei (Tokyo, Japan) and (CH3)₄NB₃H₈ (tetramethylammonium octahydridotriborate) from Alfa Johnson Matthey (Karlsruhe, Germany) were used.

RESULTS

Growth of R. erythropolis HLPM-1 on picric acid as the sole carbon, nitrogen, and energy source.

This strain was originally isolated with picric acid as the sole nitrogen source (8). In order to demonstrate that the organism could also use picric acid as the sole carbon and energy source, we used a culture that was cultivated for several months with 2,4-dinitrophenol (>2 mM) as the sole carbon, nitrogen, and energy source in mineral medium (15). To prevent any carryover of residual organic carbon, cells from the stationary-growth phase were resuspended in phosphate buffer and served as inoculum (optical density, 0.1) for a mineral medium containing 3 mM picric acid (Fig. 1). Initially, the cells formed strong conglomerates and thus prevented the monitoring of the optical density. Therefore, the cultures were sonified for 30 to 60 s prior to each measurement.

FIG. 1.

Growth of R. erythropolis HLPM-1 with picric acid as the sole carbon, nitrogen, and energy source.

After a lag phase of 5 days, the shape of the cells turned from coccoid (size, 1 μm) to elongated (5 μm), which indicated the beginning of the exponential-growth phase. The decrease in concentration of picric acid was accompanied by a transient intense orange-red color of the culture medium. This color change is in agreement with previous observations where picric acid served as a nitrogen source (8), indicating that the orange-red metabolite also plays a key role when picric acid is mineralized by R. erythropolis HLPM-1. Subcultivating the cells resulted in the following reproducible growth parameters: 43% of the nitrogen from picric acid was released as nitrite, the growth yield was 19 mg of dry weight per mmol of picric acid, and the growth rate (μ) was 0.012 h⁻¹. Picric acid and nitrite were quantified by HPLC as described in the Materials and Methods section.

Chemistry of the hydride-Meisenheimer complex of picric acid and its protonated form.

For extensive spectroscopical and physiological studies, the supposed complex [compound (I); Fig. 2] had to be isolated in sufficient amounts. Since an authentic standard has not yet been described and isolation from the biological system yielded only small amounts, a procedure to synthesize this complex chemically was developed. This was accomplished by generating and immediately precipitating the complex at low temperature. To a solution of dry picric acid in absolute acetonitrile a solution of (H₃C)₄NB₃H₈ in acetonitrile was added dropwise over a period of a few minutes at −20°C. Orange-red crystals precipitated instantly and after being dried were directly used for spectroscopic investigations. HPLC analysis revealed traces of picric acid as an impurity. Under the above procedure, the formation of 2-amino-4,6-dinitrophenol (picramic acid) as a contaminant was avoided.

FIG. 2.

Formation of the hydride-Meisenheimer complex of picric acid (I), its protonated form (II), and the possible tautomeric aci-nitro form (III).

NMR experiments in organic solvents.

¹H NMR spectra of the reaction product in deuterated dimethyl sulfoxide (DMSO-d₆) always showed formation of a second product which was generated at the expense of the first one immediately after dissolution in DMSO. Repeated recording of the UV visible spectra confirmed a uniform slow transformation (within minutes) of one substance into the other by exhibiting four isosbestic points at λ = 325, 375, 432, and 473 nm (15). The ¹H and ¹³C NMR spectra in DMSO-d₆ revealed resonances which are in accordance with the two systems (I) and (II) shown in Fig. 2. The final assignments of the signals to the hydride complex of picric acid (I) and its protonated form (II) are summarized in Table 1. It is important to mention here that these assignments were possible only after further spectroscopic investigations in deuterated acetonitrile (CD₃CN) as described below.

TABLE 1.

NMR spectroscopic data for the hydride complex of picric acid (I) and its protonated form (II) as (H₃C)₄N⁺ salt in DMSO-d₆

δ1H (ppm) (multiplicity)a	Relative intensity	J (H H) (Hz)	δ13C (ppm)
Hydride complex (I)
			C-1 = 166.5
			C-2 = 116.3b
H-3a,a′ = 3.77 (s)	2.6		C-3 = 28.5
			C-4 = 125.1b
H-5 = 8.4 (s)	1		C-5 = 129.16
			C-6 = 124.1b
(H3C)4N+ = 3.08			(H3C)4N+ = 54.39
Protonated form of the hydride complex (II)
H-2 = 5.62 (dd)	1.1	J2,3a = 7.1	C-1 = 171.9
		J2,3b = 10.7	C-2 = 88.16
H-3a = 3.52 (dd)	2.6	J3a,3b = 16.3	C-3 = 27.3
H-3b = 3.43 (ddd)		J3b,5 = 1.0	C-4 = 123.9
H-5 = 8.6 (d)	1		C-5 = 132.7
			C-6 = 118.6
(H3C)4N+ = 3.08			(H3C)4N+ = 54.39

^as, singlet; d, doublet; dd, doublet of doublet; ddd, doublet of doublet of doublet. 

^bTentative assignment. 

Due to the presence of only three protons in the structure (I), with two of them being equivalent (H-3a,a′), the ¹H NMR spectrum showed only two singlets in DMSO-d₆. In contrast, the protonated form (II) with an additional proton (H-2) lost its planarity due to a change in the hybridization of C-2. Therefore, it displays different resonances in the methylene proton region. This observation allowed us to deduce further information regarding the structure of (II). Characteristic of compound (II) is the ABX spin system which is resolved at δ = 3.5 ppm. The resonance at 3.52 ppm is assigned to H-3a, which splits into a doublet through geminal coupling with H-3b. This doublet splits again into a doublet through vicinal coupling with H-2. For the second proton at position 3 (H-3b), a doublet of doublets of doublets is found at δ = 3.43 ppm as a result of geminal coupling with H-3a, vicinal coupling with H-2, and a long-distance coupling with H-5 (data given in Table 1). In addition, ¹³C resonances confirmed two independent spin systems. There is a twofold set of absorptions with two different methylene-C at δ = 28.5 and 27.3 ppm, two carbonyl-C at δ = 166.5 and 171.9 ppm, and the characteristic nitrosubstituted C-2 of the structure (II) at 88 ppm (Table 1). These assignments were confirmed through measurement in CD₃CN. Final proof came from a ¹H, ¹H correlation spectrum (15). The ¹H NMR spectral characteristics in CD₃CN are listed in Table 2.

TABLE 2.

¹H NMR spectroscopic data for the hydride complex of picric acid (I) and its protonated form (II) in CD₃CN

δ1H (ppm) (multiplicity)a	Relative intensity	J (H H) (Hz)
Hydride complex (I)
H-3a,a′ = 3.82 (s)	2.4
H-5 = 8.49 (s)	1
(H3C)4N+ = 3.07
Protonated form of the hydride complex (II)
H-2 = 5.4 (multiplet)	1	J2,3a = 8.6
H-3a, H-3b = 3.59–3.47 (multiplet)	2.4	J2,3b = 9.7
H-5 = 8.7 (s)	1
(H3C)4N+ = 3.07

^aAbbreviations for multiplicity designations are as given in footnote a to Table 1. 

Spectroscopic investigations in aqueous systems.

After the structural properties had been characterized in organic solvent systems, the question of whether the hydride complex of picric acid (I) could also be identified in aqueous systems remained. This was particularly important with respect to the formation of the protonated form (II), which could be the favorable intermediate for enzymatic elimination of nitrite generating 2,4-dinitrophenol.

The ¹H NMR spectrum in D₂O-phosphate buffer at pH 8.2 revealed significant differences compared to the systems in organic solvents described above. At this pH, only product (I) and no protonated form (II) can be recognized (Fig. 3 and Table 3). A shift to pH 7.4 (D₂O-phosphate buffer) clearly resulted in the formation of small amounts of the protonated form (II) (data not shown). According to the proton integrals, about 5% of (I) are protonated at pH 7.4. As a consequence, it can be expected that considerable amounts of the protonated complex (II) are present under physiological conditions at pH 7 in the bacterial cell. In addition, it is important to mention here that under the conditions described, spontaneous elimination of nitrite was observed neither from the hydride complex (I) nor from its protonated form (II).

FIG. 3.

¹H NMR spectrum of the hydride complex of picric acid (I) at pH 8.2 in D₂O-phosphate buffer.

TABLE 3.

¹H NMR data for the hydride complex (I) and its protonated form (II) in different aqueous systems

Hydride complex	δ1H (ppm) (multiplicity)a	J (H H) (Hz)
(I) at pH 8.2	H-3a,a′ = 4.01 (s)
	H-5 = 8.81 (s)
	(H3C)4N+ = 3.32
(I) at pH 7.4	H-3a,a′ = 4.10 (s)
	H-5 = 8.90 (s)
	(H3C)4N+ = 3.32
(II) at pH 7.4	H-3a = 3.91 (d)	J3a,3b = 17.3
	H-3b = 3.74 (d)
(I) in D2O-DCl	H-3a,a′ = 3.66 (d)
	H-5 = 8.91 (d)
	(H3C)4N+ = 2.91
(II) in D2O-DCl	H-3a = 3.49 (d)	J3a,3b = 17.8
	H-3b = 3.32 (d)
	H-5 = 8.43 (s)
	(H3C)4N+ = 2.91

^aAbbreviations for multiplicity designations are as given in footnote a to Table 1. 

Furthermore, the ¹H NMR at pH 7.4 indicated that the proton at position 2 of structure (II) is readily exchangeable. Obviously, H-2 of structure (II) is replaced by D from D₂O. Consequently, the vicinal coupling between the protons at position 3 and 2 cannot be observed, for the H,D coupling is much smaller than the H,H coupling. Since this finding is characteristic of both the structural and the chemical properties of compounds (I) and (II), we tried to prove the transformation of structure (I) into structure (II). It was expected that the addition of DCl to a solution of (I) in D₂O would result in the formation of the deuterated form (II). This outcome is shown in Fig. 4. Through the addition of D⁺ at C-2, planarity of the ring system is lost. Therefore, the two protons at C-3 are no longer equivalent and form an AB spin system represented by two doublets, at δ = 3.49 and 3.32 ppm. Because of the deuteration at C-2, a vicinal coupling (H-3,D-2) cannot be observed. For the same reason, the resonance of D-2 is not detectable, as it was shown for H-2 in DMSO-d₆ (compare Tables 1 and 3).

FIG. 4.

¹H NMR spectrum of the hydride complex (I) and the protonated form (II) in DCl-D₂O.

This transformation of (I) into (II) yielded the essential data for the structural and chemical characterization of this complex system. For further physiological investigations, it was important to identify protonation already at pH 7.4 as described above. Due to changes in the resonance structure, protonation should also result in a significant change of the UV visible spectrum. This is shown in Fig. 5A. It is worth mentioning that protonation can be reversed by shifting the pH to values above 7.4. This is particularly important for the chromatographic analysis of culture supernatants.

FIG. 5.

UV visible spectra of the hydride complex (I) in water at different pH (A). Panel B displays two in situ spectra of the biologically and chemically formed hydride complex (I) recorded during ion pair chromatography (30% methanol, pH 8.2).

Routine HPLC analysis of the culture supernatant of picric-acid-grown cells and the reference hydride complex (I) always showed a broad signal with two unresolved peaks under acidic conditions (20% acetonitrile–water [pH 2]). According to the above-presented facts, this can be rationalized with increasing protonation during the HPLC run. The in situ UV visible spectra recorded at the peak maxima were identical for both the synthetic sample and the microbial product (15). In contrast, ion pair chromatography at pH 8.2 in 30% methanol–water with tetrabutylammonium as the counterion revealed a sharp discernible signal with a well-defined UV visible spectrum as shown in Fig. 5B. According to the proton resonance spectra (Fig. 3) this spectrum must be assigned to the hydride complex (I). It is characterized by three absorption maxima, at λ = 238, 318, and 423 nm, and a shoulder at 490 nm. The two superimposed spectra in Fig. 5B provide additional evidence for the identity of the biologically formed compound and the chemically synthesized hydride complex of picric acid (I).

Enzymatic transformation of the hydride complex of picric acid.

In order to identify the hydride complex (I) as a key metabolite of picric acid metabolism, it was necessary to prove that it can serve as a substrate for further enzymatic transformation. Therefore, R. erythropolis HLPM-1 was cultivated with 0.5 mM picric acid as the sole nitrogen source and 10 mM succinate as the carbon and energy source. Washed cells were disrupted by passage through a French press, and the cell extract obtained by centrifugation at 100,000 × g was used for the following test. To a solution of the chemically synthesized hydride complex of picric acid (I) in 50 mM phosphate buffer (pH 7.4) cell extract was added to start the transformation. As shown in Fig. 6, repeated recordings of the UV visible spectrum illustrate continuous transformation of complex (I) into 2,4-dinitrophenol (17), the identity of which is demonstrated by the characteristic spectrum of 2,4-dinitrophenol with λ_max = 355 nm. HPLC analyses confirmed the formation of 2,4-dinitrophenol (specific activity = 2.2 nmol of 2,4-dinitrophenol min⁻¹ mg of protein⁻¹) with concomitant liberation of nitrite. From the ¹H NMR data presented above, it can be concluded that elimination of nitrite proceeds via the protonated form (II), which is already present at a ratio of 5% (at pH 7.4). As expected, turnover of the biologically produced hydride complex revealed the same spectral changes as described above. The turnover was strictly dependent on complete cell extract. Attempts to fractionate the cell extract by column chromatography in order to identify the responsible components resulted in complete loss of activity. Control experiments without extract or with denatured extract did not result in the transformation of (I).

FIG. 6.

Enzymatic turnover of the chemically synthesized hydride complex of picric acid with formation of 2,4-dinitrophenol (λ_max = 355 nm) by cell extract of R. erythropolis HLPM-1 at pH 7.4. The arrows indicate the disappearance of the characteristic absorption of the hydride complex (I). Spectra were recorded at intervals of 10 min.

DISCUSSION

In picric acid, the three nitro groups cause a particular electron-deficient aromatic system due to their negative mesomeric and negative inductive effect. Consequently, it is not surprising that studies on aerobic microbial degradation of picric acid did not reveal any electrophilic oxygenation as an initial transformation reaction (16). Instead, an initial reductive attack with formation of a hydride-Meisenheimer complex of picric acid (I) in vivo was proposed (8). This was also demonstrated by in vitro studies. These studies also indicate that NADPH is the potential hydride source, since the addition of this cofactor restored activity (15, 17). Therefore, spectroscopic and chemical data were required for unequivocal identification of this novel complex, which could be the key intermediate in the catabolic pathway of trinitroaromatic compounds.

During growth studies with R. erythropolis HLPM-1, originally isolated as a mutant of a 2,4-dinitrophenol-degrading organism (8), we were able to cultivate this strain with picric acid as the sole nitrogen, carbon, and energy source. The initial transient color change of the culture medium to orange-red corresponded to observations in cultures containing picric acid as the sole nitrogen source (8). It indicated the potential key function of this metabolite in the catabolic pathway of picric acid. In contrast to these findings, a color change was not (19) or was only occasionally (14) reported in recent publications on aerobic degradation of picric acid as the sole carbon, nitrogen, and energy source by different isolates. With respect to toxicity, it is important to mention that picric acid was utilized by R. erythropolis HLPM-1 at concentrations up to 3.4 mM. Under batch conditions, the organism tolerated picric acid at concentrations up to 14 mM (1).

It was expected that the characterization of the orange-red metabolite would provide insight into the mechanism of the reduction of picric acid and the subsequent elimination of the first nitro group. Because the amount of this metabolite from culture was insufficient for thorough spectroscopic analysis, it was necessary to synthesize the complex. From the literature, it was known that hydride complexes of polynitroaromatic compounds can be generated by hydride transfer by using NaBH₄, Bu₃SnH, (H₃C)₄NBH₄, or (H₃C)₄NB₃H₈ as donor (2, 4, 7, 9, 18, 21). However, none of these studies included a strong acid, such as picric acid (with a pK_a of 0.38, according to reference 3). It turned out that the reaction of (H₃C)₄NB₃H₈ with picric acid must be performed under kinetic control. Under these conditions, the reaction product was free of the red 2-amino-4,6-dinitrophenol which is formed on prolonged reaction time or at elevated temperatures, indicating it to be the thermodynamically preferred product. Possibly due to disproportionation reactions in stored samples of the crystalline hydride complex (I), picric acid and 2-amino-4,6-dinitrophenol were found occasionally as impurities.

The ¹H NMR studies of the synthetic complex (I) in DMSO revealed formation of the protonated complex (II) immediately after dissolution. Although the NMR spectrum became more complex, the observation of the protonated form (II) provided the final structural proof for compounds (I) and (II). Since compound (I) bears only three protons, two of them being magnetically equivalent, the ¹H and ¹³C NMR was not sufficiently expressive for structural analysis. Through protonation at C-2, the protons in position 3 became magnetically different.

The question of to what extent nitro-aci-nitro tautomerism of complex (II) (Fig. 2) contributes to the resonance signals remained. If the proton at position 2 is added to the nitro group at position 2, the so-called aci-form (III) is generated (Fig. 2). This form shows high structural conformity with complex (I) and would be difficult to distinguish from it by ¹H NMR. The aci-form (III) could not be correlated with the signals found because a corresponding hydroxyl resonance is missing and because this form (III) should become dominant under acidic conditions. This is in contrast to our findings with compound (II), being the preferred species under acidic conditions. The ¹³C NMR confirmed the structural characteristics of compound (I) and (II) with two different carbonyl-C (C-1), two different methylene-C (C-3), and a single nitro-substituted methylene-C (C-2 of II) (Table 1).

NMR studies in aqueous systems clearly revealed the existence and stability of the hydride complex of picric acid (I) and the pH dependence of the protonation reaction. For the biological system, this is of major importance because enzymatic formation of (I) is not followed by spontaneous rearomatization to form 2,4-dinitrophenol. In contrast, the formation of compound (I) is followed by subsequent protonation beginning at pH 7.4. The proton H-2 is readily exchanged as shown by the H-D experiment whose results are shown in Fig. 4. Even this protonation did not result in spontaneous rearomatization, with liberation of nitrite occurring under neither physiological nor acidic (pH > 2) conditions.

Therefore, it was indispensable to prove enzymatic turnover of the isolated hydride complex (I) by cell extracts of R. erythropolis HLPM-1. As shown in Fig. 6, the hydride complex (I) is transformed in an enzyme-dependent reaction generating 2,4-dinitrophenol with concomitant liberation of nitrite. The recovery of nitrite was about 72% of the theoretical amount. Obviously, some of the nitrite was lost by biological oxidation or interaction with the cell extract.

Elimination of the nitro group from C-2 of the hydride complex (I) is unfavorable because the system is stabilized by mesomeric effects. This elimination would require an intramolecular hydrogen migration from C-3 to C-2. Gold et al. (4) described this so-called vicinal attack and observed spontaneous elimination of nitrite from the 3-H⁻-complex of 2,4-dinitroanisole. In contrast to these studies, spontaneous elimination of nitrite was never observed with compound (I). Consequently, we conclude that the formation of the protonated species (II), which is found at physiological pH, provides the reactive intermediate which allows the enzymatic elimination of nitrite.

The results presented here justify the conclusion that the key reaction of picric acid degradation is the enzymatic hydride transfer to the aromatic nucleus, followed by protonation with subsequent enzyme-dependent elimination of nitrite.