N-Acetylanthranilate Amidase from Arthrobacter nitroguajacolicus Rü61a, an α/β-Hydrolase-Fold Protein Active towards Aryl-Acylamides and -Esters, and Properties of Its Cysteine-Deficient Variant

Bacterial strains, media, and growth conditions.The strains and plasmids used in this study are listed in Table 1. For the isolation of genomic DNA, A. nitroguajacolicus Rü61a was grown in mineral salts medium with 8 mM sodium benzoate as the source of carbon and energy at 30°C (52). The same mineral salts medium with 2 mM propanil, and with 2 mM propanil and 2 mM 1H-4-oxoquinaldine, was used to assess utilization and cometabolic conversion of propanil. E. coli strains were grown in lysogeny broth (62) at 37°C. E. coli DH5α (25) harboring pWFAAM and derivatives and E. coli M15(pREP4)(pWFAAM) strains producing His₆Amq or His₆Amq protein variants were grown in the presence of ampicillin (100 μg/ml) and ampicillin (100 μg/ml) plus kanamycin (25 μg/ml), respectively. Expression of amq was induced by addition of 0.1 mM isopropyl-β-d-thiogalactopyranoside (IPTG) at an optical density at 600 nm of 0.9 to 1.0. Cells were harvested after 3 h of induction at an optical density at 600 nm of about 3.5 by centrifugation at 10,000 × g and 4°C for 10 min. The yield of wet biomass of E. coli M15(pREP4)(pWFAAM) from a 5-liter bioreactor was about 20 g.

DNA isolation and gene cloning.Total DNA of A. nitroguajacolicus Rü61a was isolated according to Rainey et al. (57). Plasmid DNA was isolated with the E.Z.N.A. Plasmid Miniprep kit (peqlab, Erlangen, Germany). Gel extraction of DNA fragments from agarose gels was performed with the Perfectprep gel cleanup kit (Eppendorf, Hamburg, Germany). For cloning purposes, DNA fragments were purified with the High Pure PCR Product Purification kit (Roche, Grenzach-Wyhlen, Germany). Standard protocols were used for agarose gel electrophoresis, restriction digestion, and DNA ligation (63). The amq gene (accession no. AJ537472; nucleotides 13840 to 14721) was amplified by PCR with Pfu polymerase (Promega, Mannheim, Germany) using total DNA of A. nitroguajacolicus Rü61a as template and the primers 5′-GCGGTACCGGGGTGCGAGGGGTTCACCAGAT-3′ (forward) and 5′-ACGACAGGACGGGAGGACACCGCCACGAGG-3′ (reverse). The PCR product was ligated into the StuI/KpnI restriction sites of pQE-30 Xa (QIAGEN, Hilden, Germany), generating pWFAAM. Correct insertion was verified by sequencing the insert as well as flanking regions (MWG Biotech, Ebersberg, Germany). Competent cells of E. coli DH5α and E. coli M15(pREP4) were generated as described by Hanahan (27). For production of recombinant His₆Amq protein or variants thereof, E. coli M15(pREP4) was transformed with pWFAAM or derivatives isolated from E. coli DH5α clones.

Site-directed mutagenesis.Site-directed mutagenesis was performed according to the protocol of the QuikChange site-directed mutagenesis kit (Stratagene, Heidelberg, Germany) using pWFAAM as template, Pfu polymerase for amplification, and the mismatch primer pairs listed in the table in the supplemental material. Sequencing of the inserts and flanking regions of all mutant plasmids was performed by MWG Biotech.

Preparation of cell extracts and purification of recombinant His₆Amq proteins.Twenty grams of cells (wet biomass) of E. coli M15(pREP4) containing pWFAAM (or derivatives) was suspended in 40 ml 50 mM sodium phosphate buffer containing 10 mM imidazole and 300 mM NaCl (pH 8.0). Crude extract containing soluble proteins was obtained by sonication (sonifier UP200s; Dr. Hilscher GmbH, Stuttgart, Germany) and subsequent centrifugation at 40,000 × g and 4°C for 40 min. The supernatant was applied to an Ni-Sepharose High Performance column (GE Healthcare Europe, Munich, Germany; 10-ml bed volume in Bio-Scale MT10 column from Bio-Rad Laboratories, Munich, Germany), equilibrated in 50 mM sodium phosphate buffer containing 10 mM imidazole, 3 mM dithiothreitol (DTT), and 300 mM NaCl (pH 8.0). After washing with the same buffer, His₆Amq was eluted with a linear gradient (30 ml) of 20 to 250 mM imidazole in equilibration buffer. Fractions containing His₆Amq were pooled and washed in 40 mM Britton-Robinson (BR) buffer (58) (pH 7.5; 12.5% [vol/vol] glycerol added) by ultrafiltration (Vivaspin 20; molecular weight cutoff, 10,000; Vivascience, Hannover, Germany). The concentrated protein solution was diluted in 50 mM Tris-HCl buffer (pH 8) and loaded onto an UNOsphere Q column (2-ml bed volume; Bio-Rad) equilibrated in the same buffer. After a washing step, His₆Amq was eluted in a linear gradient (32 ml) from 0 to 300 mM NaCl in 50 mM Tris-HCl buffer (pH 8.0). Fractions containing the enzyme were washed and concentrated in BR buffer (pH 7.5, with 12.5% [vol/vol] glycerol) by ultrafiltration and stored at −20°C. Protein samples for CD spectroscopy were dialyzed for 15 h against 10 mM sodium phosphate buffer, pH 7.5.

For enrichment of His₆AmqS155A, His₆AmqE235A, and His₆AmqH266A, 0.1 mM 4-(2-aminoethyl) benzenesulfonyl fluoride hydrochloride (AEBSF) was added to the cell suspension prior to sonication.

Removal of the His₆ tag.The N-terminal His₆ tag of His₆Amq was removed by treatment with Factor Xa protease (QIAGEN) according to the manufacturer's instructions (The QIAexpressionist; QIAGEN GmbH, Hilden, Germany). Optimal cleavage without formation of unspecific degradation products was achieved by incubation of 10 μg His₆Amq (0.25 μg/μl) with 0.2 U Factor Xa protease for 2.5 h at room temperature. The His₆ tag and residual His₆Amq was removed by loading the reaction mix on an Ni-nitrilotriacetate Spin Column (QIAGEN) and centrifugation at 8,000 × g.

Enzyme assays and kinetics.Stock solutions of potential substrates (50 mM) and their hydrolysis products were prepared in ethanol. Since incubation of His₆Amq (0.11 μM) with up to 5 mM EDTA did not inhibit its activity, the amidase assays were routinely performed with buffer containing 100 μM EDTA to protect the enzyme from divalent heavy-metal cations. In the spectrophotometric standard assay, carried out in BR buffer (40 mM, pH 8) with 100 μM EDTA at 25°C, enzyme-catalyzed formation of anthranilate from N-acetylanthranilate (5 mM) was measured at 325 nm (ε₃₂₅ = 2.012 mM⁻¹ cm⁻¹). In preliminary tests for His₆Amq-catalyzed conversion of other (aromatic) amides and esters, UV/Vis spectra (250 to 600 nm) were taken at appropriate time intervals of an assay mixture that contained 5 mM of the potential substrate and 0.06 μM of His₆Amq. Quantitative spectrophotometric determinations of products generated by enzyme-catalyzed hydrolysis of aryl-acylamides or -esters were based on the following molar extinction coefficients (in assay buffer): 2-nitroaniline, ε₄₂₀ = 4.438 mM⁻¹ cm⁻¹; 2-chloroaniline, ε₂₉₀ = 1.670 mM⁻¹ cm⁻¹; 2-bromo-4-methylaniline, ε₂₉₅ = 1.966 mM⁻¹ cm⁻¹; aniline, ε₂₈₅ = 1.152 mM⁻¹ cm⁻¹; phenol, ε₂₇₀ = 1.068 mM⁻¹ cm⁻¹; p-nitrophenol, ε₄₀₀ = 14.847 mM⁻¹ cm⁻¹; salicylic acid, ε₂₉₈ = 3.162 mM⁻¹ cm⁻¹; 3,4-dichloroaniline, ε₃₀₅ = 1.99 mM⁻¹ cm⁻¹. One unit was defined as the amount of enzyme that catalyzes the formation of 1 μmol of product in 1 min under the conditions described. For the determination of kinetic constants K_m and k _cat, concentrations of 1 to 5 mM were used for all substrates except phenylacetate and p-nitrophenylbutyrate, which were assayed at 0.05 to 2 mM and 0.05 to 1.5 mM, respectively. Apparent kinetic constants were deduced from Hanes plots (28). Assays were done at least in triplicate.

The activity of His₆Amq was assessed in the presence of some organic solvents. Their partition coefficient log P was calculated by the java applet Marvin with default values (ChemAxon Ltd.; http://intro.bio.umb.edu/111-112/OLLM/111F98/newclogp.html ).

Hydrolysis of N-acetylated amino acids (5 mM) by His₆Amq (0.06 μM) was measured as release of acetic acid in the phenol red assay described by Holloway et al. (32). This assay was sensitive enough to detect the release of 0.4 mM acid from 5 mM phenylacetate. Hydrolysis of primary amides (5 mM) by His₆Amq (0.06 μM) was examined by determining release of ammonia (69).

To assess the potential of His₆Amq for hydrolysis of the natural polyamide cyanophycin, a suspension of cyanophycin was treated with 6 μM enzyme in BR buffer and degradation products carrying amino groups were determined by high-performance liquid chromatography (HPLC) and fluorescence detection after precolumn derivatization with o-phthaldialdehyde, as described in Aboulmagd et al. (1). Proteolytic activity of His₆Amq was tested by incubation of the protein (0.6 μM and 1.2 μM) with a suspension of azocoll (1.5 mg) for 30 min at 30°C and subsequent spectrophotometric detection of soluble dye-coupled peptides in the supernatant at 520 nm (13).

To test for triacylglycerol hydrolase activity, 10 μl of His₆Amq (67 μM) was applied on a tributyrin agarose plate (1% [wt/vol] agarose, 1% [vol/vol] Tween 80, 0.25% [vol/vol] tributyrin in BR buffer, pH 8) and incubated for 24 h at 25°C. Application of o-nitroacetanilide (5 mM in BR buffer, pH 8) to the tributyrin plate after the incubation period resulted in formation of yellow 2-nitroaniline by His₆Amq, indicating that the enzyme was still functional. As a positive control, 2.5 μM of an esterase (accession no. AJ537472; ORF7) active towards tributyrin was applied to the same plate, which formed a spot of clearance in the opaque layer.

Protein analysis.Protein concentrations were estimated by the Bradford method as modified by Zor and Selinger (72). For CD spectroscopy, concentrations of electrophoretically pure His₆Amq and His₆AmqC22A/C63A were deduced from the theoretical molar extinction coefficients at 280 nm, calculated according to Pace et al. (51): ε₂₈₀ = 1.278 liter g⁻¹ cm⁻¹ for His₆Amq and ε₂₈₀ = 1.277 liter g⁻¹ cm⁻¹ for the cysteine-deficient protein. Denaturing sodium dodecyl sulfate-polyacrylamide gel electrophoresis (SDS-PAGE) was performed as described by Laemmli (39), using an overall acrylamide concentration of 10.8% and a cross-linker concentration of 2.6% in the separating gels. For nondenaturing (native) PAGE, the Laemmli method was used, omitting SDS from the buffers. Polyacrylamide gels were stained with Coomassie blue R-250 (0.1% [wt/wt] Coomassie blue R-250, 50% [wt/wt] trichloroacetic acid in H₂O) and destained in an aqueous solution of 30% (vol/vol) methanol and 10% (vol/vol) acetic acid. Transfer of proteins (67) from gels to polyvinylidene fluoride membranes (Carl Roth, Karlsruhe, Germany) was performed according to the protocol of QIAGEN (QIAexpress). Immunodetection of His₆-tagged proteins on blots was performed using primary antibodies [anti-(His)₅ mouse immunoglobulin G1; QIAGEN], secondary antibodies (horseradish peroxidase-conjugated sheep anti-mouse immunoglobulin G; Chemicon International), and p-nitrotetrazolium blue and 5-bromo-4-chloro-3-indolyl phosphate for colorimetric detection (QIAexpress protocol; QIAGEN).

Gel filtration of His₆Amq for determination of its native molecular mass was performed on a Bioprep SE-1000/17 column (Bio-Rad Laboratories) in 50 mM sodium phosphate buffer containing 150 mM NaCl (pH 7.5). A gel filtration standard marker from Bio-Rad was used for calibration of the column.

Shelf stability of His₆Amq and pH and temperature optima of activity.Stability of His₆Amq was measured in 20 mM Tris-HCl, 50 mM sodium phosphate, and 40 mM BR buffer in a pH range of 6.0 to 8.0 at temperatures of −20°C, 4°C, and 25°C. Residual enzyme activity was determined in the standard enzyme assay with aliquots of 0.06 μM His₆Amq. The activity of His₆Amq at different pHs was tested in the same buffers (Tris-HCl, pH 7 to 9; sodium phosphate, pH 6.2 to 8.2; BR, pH 4.5 to 11). Temperature dependence of His₆Amq activity was determined in the buffer of the standard assay.

Chemical modification of amino acid residues and potential effectors of His₆Amq.Since conserved serine and histidine residues of Amq were hypothesized to be catalytically relevant, the influence of phenylmethylsulfonyl fluoride (PMSF) and diethyl pyrocarbonate (DEPC) on enzyme activity was measured in concentrations of 0 to 70 μM and 0 to 2 μM, respectively. Additionally, the effect of HgCl₂ (0 to 0.2 μM) was tested. Samples of His₆Amq (6 μM) in 40 mM BR buffer (without EDTA) were preincubated with potential inhibitor for 10 min at room temperature, and residual activity towards N-acetylanthranilate was measured spectrophotometrically in the same buffer using aliquots of 0.06 μM His₆Amq.

The influence of EDTA (5 μM to 30 mM) and DTT (1 to 50 mM) on His₆Amq activity was analyzed in the standard enzyme assay, omitting EDTA from the BR buffer. To assess the effect of divalent metal cations, N-acetylanthranilate amidase activity was determined after incubation of 0.11 μM His₆Amq with metal salts (10 and 100 μM) in BR buffer (40 mM, pH 8, without EDTA) for 5 min.

CD spectroscopy.CD spectra were recorded using a Jobin-Yvon (Paris, France) Spectropolarimeter model CD6 equipped with a peltier-thermostated cell holder, which was constructed by the machine shop of the Institut für Physikalische Chemie of the University of Münster. Temperature was controlled to a precision of ±0.1 K. Quartz cells of 0.01-cm optical path length were used with protein concentrations between 0.8 and 2.0 mg/ml; a path length of 0.1 cm was employed with concentrations between 0.1 and 0.8 mg/ml. The results were expressed as mean residue ellipticity (MRE): [Θ]_MRE = (MRW θ_obs)/(cd), where θ_obs is the observed ellipticity at the respective wavelength, MRW is the mean residue weight of the protein (108,852 g/mol for His₆Amq and 108,647 g/mol for His₆AmqC22A/C63A), d is the optical path length of the cell, and c is the specific concentration (in milligrams/milliliter) of the samples. Deconvolution of CD spectra was done using the programs K2D of Andrade et al. (8) and CDNN of Böhm et al. (11).

Amino acid sequence analysis.BLASTP analysis (7) of the Amq protein (293 amino acids [aa], calculated molecular mass of 32.8 kDa) revealed significant sequence similarity to putative esterases from Bordetella spp. (CAE40678; 39% aa identity, 47.7% aa similarity) and Ralstonia eutropha JMP134 (46) (AAZ64529; 32.7% aa identity, 41.6% aa similarity) and a hypothetical protein from Pseudomonas sp. strain CA10 (BAB32459.1; 39.9% identity). Other related proteins include, for example, carboxylesterase from Xanthomonas axonopodis pv. citri 306 (15) (AAM37245; 26% aa identity) and kynurenine formamidase from Mus musculus (50) (AAM62284; 28% aa identity).

Secondary structure prediction (60, 61) for Amq (Fig. 2) suggested a pattern of β-strands and α-helices which matches that of the “canonical” α/β-hydrolase fold (29, 45, 47). Using core regions of bacterial heroin esterase (Her) (1LZL_A, 1LZK_A) and esterase Est2 from Alicyclobacillus acidocaldarius (1U4N_A, 1EVQ_A, 1QZ3_A) as templates, which show a degree of sequence identity to the corresponding region of Amq (aa 71 to 175) of 40.15% (Her) and 35.85% (Est2), the program ProModII (64) calculated a three-dimensional model for this region of Amq. Validation analysis with WHATCHECK (33) revealed an exceedingly good Ramachandran plot appearance Z-score of −0.175. The three-dimensional model comprises the elements β3, αA, β4, αB-αB′, β5, and αC (Fig. 2) and thus supports the secondary structure prediction for this central segment.

• Open in new tab
• Download powerpoint

FIG. 2.

Multiple alignment of Amq (EMBL accession no. CAD61042) and related proteins, performed with the Clustal W algorithm (2). COest, carboxylesterase from Xanthomonas axonopodis pv. citri strain 306 (15) (AAM37245); KFase, kynurenine formamidase from Mus musculus (50) (AAM44406); RAest, hypothetical esterase from Ralstonia eutropha JMP134 (46) (AAZ64529). Residues conserved throughout are marked with an asterisk, while residues marked with a colon and dot indicate conserved and semiconserved substitutions, respectively. Residues of the putative catalytic triad are highlighted in gray and are marked with a diamond. The conserved motif (G-X-S-X-G-G/A) surrounding the catalytic nucleophile of the α/β hydrolase-fold enzymes is enclosed in a box. α-Helical regions and β-strands predicted for Amq by the program PredictProtein (61) are marked with wavy shaded boxes and large white arrows, respectively. Amino acid sequences of Amq that are underlined with continuous and broken lines indicate α-helices and β-strands, respectively, as shown in the three-dimensional model of the core region of Amq (aa 71 to 175, framed by vertical arrows), calculated with ProModII 3.70 (64).

The catalytic residues of the α/β-hydrolases constitute a highly conserved catalytic triad consisting of a nucleophile, an acidic residue, and an absolutely conserved histidine. The nucleophile is located at the top of a sharp turn, called the nucleophile elbow, which is characterized by the consensus sequence Sm-X-Nu-X-Sm-Sm (Sm, small residue; X, any residue; and Nu, nucleophile) (47); this motif is clearly conserved in Amq (¹⁵³GSSAGG¹⁵⁸). The signature GxSAG, found in acetylcholinesterase and butyrylcholinesterase, is thought to be typical for many carboxylesterases (54). The presumed nucleophile of Amq (S155) is predicted to be situated between β5 and αC (Fig. 2), and the model calculated with ProModII positions its side chain at the apex of a sharp turn, consistent with the canonical fold (29, 45, 47). If E235 and H266 of Amq represent the other triad residues (Fig. 2), Amq shares with acetylcholine esterases and p-nitrobenzyl esterase of Bacillus subtilis (71) the use of glutamate instead of the more frequent aspartate as the active-site carboxylate of serine hydrolases. The topological positions of E235 and H266 after predicted strand β7 and between β8 and helix αF, respectively, match the criteria of the α/β-hydrolase fold (29, 45, 47).

The oxyanion hole of most α/β-hydrolases comprises two residues which donate their backbone amide protons to stabilize the negative charge of the transition state. One residue is located adjacent to the nucleophile (Nu + 1), whereas the other is situated in a turn between β3 and αA in the canonical fold (10, 47). Amq, like many α/β-hydrolases, contains an HGX motif (⁸²HGG⁸⁴) in this region (10, 17, 54) and thus belongs to the “GX-type hydrolases,” with X presumed to be the second oxyanion hole residue; the “GX class” of hydrolases comprises many esterases and lipases. In contrast, acetylcholine esterases and other carboxylesterases as well as members of the hormone-sensitive lipase/esterase family belong to the “GGGX class”; in Torpedo californica acetylcholinesterase, the peptidic NH groups of G118 and G119 together with the backbone NH group of the Nu + 1 residue A201 actually form a tridentate oxyanion hole (48). The GGGX motif forms a flexible “glycine loop” which is involved in determining the active-site architecture and thus its reactivity towards substrates and inhibitors (48). However, in the absence of structural data, it is unclear whether the ⁸²HGGY⁸⁵ motif of Amq might also represent a short glycine loop.

A search for conserved domains (41) in Amq revealed significant similarity with part of the esterase/lipase domain cd00312 (aa 58 to 158 of Amq, 23.3% aligned). This domain family includes large enzymes with aryl-acylamidase/-esterase activity, namely, acetylcholinesterases, p-nitrobenzyl hydrolase (71), and phenylcarbamate hydrolase from Arthrobacter oxydans P52 (55). Amq indeed shows similarity to acetylcholinesterase (e.g., from Torpedo californica), especially in the region surrounding the nucleophile elbow; however, acetylcholinesterase has a large insertion in the region between β6 and αD, and its C-terminal region (aa 434 to 586 of P04058) does not align with Amq. A segment comprising aa 75 to 202 of Amq is similar to part of a conserved domain of the Aes (acetylesterase) family (COG0657.1; esterase/lipase; 40.1% aligned), and parts of the related consensus sequences of pfam00135 (carboxylesterases) and COG2272 (type B carboxylesterase) also align with a region of Amq (aa 58 to 163 and aa 35 to 168, respectively). Since the physiological role of Amq in strain Rü61a clearly is that of an amidase involved in the quinaldine degradation pathway, catalyzing the hydrolysis of N-acetylanthranilate to anthranilate and acetate, its marked similarity with esterases is remarkable and prompted us to study its substrate specificity.

Purification of His₆Amq and effect of the His₆ tag on amidase activity.Highest yields of His₆Amq from E. coli M15(pREP4)(pWFAAM) were obtained from cells grown at 37°C and harvested after 3 h of induction with IPTG. From the soluble fraction of cell extract, His₆Amq was purified 74-fold in a two-step procedure to electrophoretic homogeneity (Table 2; Fig. 3A and B). Detection in Western blots of His₆-tagged proteins showed only small amounts of protein in the insoluble fraction of cell extract (redissolved in 10% SDS). A molecular mass of 34 kDa was deduced from SDS-PAGE, which matches the calculated mass of the tagged protein (313 aa) of 34,070 Da. Gel filtration indicated a molecular mass of 32.2 kDa, suggesting that His₆Amq is a monomeric protein in its native state.

• Open in new tab
• Download powerpoint

FIG. 3.

Purification of His₆Amq from E. coli M15(pREP4)(pWFAAM) and electrophoretic properties of His₆Amq and His₆AmqC22A/C63A. (A) SDS-PAGE and (B) corresponding Western blot of proteins after each purification step. Lane M, protein standard; lane 1, crude extract from E. coli M15(pREP4)(pWFAAM) (30 μg protein); lane 2, His₆Amq after Ni²⁺ chelate affinity chromatography (5 μg protein); lane 3, His₆Amq after anion exchange chromatography (5 μg protein). (C) Western blot of His₆Amq after Ni²⁺ chelate affinity chromatography in the absence (lane 4) and presence (lane 5) of DTT (20 μg protein in each lane). (D) SDS-PAGE of His₆AmqC22A/C63A (lane 6, 8 μg protein) and His₆Amq (lane 7, 10 μg protein) in the absence of DTT; the arrow indicates dimers of His₆Amq.

Removal of the His₆ tag by treatment with Factor Xa protease and trapping of the tag and of residual uncleaved His₆Amq yielded electrophoretically pure Amq protein. His₆Amq and Amq showed the same K_m s for N-acetylanthranilate, but the apparent k _cat of Amq was about 65% of that of His₆Amq (see Table 5), resulting in a catalytic efficiency k _cat/K_m of 31 mM⁻¹ s⁻¹ for His₆Amq and 20 mM⁻¹ s⁻¹ for Amq. Very similar results were obtained for K_m and k _cat of Amq and His₆Amq for the ester acetylsalicylic acid. The apparent K_m was hardly affected, whereas k _cat of Amq was 58% of that of His₆Amq (see also Table 5). Such decrease in apparent k _cat of Amq may well be caused by the conditions during Factor Xa protease treatment of His₆Amq, which involves protease digestion for 2.5 h in 20 mM Tris-HCl (pH 6.5) at 23°C. Incubation of His₆Amq in this buffer for the same time period resulted in significant loss of specific activity.

Shelf stability and conditions for activity of His₆Amq.Among the buffers, pHs, and temperatures tested, stability of His₆Amq was highest in BR buffer (pH 7.5) at −20°C. Storage for 160 h in this buffer at −20°C retained 91% of activity, whereas losses of 58% and 55% of activity were observed when the enzyme was kept frozen for the same time in 50 mM Tris-HCl (pH 7.5) and 50 mM sodium phosphate buffer (pH 7.5), respectively. Glycerol (12.5%) was added to the storage buffer for long-term freezing; a negative impact of glycerol on enzyme stability and activity was not observed. Highest activity of His₆Amq was measured in BR buffer at pH 8.0 and at a temperature of 30°C (Fig. 4). In Tris buffer (20 mM Tris/HCl, pH 8.0) in the absence of reducing agents, purified His₆Amq existed in two monomeric forms with different electrophoretic mobilities, and it formed dimers (Fig. 3C and D). When preparations after Ni²⁺ affinity chromatography were concentrated in Tris buffer in the absence of DTT, even multimers were observed (Fig. 3C, lane 4). Treatment with DTT eliminated both the second monomeric form and the multimers (Fig. 3C, lane 5), suggesting that they resulted from oxidative formation of intra- and intermolecular disulfide bonds, respectively. Notably, formation of additional protein forms due to disulfide bond formation and increase in activity by addition of DTT were not observed for His₆Amq stored in the BR buffer. The Amq protein contains only two cysteine residues, which were both replaced by alanine. In contrast to His₆Amq, the His₆AmqC22A/C63A protein was stable as a single monomeric form (Fig. 3D, lane 6).

• Open in new tab
• Download powerpoint

FIG. 4.

Effects of pH (A) and temperature (B) on His₆Amq activity towards the physiological substrate N-acetylanthranilate.

Amidase activity was significantly influenced by the presence of organic solvents. For instance, relative residual activities of 4% and 46% were observed in the presence of 5% (vol/vol) trichloromethane (log P = 1.61) and 5% acetonitrile (log P = −0.33), respectively. Dimethyl sulfoxide (log P = −1.43) and dimethyl formamide (log P = −0.64) had a less drastic effect (residual activities of 90% and 72%, respectively, at 5% solvent concentration). A clear correlation between log P values and activity of His₆Amq was, however, not obvious (38).

Putative catalytic triad residues of Amq.To assess the hypothesis of a putative catalytic triad, PMSF and DEPC were used to modify serine and histidine, respectively, and the conserved residues S155, E235, and H266 were replaced by site-directed mutagenesis. Both PMSF and DEPC at micromolar concentrations decreased the apparent V _max of His₆Amq without changing its apparent K_m , suggesting partial irreversible inactivation of the enzyme (data not shown). It is interesting to note that the serine hydrolase inhibitor AEBSF in concentrations up to 0.5 mM did not affect the activity of His₆Amq.

Replacement of the potential triad residues (Fig. 2) by alanine resulted in drastically decreased yields of protein in cell extracts. Addition of the protease inhibitor AEBSF to the cell suspension prior to sonication was necessary to prevent loss of the proteins. SDS-PAGE and Western blot analysis of both soluble and insoluble fractions of the cell extracts after resuspension in 10% SDS indicated that the respective recombinant E. coli cells contained only small amounts of His₆AmqH266A, His₆AmqE235A, or His₆AmqS155A protein in the soluble fractions. The immunodetection of His₆-tagged proteins on Western blots failed to show any signal for the proteins of the insoluble fractions. We thus assume that loss of these protein variants is due to their high susceptibility to proteolytic degradation rather than formation of inclusion bodies. Some enrichment by Ni²⁺ affinity chromatography was possible, but subsequent anion exchange chromatography led to complete loss of the proteins. In standard assays containing up to 0.17 mg of protein prepared by metal chelate chromatography, N-acetylanthranilate amidase activity of the His₆AmqE235A and -S155A preparations was below the detection limit of the spectrophotometric assay (more than 212-fold decrease in activity). Replacement of H266 resulted in major reduction but not abolition of activity. A residual activity of 0.11 U/mg was detected in the preparation of His₆AmqH266A, which contained many contaminating proteins. Such residual activity was unexpected, since substitution of the active-site histidine of α/β-hydrolases generally resulted in inactive proteins, consistent with the role of this residue as an essential catalytic base (4, 9, 17, 24, 35, 44, 53, 56, 65). However, reports on protein variants of C-C hydrolase MhpC (40) and Agrobacterium radiobacter epoxide hydrolase (59) show that the observation of minor residual activity after replacement of the triad's histidine is not unprecedented. Such residual activity may be due to specific base catalysis by solvent OH⁻ (40).

Role of cysteine residues of Amq.As shown in Table 3, incubation of His₆Amq with an excess of Zn²⁺, Hg²⁺, and Cd²⁺ resulted in drastic decreases in N-acetylanthranilate amidase activity. These effects were reversible by EDTA. When Amq activity was measured in the presence of micromolar concentrations of HgCl₂ (without EDTA), the apparent V _max of the enzyme was decreased while its apparent K_m was not affected, suggesting partial inactivation of the enzyme. Since Cd²⁺ and Hg²⁺ preferably bind to sulfhydryl groups, this observation might suggest that cysteine is involved in catalysis. However, since sequence analysis of Amq predicted an α/β-hydrolase-fold protein with a canonical catalytic triad, the potential role of cysteine in substrate turnover was enigmatic. Actually, inhibition of His₆AmqC22A/C63A (0.025 to 0.05 μM) by 10 and 100 μM Hg²⁺, Cd²⁺, and Zn²⁺ was similar as observed for His₆Amq (data not shown), indicating unspecific inactivation rather than group-specific cysteine modification.

Substrate specificity of Amq.Release of ammonia from the primary amides tested (acetamide, propionamide, acrylamide, benzamide, and nicotinamide) and hydrolysis of urea or thiourea were not observed after incubation of 1 U (N-acetylanthranilate amidase activity) of His₆Amq with up to 25 mM of potential substrate for 18 h at 30°C. Ammonia quantification would have been sensitive enough to detect a relative activity of 0.18% of N-acetylanthranilate amidase activity. In contrast to acetanilides with an acidic or polar substituent in ortho position (see below), conversion of the following anilides was not detected: o-aminoacetanilide, m-nitroacetanilide, m-hydroxyacetanilide (3-acetamidophenol), p-nitroacetanilide, and benzanilide. Compounds that also were not hydrolyzed include N-phenyl urea, 1,1-carbonyldiimidazol, the benzoate derivatives benzohydroxamate and hippurate, the lactams penicillin G, ampicillin, and ε-caprolactame, the cyclic compounds succinimide, barbiturate, 5-bromouracil, and imidazolidine-2-thione, the polyamide cyanophycin, the chromogenic protein azocoll, and the ester acetylcholine chloride. Hydrolysis of tributyrin was not evident in a qualitative plate assay. Release of acid from N-acetylated amino acids (N-acetyl-glycine, N-acetyl-dl-leucine, N-acetyl-dl-glutamate, N-acetyl-dl-valine, N-acetyl-dl-alanine, N-acetyl-dl-proline, N-acetyl-dl-tryptophan, N-acetyl-dl-serine, N-acetyl-dl-phenylalanine, N-acetyl-l-cysteine, and N-acetyl-dl-methionine) also was not observed.

The activity of Amq towards N-acetylanthranilate and other anilides suggested that acylanilide pesticides also might be accepted as substrates. Actually, the herbicide propanil (3′,4′-dichloropropionanilide) was hydrolyzed by His₆Amq with a specific activity of 0.95 U/mg, i.e., about 1% of N-acetylanthranilate amidase activity. However, A. nitroguajacolicus Rü61a neither grows on nor cometabolically converts propanil. Hydrolysis of propachlor (2′-chloro-N-isopropylacetanilide), which carries a tertiary amine, was not detected when up to 0.3 μM of His₆Amq was used in the phenol red assay or in UV/Vis spectral analyses. Hydrolysis of the urethane herbicides propham (isopropyl phenylcarbamate), carbaryl (1-naphthyl-N-methylcarbamate), and isoprocarb (2-isopropylphenyl-N-methylcarbamate) likewise was not observed.

With the substrates tested, His₆Amq showed typical Michaelis Menten kinetics. Its kinetic parameters for a series of secondary amides and esters are shown in Table 4. Replacement of the acetyl group of the physiological substrate by a formyl moiety (N-formylanthranilate) increased the K_m about 10-fold (while increasing k _cat by a factor of about 2). The aryl-acetylester p-nitrophenylacetate was preferred over the corresponding butyrate ester. Thus, Amq may be described as an aryl-acetylamidase/-esterase. Among the acetanilides, Amq converts derivatives with a polar (o-nitroacetanilide, o-chloroacetanilide) or acidic (N-acetylanthranilate) substituent in ortho position with high activity. Lack of activity towards o-aminoacetanilide suggests that a basic substituent in ortho position prevents conversion. Acetanilide derivatives monosubstituted in meta or para position (m- and p-nitroacetanilide) were not hydrolyzed, whereas the disubstituted compound 2′-bromo-4′-methylacetamide was converted, albeit with low efficiency (K_m and k _cat of His₆Amq of about 12 mM and 9.5 s⁻¹, respectively). Among the amides listed in Table 4, hydrolysis of the nonsubstituted acetanilide was catalyzed with lowest efficiency, mainly due to a k _cat of only 0.8 s⁻¹. Remarkably, the corresponding ester phenylacetate was converted best, with a very low K_m and high k _cat and an almost sevenfold-increased catalytic efficiency compared with turnover of the physiological substrate. The notion that Amq prefers different substitution patterns for aryl-acetylamides and aryl-acetylesters is supported by the kinetic parameters observed for para-substituted substrates. While p-nitroacetanilide was not converted, hydrolysis of p-nitrophenylacetate was catalyzed at a high rate (Table 4).

The preferences of the protein variant His₆Amq-C22A/C63A for differently substituted aryl-acylamides and -esters roughly follow those of His₆Amq for most, but not all, substrates tested (Table 4). A drastically decreased k _cat of His₆AmqC22A/C63A was observed with N-formylanthranilate. Remarkably, the cysteine-deficient protein showed decreased K_m and increased k _cat towards phenylacetate and the ortho-carboxy-substituted compounds N-acetylanthranilate and acetylsalicylate, resulting in pronounced increases in catalytic efficiency compared to that of His₆Amq. Actually, about three- to eightfold increases in turnover numbers of His₆AmqC22A/C63A were observed for all aryl-acetylesters and aryl-acetylamides tested, except o-chloroacetanilide. However, the increases in k _cat towards p-nitrophenylacetate, o-nitroacetanilide, and acetanilide were compensated by increases in K_m , resulting in somewhat similar catalytic efficiencies of His₆Amq and His₆AmqC22A/C63A towards these compounds.

During protein purification, the increase in specific activity of His₆AmqC22A/C63A towards the physiological substrate N-acetylanthranilate correlated with a similar increase in specific activity towards the ester analogue acetylsalicylate (Table 5), suggesting that esterase activity of the protein preparation indeed is a property of N-acetylanthranilate amidase rather than a contaminating esterase. Analogous results were obtained for His₆Amq (data not shown).

Secondary structure estimates from CD spectra.CD spectra of His₆Amq and His₆AmqC22A/C63A show some differences, especially in the wavelength range between 205 and 225 nm. The cysteine-deficient protein shows a significant minimum at 208 nm, which is characteristic for α-helical proteins and does not appear in the His₆Amq protein. Furthermore, the typical α-helical minimum at 222 nm, which both proteins show, is more pronounced in the protein carrying the cysteine-to-alanine replacements than in His₆Amq. These findings suggest the occurrence of a higher α-helical content in His₆AmqC22A/C63A than in the tagged wild-type protein.

Deconvolution of the spectra using the program K2D by Andrade et al. (8) results in the same amounts of α-helices and β-sheet for either type of the protein; the values are 37% α-helix, 26% β-sheet, and 38% aperiodic structures. The results obtained with the more sophisticated software of Böhm et al. (11) are summarized in Table 6. The parameters between 210 and 260 nm are probably most reliable due to highest experimental accuracy in that wavelength range, so only these values were taken into account for the deconvolution. The result of the deconvolution clearly shows that the α-helical content increases with the cysteine-to-alanine exchanges in positions 22 and 63 at the expense of β-sheets and aperiodic structures. Both deconvolution programs unambiguously indicate α/β-folding motifs. The CDNN program uses as its basis a set of 33 spectra of known secondary structures, whereas the K2D program uses only 15. Because of this broader basis of reference sets, we consider the results of the CDNN software more reliable. Its application supports quantitatively the nonidentity of the CD spectra shown in Fig. 5. The most pronounced difference is evidently a higher degree of α-helicity in the cysteine-deficient protein variant.

• Open in new tab
• Download powerpoint

FIG. 5.

CD spectra of His₆Amq and His₆AmqC22A/C63A. The spectra were recorded in a 0.1-mm cuvette. deg, degree.