The Role of the Conserved Residues His-246, His-199, and Tyr-255 in the Catalysis of Catechol 2,3-Dioxygenase from Pseudomonas stutzeri OX1*□S

Catechol 2,3-dioxygenase (C2,3O) from Pseudomonas stutzeri OX1, which is able to grow on various aromatic substrates as the sole source of carbon and energy, has been expressed in Escherichia coli, purified, characterized, and found to be very similar to other dioxygenases from Pseudomonas species. Interestingly, the activity of the protein shows a rather unusual pH dependence when assayed on catechol. A model of the catalytic mechanism was developed that is able to reproduce the catalytic behavior of the protein as a function of the pH. The model includes multiple equilibria and four productive intermediates with different ionization states of the enzyme-substrate complex. The fitting of the theoretical curve to the experimental data suggests that a tyrosine and two histidine residues are involved in catalysis. Mutants (H246N)-, (H246A)-, (H199N)and (Y255F)-C2,3O were produced to investigate the role of highly conserved His-199, His-246, and Tyr-255. The strongly reduced activity of the mutants suggests a primary catalytic role for each of these residues. Moreover, mutants at positions 199 and 246 display pH profiles different from that of the wild-type protein, thus indicating that residues His-246 and His-199 play a role in determining the unusual pH dependence of the enzyme. In addition, electron-withdrawing groups on catechol, which increase the acidity of the phenolic hydroxyl group, are able to counterbalance the effect of the mutation H246N in reducing catalytic activity but cause a further reduction of the activity of (H199N)-C2,3O. This finding suggests that His-246 is involved in the initial catechol deprotonation, whereas His-199 promotes the reaction between oxygen and the aromatic ring.

Many soil bacteria can grow on aromatic molecules as the sole source of carbon and energy, revealing an interesting potential for applications in bioremediation strategies (1). Aromatic compounds are first transformed by upper pathways into dihydroxylated intermediates. These intermediates are subse-quently metabolized into Krebs cycle intermediates by a limited number of lower pathways (2), which are initiated by ring cleavage reactions that are generally catalyzed by ring cleavage dioxygenases. These enzymes catalyze the addition of two atoms from molecular oxygen into the aromatic ring with a substrate specificity more restricted with respect to that of the upper pathway enzymes. As a consequence, ring cleavage enzymes guide and reduce the metabolic flow of aromatic compounds that can be degraded, thus playing a central role in their catabolism (3).
Ring cleavage dioxygenases acting on catechol and substituted catechols are classified into two groups (4), intradiol dioxygenases and extradiol dioxygenases, with the latter producing 2-hydroxymuconic semialdehydes. The majority of known extradiol dioxygenases belong to a single family of evolutionarily related proteins widespread among prokaryotes (5).
Extradiol dioxygenase activity depends on the presence at the catalytic center of a Fe(II) ion, with the few exceptions of Mn 2ϩ -containing dioxygenases (6,7). The catalytic metal is bound to two histidine residues and one glutamate residue that are well conserved among the members of the family (5). These residues are essential for metal binding and activity (8,9).
Multiple alignments show that three other residues also present at the active site, His-246, His-199, and Tyr-255 (P. putida mt2 C2,3O numbering; Fig. 1, A and B), are well conserved (5,13). Thus, it might be inferred that these residues play a role in substrate binding and/or catalysis. DHBD from Pseudomonas sp. KKS102 (8) and B. cepacia LB400 (11) and HPCD from A. globiformis and B. fuscum (12) have been crystallized with substrates bound in the active site (Protein Data Bank codes 1KW9, 1KMY, 1KND, 1KNF, 1F1V, and 1Q0C). It * The costs of publication of this article were defrayed in part by the payment of page charges. This article must therefore be hereby marked "advertisement" in accordance with 18 U.S.C. Section 1734 solely to indicate this fact. □ S The on-line version of this article (available at http://www.jbc.org) contains supplemental material (including Table S1) that elaborates on the reaction scheme in Fig. 3.
The nucleotide sequence(s) reported in this paper has been submitted to the GenBank TM /EBI Data Bank with accession number(s) AJ496739.
§ These authors contributed equally to this work. ʈ To whom correspondence should be addressed. should be noted that in all of the structures the oxygen atom of the substrate that is on the same side of Tyr-249 (Tyr-257 in HPCD and Tyr-255 in C2,3O) is closer to the iron ion than is the other catechol oxygen atom (Fig. 1A), suggesting that only this hydroxyl group is in the anionic form (8,11). This hypothesis is confirmed by Raman and electronic absorption studies (11). Moreover, extended x-ray absorption fine structure studies indicate that P. putida C2,3O also binds catechol as a monoanion (14). It should also be noted that Tyr-249 has been proposed to be H-bonded to the anionic oxygen of the catechol ring, acting as H donor, whereas His-194 (His-199 in C2,3O from P. putida mt2) is H-bonded to the non-dissociated hydroxyl group, acting as H-bond acceptor (8). Finally, studies on synthetic catalysts that mimic extradiol dioxygenases (15) and studies on the manganese-dependent HPCD (16) indicate that the first deprotonation of catechols is the first crucial step of extradiol cleavage reactions. According to these observations, a mechanism was proposed (8, 10, 11, 17) (Fig. 1C) in which conserved His-246 and Tyr-255, or their homologous residues, act in a concerted way to abstract the first proton from catechol, FIG. 1. A scheme of the active site of DHBD from Pseudomonas sp. KKS102 with DHBP bound in the active site (PDB code 1EIM) (A) and a scheme of P. putida C2,3O with an acetone molecule bound to iron (PDB code 1MPY) (B). The active site of Burkholderia sp. LB400 DHBD is identical to that shown in panel A, but residue numbers must be increased by one unit. Residues in boldface are completely conserved in dioxygenases. H-bonds are shown as dotted lines. Distances are expressed in Ångstroms. His-240 (in panel A) and His-246 (in panel B) are shown in the acidic form in order to show the two possible H-bonds. His-240 is located under the hydroxylated ring of the substrate (i.e. under the plane of the sheet), and the line with the round ends shows the average distances between the two rings. Panel C depicts the catalytic mechanism proposed for extradiol-cleaving dioxygenases. The scheme shows both the hypothesis of a radical attack of O 2 on the substrate (Step 3) as proposed for DHBD (8,11) and that of a nucleophilic attack (Step 3Ј) as proposed for C2,3O (10,39). His-199 acts as the base catalyst to promote step 3.
whereas His-199, at a later stage, acts as a basic catalyst by accepting the second proton from the substrate. Finally, the position of O 2 attack and the mechanism of its insertion in the aromatic ring have been addressed by density functional theory studies (18) and by studies with model compounds and analogues of the hypothetical reaction intermediates (19 -21). However, despite the large number of studies, direct evidence of the roles of His-199, His-246, and Tyr-255 in catalysis is still lacking, with the exception of data available on the involvement of His-200 of the iron-dependent HPCD from B. fuscum (22).
In this paper we report the overexpression in Escherichia coli and the purification and characterization of the catechol 2,3dioxygenase from Pseudomona stutzeri OX1 and four active site mutants designed for obtaining information on ionizable groups involved in catalysis. The catalytic parameters of C2,3O and the mutants were studied on catechol as a function of pH and by using three acidic catechol derivatives as substrates. The data we have collected confirm the hypothesis that His-246, His-199, and Tyr-255 perform major roles in catalysis.

MATERIALS AND METHODS
Bacterial Strains and Plasmids-E. coli strain BL21(DE3) and plasmid pET22b(ϩ) were purchased from Novagen. Plasmid DNA purification was performed using the purification kit from Qiagen. Bacterial transformation was carried out according to Sambrook et al. (23). Plasmid pBZ3120 containing a 2-kb fragment coding for a dioxygenase activity isolated from P. stutzeri OX1 was kindly provided by Dr. P. Barbieri (Dipartimento di Biologia Strutturale e Funzionale, Università dell'Insubria, Varese, Italy).
Construction of the Expression Vector-Vector pET22b(ϩ)⌬XN was obtained by digesting vector pET22b(ϩ) with XbaI and NcoI. Sticky overhangs were blunted with T4 DNA polymerase and ligated. This deletion removes the endogenous ribosome binding site together with the sequence coding for the pelB leader. The sequence of the vector was verified by DNA sequencing.
To excise the C2,3O coding sequence, plasmid pBZ3120 was digested with BamHI and EagI. The fragment was purified by electrophoresis and ligated to vector pET22b(ϩ)⌬XN cut previously with the same enzymes. The resulting recombinant plasmid was named pET22b(ϩ)⌬XN/C2,3O.
Expression and Purification of C2,3Os-E. coli strain BL21(DE3), transformed with pET22b(ϩ)⌬XN containing the sequences coding for wild-type and the mutant C2,3Os, was grown at 37°C in Luria-Bertani medium (23) containing ampicillin (50 g/ml) until optical density at 600 nm was 0.6 -0.7. Induction was performed by adding isopropyl-1thio-␤-D-galactopyranoside and Fe(NH 4 ) 2 (SO 4 ) 2 at a final concentration of 0.5 mM and 100 M, respectively. After 3 h, cells were harvested by centrifugation at 4,000 ϫ g for 20 min at 4°C, washed with 50 mM Tris/HCl (pH 7.5), and collected by centrifugation. Cell paste was stored at Ϫ80°C until use. An SDS-PAGE analysis of induced and non-induced cell extracts showed the presence of a protein with a molecular mass of ϳ35 kDa only in induced cells (data not shown).
The bacterial pellet was suspended in buffer A (50 mM Tris/HCl (pH 7.5) containing 0.08 M NaCl, 10% (v/v) glycerol, 10% (v/v) ethanol, and 2 mM dithiothreitol) in the presence of 100 M Fe(NH 4 ) 2 (SO 4 ) 2 and an EDTA-free protease inhibitor mixture at the concentration suggested by the manufacturer and sonicated in a cell disruptor (10 ϫ 1 min cycle, on ice). Cell debris was removed by centrifugation at 12,000 ϫ g for 1 h at 4°C. The supernatant was loaded at 4°C onto a 15-ml Q-Sepharose column equilibrated previously in buffer A and washed with 30 ml of the equilibration buffer. Elution was carried out with 260 ml of buffer A with an NaCl linear gradient ranging from 0.08 to 0.5 M at a flow rate of 13 ml/h. Fractions were analyzed by SDS-PAGE and an activity assay on catechol. Fractions containing dioxygenase activity were collected and stored at Ϫ80°C under a nitrogen atmosphere.
Enzyme Assays-All assays were performed at 25°C in 50 mM Tris/ HCl (pH 7.5) or 50 mM sodium phosphate (pH 7.5) in a final volume of 500 l by spectrophotometric determination of the product of the reaction. Wild-type and mutant C2,3Os were used to start the reaction.
Kinetic parameters were determined by the program GraphPad Prism (GraphPad Software; www.graphpad.com). One unit of enzyme activity was defined as the amount of enzyme required to form 1 mol of the meta cleavage product per minute under the assay conditions. Specific activity is given as units per milligram of protein. Partition ratios were determined as the ratios between the concentration of product formed before the complete inactivation of the enzyme and the concentration of the enzyme in the assay (26).
Determination of M r Value-The M r value of the native wild-type enzyme was determined by gel filtration on a Superose 12 PC 3.2/30 (3.2 ϫ 300 mm) column equilibrated in 50 mM Tris/HCl (pH 7.5) containing 0.2 M NaCl, 10% ethanol (v/v), and 5% glycerol (v/v) using a SMART system (Amersham Biosciences). The molecular mass markers used as standards were ␣-amylase (M r , 200,000 Da), yeast alcohol dehydrogenase (M r , 150,000 Da), and aspartate aminotransferase (M r , 90,000 Da).
Determination of Titration Curves and pK a Values-Titration curves were determined spectrophotometrically as described (28). The theoretical titration curves were fitted to experimental data by the program GraphPad Prism that was also used to calculate the pK a values. Titration was followed at 238 nm for the determination of 3-formylcatechol pK a , 347 nm for the determination of 4-formylcatechol pK a , and 430 nm for the determination of 4-nitrocatechol pK a .
Determination of k cat Values as a Function of pH-The variations of k cat values as a function of pH were determined spectrophotometrically. When catechol was used as substrate, the reaction was followed at 375 nm (pH 5-10). When 4-formylcatechol was used, the reaction was followed at 400 nm (pH 5-10). When 3-formylcatechol was used, the reaction was followed at 300 nm (pH 5-6.8), 425 nm (pH 6.8 -7.5), and 450 nm (pH 7.5-10). ⌬⑀ values as a function of pH were calculated from the differences between the spectrophotometric titration curves of the product and the substrate. Spectrophotometric titration curves were determined as described above under "Determination of Titration Curves and pK a Values." All k cat values were determined at saturating concentrations of the substrate. Saturation was verified at each pH value by measuring the rate of the reaction at three increasing substrate concentrations.
Protein Sequence Determination and Mass Spectrometry Measurements-Protein sequencing, electrospray mass spectrometry measurements, and matrix-assisted laser desorption ionization mass spectrometry (MALDI/MS) analysis of peptide mixtures were performed as described previously (29).
Verification of C2,3O Primary Structure-The complete primary structure of wt-C2,3O was verified by peptide mapping. Aliquots of the high pressure liquid chromatography-purified protein were digested with either trypsin or Lys-C, and the resulting peptide mixtures were analyzed by MALDI/MS. The mass signals recorded in the spectra were mapped onto the predicted sequence of C2,3O on the basis of their mass value and the specificity of the protease.
Analytic Methods-SDS-PAGE was carried out according to Laemmli (30). Protein concentration was determined colorimetrically with the Bradford reagent (31) using bovine serum albumin as a standard. Total iron content was determined colorimetrically by complexation with Ferene S (32).
Chemicals-All chemicals were of the highest grade available and were from Amersham Biosciences, Promega, New England Biolabs, Sigma, ABCR GmbH, Fluka, or Applichem.

RESULTS
wt-C2,3O, Sequence of the Gene, Recombinant Protein Purification, and Characterization-The nucleotide sequence of the 2-kb BamHI-XhoI fragment isolated from a genomic library of P. stutzeri OX1 (33) cloned in pBZ3120 has been submitted to the EMBL data base (accession number AJ496739). This sequence was used to search protein databases (Swiss-Prot, TrEMBL, and TrEMBLnew) with the BLASTX program. Also, the deduced amino acid sequences were used to search for homologous sequences using the BLAST and FASTA programs. These analyses allowed the identification of two complete open reading frames, one coding for a plant-type [2Fe-2S] ferredoxin (34), the other coding for a catechol 2,3-dioxygenase (Table I). The open reading frame for C2,3O is 921 base pairs coding for a 307-amino acid polypeptide. Multiple amino acid alignments with other extradiol dioxygenases (Table I) indicate that C2,3O from P. stutzeri belongs to the major family of extradiol ring cleavage dioxygenases that have the substrate binding domain at the C terminus (5). Its sequence is 95% identical to that of the well characterized C2,3O from P. putida mt2; moreover, the two sequences have no insertions or deletions. Only one point mutation is located in the C-terminal domain containing the active site, with Val-225 in the P. stutzeri enzyme replacing Leu-225 in the P. putida enzyme. As this residue is 14 Å away from the catalytic iron ion in the P. putida C2,3O structure, it can likely be concluded that the active sites of the two C2,3Os are identical.
The purification of the enzyme, carried out as described under "Materials and Methods," yielded ϳ120 mg of protein per liter of culture with a specific activity on catechol of ϳ90 Ϯ 10 units/mg, depending on the preparation.
The molecular mass of denatured C2,3O was determined by electrospray mass spectroscopy and found to be 35,107.2 Ϯ 3 Da, in agreement with the value of 35,105.5 Da expected from the deduced amino acid sequence. The primary structure de-duced from the gene sequence was verified by peptide mapping as described under "Materials and Methods." The mapping revealed that recombinant C2,3O is expressed in E. coli with the initial Met residue.
The native molecular mass of wt-C2,3O was determined by gel filtration on a Superose 12 PC 3.2/30 column. The apparent molecular mass was found to be 117,000 Ϯ 20,000 Da. This value is consistent with the hypothesis that C2,3O is a homotetramer, as expected by its homology with other catechol dioxygenases (10).
Sensitivity to oxidation of the wild-type enzyme was verified by incubation in 0.04% H 2 O 2 for 30 min. This treatment resulted in the loss of ϳ98% of the activity. Incubation of the H 2 O 2 -inactivated enzyme in Fe(NH 4 ) 2 (SO 4 ) 2 solutions resulted in the recovery of ϳ35% of the activity. When Mn 2ϩ , Co 2ϩ , and Fe 3ϩ salts were used instead, no recovery of the activity was ever detected, thus indicating the specificity of the enzyme for Fe 2ϩ . The ratio of moles of iron to moles of a monomer of C2,3O was found to be 1.1 Ϯ 0.02. Moreover, the Ferene S assay carried out with and without the addition of ascorbic acid yielded the same value, thus indicating that iron is present as Fe 2ϩ .
The rate of ring fission of wt-C2,3O was determined on catechol and several substituted catechols (Table II). Catechol is the best substrate, followed by 4-methylcatechol and 3-methylcatechol. Larger groups at positions 3 or 4 cause further reduction of k cat values.
When 4-methylcatechol, 4-ethylcatechol, 3-metoxycatechol, 4-formylcatechol, and 4-nitrocatechol were used as substrates, a rapid inactivation of C2,3O was observed. In fact, the initial reaction rate rapidly decreased to zero. When fresh enzyme was added, the reaction started again, and the total amount of substrate converted was found to be proportional to the amount of the enzyme added. This behavior has already been observed for other C2,3Os and has been attributed to inactivation mediated by the substrate (35)(36)(37). The ability of the substrate to  induce inactivation of the enzyme was measured by the ratio between the amount of substrate converted before complete inactivation and the number of micromoles of the enzyme. These ratios, called "partition ratios" (26), are reported in Table  II. Note that 4-methylcatechol is a better substrate than 3-methyl and 3-formylcatechol, but it is also more inactivating. It should also be emphasized that all of substituents tested caused an increase in the ability of the substrate to inactivate the enzyme. Moreover, larger groups and groups at position 4 had a more pronounced effect. Kinetic Parameters of wt-C2,3O-To study the influence of substrate ionization on catalysis and the influence of substituents on the substrate, kinetic parameters of C2,3O were determined both on catechol and on substituted catechols bearing different electronegative groups that increase the acidity of the aromatic hydroxyl groups. pK a1 values of substituted catechols were determined spectrophotometrically as described under "Material and Methods" and are reported in Table III.  Table III shows the kinetic parameters determined for wt-C2,3O on catechol and acidic catechols at pH 7.5. It is notable that the k cat of wt-C2,3O decreases as substrate acidity increases, i.e. as the electron-withdrawing effect of the substituent group increases. In fact the k cat values for 3-formyl, 4-formyl, and 4-nitrocatechol are 23, 8, and 0.7% of the k cat value on catechol.
To gain information on the ionizable groups involved in catalysis, k cat values on catechol and formylcatechols were also determined at pH values between 5 and 10. Fig. 2A shows the apparent k cat value of wt-C2,3O on catechol as a function of pH. The curve is markedly asymmetric. In fact, it shows a maximum at pH 6, whereas it decreases at lower and higher pH values with very different degrees of steepness. Activity drops quickly at pH values Ͻ6, and at pH 5 activity is already near zero. On the other hand, at pH values Ͼ6, activity decreases very slowly and a shoulder is present between pH 7.5 and 8.5.
To exclude an effect of buffer ions between pH 6.5 and 9, activity was measured using at least two different buffer systems for each pH, i.e. MOPS, sodium acetate, Tris-HCl, or CHES/NaOH. Identical k cat values were determined (data not shown), thus excluding any buffer effect. Fig. 2, B and C show the variation of k cat values on 3-and 4-formylcatechol as a function of pH. The two curves are very similar and bell-shaped, in contrast with the curve obtained using catechol as substrate. For both formylcatechols, k cat values show a large maximum between pH 7 and 8 that drops at pH values Ͼ8, whereas they slowly decrease at pH values Ͻ7. This decrease is slower for 4-formylcatechol than for 3-formylcatechol.
To obtain an estimate of the pK a values of the residues responsible for the shape of the curves shown in Fig. 2, A-C, we developed a model for the reaction of C2,3O with its substrates and derived an approximate mathematical solution that was used to fit the experimental data. The model, shown in Fig. 3, hypothesizes the presence of two different enzyme active forms, EH 2 and EH, which depend on the ionization state of three ionizable groups on the enzyme. Each of these two species can productively bind two forms of the substrate, SH and S Ϫ with different affinities (K EHi SH and K EHi S ). The model also assumes that productive reactions have different k cat values and that EH i SH and EH i S Ϫ species do not directly convert one into the other (Fig. 3).
The derivation of the equation and its fitting to experimental data, described in detail in the supplemental material available in the on-line version of this article, allowed us to obtain three sets of values for the three pK a values of the groups that influence catalysis. The pK a values of the two most acidic residues are 5.5 Ϯ 0.1 and 6.6 Ϯ 0.15. These values suggest the involvement in catalysis of two histidine residues likely positioned in different environments. The analysis of the structure of P. putida mt2 C2,3O showed that, excluding the His-153 and the His-214 that bind the iron ion, only two other histidines are placed in the proximity of the active site cavity, His-199 and His-246 (Fig. 1). Thus, these residues can be tentatively identified as the residues with pK a values of 5.5 and 6.6.
The third group is a weaker acid, the pK a value of which was determined with lower accuracy and falls in the range between pH 8.9 and 9.9. In principle, this pK a value could be assigned to a tyrosine, a lysine, or a cysteine residue. Residue Tyr-255, at Ͻ4 Å from the iron ion, is a likely candidate.
The fitting procedure also allowed us to obtain information on the k cat values of the different active forms of the enzyme. When catechol is used as substrate, reaction 2, between EH 2 and S Ϫ , gives a negligible contribution because the EH 2 form is present only at pH values in which catechol is completely in its neutral form ( Figs. 2A and 3). Therefore the shape of the activity curve on catechol is determined only by the values of k cat1 , k cat3 , and k cat4 . The shoulder at the alkaline pH values depends only on the EH form ( Fig. 2A, dotted lines, and Fig. 3).
It should be noted that k cat3 , and k cat4 values, i.e. the catalytic constants of EH species, are ϳ2.6 -3 times lower than that of k cat1 (supplemental Table S1, available in the on-line version of this article), indicating that the EH 2 form is more active than the EH form. A different picture emerges from the fitting of the model to experimental data collected using formylcatechols as substrates. In this case, the EH form has k cat values higher than those of the EH 2 form (Table S1 and the dotted lines in Fig. 2, B and C), indicating that the EH form is more active on these substrates than the EH 2 form. Kinetic Parameters of Mutated C2,3O-To investigate the role of the conserved residues His-199, His-246, and Tyr-255 in the catalytic process of P. stutzeri C2,3O, four mutants were produced by site-directed mutagenesis as described under "Materials and Methods," namely (H199N)-, (H246N)-, (H246A)-, and (Y255F)-C2,3O. These mutants were expressed, purified, and characterized. All mutated proteins showed an Fe 2ϩ content similar to that of wt-C2,3O within the experimental error. Table III shows the kinetic parameters determined at pH 7.5 using catechol or formylcatechols as substrates. Mutation H199N leads to a 400-fold decrease of the k cat value on catechol, whereas the K m value is increased 290 times. Moreover, the inverse correlation between k cat values and the substrate acidity already observed for the wild-type enzyme is much more evident in the case of the mutant (H199N)-C2,3O. In fact, the k cat values for 3-formyl and 4-formylcatechol are 6.5 and 1% of those measured for catechol.
On the other hand, the mutation H246N causes a slightly less pronounced decrease of the k cat value for catechol, which is 125-fold lower than that of wt-C2,3O, but a larger increase of the K m value, which is 480-fold higher than that of wt-C2,3O. It is noteworthy that the relationship between the k cat and the substrate acidity of (H246N)-C2,3O is considerably different from that observed in the case of wt-and (H199N)-C2,3O (Table III). In fact, the k cat value of (H246N)-C2,3O for 3-formylcatechol is ϳ4-fold higher than that for catechol, whereas wtand (H199N)-C2,3O show k cat values for catechol that are 5-to 10-fold higher than those for acidic catechols. Moreover, the k cat value of (H246N)-C2,3O for 4-formylcatechol is only 3.7fold lower than that for catechol, in contrast with the 12.5-and 100-times reduction measured in the case of the wild-type protein and the (H199N)-C2,3O mutant, respectively.
As for mutation H246A, it has a dramatic effect on catalysis; the K m value for catechol increases Ͼ3000-fold, and the k cat value decreases ϳ450-fold. The mutation Y255F causes a limited effect on the K m value, which increases only ϳ3-fold, and a larger effect on the k cat value, which decreases 25-fold. The relationship between the k cat values and substrate acidity is very similar to that of wt-C2,3O (Table III).
The dependence of the k cat values of mutated proteins on pH was studied using catechol as the substrate. The mutants (H246N) and (H246A)-C2,3O show pH profiles that are different from that of the wild-type enzyme and different from each other (Fig. 2D). Mutant (H246N)-C2,3O shows a maximum at a pH near 8 with a shoulder between pH 8 and 9. Moreover, the maximum at pH 6 is absent. Mutant (H246A)-C2,3O shows a simpler profile with only a wide but very low maximum at pH 8.5-9.
The C2,3O from P. stutzeri that we purified is very similar to other well characterized C2,3Os (see Table I), particularly to that from P. putida MT2, with ϳ95% sequence identity. The differences are mainly located in the N-terminal domain, whereas in the C-terminal catalytic domain of the two proteins only one conservative mutation is present, 14 Å away from the catalytic iron. Thus, P. stutzeri C2,3O also retains the conserved active site residues His-246, His-199, and Tyr-255 for which a role in the catalytic mechanism of dioxygenases has been hypothesized (8,10,11). However, it should be emphasized that no direct experimental evidence is available on the function of these residues. Only in the case of residue His-200 (His-199 in C2,3O from P. putida and P. stutzeri) of the irondependent HPCD from B. fuscum, data have been collected on the effect of its mutation to phenylalanine, causing a 99% reduction of activity on homoprotocatechuate (22).
To gain information on the ionizable groups involved in the catalytic action of P. stutzeri C2,3O, we studied the dependence of its activity on catechol as a function of pH. Surprisingly, we found that the curve describing the variation of the k cat value as a function of pH is asymmetric (Fig. 2A). It should be remembered that when two distinct ionizable groups are involved in catalysis, a bell-shaped curve is usually measured for the dependence of the activity from pH. Holman et al. (38) reported that stromelysin displays an activity profile as a function of the pH with a maximum at pH 6 and a pronounced shoulder at alkaline pH, which strongly resembles the results we obtained in the case of C2,3O. They proposed a mechanism involving three ionizable groups that generate two active forms of the enzyme, each characterized by a different catalytic efficiency (38). We found that the curve describing the dependence of C2,3O activity on catechol from pH can be explained by a similar but more complex model involving not only three ionizable enzyme groups, hence two differently active forms (see the scheme in Fig. 3 and the supplemental material in the on-line version of this article), but also one substrate-ionizable group. Moreover, our model is based on the hypothesis that each active form possesses different k cat values and different affinities for fully protonated catechol and for its monoanionic form. Thus, the experimental curve can be obtained from the sum of four distinct Gaussian curves representing the contribution of the two different enzyme forms (EH 2 and EH; see Fig.  3), each reacting with two differently ionized substrate molecules (SH and S Ϫ ; see Fig. 3). The fitting of the model to the experimental data ( Fig. 2A) led to the suggestion that the three enzyme-ionizable groups should have pK a values of ϳ5.5, 6.6, and 9 -10 (see the legend of Fig. 2), respectively. It should be noted that the two more acidic pK a values (5.5 and 6.6) agree with the involvement of two histidine residues in catalysis. Moreover, the presence of a maximum of activity at pH 6 would indicate that the enzyme species present at this pH is the most active one. This form in our model is EH 2 , which should bear a protonated and a deprotonated residue. These two residues were tentatively identified with the conserved histidine residues 199 and 246, based on the following considerations. The analysis of the structures of C2,3O (Protein Data Bank code 1MPY) and DHBD (Protein Data Bank codes 1KW9, 1KMY, 1KND, and 1KNF) from P. putida shows that the conserved His-199 and His-246 are located in two rather different environments (Fig. 1, A and B). As for His-199, its N ⑀2 atom points toward the substrate binding pocket (in DHBD structures it is H-bonded to the OH group of the substrate), whereas its N ␦1 atom points toward a narrow solvent-filled channel (Fig. 1, A  and B). As for His-246, it cannot form H-bonds with the substrate, but the possibility exists that it may be H-bonded to Glu-265 through its N ⑀2 atom and/or to the backbone CO of Ile-291 through its N ␦1 atom. Thus, it can be hypothesized that when His-246 is protonated, both H-bonds are formed (as shown in Fig. 1, A and B), whereas only one H-bond will be formed in its deprotonated form. Thus, based on the close proximity of His-246 to Glu-265 and on its potential to form two hydrogen bonds in its protonated form, it can be hypothesized that His-246 is the protonated histidine in the EH 2 species of our model. On the other hand, we suggest that His-199 is in its deprotonated form in the EH 2 species as required by the mechanism shown in Fig. 1C.
As for the third group with a pK a value in the range 9 -10, Tyr-255, located at Ͻ4 Å from the iron ion, is a likely candidate. It should be emphasized that its proximity to the iron ion could explain a pK a value Ͻ10.
To test the hypothesis that the three ionizable groups of our model are the residues His-199, His-246, and Tyr-255, we prepared and characterized mutants (H199N)-, (H246N)-, (H246A)-, and (Y255F)-C2,3O. The kinetic parameters determined at pH 7.5 (Table III) indicate that the mutation H199N influences both k cat and K m values. It should be noted that the k cat value decreases 430 times, thus indicating that His-199 is very important for catalysis, in line with the result already obtained on HPCD from B. fuscum (22). Moreover, the K m value increases from 1.5 to ϳ400 M. Assuming that K m values are a measure of substrate affinity, this finding indicates that His-199 is also important in substrate binding, likely because of the formation of an H-bond with the substrate and/or by contacting the catechol ring (Fig. 1A). Moreover, its k cat versus pH profile is flat in the entire pH range tested, and the k cat values are very low at all pH values with respect to those of the wild-type enzyme, as expected for a residue that is involved in catalysis at all pH values. Thus, we can conclude that His-199 is essential for both active forms of C2,3O (EH 2 and EH; see Fig. 3) in its deprotonated form. This finding may be suggestive of a proton acceptor role for His-199.
As for His-246, two different substitutions were made, H246N and H246A. H246N was designed based on the hypothesis that an asparagine residue could still bind the backbone of Ile-291 through an H-bond (Fig. 1), whereas this H-bond would also be lost in the case of H246A. Our data indicate that both mutations cause a decrease in the K m values and in catalytic efficiency. In fact, the K m values for the catechol of mutants (H246N)-and (H246A)-C2,3O are 470-and 3000-fold higher, respectively, than the K m value of the wild-type enzyme. Based on the structures available for two DHBDs (8,11), we can hypothesize that the ring of His-246 could provide a rigid surface for positioning the catechol molecule bound to iron (Fig.  1, A and B). This surface would be reduced in (H246N)-C2,3O and be completely removed in (H246A)-C2,3O, thus increasing the K m values of the two mutants. On the other hand, both mutations at position 246 cause a decrease in the k cat values and a surprising change in their pH dependence. In fact, the activity of H246N is almost completely abolished at a pH Ͻ7, whereas a narrow new peak is present at pH 8 with a pronounced shoulder at pH 9.5-10. Instead, mutant H246A activity profile shows only a large peak at alkaline pH, and its activity is further reduced with respect to H246N.
These findings indicate that His-246 contributes to determining both the maximum at pH 6 in the activity profile in its protonated form (the EH 2 form of the model) and the shoulder at pH 8 when deprotonated (EH form of the model). Thus, His-246 would be the residue involved in the conversion EH 2 N EH in the model of Fig. 3. Moreover, its contribution to catalysis is more important in its charged form at acidic pH values, given the higher k cat values at a pH of ϳ6 with respect to those at pH values of ϳ8.
Therefore we can conclude the following: (i) that His-246 is the major residue responsible for the unusual shape of C2,3O activity dependence from pH; and (ii) that His-246 is essential in its protonated state for determining the activity of the form EH 2 of C2,3O, whereas when unprotonated it triggers the activity of EH (Figs. 2 and 3).
As for mutant (Y255F)-C2,3O, the kinetic constants we de-termined (Table II) indicate that the hydroxyl group of Tyr-255 does not provide an important contribution to K m but that it significantly contributes to the k cat value of wt-C2,3O. In fact, the k cat value of mutant Y255F is 25-fold lower than that measured for the wild-type enzyme. Moreover, the variation of its k cat value on catechol with pH is very similar to that observed in the case of wt-C2,3O on the same substrate. Therefore, experimental k cat values measured at different pH values for this mutant were fitted to the same model developed in the case of the wild-type enzyme. The fitting procedure yielded three pK a values very similar to those obtained for C2,3O (data not shown) with a pK a3 value only slightly increased (10.2 instead of 9.9). This would allow the conclusion that Tyr-255 does not give any, or gives only a minor, contribution to the decrease of the activity of the enzyme at alkaline pH values. The above conclusions strongly support the identification of resides His-199, His-246, and Tyr-255 as catalytic residues, the functions of which are related to their ionization states. However, this finding gives no clues as to the identification of their exact roles in the catalytic mechanism, although it may be suggestive of their involvement in the extraction and/or release of a proton from/to the substrate.
To investigate this latter suggestion, the catalytic parameters and their pH dependence were measured using more acidic substrates. 3-formylcatechol (2,3-dihydroxybenzaldehyde; pK a ϭ 7.7) and 4-formylcathechol (3,4-dihydroxybenzaldehyde; pK a ϭ 7.2) were selected on the basis of their k cat values (Table  II and III), which are within one order of magnitude from the k cat value for catechol, and their low inactivating activity (Table II). As shown in Fig. 2, B and C, the curves recorded on both substrates for the dependence of the k cat of wt-C2,3O from pH have a single, very large, and slightly asymmetric peak.
When the model described in Fig. 3 was used to fit the experimental data, an interesting result was obtained. Using the fitting procedure described in the supplemental material (available in the on-line version of this article), in this case the experimental curve could also be reproduced as the sum of four distinct Gaussian curves, each representing the contribution of the two different enzyme forms on the two forms of the substrate, but with the acidic curve maximum (curves 1 of Fig. 2, B and C) lower that the basic curve maximum (curves 4 of Fig.  2, B and C). It should be emphasized that this is the opposite of the result obtained in the fitting procedure of the data obtained using catechol as a substrate ( Fig. 2A). In fact, the resolution of the equations based on the model reported in Fig. 3 as described in the supplemental material for an acidic catechol leads to the conclusion that, in contrast with the case in which catechol is the substrate, the reaction between the EH form and the deprotonated substrate provides the major contribution to the apparent k cat values. This finding supports the role of His-246 as a proton-abstracting residue because the easier ionization of the substrate (3-or 4-formylcatechol) would compensate the decrease in activity recorded at pH values higher than 6, depending on the deprotonation of His-246.
Further insight into the roles of His-246 and His-199 in catalysis comes from the comparisons of the kinetic constant values of wt-, (H246N)-, (H199N)-C2,3O, and (Y255F)-C2,3O on 3-and 4-formylcatechol at pH 7.5 (Table III). wt-C2,3O has the highest k cat value on catechol, whereas it decreases with the increase of the acidity of the substrate; the apparent k cat values on 3-formyl, 4-formyl, and 4-nitro catechol are ϳ20, 10, and 1% of the corresponding values measured for catechol, respectively (Table III). This result is surprising, because the formyl and nitro groups help the deprotonation of the substrate, and, hence, they should increase the efficiency of the reaction. The behavior observed can be explained by the hy-pothesis that these groups also reduce catecholate reactivity toward O 2 by their electron-withdrawing effect. This observation rules out the hypothesis of a nucleophilic attack of O 2 on the substrate as proposed for C2,3O (10, 39) and supports an electrophilic radical attack as proposed for DHBD (8,11).
As for the mutant H199N, its apparent k cat values on 3-formyl and 4-formyl catechol are ϳ6 and 1% of the corresponding values measured for catechol (Table III). Thus, a formyl group on the aromatic ring leads to a more reduced efficiency of the catalytic process with respect to the effect measured for the wild-type enzyme. This behavior, different from that recorded in the case of H246N (see below), would suggest that the major role of residue His-199 is not to deprotonate the second hydroxyl group of the substrate but instead to activate the catechol ring and/or O 2 for the reaction. On the basis of the structure of DHBD bound to the aromatic substrate and to NO (an analog of O 2 ), Sato et al. (8) recently suggested that His-194 (homologous to His-199 in C2,3O) has at least three different roles (Fig. 1C, Steps 3 and 4) as follows: (i) abstraction of the second hydrogen atom from the aromatic substrate; (ii) stabilization of the partially reduced oxygen intermediates to promote the oxidative attack on the aromatic substrate; and (iii) release of a proton in the late stages of the reaction. The structural data of manganese-dependent and irondependent HPCD also suggest these different roles (12). Our data, excluding the fact that the acid-base catalysis is the primary role of His-199, might indirectly support the involvement of this residue in the activation of oxygen for the attack on the aromatic ring.
On the other hand, the behavior of the mutant (H246N)-C2,3O is very different, with a k cat value on 3-formylcatechol 4-fold higher than that on catechol (Table III). Moreover, the k cat value on 4-formylcatechol is only 30% of that on catechol. This value is to be confronted with the 10% reduction measured in the case of wt-C2,3O (Table III). Several chemical and structural studies indicate that the initial deprotonation of catechol is the first crucial step in extradiol cleavage reactions (8, 11, 14 -16) and that all proposed mechanisms start from catechol ionization (8,10,11). A formyl group at position 3 favors spontaneous deprotonation of the substrate, hence compensating the reduced ability of the mutant enzyme to induce its ionization. Based on this consideration, the increase in the k cat value of the H246N mutant on 3-formylcatechol with respect to that of wt-C2,3O is a strong indication that His-246 is directly involved in this step of catalysis. On the other hand, the effect measured with 4-formylcatechol would suggest that the reduced substrate reactivity toward O 2 mediated by the formyl electron-withdrawing effect likely prevails on the facilitated deprotonation of catechol, leading to a reduction in the k cat , although less pronounced than that measured in the case of wt-C2,3O. Thus, our conclusion agrees with the hypothesis advanced by Vaillancourt et al. (11) that His-241 in DHBD, homologous to His-246 in P. stutzeri C2,3O, is involved in substrate deprotonation, likely in its deprotonated form at pH 7.5 (Fig. 4B). It should be noted that we have already provided evidence that His-246 is also essential in its protonated state. An interesting possibility also exists that positively charged His-246, together with catalytic Fe 2ϩ , both contribute to the deprotonation of the substrate, the aromatic ring of which rests on the imidazole ring of His-246 as discussed above, by lowering its pK a , hence facilitating its deprotonation (Fig. 4A). The proton acceptor could be a solvent molecule bound to Fe 2ϩ (Fig. 4A) as initially proposed by Kita et al. (10) and more recently by Vetting et al. (12).
Finally, the k cat values on formylcatechols of mutant (Y255F)-C2,3O indicate that the increase in substrate acidity does not compensate for the loss of the hydroxyl group at position 255. Therefore, unlike His-246, Tyr-255 does not seem to be directly involved in substrate deprotonation. The role of the hydroxyl group could be that of binding the negative oxygen of catecholate (Figs. 1A and 4), thus stabilizing the EH (2) S Ϫ complex. Moreover, the hypothesis can be advanced that Tyr-255 and His-199 play the role of inducing asymmetry in catechol through hydrogen bonds, thus favoring specific deprotonation of the hydroxyl group, which is on the same side of Tyr-255 (Figs. 1A and 4). Instead, His-246 would be more directly involved in proton abstraction.