Characterization of the active site of histidine ammonia-lyase from Pseudomonas putida Dagmar Ro¨ ther 1 ,La ´ szlo ´ Poppe 2 , Sandra Viergutz 1 , Birgid Langer 1 and Ja ´ nos Re ´ tey 1 1 Institute for Organic Chemistry, University of Karlsruhe, Germany; 2 Institute for Organic Chemistry, Budapest University of Technology and Economics, Hungary Elucidation of the 3D structure of histidine ammonia-lyase (HAL, EC 4.3.1.3) from Pseudomonas putida by X-ray crystallography revealed that the electrophilic prosthetic group at the active site is 3,5-dihydro-5-methylidene-4H-i- midazol-4-one (MIO) [Schwede, T.F., Re ´ tey, J., Schulz, G.E. (1999) Biochemistry, 38, 5355–5361]. To evaluate the importance of several amino-acid residues at the active site for substrate binding and catalysis, we mutated the following amino-acid codons in the HAL gene: R283, Y53, Y280, E414, Q277, F329, N195 and H83. Kinetic measurements with the overexpressed mutants showed that all mutations resulted in a decrease of catalytic activity. The mutants R283I, R283K and N195A were < 1640, 20 and 1000 times less active, respectively, compared to the single mutant C273A, into which all mutations were introduced. Mutants Y280F, F329A and Q277A exhibited < 55, 100 and 125 times lower activity, respectively. The greatest loss of activity shown was in the HAL mutants Y53F, E414Q, H83L and E414A, the last being more than 20 900-fold less active than the single mutant C273A, while H83L was 18 000-fold less active than mutant C273A. We propose that the carboxylate group of E414 plays an important role as a base in catalysis. To investigate a possible participation of active site amino acids in the formation of MIO, we used the chromophore formation upon treatment of HAL with L-cysteine and dioxygen at pH 10.5 as an indicator. All mutants, except F329A showed the formation of a 338-nm chromophore arising from a modified MIO group. The UV difference spectra of HAL mutant F329A with the MIO-free mutant S143A provide evidence for the presence of a MIO group in HAL mutant F329A also. For modelling of the substrate arrangement within the active site and protonation state of MIO, theoretical calculations were performed. Keywords: histidine ammonia-lyase; HAL 3,5-dihydro- 5-methylidene-4H-imidazol-4-one; MIO; site-directed mutagenesis. Histidine ammonia-lyase (HAL, EC 4.3.1.3) is the first enzyme in the nonoxidative degradation pathway of L-histidine. The enzymic catalysis begins with a Friedel– Crafts-type reaction, which helps to transform L-histidine to trans-urocanate (reviewed in [1]). An analogous mechanism was proposed for the reaction catalysed by the homologous enzyme phenylalanine ammonia-lyase (PAL, EC 4.3.1.5) which converts L-phenylalanine into trans-cinnamic acid, a precursor of a great variety of phenylpropanoids [2]. Approximately 30 years ago it was postulated that a dehydroalanine residue at the active site of both enzymes acted as electrophilic prosthetic group [3–5]. Mutagenesis experiments showed that this dehydroalanine is post- translationally formed from serines 143 and 202 of HAL and PAL, respectively [6,7]. More recently, the X-ray structure of HAL was solved at 2.1 A ˚ resolution [8]. It was shown that the prosthetic group is not dehydroalanine but a 3,5-dihydro-5-methylidene-4H-imidazol-4-one (MIO). It was proposed that this group is formed by cyclization of an intramolecular 142ASG144 tripeptide followed by subsequent elimination of two molecules of water (Fig. 1). A similar mechanism was proposed for the formation of the p-hydroxy-benzylidene-imidazol-5-one fluorophore of the green fluorescent protein from Aequorea victoria [9]. Co-crystallization was possible neither with the substrate nor with an inhibitor and therefore the exact binding mode for L-histidine could not be solved by crystal structure analysis. In this work we describe the preparation of HAL mutants in which a number of active site amino-acid residues have been changed to evaluate their importance in substrate binding or catalysis. Irreversible inhibition with L-cysteine and formation of a 338-nm chromophore [10,11] and UV difference spectra [12] were also measured to see whether a MIO group is present at the active sites of the enzyme variants. MATERIALS AND METHODS Bacterial strains and plasmids HAL was overexpressed in E. coli BL21 (DE3) cells. The gene for HAL from Pseudomonas putida was subcloned in the expression vector pT7-7 [6]. Correspondence to J. Re ´ tey, Institute of Organic Chemistry, University of Karlsruhe, Richard-Willsta ¨ tter-Allee, D-76128 Karlsruhe. Fax: 1 49 721 6084823, Tel. 1 49 721 6083222, E-mail: biochem@ochhades.chemie.uni-karlsruhe.de Enzymes: histidine ammonia-lyase (EC 4.3.1.3; Swiss-Prot accession no. P21310); phenylalanine ammonia-lyase (EC 4.3.1.5; Swiss-Prot accession no. P24481). Dedication: dedicated to Professor Wolfgang Buckel on the occasion of his 60th birthday. (Received 28 June 2001, accepted 5 September 2001) Abbreviations: HAL, histidine ammonia-lyase; PAL, phenylalanine ammonia-lyase; MIO, 3,5-dihydro-5-methylidene-4H-imidazol-4-one. Eur. J. Biochem. 268, 6011–6019 (2001) q FEBS 2001 Site-directed mutagenesis Mutagenesis was carried out in a C273A mutated gene for HAL from Pseudomonas putida to permit a subsequent crystallization without forming polymeric forms of enzyme [13]. HAL mutants R283I, R283K, H83L, N195A, E414A, E414Q, Q277A and F329A were performed following the QuickChange TM site-directed mutagenesis system (Stratagene) [14]. The oligonucleotides used in the mutagenesis reactions were: HAL-R283I(1): 5 0 -CGTACTCGCTGATCTGCCAGCCG- 3 0 ; HAL-R283I( –): 5 0 -CGGCTGGCAGATCAGCGAGTA CG-3 0 ; HAL-R283K(1): 5 0 -CGTACTCGCTGAAATGC CAGCCG-3 0 ; HAL-R283K(–): 5 0 -CGGCTGGCATTTCA GCGAGTACG-3 0 ;HAL-H83L(1): 5 0 -GTGCTGTCCC TGGCCGCTGG-3 0 ; HAL-H83L( – ): 5 0 -CCAGCGGCCA GGGACAGCAC-3 0 ; HAL-N195A(1): 5 0 -GCCCTGCTCG CCGGCACCCAG-3 0 ; HAL-N195A(–): 5 0 -CTGGGTGCC G GCGAGCAGGGC-3 0 ; HAL-E414A(1): 5 0 -GCCAA CCAGG CAGACCACGTATCG-3 0 ; HAL-E414A(–): 5 0 - CGATACGTGGTCT GCCTGGTTGGC-3 0 HAL-E414Q(1): 5 0 -GCCAACCAGCAAGACCACGT ATCG-3 0 ; HAL-E414Q(–): 5 0 -CGATACGTGGTCTTG CTGGTTGGC-3 0 ; HAL-Q277A(1): 5 0 -CGACAAGGT C GCGGACCCGTACTCG-3 0 ; HAL-Q277A(–): 5 0 -CGA GTACGGGTCC GCGACCTTGTCG-3 0 ; HAL-F329A(1): 5 0 -CGGTGGCAACGCCCACGCAGAACC-3 0 ; HAL- F329A(–): 5 0 -GGTTCTGCGTGGGCGTTGCCACCG-3 0 . HAL mutants Y53F and Y280F were constructed following a method described by Olsen et al. [15]. The oligonucleotides used in these mutagenesis reactions were: HAL-Y53F: 5 0 -CGCACTGCCTTCGGCATCAAC-3 0 ; HAL-Y280F: 5 0 -CCAGGACCCGTTCTCGCTGCGC-3 0 . The mutations were checked by sequence analysis using the dideoxynucleotide chain-termination method [16]. Protein expression and purification E. coli BL21 (DE3) cells carrying the plasmids with the genes for wild-type HAL and HAL mutants were cultured and HAL was purified as described previously [6]. SDS/PAGE and Western blot analysis SDS/PAGE was carried out according to Laemmli [17] using 10% polyacrylamide gels. The gels were stained with Coomassie Brillant Blue R250. Western Blot analysis was performed following a previously described method using nitrocellulose blotting filters [18,19]. Wild-type HAL and mutants were detected with rabbit polyclonal antibodies raised against HAL from Pseudomonas putida (the antibody was a generous gift from G. Mu ¨ nscher, Behringwerke AG, Marburg, Germany). Enzyme assay and protein determination HAL activity was measured spectrophotometrically at 25 8C following the formation of trans-urocanate at 277 nm. The assay was performed in 1-cm quartz cuvettes by modifi- cation of the method described in [20] with enzyme concentrations varying between 1 and 25 mg for active enzymes and between 0.1 and 1 mg for less active mutants. The enzyme was preincubated at 25 8C for 5 min in 950 mL 0.1 M sodium pyrophosphate pH 9.3 supplemented with 10 m M ZnCl 2 and 2 mM glutathione. Reaction was started by adding 50 mL of a 0.5- ML-histidine solution. Wild-type enzyme and moderately active mutant enzymes were measured in intervals of 1 min for 5 min, less active enzyme mutants were measured in intervals of 5 min for 20 min. For deter- mination of K m and V max L-histidine concentrations were varied from 0.5 to 35 m M. Kinetic parameters (K m , V max ) were determined using a double reciprocal plot [21]. Because we used pure enzyme fractions it was possible to measure the turnover numbers (k cat values) accounting for a molecular mass of M r ¼ 214.372 of the tetrameric HAL. Determination of the protein concentration was carried out according to Warburg & Christian [22,23], Murphy & Kies [24] and Groves et al. [25] and Smith et al. [26]. As a reference protein for the measurements we used bovine serum albumin (BSA). Irreversible inactivation with L-cysteine Irreversible inactivation by L-cysteine was carried out in 1 cm quartz cuvettes as described earlier [10,11,27]. A total Fig. 1. Mechanism for the formation of MIO by cyclization of an intramolecular 142ASG144 tripeptide. 6012 D. Ro ¨ ther et al. (Eur. J. Biochem. 268) q FEBS 2001 of 0.75 mg (3.5 nmol) enzyme was dissolved in 1.0 mL 50 m M NaHCO 3 /Na 2 CO 3 buffer pH 10.5. Inhibition was started by addition of L-cysteine to a final concentration of 10 m M. Inactivation was controlled spectrophotometrically in a Cary 3E spectrophotometer (Varian), following the increase in absorbance at 338 nm during 50 min in intervals of 10 min to get repetitive overlays of the absorption spectra. For determining the time dependence of the inactivation by activity measurements 12 nmol enzyme was dissolved in 2mL50m M NaHCO 3 /Na 2 CO 3 buffer pH 10.5. Inactivation of the enzyme was started by addition of L-cysteine to a final concentration of 10 m M. Every 10 min, we used a 200-mL aliquot of the reaction mixture for the enzyme assay. After 60 min the residual mixture was dialysed against 50 m M NaHCO 3 /Na 2 CO 3 buffer pH 10.5 to remove L-cysteine followed by an enzyme assay. UV difference spectroscopy UV difference spectra were measured at enzyme concen- trations of 2 mg (9 nmol) in 1 mL 10 m M Tris/HCl pH 7.2 from 240 to 360 nm using 1-cm quartz cuvettes [12]. First a blank with the MIO-free HAL mutant S143A was measured followed by a scan of the wild-type enzyme and various active site mutants of HAL. Substrate fit and optimization within the active area Calculations were performed on 300–500 MHz Pentium II computers running under WINDOWS 95 or WINDOWS 98. For molecular mechanics, a switched smoothing function which gradually reduced nonbonding interactions to zero from 10 A ˚ inner radius to 14 A ˚ outer radius, was applied. Otherwise, all calculations were performed by using default settings of the program packages. Analysis of the X-ray structure of the HAL homotetramer (PDB code: 1B8F) showed that Ser143 is fully covered by residues of three monomer subunits within a global area of 25 A ˚ radii. This part (representing 475 amino-acid residues, a number which is comparable to the 509 amino-acid resi- dues size of a monomeric HAL unit, together with structure waters, a glycerol molecule and a sulfate anion) was cut off from the full HAL homotetramer structure and used for modelling the substrate free and substrate incorporating states of the active site by MM1 calculations of the HYPERCHEM package [28]. All the mutated residues were found within 12 A ˚ radii around the methylene carbon of MIO formed from Ser143. Therefore the outside sphere between 12 and 25 A ˚ of the whole 25 A ˚ radii globe was kept ‘frozen’ during the calculations. The calculations were performed on 1232 atoms within the 12 A ˚ inside area. Conformational analysis of L-histidine in its zwitterionic state was performed by PM3 calculations in the PC SPARTAN PRO package [29] similarly as reported for the zwitterionic form of L-phenylalanine [30]. The appropriate zwitterionic L-histidine structure was docked to the substrate-free X-ray structure of HAL by applying the following considerations. (a) The C5 position of the L-histidine imidazole ring should be close enough to the methylene of the MIO to perform the nucleophilic addition to the C¼C double bond. (b) The NH 3 1 Fig. 2. Calculated models for the zwitterionic L-histidine binding (A), for the cationic intermediate containing (B), and for the trans-urocanate/ammonia binding (C) state of HAL’s active site. q FEBS 2001 Active site mutants of histidine ammonia-lyase (Eur. J. Biochem. 268) 6013 and the pro-R b-H should be nearly antiperiplanar. (c) The best zwitterionic conformation of L-histidine fulfilling these requirements was aligned by RMS fit over two water mol- ecules (hydrogen bonded to the imidazole N of H83 and to the carbonyl O of Asn195) and the sulfate anion, all of which are present in the experimental X-ray structure of HAL in the close vicinity of the MIO methylidene moiety. The atomic pairs used for this fit were: H83 coordinating water O , imidazolyl-N1 of the histidine, Asn195 coordinating water O , NH 3 1 of histidine, and S atom of the sulfate anion , carboxylate C of histidine. After docking, the sulfate ion and the two water molecules were deleted and the structure containing the zwitterionic L-histidine substrate was optimized using the MM1 method of the HYPERCHEM [28] program (Fig. 2A). The cationic intermediate state was obtained by con- structing a single bond between the L-histidine imidazole C5 and MIO methylidene C atoms, correcting the atom and bond types and orders, and relaxing the structure by MM1 optimization (Fig. 2B). The trans-urocanate/ammonia binding model was obtained from the cationic intermediate model by breaking the appropriate bonds, correcting the atom and bond types and orders, and optimizing the structure by the MM1 method (Fig. 2C). Calculations of electronic spectra of different forms of a truncated MIO model Full PM3 geometry optimization on a truncated MIO model compound and on its possible protonated forms (Fig. 3) followed by single point calculations for their excited states were performed in the HYPERCHEM [28] package. RESULTS AND DISCUSSION Construction, overproduction and purification of active site mutants On the basis of the recently elucidated X-ray structure of HAL [8], several active site amino-acid residues can be identified. These are R283, Y53, E414, Y280, N195, H83, Q277 and F329 (see Fig. 4). To evaluate the importance of these residues, HAL mutants were constructed at the corresponding sites using the QuickChange TM site-directed mutagenesis kit and the method of Eckstein [14,15]. The results of the mutagenesis experiments were verified by sequence analysis. Overproduction and purification of the HAL mutants were carried out as described by Langer et al. [6]. Crude extracts of bacterial cells producing wild-type enzyme and the mutated variants were separated by SDS/ PAGE to compare the expression rates and the sizes of the recombinant proteins. In all cases, high quantities of recombinant enzyme were produced showing the same Fig. 3. Calculated UV absorptions of a truncated MIO model compound and of its protonated forms. Calculations were performed by PM3 method. Symbols indicate the calculated relative oscillator strengths: (s), strong; (m), medium; (w), weak. Fig. 4 Active site of HAL. 6014 D. Ro ¨ ther et al. (Eur. J. Biochem. 268) q FEBS 2001 monomeric size. Western Blot analysis showed that all enzyme variants were detected by the anti-HAL Ig. E. coli BL21 (DE3) used as host did not show any HAL activity. A search in the Swiss-Prot data bank for sequences homologous to HAL from various sources was negative. Purification of wild-type HAL and HAL mutants resulted in yields varying from 5 to 80 mg pure enzyme per L cell culture. After purification the turnover number or k cat of recombinant wild type HAL was 86 s 21 which is in agreement with a specific activity of 24 U : mg 21 previously reported [6]. Characterization of the mutants by kinetic measurements Steady state kinetic parameters of HAL mutants were measured at substrate concentrations varying from 0.5 to 35 m ML-histidine. Comparison of the K m values revealed that all mutants have similar affinities for L-histidine. This indicates that several residues are responsible for binding of the substrate and mutagenesis of a single residue affects the K m value very little. HAL mutant C273A which was constructed to achieve better crystals [13] and the double mutant C273A/R283I showed a somewhat higher K m (18 mM) than other mutants or wild-type enzyme (K m ¼ 3–8 mM), pointing to the relative importance of R283 for substrate binding. HAL mutant C273A showed a fivefold lower k cat compared to wild-type HAL as was recently described [13]. In Table 1 the K m and k cat values of the HAL variants and the factors k catC273A : k catmut (k cat of single mutant C273A divided by k cat of HAL double mutants) are listed. The factors show to what extent the double mutants are less active in relation to the single mutant C273A. HAL mutant C273A/R283K shows < 20 times lower activity compared to HAL mutant C273A, whereas a substitution of this arginine by isoleucine leads to a larger decrease of activity (1640 times less active than mutant C273A). This indicates that a noncationic residue at that position results in a more severe decrease of activity. Substitution of Y53, which is positioned in the neighbour- hood of R283 at the active site, leads to more dramatic effects. Exchange to phenylalanine results in a 2650-fold less activity compared to the single mutant C273A. These data indicate that this region in the active site may be responsible for coordination of a cationic group of L-histidine that is located near an anionic group of the substrate. We propose therefore that the neighbouring resi- dues R283 and Y53 coordinate the carboxylic and amino group of the substrate L-histidine, respectively. Based on the X-ray structure of HAL (Fig. 4) in Fig. 2, models for binding of L-histidine (Fig. 2A), the cationic intermediate formed by attack of C5 of the imidazole moiety of L-histidine at the methylidene carbon of MIO (Fig. 2B), and trans-urocanate and ammonia (Fig. 2C) at the active site of HAL are shown which explain possible functions of some active site residues. Residues Q277 and F329 (see Fig. 4) were both converted into alanine which resulted in a 125 and 100 times lower activity, respectively (Table 1). The cationic intermediate-binding model (Fig. 2B) indicates Fig. 5. Model for the binding of L-histidine at the active site of HAL. Table 1. Kinetic constants of wild-type HAL and active site mutants. HAL activity was measured by following the formation of trans-urocanate at 277 nm in the presence of purified enzyme. The enzyme was preincubated at 25 8C in 0.1 M sodium pyrophosphate (pH 9.3) supplemented with 10 m M ZnCl 2 and 2 mM glutathione. Reaction was started by addition of a 0.5-ML-histidine solution. The L-histidine concentrations were varied from 0.5 to 35 m M. The kinetic constants K m (mM) and V max (U : mg 21 or mmol : min 21 : mg 21 ) were determined using a double reciprocal plot [21]. Turnover numbers or k cat values (s 21 ) were determined with the molecular mass M r ¼ 53 593 for one subunit of the tetrameric HAL. Determination of protein concentration was carried out according to Warburg and Christian [22,23], Murphy and Kies [24], Groves et al. [25] and Smith et al. [26]. K m (mM) k cat (s 21 ) k catC273A /k catmut ratio Wild-type HAL 3.9 ^ 0.9 86 ^ 6 C273A HAL 18 ^ 318^ 11 C273A/R283I HAL 18 ^ 4 0.011 ^ 0.001 1640 C273A/R283K HAL 4.1 ^ 0.7 0.79 ^ 0.03 20 C273A/Y53F HAL 8 ^ 1 0.0068 ^ 0.0004 2650 C273A/E414A HAL 6.1 ^ 0.7 0.00086 ^ 0.00007 20 930 C273A/E414Q HAL 1.7 ^ 0.9 0.053 ^ 0.0025 339 C273A/Y280F HAL 8 ^ 1 0.32 ^ 0.01 55 C273A/N195A HAL 3 ^ 1 0.018 ^ 0.001 1000 C273A/Q277A HAL 7 ^ 2 0.14 ^ 0.01 125 C273A/F329A HAL 4.4 ^ 0.7 0.18 ^ 0.01 100 C273A/H83L HAL 1.2 ^ 0.4 0.001 ^ 0.0002 18 000 q FEBS 2001 Active site mutants of histidine ammonia-lyase (Eur. J. Biochem. 268) 6015 the p-stacking role of F329. This residue may stabilize the s-complex-like intermediate and prevents abstraction of the proton of the C5 carbon by excluding any basic group. A dramatic loss in activity was achieved by substitution of N195 and E414. The measured k cat values were 1000– 20 930 times, respectively, lower than that of the wild-type enzyme. HAL mutant H83L showed almost no activity. We propose that these residues have important functions in the enzyme HAL which is shown in Fig. 5. H83 is possibly involved in binding and orienting of the imidazolyl moiety of L-histidine at the active site (Fig. 2A) and stabilization of the cationic intermediate arising from the imidazolyl moiety of the substrate (Fig. 2B). Because of the larger distance (4.01 A ˚ , Table 2) the interaction of H83 with the substrate imidazole group could be mediated by a H 3 O 1 molecule or, alternatively, by coordination with a metal ion such as Mn 21 or Zn 21 (Fig. 6). The calculated steric arrangement of the cationic intermediate suggests that the imidazolyl N1-H of the substrate may polarize the MIO group by partial protonation of its carbonyl oxygen (Fig. 2B). This partial protonation facilitates the electrophilic attack at the methyl- ene moiety of MIO by decreasing of the electron density in the p system of the C¼C double bond (Fig. 5). Calculation of the UV spectra of MIO and the energy of putative intermediate states For estimating the degree of polarization of the MIO moiety in the substrate free state of HAL by partial protonation of Fig. 6. Arrangement of H83 and substrate imidazole in the cationic intermediate model of HAL compared to experimental Zn 21 com- plex found in human carbonic anhydrase II (PDB code: 1CRA) [31]. Table 2. Distances in models of the HAL active site. Selected distances (measured in A ˚ ) in models for the zwitterionic L-histidine binding (a), for the cationic intermediate containing (b), and for the trans-urocanate/ammonia binding (c) state of HAL’s active site are listed. Atomic pairs Model a Model b Model c S142 C3 –His C4 0 4.43 1.53 3.89 S142 O1 –His N1 0 4.79 2.61 3.21 H83 N1 0 –His N3 0 4.01 4.15 5.25 E414 O1 –His C3 5.54 3.66 3.22 N195 O1 –His N2 2.89 3.55 3.04 Y53 OH –His N2 4.85 4.05 3.72 Y53 OH –His O1 2.84 2.87 4.55 Q277 N4 –His O1 4.31 4.04 6.32 R283 NH1 –His O 0 1 4.39 4.04 4.79 R283 NH2 –His O 0 1 3.68 3.16 3.00 Fig. 7. Mechanism for the formation of the 338 nm chromophore by irreversible inactivation of HAL with L-cysteine. 6016 D. Ro ¨ ther et al. (Eur. J. Biochem. 268) q FEBS 2001 polar amino-acid residues, the electronic spectrum of this chemically unprecedented chromophore was calculated in a truncated model at the PM3 level of theory (Fig. 3). Spectra for the protonated MIO structures (fully protonated at the carbonyl O, at N1 and at N3, respectively) were calculated similarly. The calculated absorption maximum at 303 nm for the nonprotonated MIO model is in good agreement with the experimentally determined maximum around 302 nm in the UV difference spectra obtained by substracting the spectra of MIO-lacking mutants from spectrum of MIO- containing HAL. On the other hand, all the three fully protonated MIO models gave significantly different calculated UV spectra. This indicates that the MIO moiety in the substrate free state of HAL is not substantially protonated. Consequently, the substrate itself activates the MIO by partial protonation upon approaching it. The fact, that methylation either at N1 or N4 of the imidazolyl moiety of L-histidine resulted in compounds which are neither substrates nor inhibitors of HAL (S. Viergutz & J. Re ´ tey, unpublished results), indicates the importance of the orienting effect of H83 (N1 methylated analogue) and the partial protonation effect on MIO (N4 methylated analogue). The carbonyl group of N195 may be involved in hydrogen bonding to the ammonium function of the substrate (Fig. 2C) and also to the leaving ammonia molecule (Fig. 2C). The active site residue E414 should have a key role in catalysis and with the assistance of Y280 may provide the enzymic base designed to abstract the activated b proton of the substrate. Y53, R283 and Q277 might be involved in anchoring the carboxylate moiety of the substrate or the cationic intermediate (Fig. 2A,B). Inspection of the whole HAL tetramer reveals that residue Y53, which showed the most dramatic effect on the reaction rate among these three residues, is located at the edge of a channel through which the substrate can enter into or the product can be released from the active site. The calculated total energies for the zwitterionic L-histidine binding (Fig. 2A) and cationic intermediate binding models (Fig. 2A) were similar, whereas < 46 kJ : mol 21 lower total energy was obtained for Fig. 8. Inactivation of HAL with L-cysteine to a final concentration of 10 m M, and an enzyme concentration of 0.75 mg (3.5 nmol) in 1 mL. For further details see Experimental procedures. Inactivation of HAL mutant C273A and repetitive scans between 0 and 50 min after supplementation of L-cysteine (A). Inactivation of HAL mutant C273A/ Y280F and repetitive scans between 0 and 50 min (B). Inactivation of HAL mutant C273A/Y280F and C273A/H83L 20 and 48 h after supplementation of L-cysteine, respectively (C). Fig. 9. Inactivation with L-cysteine and enzyme assay between 0 and 60 min after supplementation of L-cysteine. For experimental details see Legend to Fig. 8 and Experimental procedures. After 60 min the residual reaction mixture was dialysed to remove unbound cysteine. (A) Inactivation of HAL mutant C273A/F329A with L-cysteine. (B) Inactivation of HAL mutant S143A with L-cysteine. Fig. 10. UV difference spectra of HAL mutants S143A and C273A (solid line) and HAL mutants S143A and C273A/F329A (dotted line). These were measured at enzyme concentrations of 2 mg (9 nmol) per ml in 10 m M TrisHCl buffer pH 7.2. For further details see Experimental procedures. q FEBS 2001 Active site mutants of histidine ammonia-lyase (Eur. J. Biochem. 268) 6017 the relaxed trans-urocanate/ammonia binding structure (Fig. 2C). In this structure, as a consequence of a < 608 flop of the carboxylate group of the urocanate caused by its conjugation, i.e. coplanarity to the p system of the forming C¼C double bond, the carboxylate moiety is displaced from the vicinity of the Y53, R283 and Q277 triad. The calculated distances between particular atoms of the substrate histidine (His) and of important active site amino acids are listed in Table 2. Identification of the MIO group by inhibition with L-cysteine and UV difference spectroscopy To determine whether the various mutant enzymes have a prosthetic MIO group at their active sites, we performed irreversible inhibition with L-cysteine to a final concen- tration of 10 m M in slightly basic solution and in the presence of O 2 . Under these conditions, both wild-type HAL and HAL mutant C273A show an increase in absorbance at 338 nm during 50 min as previously described [11]. The chromophore is generated by nucleophilic attack of the thiolate anion of cysteine at the MIO group followed by oxidation and intramolecular S-to-N rearrangement as recently proposed (Fig. 6) [32,33]. In Fig. 8A repetitive scans of the single mutant C273A following the inactivation with L-cysteine are shown. During 50 min, an absorbance maximum develops that is located around 338 nm. Some double mutants showed different behaviour upon treatment with L-cysteine. The mutant C273A/F329A did not show an absorbance maximum upon treatment with L-cysteine even after 24 h of incubation. In the case of the HAL mutant C273A/Y280F and mutant C273A/H83L there was a slower increase in absorbance but after 20 and 48 h, respectively, a chromophore around 338 nm appeared also in these cases (Fig. 8C). These results indicate the presence of a MIO at their active sites, but in a less reactive form. Concomittant with the formation of a new chromophore the activity of the enzyme decreases irreversibly. After addition of L-cysteine, the activity of the enzymes dropped very quickly and in most cases there was no activity at all after 60 min. Removal of excess cysteine by dialysis did not restore activity (Fig. 9A). The MIO-free HAL mutant S143A also showed inactivation, but 60 min after treatment with L-cysteine the remaining solution regained almost 100% of its original activity upon dialysis (Fig. 9B). In this case L-cysteine does not bind irreversibly to the enzyme. With HAL mutants C273A, C273A/F329A and C273A/ H83L we carried out UV difference spectroscopic measure- ments, because this method is an excellent means to show the presence of the MIO group [12] (Fig. 10). HAL mutants C273A and C273A/F329A both showed a maximum around 302 nm in the UV difference spectra with HAL mutant S143A which lacks an intact MIO group (Fig. 9). This result indicates that HAL mutant C273A/F329A contains a MIO group at their active site but is not able to form a 338-nm chromophore with L-cysteine and dioxygen. It is noteworthy that reduced glutathione of high purity, present in the enzyme essay, does not inhibit HAL. However under certain conditions glutathione may release L-cysteine which is an inhibitor. It may be better to use dithiothreitol to keep some cysteine residues in the reduced form. The role of Zn 21 and other metal ions carrying two positive charges has been thoroughly investigated [34]. The decrease of the enzymatic activity by removal of such ions and the slight activation at their presence might be explained by assuming their interaction with H83 and the substrate histidine (Fig. 6). In contrast no metal-ion effect has ever been observed on the PAL reaction. This is in agreement with the lack of histidine in a similar position in all PAL sequences. ACKNOWLEDGEMENTS We thank Prof. G. E. Schulz and Dr T. F. Schwede (University of Freiburg, Germany) for the cooperation in the work on HAL and PAL and the production of two HAL mutants. The work was supported by the Deutsche Forschungsgemeinschaft and the Fonds der Chemischen Industrie. D. R. thanks the Land Baden-Wu ¨ rttemberg for a scholarship for graduate students. L. P. thanks the Hungarian OTKA (T-033112) for financial support. We thank A. Sigrist for help with the figures and S. Vollmer for technical assistence. REFERENCES 1. Re ´ tey, J. (1996) Enzymatic catalysis by Friedel –Crafts-type reactions. Naturwissenschaften 83, 439–447. 2. Hanson, K.R. & Havir, E.A. (1978) An introduction to the enzymology of phenylpropanoid biosynthesis. Rec. Adv. Phyto- chem. 12, 91–137. 3. Wickner, R.B. (1969) Dehydroalanine in histidine ammonia-lyase. J. Biol. Chem. 244, 6550–6552. 4. Givot, I.L., Smith, T.A. & Abeles, R.H. (1969) Studies on the mechanism of action and the structure of the electrophilic center of histidine ammonia-lyase. J. Biol. Chem. 244, 6341 –6353. 5. Hanson, K.R. & Havir, E.A. (1970) L-Phenylalanine ammonia- lyase. IV. Evidence that the prosthetic group contains a dehydroalanyl residue and mechanism of action. Arch. Biochem. Biophys. 141, 1– 17. 6. Langer, M., Reck, G., Reed, J. & Re ´ tey, J. (1994) Identification of serine-143 as the most likely precursor of dehydroalanine in the active site of histidine ammonia-lyase. A study of overexpressed enzyme by site-directed mutagenesis. Biochemistry 33, 6462–6467. 7. Schuster, B. & Re ´ tey, J. (1994) Serine-202 is the putative precursor of the active site dehydroalanine of phenylalanine ammonia lyase. FEBS Lett. 349, 252–254. 8. Schwede, T.F., Re ´ tey, J. & Schulz, G.E. (1999) Crystal structure of histidine ammonia-lyase revealing a novel polypeptide modifi- cation as the catalytic electrophile. Biochemistry 38, 5355– 5361. 9. Ormo ¨ , M., Cubitt, A.B., Kallio, K., Gross, L.A., Tsien, R.Y. & Remington, S.J. (1996) Crystal structure of the Aequorea victoria green fluorescent protein. Science 273, 1392–1395. 10. Klee, C.B. (1970) Reversible polymerization of histidine ammonia- lyase. The role of sulfhydryl groups in the activity and polymeric state of the enzyme. J. Biol. Chem. 245, 3143 –3152. 11. Klee, C.B. (1974) Stereospecific irreversible inhibition of histidine ammonia-lyase by L-cysteine. Biochemistry 13, 4501 –4507. 12. Ro ¨ ther, D., Merkel, D. & Re ´ tey, J. (2000) Spectroscopic evidence for a 4-methylidene imidazol-5-one in histidine and phenylalanine ammonia-lyases. Angew. Chem. Int. Ed. 39, 2462–2464. 13. Schwede, T.F., Ba ¨ deker, M., Langer, M., Re ´ tey, J. & Schulz, G.E. (1999) Homogenization and crystallization of histidine ammonia- lyase by exchange of a surface cysteine residue. Protein Eng. 12, 151–153. 14. Braman, J., Papworth, C. & Greener, A. (1996) Site-directed mutagenesis using double-stranded plasmid DNA templates. Methods Mol. Biol. 57, 31–44. 15. Olsen, D.B., Sayers, J.R. & Eckstein, F. (1993) Site-directed mutagenesis of single-stranded and double-stranded DNA by phosphorothioate approach. Methods Enzymol. 217, 189–217. 6018 D. Ro ¨ ther et al. (Eur. J. Biochem. 268) q FEBS 2001 16. Sanger, F., Nicklen, S. & Coulson, A. (1977) DNA sequencing with chain-terminating inhibitors. Proc. Natl Acad. Sci. USA 78, 177–181. 17. Laemmli, U.K. (1970) Cleavage of structural proteins during the assembly of the head of bacteriophage T4. Nature 277, 680–685. 18. Twobin, H., Staehlin, T. & Gordon, A. (1979) Electrophoretic transfer of proteins from acrylamide gels to nitrocellulose sheets: procedures and applications. Proc. Natl Acad. Sci. USA 76, 4350–4354. 19. Burnette, W.N. (1981) ’Western-blotting’: electrophilic transfer of proteins from sodium dodecyl sulfate-polyacrylamide gels to unmodified nitrocellulose and radiographic detection with antibodies and radioionated protein A. Anal. Biochem. 112, 195–203. 20. Tabor, H. & Mehler, A.H. (1955) Histidase and urocanase. Methods Enzymol. 2, 228– 233. 21. Lineweaver, H. & Burk, D. (1934) The determination of enzyme dissociation constants. J. Am. Chem. Soc. 56, 638–666. 22. Warburg, O. & Christian, W. (1942) Isolierung und Kristallisation des Ga ¨ rungsfermentes Enolase. Biochem. Z. 310, 384–421. 23. Layne, E. (1957) Spectrophotometric and turbidimetric methods for measuring proteins. Methods Enzymol. 3, 447–454. 24. Murphy, J.B. & Kies, M.W. (1960) Note on spectrophotometric determination of proteins in dilute solutions. Biochim. Biophys. Acta 45, 382–384. 25. Groves, W.E., Davis, F.C. Jr & Sells, B.H. (1968) Spectro- photometric determination of microgram quantities of protein without nucleic acid interference. Anal. Biochem. 22, 195–210. 26. Smith, P.K., Krohn, R.I., Hermanson, G.T., Mallia, A.K., Gartner, F.H., Provenzano, M.D., Fujimoto, E.K., Goeke, N.M., Olson, B.J. & Klenk, D.C. (1985) Measurement of protein using bicinchoninic acid. Anal. Biochem. 150, 76 –85. 27. Langer, M., Lieber, A. & Re ´ tey, J. (1994) Histidine ammonia-lyase mutant S143C is posttranslationally converted into fully active wild-type enzyme. Evidence for serine 143 to be the precursor of active site dehydroalanine. Biochemistry 33, 14034 –14038. 28. Hypercube Inc. (2001) http://www.hyper.com/products/ description/hyper6.htm Hyperchem 6. Hypercube Inc., Gainesville, FL, USA. 29. Wavefunction Inc. (2001) http://www.wavefun.com/software/ pc_spartan_pro/pcpro_main.html PC Spartan Pro. Wavefunction Inc., Irvine, CA, USA. 30. Gloge, A., Zon, J., Ko ´´ va ´ ri, A ´ ., Poppe, L. & Re ´ tey, J. (2000) Phenylalanine ammonia-lyase: the use of its broad substrate specificity for mechanistic investigations and biocatalysis. Synthesis of L-arylalanines. Chem. Eur. J. 6, 3386– 3390. 31. Mangani, S. & Liljas, A. (1993) Crystal structure of the complex between human carbonic anhydrase II and the aromatic inhibitor 1,2,4-triazole. J. Mol. Biol. 232, 9–14. 32. Galpin, J.D., Ellis, B.E. & Tanner, M.E. (1999) The inactivation of histidine ammonia-lyase by L-cysteine and oxygen: modification of the electrophilic center. J. Am. Chem. Soc. 121, 10840–10841. 33. Merkel, D. & Re ´ tey, J. (2000) Further insight into the mechanism of the irreversible inhibition of histidine ammonia-lyase by L-cysteine and dioxygen. Helv. Chim. Acta 83, 1151–1160. 34. Klee, C.B. (1972) Metal activation of histidine ammonia-lyase. Metal ion-sulfhydryl group relationship. J. Biol. Chem. 247, 1398–1406. q FEBS 2001 Active site mutants of histidine ammonia-lyase (Eur. J. Biochem. 268) 6019 . formation of a new chromophore the activity of the enzyme decreases irreversibly. After addition of L-cysteine, the activity of the enzymes dropped very quickly. Hungary Elucidation of the 3D structure of histidine ammonia-lyase (HAL, EC 4.3.1.3) from Pseudomonas putida by X-ray crystallography revealed that the electrophilic