Hydroxyatrazine N-Ethylaminohydrolase (AtzB): an Amidohydrolase Superfamily Enzyme Catalyzing Deamination and Dechlorination

Abstract

Hydroxyatrazine [2-(N-ethylamino)-4-hydroxy-6-(N-isopropylamino)-1,3,5-triazine] N-ethylaminohydrolase (AtzB) is the sole enzyme known to catalyze the hydrolytic conversion of hydroxyatrazine to N-isopropylammelide. AtzB, therefore, serves as the point of intersection of multiple s-triazine biodegradative pathways and is completely essential for microbial growth on s-triazine herbicides. Here, atzB was cloned from Pseudomonas sp. strain ADP and its product was purified to homogeneity and characterized. AtzB was found to be dimeric, with subunit and holoenzyme molecular masses of 52 kDa and 105 kDa, respectively. The k_cat and K_m of AtzB with hydroxyatrazine as a substrate were 3 s⁻¹ and 20 μM, respectively. Purified AtzB had a 1:1 zinc-to-subunit stoichiometry. Sequence analysis revealed that AtzB contained the conserved mononuclear amidohydrolase superfamily active-site residues His74, His76, His245, Glu248, His280, and Asp331. An intensive in vitro investigation into the substrate specificity of AtzB revealed that 20 of the 51 compounds tested were substrates for AtzB; this allowed for the identification of specific substrate structural features required for catalysis. Substrates required a monohydroxylated s-triazine ring with a minimum of one primary or secondary amine substituent and either a chloride or amine leaving group. AtzB catalyzed both deamination and dechlorination reactions with rates within a range of one order of magnitude. This differs from AtzA and TrzN, which do not catalyze deamination reactions, and AtzC, which is not known to catalyze dechlorination reactions.

Microorganisms are the ultimate recyclers in the environment, often catabolizing recalcitrant and sometimes unfavorable compounds to more-benign products. The ability of these microbes to evolve tailored enzymes over relatively short periods of time makes them the first line of defense against the environmental accumulation of many substances. This role is especially true for compounds that are directly released into the environment, such as herbicides. One class of herbicides whose environmental fate is of considerable interest is the s-triazine class, which includes atrazine, simazine, ametryn, and prometryn.

s-Triazine herbicide biodegradation has been extensively studied over the past half century (4, 5, 7, 12, 16, 17, 23). Multiple proteins with differing specificities initiate the biodegradative process. For instance, both AtzA and TrzN catalyze the hydrolytic dechlorination of atrazine to yield hydroxyatrazine [2-(N-ethylamino)-4-hydroxy-6-(N-isopropylamino)-1,3,5-triazine] (Fig. 1). TrzN, but not AtzA, hydrolyzes ametryn to hydroxyatrazine. However, only one enzyme, hydroxyatrazine N-ethylaminohydrolase (AtzB), has been identified that is capable of transforming hydroxyatrazine to N-isopropylammelide (6). AtzB, therefore, acts to funnel s-triazine metabolites into a common biodegradative pathway, resulting in its eventual transformation to cyanuric acid, which is the central intermediate in s-triazine ring metabolism by many bacteria (7). There are reports of Nocardia and Nocardioides having the ability to convert hydroxyatrazine to N-ethylammelide, but no protein with this activity has been identified (34, 38).

FIG. 1.

AtzB represents the convergence of multiple s-triazine degradative pathways. Reactions catalyzed by AtzA and TrzN funnel numerous s-triazines through the common metabolite hydroxyatrazine. AtzB is present in both of these pathways to further degrade hydroxyatrazine to N-isopropylammelide.

From an evolutionary perspective, there may be a selective advantage to having multiple initiating proteins with different specificities that allow for the degradation of a broader range of compounds. But for substrates like atrazine (Fig. 1), no nutrients supporting bacterial growth are released in the first hydrolytic reaction. Only with the AtzB-catalyzed reaction, which hydrolytically cleaves an N-ethylamino substituent to produce N-isopropylammelide and ethylamine, is nitrogen released from the s-triazine ring. The released nitrogen is subsequently used by organisms capable of alkylamine utilization, as has been demonstrated most directly with Arthrobacter aurescens TC1 (36). Nitrogen release provides a selective advantage and likely facilitates the establishment and maintenance of these genes in soil populations of degradative bacteria. Furthermore, the existence of triazine genes on self-transmissible plasmids (18) further facilitates the propagation of these genes in the microbial community.

With the exception of AtzB, all of the enzymes involved in the conversion of s-triazine herbicide to cyanuric acid have been well characterized (9, 11, 24, 26, 27, 31). The atzB gene from Pseudomonas sp. strain ADP was previously cloned and its product sequenced (6), but poor expression prevented protein purification and detailed enzyme characterization. The work described here outlines a novel approach for cloning the atzB gene that allowed for further enzyme characterization, including a detailed study on substrate specificity.

MATERIALS AND METHODS

Chemicals.

Fluoroatrazine [2-fluoro-4-(N-ethylamino)-6-(N-isopropylamino)-1,3,5-triazine], azidoatrazine [2-azido-4-(N-ethylamino)-6-(N-isopropylamino)-1,3,5-triazine], cyanoatrazine [2-cyano-4-(N-ethylamino)-6-(N-isopropylamino)-1,3,5-triazine], atratone, aminoatrazine [2-amino-4-(N-ethylamino)-6-(N-isopropylamino)-1,3,5-triazine], 2-isopropylamino-4-ethylamino-1,3,5-triazine, and 2-mercapto-4-isopropylamino-6-ethylamino-1,3,5-triazine were synthesized as previously described (26, 33). Compounds synthesized for this study were analyzed for purity by gas chromatography-mass spectrometry (MS) using an HP 6890/5973 instrument (Hewlett-Packard, San Fernando, CA).

Hydroxyatrazine analogs containing different substituents in place of the hydroxyl group were synthesized as previously described (26). 2-Chloro-4-amino-6-hydroxy-s-triazine derivatives were prepared as described by Seffernick and coworkers (28). Ammelide and ammeline derivatives were prepared as follows. First, 2,4-dichloro-6-amino-1,3,5-triazine and 2,4-dichloro-6-(N-substituted-amino)-1,3,5-triazines were prepared from cyanuric chloride as described by Thurston et al. (37), and then the amino-dichloro compounds were hydrolyzed to give the parent ammelides as described by Smolin and Rapoport (35). Ammeline derivatives were prepared as follows. 2-Chloro-4,6-diamino-1,3,5-triazine, purchased from Aldrich Chemical Co. (Milwaukee, WI), was used to prepare (N-substituted-amino)-amino-chloro-s-triazines as described by Thurston et al. (37). These chloro compounds were then hydrolyzed to the ammelines in 20% (vol/vol) acetic acid, catalyzed with potassium bromide, and buffered with sodium acetate.

Cloning and expression of AtzB.

Initially, the atzA and atzB genes from Pseudomonas sp. strain ADP were independently cloned into two different locations within plasmid pACYC184 (6, 9). The atzA gene, cloned into the AvaI site on a 1.9-kb fragment (resulting in pMD4), expressed large amounts of protein, while the atzB gene cloned into the ClaI site on a 4-kb fragment (pATZB2) did not. The upstream regions of atzA and atzB were identical except for the eight nucleotides directly adjacent to the genes, namely, GACATATC and TAACCACC, respectively. This region was thought to be part of a putative Shine-Dalgarno sequence (6, 9). To facilitate the expression of the AtzB protein, a 1.65-kb AflIII fragment from pATZB2, containing the atzB gene, was inserted into the pMD4 AflIII site of pMD4. The resulting plasmid expressed both the AtzA and AtzB protein. Digestion of this plasmid with BssHII deleted a 1.25-kb fragment containing the upstream and 5′ end of atzA, producing pAAJLS2. Escherichia coli DH5α containing pAAJLS2 only expressed AtzB activity. In this plasmid, the eight nucleotides upstream of the atzB gene remained unchanged relative to the initial pATZB2 vector. A QuikChange mutagenesis kit (Stratagene, La Jolla, CA) was used to change five of the eight nucleotides directly upstream of the atzB gene to the corresponding residues found in the atzA gene Shine-Dalgarno sequence. The resulting plasmid, pAAJLS3, was used for all expression work throughout this paper.

Enzyme assay.

Routine enzymatic activity was measured by monitoring the decrease in the absorbance of hydroxyatrazine at 242 nm with a Beckman DU 640 spectrophotometer (Beckman Coulter, Inc., Fullerton, CA). Reactions were carried out at 23°C in 1 ml of 50 mM phosphate buffer, pH 7.0, containing 20 μM hydroxyatrazine. The reactions were initiated by the addition of enzyme.

Protein purification.

E. coli DH5α(pAAJLS3) was grown overnight at 37°C in super broth (12 g/liter tryptone; 14 g/liter yeast extract; 5 ml/liter glycerol; 3.8 g/liter KH₂PO₄; 12.5 g/liter K₂HPO₄) containing chloramphenicol (30 μg/ml). Cells were harvested by centrifugation at 10,500 × g for 20 min at 4°C, and the cell pellet was resuspended in 25 mM morpholinepropanesulfonic acid (MOPS) buffer (pH 7.0) containing phenylmethylsulfonyl fluoride (100 μg/ml). Crude cell extract was obtained by lysing cells using an AMINCO French pressure cell (Silver Spring, MD) at 10,000 lb/in² at 4°C, followed by centrifugation at 18,000 × g for 100 min. Solid (NH₄)₂SO₄ was added, with stirring over ice, to attain 30% saturation. The chilled solution was stirred for an additional 30 min and centrifuged at 18,000 × g for 30 min. The precipitated material was resuspended in 25 mM MOPS buffer (pH 7.0) and dialyzed overnight at 4°C against 25 mM MOPS buffer, pH 7.0.

The dialyzed protein was loaded onto a 1.75- by 15-cm CHT ceramic hydroxyapatite type I column (Bio-Rad Laboratories, Hercules, CA), equilibrated with 25 mM MOPS buffer, pH 7.0, and separated by using a fast-performance LC (FPLC) system (Pharmacia, Uppsala, Sweden). The protein was eluted using a 0 to 40 mM potassium phosphate gradient, pH 7, at a flow rate of 1 ml per min. Protein eluting from the column was monitored at 280 nm with a Pharmacia UV protein detector and assayed for AtzB activity. The protein was analyzed for purity by sodium dodecyl sulfate-polyacrylamide gel electrophoresis (SDS-PAGE), and the relative amount and percent purity of AtzB were determined using ImageJ software (1).

Determination of the subunit size.

The subunit molecular mass was determined by SDS-PAGE after heating the samples at 95°C for 3 min in the presence of 0.15% (wt/vol) SDS and 5 mM dithiothreitol. Electrophoresis was performed using 12% acrylamide gels, and the subunit size was determined by comparison with the known standards (Bio-Rad Laboratories, Hercules, CA) phosphorylase B (103,774 Da), bovine serum albumin (81,083 Da), ovalbumin (47,753 Da), carbonic anhydrase (35,833 Da), soybean trypsin inhibitor (27,084 Da), and lysozyme (19,333 Da).

Gel filtration chromatography.

A Superose 12 high-resolution gel filtration column (Pharmacia, Uppsala, Sweden) was used for the aforementioned fast-performance LC. The column was equilibrated with 25 mM MOPS buffer, pH 7.0, containing 0.15 M KCl, and the protein was eluted at a flow rate of 0.1 ml per min. The holoenzyme molecular mass was determined based on the elution times for proteins and compounds of known molecular mass (thyroglobulin, 670,000 Da; gamma globulin, 158,000 Da; chicken ovalbumin, 44,000 Da; horse myoglobin, 17,000 Da; and vitamin B-12, 1,350 Da).

Isoelectric focusing.

Analytical isoelectric focusing was performed in broad-range (pH 3 to 10) polyacrylamide gels (Bio-Rad Laboratories, Hercules, CA), as described by the manufacturer. The following protein standards (Bio-Rad Laboratories, Hercules, CA) were used: phycocyanin (pI, 4.45 to 4.75), β-lactoglobulin B (pI, 5.1), bovine carbonic anhydrase (pI, 6.5), equine myoglobin (pI, 6.8 to 7.0), human hemoglobin A (pI, 7.1), human hemoglobin C (pI, 7.5), lentil lectin (pI, 7.8 to 8.2), and cytochrome c (pI, 9.6).

Protein concentration determination.

The protein concentration for quantitative studies was determined through amino acid analysis by Scientific Research Consortium, Inc. (SRC, St. Paul, MN), using a model 7300 amino acid analyzer (Beckman Coulter, Inc.). The Bio-Rad protein reagent was used for routine protein concentration determinations, using bovine serum albumin as a standard.

Influence of Zn(II) concentration in growth media on AtzB activity.

Due to the variability in metal content and specific activity of various enzyme preparations, the effects on enzyme activity of Zn(II) supplementation in the growth medium were investigated. E. coli DH5α was grown in super broth in the presence and absence of 500 μM zinc sulfate. The enzyme was purified, and the specific activity and metal content were determined as described above and below, respectively.

Metal analysis.

Purified AtzB (3 to 10 mg) was hydrolyzed with an equal volume of metal-free concentrated hydrochloric acid under vacuum at 115°C for 21 to 24 h (SRC). The sample was diluted 10-fold, and the metal content was determined by inductively coupled plasma emission spectroscopy at the University of Minnesota Soils Analytical Laboratory (St. Paul, Minnesota).

Effect of chelators on AtzB activity.

Purified AtzB was incubated separately with 15 chelators (each at 5 mM) at 4°C, and the enzymatic activity was monitored over a 24-h period. The chelators tested included 1,10-phenanthroline (Aldrich, Milwaukee, WI), bathophenanthrolinedisulfonic acid (Sigma, St. Louis, MO), 8-hydroxyquinoline (Sigma-Aldrich, St. Louis, MO), 8-hydroxyquinoline-5-sulfonic acid (Sigma, St. Louis, MO), 2,2′-bipyridine (Lancaster Synthesis, Inc., Pelham, NH), Chelex-100 resin (Bio-Rad Laboratories, Hercules, CA), nitriloacetic acid (Sigma, St. Louis, MO), diethylenetriaminepentaacetic acid (Aldrich, Milwaukee, WI), EDTA (Sigma-Aldrich, St. Louis, MO), dipicolinic acid (Aldrich, Milwaukee, WI), thioglycolic acid (Sigma-Aldrich, St. Louis, MO), oxalic acid (J. K. Baker Chemical Co., Phillipsburg, NJ), d-penicillamine (MP Biomedicals Inc., Illkirch, France), histidine (Sigma, St. Louis, MO) and N,N,N′,N′-tetrakis(2-pyridylmethyl) ethylenediamine (Fluka, Germany). Chelex-100 treated buffers were used throughout these experiments to ensure trace-metal removal.

Steady-state state kinetics analysis of substrates and analogs.

The steady-state kinetics parameters were determined for hydroxyatrazine and substrate analogs containing N-alkyl or chlorine substituents. The data were obtained from initial velocity measurements using the standard spectrophotometric assay. The extinction coefficient of the respective substrate or analog was monitored on a Beckman DU64 spectrophotometer to give the following values: 2-(N-ethylamino)-4-hydroxy-6-(N-isopropylamino)-1,3,5-triazine (hydroxyatrazine), ɛ₂₄₂ = 12.8 mM⁻¹ cm⁻¹; 2,4-diisopropylamino-6-hydroxy-1,3,5-triazine, ɛ₂₄₂ = 5.8 mM⁻¹ cm⁻¹; 6-hydroxy-4-(N-isopropylamino)-2-(N-methylamino)-1,3,5-triazine, ɛ₂₄₂ = 13.2 mM⁻¹ cm⁻¹; 2-chloro-4-(N-ethylamino)-6-hydroxy-1,3,5-triazine, ɛ₂₅₆ = 2.4 mM⁻¹ cm⁻¹; and 2-chloro-4-hydroxy-6-(N-isopropylamino)-1,3,5-triazine, ɛ₂₅₈ = 2.5 mM⁻¹ cm⁻¹. The kinetic parameters of AtzB were calculated using the Hanes-Woolf equation: [S]/V₀ = [S]/V_max + K_m/V_max (30), where S is the substrate concentration, V₀ is the rate of a fixed substrate concentration, K_m is the Michaelis-Menten constant, and V_max is the rate of saturating substrate concentration. Linear regression analyses were used to determine the V_max and K_m parameters.

Substrate specificity analysis.

Fifty s-triazine and pyrimidine substrate analogs were tested as substrates for AtzB, and when enzymatic activity was observed, the products were identified (see Table 3). Each of these compounds, at 100 μM in 50 mM phosphate buffer (pH 7.0), were incubated with purified AtzB for 48 h at 23°C. The enzymatic reactions were stopped by heating for 2 min at 95°C. The solutions were filtered through a 0.2-μm Acrodisc CR 13-mm syringe filter (Pall Gelman Laboratory, Ann Arbor, MI) prior to high-pressure liquid chromatography (HPLC) analysis. Control samples without enzyme were handled in parallel with the enzyme-treated samples. Dechlorination and deamination reactions do not occur nonenzymatically to any measurable extent under the conditions of the experiments conducted here. The samples were analyzed by HPLC, using a Hewlett-Packard HP 1100 HPLC system equipped with a photodiode array detector interfaced to a Hewlett-Packard ChemStation. An Adsorbosphere C₁₈ 5-μm column (250 mm by 4.6 mm) (Alltech, Deerfield, IL) was used to separate alkylated triazines and pyrimidines with an acetonitrile-water linear gradient as previously described (6). Hydroxyatrazine and its analogs were separated by using a Waters IC-PaK anion high-capacity column (150 by 4.6 mm) with an isocratic 5 mM phosphate buffer (pH 8.0) mobile phase and 0.5 ml/min flow rate. Hydroxyatrazine analogs were monitored at 224 nm. 2-Amino-4-chloro-6-hydroxy-1,3,5-triazine, 2-chloro-4-(N-ethylamino)-6-hydroxy-1,3,5-triazine, and 2-chloro-4-hydroxy-6-(N-isopropylamino)-1,3,5-triazine were detected at 200 nm with the Waters IC-PAK column, with an isocratic mobile phase consisting of 95% 10 mM phosphate buffer (pH 7.0) and 5% acetonitrile mobile phase at a flow rate of 1 ml/min. Liquid chromatography (LC)-MS analysis was carried out at the Center for Mass Spectroscopy and Proteomics at the University of Minnesota.

TABLE 3.

Compounds used throughout this study

Substrate	Chemical name	Product(s)a
IEOT	2-(N-Ethylamino)-4-hydroxy-6-(N-isopropylamino)-1,3,5-triazine	IOOT, EOOT
MMOT	2,4-Dimethylamino-6-hydroxy-1,3,5-triazine	MOOT
EEOT	2,4-Diethyamino-6-hydroxy-1,3,5-triazine	EOOT
IIOT	2,4-Diisopropylamino-6-hydroxy-1,3,5-triazine	IOOT
(iB)2OT	2,4-Di(N-isobutylamino)-6-hydroxy-1,3,5-triazine	(iB)OOT
(sB)2OT	2,4-Di(N-secbutylamino)-6-hydroxy-1,3,5-triazine	(sB)OOT
(tB)2OT	2,4-Di(N-terbutylamino)-6-hydroxy-1,3,5-triazine	(tB)OOT
EMOT	4-(N-Ethylamino)-6-hydroxy-2-(N-methylamino)-1,3,5-triazine	EOOT, MOOT
IMOT	6-Hydroxy-4-(N-isopropylamino)-2-(N-methylamino)-1,3,5-triazine	IOOT, MOOT
AAOT	2,4-Diamino-6-hydroxy-1,3,5-triazine	AOOT
EAOT	2-Amino-6-(N-ethylamino)-4-hydroxy-1,3,5-triazine	EOOT, AOOT
IAOT	2-Amino-4-hydroxy-6-(N-isopropylamino)-1,3,5-triazine	IOOT, AOOT
(oE)2OT	2-Hydroxy-4,6-di(N-hydroxyethylamino)-1,3,5-triazine	(oE)OOT
(oE)EOT	2-(N-Ethylamino)-4-hydroxy-6-(N-hydroxyethylamino)-1,3,5-triazine	(oE)OOT, EOOT
(oE)AOT	2-Amino-4-Hydroxy-6-(N-hydroxyethylamino)-1,3,5-triazine	(oE)OOT, AOOT
I(3MoP)OT	2-Hydroxy-4-(N-isopropylamino)-6-N-(3-methoxypropylamino)-1,3,5-triazine	(3MoP)OOT, IOOT
(EM)EOT	2-(N-ethyl-N-methylamino)-4-ethylamino-6-hydroxy-1,3,5-triazine	EOOT
CAOT	2-Amino-4-chloro-6-hydroxy-1,3,5-triazine	AOOT
CEOT	2-Chloro-4-(N-ethylamino)-6-hydroxy-1,3,5-triazine	EOOT
CIOT	2-Chloro-4-hydroxy-6-(N-isopropylamino)-1,3,5-triazine	IOOT
BBOT	2,4-Di(N-Butylamino)-6-hydroxy-1,3,5-triazine	NS
AOOT	2-Amino-4,6-dihydroxy-1,3,5-triazine	NS
MOOT	2,4-Dihydroxy-6-(N-methylamino)-1,3,5-triazine	NS
EOOT	2,4-Dihydroxy-6-(N-ethylamino)-1,3,5-triazine	NS
IOOT	2,4-Dihydroxy-6-(N-isopropylamino)-1,3,5-triazine	NS
OOOT	2,4,6-Trihydroxy-1,3,5-triazine	NS
(MM)2OT	2,4-Bis(dimethylamino)-6-hydroxy-1,3,5-triazine	NS
(EE)MOT	2-Diethylamino-4-methylamino-6-hydroxy-1,3,5-triazine	NS
(EE)EOT	2-Diethylamino-4-ethylamino-6-hydroxy-1,3,5-triazine	NS
(Mo)2OT	2,4-Dimethoxy-6-hydroxy-1,3,5-triazine	NS
C(MM)OT	2-Chloro-4-dimethylamino-6-hydroxy-1,3,5-triazine	NS
CIET	2-Chloro-4-(N-ethylamino)-6-(N-isopropylamino)-1,3,5-triazine	NS
CEET	2-Chloro-4,6-di(N-ethylamino)-1,3,5-triazine	NS
CIIT	2-Chloro-4,6-di(N-isopropylamino)-1,3,5-triazine	NS
FIET	2-Fluoro-4-(N-ethylamino)-6-(N-isopropylamino)-1,3,5-triazine	NS
(N3)IET	2-Azido-4-(N-ethylamino)-6-(N-isopropylamino)-1,3,5-triazine	NS
(CN)IET	2-Cyano-4-(N-ethylamino)-6-(N-isopropylamino)-1,3,5-triazine	NS
(OMe)IET	2-Methoxy-4-(N-ethylamino)-6-(N-isopropylamino)-1,3,5-triazine	NS
(SMe)IET	2-Methylthiol-4-(N-ethylamino)-6-(N-isopropylamino)-1,3,5-triazine	NS
(NH2)IET	2-Amino-4-(N-ethylamino)-6-(N-isopropylamino)-1,3,5-triazine	NS
SIET	2-Ethylamino-4-mercapto-6-isopropylamino-1,3,5-triazine	NS
Cyanazine	2-Chloro-4-ethylamino-6-methylpropionitrile-1,3,5-triazine	NS
EEMT	2,4-Diethylamino-6-methyl-1,3,5-triazine	NS
HAAT	2,4-Diamino-1,3,5-triazine	NS
IEHT	2-(N-Ethyamino)4-(N-isopropylamino)-1,3,5-triazine	NS
AO(MT)	2-Amino-4-hydroxy-6-methyl-1,3,5-triazine	NS
AO(PT)	2-Amino-4-hydroxy-6-propyl-1,3,5-triazine	NS
AOOP	2,4-Diamino-6-hydroxy-1,3-pyrimidine	NS
NOOP	4,6-Dihydroxy-5-nitro-1,3-pyrimidine	NS
SOOP	4,6-Dihydroxy-2-mercapto-1,3-pyrimidine	NS
OOOP	2,4,6-Trihydroxy-1,3-pyrimidine	NS

^aNS, not a substrate for AtzB.

Leaving group selectivity.

Individual asymmetric hydroxy-s-triazines with different N-alkyl substituents were tested for leaving group selectivity by AtzB. AtzB was incubated with 100 uM of each compound tested, and the reaction was allowed to proceed until all of the starting compound was consumed, typically 48 h. The products were analyzed by HPLC as described above, and the ratios of the products were determined.

RESULTS

Protein expression and purification.

The atzB gene from Pseudomonas sp. strain ADP was initially cloned by inserting a 4.0-kb ClaI fragment into pACYC184 to produce pATZB2 (6). Although its level of expression was sufficient for initial studies, a higher level of expression was required for the present work. Initial attempts in this study used high copy number expression vectors that resulted in the majority of AtzB occurring in inclusion bodies and having low soluble activity (data not shown). Thus, we followed a strategy of using a low copy vector but modifying the region upstream of the atzB gene.

The upstream region of the atzB gene is identical to that of atzA except for the eight nucleotides adjacent to the start codons (5′-TAACCACC-3′ in atzB and 5′-GACATATC-3′ in atzA). The atzA gene, which was cloned as a 1.9-kb AvaI fragment into pACYC184, was known to express extremely well (9). Using that knowledge and a previous construct, a plasmid was engineered in which the atzB gene was inserted into pMD4 in place of the atzA gene. The resulting plasmid, pAAJLS3, produced AtzB at 15 to 20% of the total cell protein, and the specific activity of AtzB in crude extracts was 10-fold higher than that obtained from the previously published clone (6).

Ammonium sulfate fractionation (0 to 30% saturation) and ceramic hydroxyapatite type I column chromatography were used to purify AtzB to homogeneity. By this procedure, AtzB was purified 6.6-fold, with a 26% yield (Table 1). The purified protein produced a single band on SDS-PAGE gels (Fig. 2).

TABLE 1.

AtzB purification from E. coli DH5α(pAAJLS3)^a

Purification step	Total protein (mg)	Total activity (U)b	Sp. act. (U mg−1)	Recovery (%)	Purification (fold)
Crude extract	2322	375	0.16	100	0
Ammonium sulfate (0-30% saturation)	582	191	0.33	51	2.1
Ceramic hydroxyapatite type I column	94	99	1.05	26	6.6

^aGrowth medium was supplemented with 500 μM zinc sulfate.

^bOne unit of enzyme activity is defined as the conversion of 1 μmol of hydroxyatrazine to N-isopropylammelide per min.

FIG. 2.

SDS-PAGE analysis of protein samples at various stages of AtzB purification. Lanes: 1, crude cell extract; 2, 0 to 30% ammonium sulfate fraction; 3, peak from ceramic hydroxyapatite column; 4, standard molecular mass markers. Numbers to the right of the gel are in kilodaltons.

AtzB characterization.

According to the results of SDS-PAGE, AtzB had a subunit molecular mass of 52 kDa (Fig. 2). This corresponded to the calculated molecular mass derived from the translated gene sequence. The holoenzyme's molecular mass, determined by gel filtration chromatography, was estimated to be 105 kDa, suggesting a dimeric oligomerization state. The theoretical pI of AtzB, based on the amino acid sequence, was 5.7, while experimental determination yielded an isoelectric point of 5.0. Buffer and pH studies, done over a pH range of 6.0 to 10.0, indicated that the maximum activity occurred in 50 mM phosphate buffer at a pH range of 6.5 to 7.5.

AtzB's homology to the amidohydrolase superfamily.

The atzB genes from a variety of hydroxyatrazine-degrading bacteria were previously sequenced and found to have >99% identity to the gene from Pseudomonas sp. strain ADP (8). A GenBank search using BLASTP identified additional bacteria having AtzB homologs with a similar degree of sequence identity in Arthrobacter aurescens TC1, Herbaspirillum sp. strain B601, and Betaproteobacterium strain CDB21. All of these bacteria were isolated for their s-triazine herbicide-degradative ability, and these proteins are considered isofunctional with AtzB from Pseudomonas sp. strain ADP. In addition, BLASTP analysis identified gi86360134 (Rhizobium etli CFN 42) and gi23012454 (Magnetospirillum magnetotacticum MS-1) as having 61% and 59%, respectively, sequence identity to AtzB. These proteins with an intermediate level of sequence identity have unknown functions at this time. All other sequences shared less than 40% sequence identity with AtzB. The highest-scoring proteins with functional assignments from this group included TrzN (Arthrobacter aurescens, gi42558845, 31%), AtzA (Pseudomonas sp. strain ADP, gi3766246, 23%), and guanine deaminase (E. coli K12, gi16130785, 24%). These proteins have been identified as belonging to the amidohydrolase superfamily (27, 32). The conservation of the putative metal binding ligands His74, His76, His245, and Asp331 suggested that AtzB may have the ability to bind a metal. The conservation of an additional histidine residue at position 280 further provided evidence that this enzyme likely shared a common amidohydrolase mechanism that utilizes a metal-activated water for hydrolysis.

Influence of Zn(II) concentration in growth medium on AtzB activity.

The isolation of AtzB from different batches of E. coli DH5α(pAAJLS3) cells showed high variability in its activity and metal content. The only significant metals observed were iron and zinc, but their ratios were batch dependent. Summation of the iron and zinc content, however, resulted in a 1:1 stoichiometry with the number of protein subunits. This phenomenon is not uncommon with amidohydrolase superfamily enzymes (20, 21, 32). To alleviate batch-to-batch discrepancies, 500 μM zinc sulfate was added to the bacterial growth medium. The resulting protein maintained 1:1 stoichiometry of zinc to protein subunits, with negligible iron content.

Effect of chelators on AtzB activity.

Fifteen chelators were tested for their ability to inactivate AtzB. Though long incubation times were required, decreased activity was observed with 1,10-phenanthroline, 8-hydroxyquinoline-5-sulfonic acid, dipicolinic acid, and Chelex-100 resins. The chelator which was most effective at reducing activity with the shortest incubation time was 1,10-phenanthroline. After 100 min, 1,10-phenanthroline-treated AtzB showed a 28% loss in activity. The lowest activity occurred with 1,10-phenanthroline after a 24-h incubation, resulting in an 82% loss of activity.

The removal of 1,10-phenanthroline by size exclusion chromatography or dialysis resulted in the restoration of AtzB activity, even in the presence of Chelex-100-treated buffers. Inductively coupled plasma emission spectroscopy analyses revealed that AtzB retained 40 to 70% of its metal content after chelator treatment, suggesting that the 1,10-phenanthroline inactivation was not due to metal removal. Wagner, et al. observed similar results with human liver alcohol dehydrogenase, concluding that an enzyme-metal-chelator complex was formed which inhibited activity by blocking substrate access to the active site (39).

Deamination versus dechlorination.

Previously, AtzB was shown to catalyze both dechlorination and deamination reactions (28). With the substrate 2-amino-4-chloro-6-hydroxy-1,3,5-triazine, dechlorination occurred exclusively, in preference to deamination reactions. Here, we investigated these two different hydrolytic activities in more detail, establishing kinetic parameters for five different substrates by monitoring substrate disappearance (Table 2). The k_cat values for these two types of reactions were not greatly different, with the dechlorination and deamination values overlapping from 0.8 to 3.2 s⁻¹. The K_m values, however, differed by 3- to 12-fold, ranging from 20 to 40 μM for deamination and 120 to 230 μM for dechlorination. This corresponded to k_cat/K_m values that were 1.6- to 23-fold greater for deamination than dechlorination. AtzB is therefore more efficient in catalyzing deamination reactions. Of the deamination reactions examined, AtzB was most efficient in the deamination of hydroxyatrazine. The products of all dechlorination and deamination reactions as determined by HPLC analysis are listed in Table 3.

TABLE 2.

Kinetic values for substrate hydrolysis by AtzB

Kinetic measure	Value for type of reaction on substrate (mean ± SD)
	Deamination			Dechlorination
	IEOT	IMOT	IIOT	CIOT	CEOT
Km (μM)	20 ± 2	40 ± 2	29 ± 3	120 ± 8	230 ± 20
kcat (s−1)	3.2 ± 0.2	1.4 ± 0.1	0.8 ± 0.1	1.9 ± 0.1	1.6 ± 0.1
kcat/Km (mM−1 s−1)	160 ± 20	35 ± 2	28 ± 6	16 ± 1	7 ± 1

Substrate specificity analysis.

The substrate specificity of AtzB was probed using a wide array of s-triazine and pyrimidine compounds. Of the 51 compounds tested, 20 were substrates for AtzB (Table 3). The most salient feature common to all AtzB substrates was the requirement for a monohydroxylated s-triazine ring (Fig. 3). For purposes of discussion, the enzymatic site of oxygen recognition is referred to here as the oxygen binding site.

FIG. 3.

Substrate analogs of hydroxyatrazine. (A) Dechlorination by AtzB: substrate specificity with 2-chloro-4-hydroxy-s-triazines. (B) Deamination by AtzB: substrate specificity with 2-hydroxy-s-triazine analogs. R₁ and R₂ refer to the 4 and 6 positions of the s-triazine ring, in which the analysis of substrates suggests that AtzB requires two amino groups for deamination activity. (C) Tested compounds lacking a hydroxy substituent that were not AtzB substrates. (D) Tested compounds with two hydroxy substituents that were not AtzB substrates.

The majority of the monohydroxylated s-triazine compounds tested were designed to examine the identity, size, and shape of the two other s-triazine ring substituents. N-alkyl alcohols and N-alkyl ether-containing compounds [2-hydroxy-4,6-di(N-hydroxyethylamino)-1,3,5-triazine, 2-(N-ethylamino)-4-hydroxy-6-(N-hydroxyethylamino)-1,3,5-triazine, 2-amino-4-hydroxy-6-(N-hydroxyethylamino)-1,3,5-triazine, and 2-hydroxy-4-(N-isopropylamino)-6-N-(3-methoxypropylamino)-1,3,5-triazine] were substrates; therefore, the side chains are not limited to completely hydrophobic groups. However, compounds containing ether (2,4-dimethoxy-6-hydroxy-1,3,5-triazine) or alkyl (2-amino-4-hydroxy-6-methyl-1,3,5-triazine and 2-amino-4-hydroxy-6-propyl-1,3,5-triazine) groups in place of the amino side chains failed to be substrates. These data suggest that AtzB requires two amino groups in the 4 and 6 positions of the s-triazine ring for deamination activity (Fig. 3B).

A series of N-alkyl compounds were synthesized to identify the size limitations of the N-alkyl side chains. Amino-, N-methyl-, N-ethyl-, N-isopropyl-, N-isobutyl-, N-secbutyl-, and N-tertbutyl-containing compounds were substrates, while linear N-butyl was not. The limited water solubility of compounds containing more carbons than the N-butyl series precluded effective activity measurements. Even though the longest chain of the branched iso- and sec-butyl compounds resembles the linear propyl group, AtzB differentiates between them, accepting the branched N-alkyl-containing compounds as substrates. In addition, AtzB differentiates between straight-chained N-alkyls and N-alkyl ethers like N-propyl-3-methoxy side chains [2-hydroxy-4-(N-isopropylamino)-6-N-(3-methoxypropylamino)-1,3,5-triazine]. Purse and Rebek found that alkyl group conformational changes, such as anti to gauche transitions, occurred readily in response to space limitations (22). These transitions occur if the binding energy supplied by attractions to the container's inner surface compensates for any conformational energetics. This finding might suggest that the AtzB binding site has interactions with the branched N-alkyl and N-alkyl ether side chains to provide additional binding energy to overcome the conformational changes required for binding substrates with longer side chains.

To further investigate the dependence on nitrogen-containing side chains, a series of compounds containing tertiary amine side chains were synthesized. Compounds with two tertiary amine side chains [2,4-Bis(dimethylamino)-6-hydroxy-1,3,5-triazine] were not substrates. When a tertiary amine was in conjunction with a secondary amine, the size of the tertiary amine seemed highly significant. Tertiary amines containing two N-ethyl groups (2-diethylamino-4-methylamino-6-hydroxy-1,3,5-triazine and 2-diethylamino-4-ethylamino-6-hydroxy-1,3,5-triazine) were not substrates, while changing one of these N-ethyl groups to an N-methyl [2-(N-ethyl-N-methylamino)-4-ethylamino-6-hydroxy-1,3,5-triazine] resulted in AtzB activity. The product of AtzB hydrolysis of 2-(N-ethyl-N-methylamino)-4-ethylamino-6-hydroxy-1,3,5-triazine was confirmed by LC-MS analysis. The tertiary N-ethyl-N-methylamino group was the leaving group, showing that tertiary amine leaving groups are acceptable for AtzB, depending heavily on the sizes of the teriary amine groups and the identities of the other ring substituents.

Another compound that was not an AtzB substrate was a choro-hydroxyl-s-triazine containing a tertiary amine in the third position (2-chloro-4-dimethylamino-6-hydroxy-1,3,5-triazine). These data suggest that dechlorination is prevented by having the tertiary amine in the non-oxygen binding, non-leaving group site.

Dihydroxylated s-triazines (substrates for AtzC) and hydroxyatrazine analogs with the hydroxyl group replaced by chloro, fluoro, azido, cyano, methoxy, methylthio, mercapto, amino, and hydride groups (substrate and substrate analogs for AtzA and TrzN) were not substrates for AtzB (Fig. 3C and D). The substrate specificity displayed by each of the enzymes involved in the conversion of atrazine to cyanuric acid suggests that each enzyme evolved to catalyze a consecutive step in the degradation and that a set of three enzymes are required for this transformation to occur.

N-alkyl chain discrimination.

Some s-triazine herbicides, such as simazine [2-chloro-4,6-di(N-ethylamino)-1,3,5-triaizne], undergo dechlorination by AtzA to generate a product that is a symmetrical substrate for AtzB with respect to the N-ethyl groups. However, hydroxyatrazine presents an asymmetric substrate for AtzB, and a previous study indicated a strong preference for displacing the N-ethyl group (6). To more fully delineate the leaving group selectivity of AtzB, monohydroxylated s-triazine substrates bearing two distinct amine substituents [hydroxyatrazine, 4-(N-ethylamino)-6-hydroxy-2-(N-methylamino)-1,3,5-triazine, 6-hydroxy-4-(N-isopropylamino)-2-(N-methylamino)-1,3,5-triazine, 2-hydroxy-4-(N-isopropylamino)-6-N-(3-methoxypropylamino)-1,3,5-triazine, 2-(N-ethylamino)-4-hydroxy-6-(N-hydroxyethylamino)-1,3,5-triazine, 2-amino-4-hydroxy-6-(N-hydroxyethylamino)-1,3,5-triazine, 2-amino-6-(N-ethylamino)-4-hydroxy-1,3,5-triazine, and 2-amino-4-hydroxy-6-(N-isopropylamino)-1,3,5-triazine] were investigated. AtzB removed only one amine per molecule. The selectivity was assessed by allowing reactions to go to completion and measuring the ratio of the two products (Fig. 4). In most cases, AtzB preferentially hydrolyzed the smaller side chain. Based on the substrates tested, there appear to be three categories of substrates: (i) those with virtually complete selectivity (>95% one product), (ii) those with intermediate selectivity (∼80% one product), and (iii) those with modest selectivity (∼60% major product). The first category seems to be characteristic of substrates with small N-alkyl groups (N-ethyl, N-methyl, or amino) on one side and large N-alkyl groups (N-isopropyl or larger) on the other. The second category is comprised of substrates with 2-hydroxy-N-ethylamine substituents with smaller groups on the preferred side, and the third category has substrates with either two small groups or two large groups. These results indicate that both enzyme pockets can accept larger groups but that the catalytic pocket may be more sensitive to size, preferring smaller or more branched groups. The intermediate specificity with 2-hydroxy-N-ethylamine-containing substrates needs to be investigated further, but may be due to a combination of preference for the smaller leaving group and some substrate-protein interaction, causing a higher-than-expected removal of the hydroxylated side chain. This interaction could be a favorable one in the leaving group enzyme pocket or an unfavorable one in the non-oxygen binding, non-leaving group pocket.

FIG. 4.

Selectivity of AtzB for N-alkyl groups removed from asymmetric 2-hydroxy-4,6-di-N-alkyl-1,3,5-triazines. The preferentially removed groups are surrounded by dashed lines.

DISCUSSION

AtzB is the only enzyme known to transform hydroxyatrazine to N-isopropylammelide. Its significance in microbial s-triazine degradation is underscored by its position at the intersection of multiple s-triazine degradative pathways (Fig. 1). Furthermore, unlike AtzA and TrzN, which largely partition between gram-negative and gram-positive bacteria, respectively, AtzB has been found in phylogenetically diverse bacteria (10). In this study, a novel cloning method was used to improve the expression of AtzB, which enabled protein purification and characterization, including a detailed study on substrate specificity. These results provide the first insights into the catalytic features of AtzB.

Previously, the catalytic promiscuity of AtzB was shown by its ability to catalyze both deamination and dechlorination reactions (28). In this study, the dual catalytic activity of AtzB was investigated in greater detail. Kinetic analyses revealed that AtzB is more efficient at catalyzing deamination than dechlorination reactions, with 1.6- to 23-fold greater k_cat/K_m; the largest differences appear to be due to K_m rather than k_cat values.

Although the amidohydrolase superfamily generally catalyzes the hydrolysis of amide and amine bonds, dechlorination reactions are known, both as major reactions and as fortuitous reactions of deaminating enzymes (2, 19, 25, 28, 29). In the case of AtzA and TriA, two proteins that share 98% sequence identity but that have distinct dechlorination and deamination activities, respectively, a single amino acid was identified as being responsible for switching the catalytic activity (29). Triazine hydrolase (TrzA), known to catalyze the deamination of nonalkylated triazines and dechlorination of mono-N-alkylated triazines, has efficiency values that are dependent upon whether or not a substrate is chlorinated. The k_cat for the deamination or dechlorination of any chlorinated substrate (0.16 to 0.45 mM⁻¹ s⁻¹) was an order of magnitude lower than that seen for the deamination of melamine (2,4,6-triamino-1,3,5-triazine) (4.8 mM⁻¹ s⁻¹) (19). Other enzymes, like adenosine deaminase, are known to catalyze physiologically significant deamination reactions but also catalyze fortuitous dechlorination reactions (2, 3). These data suggest that the ability of an amidohydrolase superfamily enzyme to catalyze both dechlorination and deamination reactions and the efficiency of each of these reactions is enzyme specific and not a property of the superfamily.

Sequence analyses and experimental evidence confirmed that AtzB is a metalloprotein belonging to the highly diverse amidohydrolase superfamily. Superfamily analysis provided insights into the three-dimensional structure and catalytic features of AtzB. Superfamily members, generally, have a conserved (βα)₈ barrel structure that binds one or two divalent metal ions at the C-terminal end of the barrel (13). The metal serves to activate water for nucleophilic attack on the substrate. The experimental results presented here indicate that there is a mononuclear active site in AtzB, similar to sites in other s-triazine-degrading and nucleotide-processing enzymes like adenosine deaminase and cytosine deaminase (15, 27, 31, 32, 40). AtzB maintains the conserved amidohydrolase superfamily metal binding ligands and active site residues at His74, His76, His245, His280, and Asp331, further supporting a superfamily-type mechanism. The crystal structures of mononuclear amidohydrolase enzymes, like adenosine deaminase and cytosine deaminase, show a conserved glutamate located three residues downstream of the third histidine metal binding ligand, which serves to form a hydrogen bond to a ring nitrogen, assisting in nonaromatic intermediate stabilization (14, 40). This glutamate is likewise conserved in the triazine-degrading enzymes: Glu 248 in AtzB, Glu246 in AtzA and TriA, Glu222 in TrzC, and Glu241 in TrzN.

Extensive substrate studies with synthetic substrates allowed us to further develop our understanding of AtzB substrate specificities. The identities of the substituents in the 2, 4, and 6 positions of the s-triazine ring substrate are constrained. The first position requires a single hydroxyl group on the s-triazine ring. This site may serve to assist in substrate orientation for catalysis. The second position involves the identity of the leaving group, putatively close to the metal ligands and active site residues mentioned above. Substrate studies revealed that primary, secondary, or tertiary amines or a chloride are acceptable leaving groups. The third position, which we term the non-oxygen binding, non-leaving group site, will tolerate primary or secondary amines but not tertiary amines or ether or alkyl groups. These results suggest that this third position requires a nitrogen with a minimum of one proton for deamination reactions. Furthermore, neither the chloride nor the tertiary amine group was hydrolyzed by AtzB when in combination in the analog 2-chloro-4-dimethylamino-6-hydroxy-1,3,5-triazine, even though both groups independently are potential leaving groups. Neither nonoxygen side chain contains a proton which may potentially be involved in orientation in the non-oxygen binding, non-leaving group site, suggesting that the dechlorination and deamination substrates have similar requirements for substrate-enzyme interactions.

The results presented here also provide the first glimpse into the AtzB active site. From this evidence, we propose at least four different potential interactions between the AtzB protein and its substrates. The first involves recognition of the hydroxyl group on the s-triazine ring. The second is the requirement for a proton on the amino substituent in the non-oxygen binding, non-leaving group site. The requirement for a proton on this nitrogen might have been due to a need for the substrate to form a hydrogen bond to the protein, substrate tautomerization, or due to size limitations for branching at this point in the structure. The lack of activity toward compounds with an ether or alkyl substituent in this position points toward hydrogen bonding. The third is proton donation by Glu248 to a ring nitrogen adjacent to the leaving group, stabilizing a tetrahedral carbon transition state. The fourth is a potential proton donor in the active site to facilitate leaving group protonation. Further research is required to further substantiate these protein-substrate interactions and to identify the exact residues involved.