Vol. 175, No. 14 JOURNAL OF BACTERIOLOGY, July 1993, p. 4414-4426 0021-9193/93/144414-13$02.00/0 Copyright © 1993, American Society for Microbiology Purification and Characterization of Protocatechuate 2,3-Dioxygenase from Bacillus macerans: a New Extradiol Catecholic Dioxygenase SANFORD A. WOLGEL,1t JAY E. DEGE,1 PATRICIA E. PERKINS-OLSON, 2 CARLOS H. JUAREZ-GARCIA,2t RONALD L. CRAWFORD,2'3 ECKARD MUNCK,2'4 AND JOHN D. LIPSCOMB`* Department of Biochemistry, University of Minnesota, Minneapolis, Minnesota 554551; Gray Freshwater Biological Institute, University of Minnesota, Navarre, Minnesota 553922; Department of Bacteriology and Biochemistry, University of Idaho, Moscow, Idaho 838433; and Department of Chemistry, Carnegie Mellon University, Pittsburgh, Pennsylvania 152134 Protocatechuate 2,3-dioxygenase (2,3-PCD) from Bacillus macerans JJlb has been purified to homogeneity for the first time. The enzyme catalyzes proximal extradiol ring cleavage of protocatechuate (PCA) with the attendant incorporation of both atoms of oxygen from 02. The holoenzyme has a mass of 143 7 kDa as determined by ultracentrifugation and other techniques. It is composed of four apparently identical subunits with Mrs of 35,500, each containing one iron atom. Mossbauer spectroscopy of "7Fe-enriched enzyme showed that the irons are indistinguishable and are high spin (S = 2) Fe2+ in both the uncomplexed and substrate-bound enzyme. However, the quadrupole splitting, AEQ, and isomer shift, 8, of the Mossbauer spectrum changed from AEQ = 2.57 mm/s and 8 = 1.29 mm/s to AEQ = 2.73 mm/s and 8 = 1.19 mm/s upon PCA binding to the enzyme, showing that the iron environment is altered when substrate is present. The enzyme was also found to bind variable and substoichiometric amounts of Mn2", but this metal could be removed without loss of activity or stability. The inherently electron paramagnetic resonance (EPR)-silent Fe2+ of the enzyme reversibly bound nitric oxide to produce an EPR-active species (g = 4.11, 3.95; S = 3/2). The specific activity of the enzyme was found to be correlated with the amount of the S = 3/2 species formed, showing that activity is dependent on Fe2'. Anaerobic addition of substrates to the enzyme-nitric oxide complex significantly altered the EPR spectrum, suggesting that substrates bind to or near the iron. The enzyme was inactivated by reagents that oxidize the Fe2 such as H202 and K3Fe(CN)6; full activity was restored after reduction of the iron by ascorbate. Steady-state kinetic data were found to be consistent with an ordered bi-uni mechanism in which the organic substrate must add to 2,3-PCD before 02. The enzyme has the broadest substrate range of any of the well-studied catecholic dioxygenases. All substrates have vicinal hydroxyl groups on the aromatic ring except 4-NH2-3-hydroxybenzoate. This is the first substrate lacking vicinal hydroxyl groups reported for catecholic extradiol dioxygenases. 2,3-PCD is the final member of the PCA dioxygenase family to be purified. It is compared with other members of this family as well as other catecholic dioxygenases. , Protocatechuate (PCA) is one of a relatively small number of single-ring aromatic compounds that are found at the points of confluence of bacterial pathways for metabolism of complex aromatics (12, 17, 28, 37). The aromatic ring of PCA is opened in reactions catalyzed by dioxygenase enzymes that result in the incorporation of both atoms of 02 into the open chain products (14, 19, 25, 46, 65). Two of these enzymes, protocatechuate 3,4-dioxygenase (3,4-PCD) (10, 21, 25, 59, 65, 70) and 4,5-PCD (5, 18, 19, 53) were among the first of this class of enzyme to be recognized. They continue to serve as prototypical enzymes for the two major subclasses of catecholic dioxygenases termed intradiol and extradiol on the basis of the site of ring cleavage (for reviews see references 37 and 46). Intradiol dioxygenases like 3,4PCD contain Fe3" (26, 59, 70) and open the aromatic ring between the vicinal hydroxyl groups (65). In contrast, extradiol dioxygenases such as 4,5-PCD invariably contain Fe2+ (5, 67) and open the ring adjacent to one of the hydroxyl groups to form highly colored muconic semialdehyde prod- ucts (19). The analogous set of reactions also occurs in the oxidative cleavage of catechol. The relevant intradiol and extradiol dioxygenases are, respectively, catechol 1,2-dioxygenase (pyrocatechase, or 1,2-CTD) (34) and catechol 2,3dioxygenase (metapyrocatechase, or 2,3-CTD) (35, 44, 47). The intradiol and extradiol nomenclature cannot be applied to enzymes that catalyze ring cleavage of key intermediates, such as gentisate, that do not have vicinal hydroxyl groups. Nevertheless, we have shown recently that gentisate 1,2dioxygenase (1,2-GTD) contains Fe2' and behaves mechanistically like an extradiol dioxygenase (29, 30). The cleavage site of the PCA ring by dioxygenases is completely specific; i.e., extradiol dioxygenases do not cleave in an intradiol manner and vice versa. Moreover, the PCA ring cleavage site in pseudomonads is species specific and is used as a taxonomic characteristic (66). To determine whether this specificity was present in Bacillus species, Crawford investigated the pathways of 4-OH benzoate degradation in various bacilli (14, 15). Interestingly, strains of aerobic spore-forming Bacillus circulans and Bacillus macerans were identified that contained a novel extradiol PCA dioxygenase that catalyzed opening of the aromatic ring * Corresponding author. t Present address: Conta Luna Foods, Grand Forks, ND 58203. t Present address: Eastman Chemical Co., Kingsport, TN 376625150. 4414 Downloaded from http://jb.asm.org/ on January 16, 2018 by guest Received 10 March 1993/Accepted 3 May 1993 VOL. 175, 1993 PROTOCATECHUATE 2,3-DIOXYGENASE FROM BACILLUS MACERANS COOH PCA COOH °2-1 H & OH ; / j~i,-> 0\ o~gOO' COOH OHC HOOC OH OH a-Hydroxy-carboxymuconic semialdehyde 9COOH OH a-Hydroxy-8-carboxymuconic semialdehyde COOH QCOOH COOH f3-Carboxy-cis,cismuconic acid FIG. 1. Known positions of ring cleavage of PCA. Chemicals. All chemicals were of the best quality available as supplied by Aldrich or Sigma Chemical companies and were used without further purification. 3-Chlorophenol was the generous gift of Peter Chapman (Environmental Protection Agency, Gulf Breeze, Fla.). Water was deionized and glass distilled. Enzymes. Pseudomonas testosteroni 4,5-PCD (5) and Pseudomonasputida (arvilla) 2,3-CTD (35) were prepared as previously described (5, 6) and had specific activities of 200 and 320 U/mg, respectively. Purification of 2,3-PCD. B. macerans JJ1b (ATCC 35889 [16]) was grown as previously described (71). The culture procedure is straightforward, except for the fact that there is a critical time window during which the cells can be harvested or transferred. It begins when the growth substrate, 4-hydroxybenzoate, is depleted, and it lasts for less than 1 h. The purification procedures reported here were adapted from preliminary studies (16, 71). All procedures were performed at 40C, and the buffer was 50 mM MOPS (morpholine propanesulfonic acid [pH 6.9]) with 100 pM Fe(NH4)2(SO4)2. 6H20 and 2 mM cysteine (10x stock prepared daily) unless otherwise indicated. All centrifugations were at 35,000 x g for 20 min. Approximately 100 g of cell paste was suspended in 300 ml of buffer and broken by sonication (Branson Inc., model 350) at 15'C. After treatment with -3 mg each of RNase, DNase, and MgCl2, the mixture was centrifuged. The supernatant was then diluted with buffer to bring the ionic strength below that of 0.025 M NaCl in buffer and was loaded onto a DEAE-Sepharose Fast Flow (Pharmacia/LKB) column (5 by 25 cm) equilibrated with buffer. The column was washed with 1 liter of buffer, and then 2,3-PCD was eluted with a linear gradient (1 by 1 liter) from buffer to 0.6 M NaCl in buffer. The fractions exhibiting >50 and >10% of the activity of the peak fraction on the low- and high-salt sides of the peak, respectively, were pooled and concentrated to 70 ml by ultrafiltration under N2 on a YM-10 membrane (Amicon Inc.). This solution was made 1.5 M NaCl in buffer and was then applied to a phenyl-Sepharose (Pharmacia/LKB) column (4 by 10.5 cm) preequilibrated in 1.5 M NaCl in buffer. The column was washed with 250 ml of 1.5 M NaCl in buffer. Apparently homogeneous enzyme was eluted as a broad peak near the middle of a linear gradient (250 by 250 ml) from 1.5 M NaCl in buffer to buffer with no added NaCl. Fractions with >5% of the peak activity were concentrated under pressure as described above to 15 ml (>7.5 mg/ml). Enzyme was stored in liquid N2 and was stable for >2 years. Incorporation of 57Fe. The 2,3-PCD used in Mossbauer experiments was prepared from cells grown in media containing 57Fe (Icon Inc.). All glassware used in culturing the bacteria and purification of the enzyme was acid washed before use. With the exception of the culture plate growth media and the soil extract media, all of the growth media contained 1.4 mg of s7Fe metal dissolved in 6 N HCl as the only source of iron per liter. The purification of 2,3-PCD was identical to that described above, except that s7Fe dissolved in 6 N HCl was substituted for Fe(NH4)2(SO4)2. 6H20. Enzyme assay. Routine assays of 2,3-PCD activity were made by monitoring the rate of 02 uptake with a Clarke-type oxygen electrode (YSI). Enzyme was added last to the reaction chamber containing 1.3 mM PCA in air-saturated 50 mM MOPS buffer (pH 7.0) at 23°C. Calculations and electrode calibration were performed as previously described (70, 71). Downloaded from http://jb.asm.org/ on January 16, 2018 by guest between the 2 and 3 carbons (2,3-PCD) (14, 15). The enzyme partially purified from B. macerans (16). With the discovery of this enzyme, the known positions of ring cleavage of PCA can be depicted as shown in Fig. 1. Studies of 3,4-PCD and 4,5-PCD have contributed significantly to our understanding of the structures and mechanisms of intradiol and extradiol dioxygenases. Both have been crystallized (5, 25, 51, 62, 70), and the three-dimensional structure of 3,4-PCD has been solved (50). This is the only structure of a dioxygenase currently available. A wide range of spectroscopic techniques has been applied to each enzyme that revealed differences in the ligation spheres of the active-site irons as well as the manner in which they react with organic substrates and oxygen analogs (37). For example, the technique of forming the nitrosyl complex of the active site Fe2+ of mononuclear iron-containing enzymes and proteins was developed, in part, during the spectroscopic investigation of 4,5-PCD (4-6). This technique converts the normally electron paramagnetic resonance (EPR)silent Fe2+ to an EPR-active form, thus allowing the number and type of iron ligands to be assessed by transferred hyperfine coupling effects from isotopically labeled ligands. On the basis of these spectroscopic and crystallographic studies, molecular mechanisms for both intradiol and extradiol dioxygenases have been proposed (4, 37, 58, 69). For each type of enzyme, PCA is proposed to initially bind directly to the iron through both hydroxyl groups to form a chelate complex (4, 56). In the case of 4,5-PCD, 02 is then proposed to bind directly to and be activated by the iron before insertion into the PCA ring. In contrast, the Fe3+ of 3,4-PCD is not proposed to bind 02. Rather, it is thought to activate the bound PCA so that 02 can attack the substrate directly. Because of the pivotal role 3,4-PCD and 4,5-PCD have played in our understanding of the structure and mechanism of ring cleavage dioxygenases, it is important to completely characterize the one remaining PCA dioxygenase, 2,3-PCD, for comparative studies. This is particularly important because the extradiol mechanism is much less firmly established than the intradiol mechanism because of the difficulties of studying the inherently EPR-silent and colorless Fe2+ of the extradiol enzyme class. The availability of 2,3-PCD as another extradiol enzyme that utilizes the same substrate as 4,5-PCD offers the best approach to determining which aspects of the structure and spectroscopy are essential to extradiol catalysis. We describe here the first purification of this enzyme to homogeneity as well as its physical and spectroscopic characterization. was MATERUILS AND METHODS ECHO 2H 3 PCD 4415 4416 J. BACTERIOL. WOLGEL ET AL. = D- f (1) f 61 rh (2) The Stokes radius, rh in equation 2, is obtained directly from the gel filtration data, and the other terms are the usual constants (see reference 11). Do was calculated for water at 200C. The Perrin shape factor, F, is the ratio of the translational frictional coefficients, f/fo, wheref is obtained from equation 2 and fo is that of an unhydrated spherical protein given by equation 3 (24): = fo = 67 (3mv 1 ,r 1/3 (3) The molecular mass, m, is obtained by dividing the MW (from sedimentation velocity measurements) by 6 x 102. The partial specific volume, v, is approximately 0.72 cm3 g-1 (11). Calibrated gel filtration chromatography. MW was also estimated by calibrated gel filtration chromatography with a Bio-Gel A-0.5m column (1 by 90 cm) and appropriate protein standards. The total column volume was determined by locating NaCl added with the sample by using a refractometer. Plots to determine MW were made with the Stokes radius of the standard proteins calculated from the known molecular masses and diffusion coefficients. The MW of 2,3-PCD was then estimated from the determined Stokes radius and the diffusion coefficient obtained from the ultra- centrifugation experiment described above with the Einstein-Sutherland equation. SDS-PAGE. Sodium dodecyl sulfate-polyacrylamide gel electrophoresis (SDS-PAGE) was performed according to the procedure of Laemmli (36). Protein was stained overnight with 50% CH30H-10 CH3COOH-0.25% Coomassie brilliant blue and was destained with 5% CH30H-10% CH3COOH. Amino acid analysis and amino-terminal sequence. The holoenzyme was reduced and S carboxymethylated according to the method of Lundell and Howard (39). Hydrolysis and amino acid analysis procedures were performed as previously described (5, 70). Cysteine was measured as carboxymethylcysteine. Tryptophan was measured from extinction coefficients at 280 and 288 nm according to the method of Edelhoch (22). Analysis of the NH2-terminal sequence of S-carboxymethylated holoenzyme was performed with a Beckman System 890 sequenator. Highperformance liquid chromatography analyses of the phenylthiohydantoin derivatives of amino acids were performed with an Altex C-18 reverse-phase column (Ultrasphere ODS) with authentic standards. Spectra. EPR spectra were recorded on a Varian E-109 spectrometer equipped with an Oxford Instruments ESR-910 liquid helium cryostat. Data were recorded for analysis with a computer interfaced directly to the spectrometer. The Mossbauer spectrometer was of the constant acceleration type (70). All isomer shifts, bFe' are quoted relative to iron metal at room temperature. Optical spectra were recorded on a Hewlett Packard 8451A diode array spectrophotometer. Analysis of EPR and Mossbauer spectra. The EPR spectra of proteins with mononuclear iron sites are largely determined by the spin of the electronic ground state (64). The spin depends upon the oxidation state of the iron as well as the nature of the coordination environment. For Fe3+, the 5 d-electrons can combine to yield ground state spins of S = 5/2, 3/2, or 1/2. Fe2 , on the other hand, has an even number of d-electrons which combine to yield states with S = 2 or 0. Most frequently, high-spin (S = 2) ferrous ions do not permit the observation of EPR because the five spin levels exhibit splittings in a zero magnetic field that exceed the energy of the EPR microwave quantum. However, more recently some Fe2+ sites have been shown to be amenable to EPR (31). By complexing a ligand containing an unpaired electron to a site with integer spin, one can transform the site into an EPR-active form with half-integral electronic spin. It has been shown for a variety of Fe + systems that addition of nitric oxide (NO) gives a complex with S = 3/2 (5, 37). The spectra of such complexes can be interpreted in terms of the spin Hamiltonian (where S = 3/2) (42): He = D [' - 1/3 (S (S + 1)) + EID (S2 - S2 )] + goj3eS . H (4) The first term in brackets describes the splitting of the spin quartet in zero magnetic field; D and E are the axial and rhombic zero-field splitting parameters, respectively. By appropriate choice of coordinate system, one can keep E/D constrained to 0 c E/D < 1/3. In a zero magnetic field, solution of equation 4 yields two degenerate doublets separated in energy by A = 2D[1 + 3(E/D)2]1/2. For S = 3/2 NO complexes studied thus far (5, 57), D is positive, D is + 10 to 15 cm-1, and E/D is <0.10, implying that the ground doublet has magnetic quantum number M = + 1/2. The Zeeman (second) term in equation 4 describes the interaction Downloaded from http://jb.asm.org/ on January 16, 2018 by guest Steady-state kinetics. The initial velocity of 2,3-PCD steady-state reactions was determined as described in "Enzyme assay," with the following modifications. All assays were performed at 250C. The electrode was calibrated with buffer which had been equilibrated with known primary gas standards. The concentration of PCA in stock solutions was determined as E290 = 0.389 M-1 cm-' (65). The 02 concentration in assay mixtures was altered by equilibrating assay solutions with 02-N2 gas mixtures established by a precision dual flowmeter (Air Products). The solubility of 100% 02 was taken to be equal to 1.4 mM at 250C. The apparent Km and Vm. values were determined by directly fitting the Michaelis-Menten equation with a nonlinear regression program (7). Sedimentation velocity ultracentrifugation. Sedimentation coefficients were measured with a Beckman model E analytical ultracentrifuge equipped with a mechanical speed control, the RTIC temperature control, Schlieren optics, and an An-D rotor. Schlieren boundary positions were measured with a Nikon model 6C comparator. Protein solutions of 2.5 to 3.5 mg/ml in 50 mM MOPS (pH 6.9) were run in singlesector cells at 29,500 rpm and 8.3YC and were photographed at 16-min intervals. Values for sedimentation and diffusion coefficients extrapolated to standard conditions (s20.W and D°20 ,) were determined by standard procedures (13). The molecular weight (MW) of 2,3-PCD was determined from these values by the Svedberg equation (13). Calculation of diffusion coefficient and Perrin shape factor. The diffusion coefficient at infinite dilution of a two-component system is given by the Einstein-Sutherland equation (equation 1). For a sphere in "stick" boundary conditions, the translational frictional coefficient, f, in equation 1 is given by Stokes law (equation 2). VOL. 175, 1993 PROTOCATECHUATE 2,3-DIOXYGENASE FROM BACILLUS MACERANS TABLE 1. Purification of B. macerans 2,3-PCD Puri- Purification step Vol Activity Sp act Yield (ml) (U/mlr Protein (mg) (U/mg) (%) fication factor Cell supernatant" 350 DEAE-Sepharose Fast 70 Flowc plus pressure dialysis Phenyl-Sepharose plus 15 pressure dialysis 123 477 5,385 1,210 1,850 138 8 27.5 201 100 77 64 1 3.4 25 a 23C, air saturated buffer. b On the basis of 100 g (wet weight) of cells. c After concentration by pressure dialysis. Protein concentration. Protein concentrations of three homogeneous 2,3-PCD samples of known A220 were independently measured by quantitative amino acid analysis as previously described (29) and were used to calibrate a Bradford protein assay (8), from which subsequent protein concentrations were measured. The data from the accurately determined samples were also used to determine an extinction coefficient for the enzyme at 280 nm. RESULTS Purification and characterization of 2,3-PCD. The purification of 2,3-PCD is summarized in Table 1. The apparently homogeneous enzyme has a specific activity approximately sixfold greater than previously reported for a partial purification (16). This improvement in specific activity was due to altered growth procedures and the use of the stabilizing reagents ferrous ion and cysteine, which we have found useful in the purification of many nonheme iron-dependent enzymes (5, 23, 29). The amino acid composition of reduced S-carboxymethylated holoenzyme is shown in Table 2. On the basis of quantitative amino acid analysis, the extinction coefficient of the holoenzyme in cysteine-free, 50 mM MOPS (pH 7.0) was found to be C(280) = 1.4 mg- ml cm-. The physical properties of 2,3-PCD determined as described in Materials and Methods are summarized in Table 3. The holoenzyme has an MW of approximately 143,000 and a slightly nonglobular shape as indicated by the Perrin shape factor of 1.3. SDS-PAGE shows a single Coomassie bluestaining protein band with an Mr of 35,500 ± 800 (Fig. 2), suggesting that the enzyme is a tetramer with four identical subunits. To support this conclusion, automated Edman degradation was performed for nine cycles on the reduced, S-carboxymethylated holoenzyme. A single amino acid was obtained in each cycle, suggesting that the protein is homogeneous and that the subunits are identical. The sequence obtained is shown in Table 2. Over this limited sequence interval, no homology with the sequences of other dioxygenases is noted. Omitting the 2-mercaptoethanol from the incubation buffer prior to electrophoresis had no effect on the apparent MW of the subunit, suggesting that the quaternary structure is not stabilized by disulfide bonds. Steady-state kinetics. Double reciprocal plots of initial velocities observed when one substrate was varied at several fixed concentrations of the other substrate intersect, showing that the mechanism of 2,3-PCD is sequential (Fig. 3). Previous studies have indicated that dioxygenases have ordered mechanisms in which the organic substrate binds first (32). Because the mechanism is sequential and PCA Downloaded from http://jb.asm.org/ on January 16, 2018 by guest of the applied magnetic field with the electronic magnetic moment. In all cases studied thus far for iron-nitrosyl complexes, this interaction has been found to be isotropic, with go = 2.0 (55). For EID = 0, equation 4 predicts for the ground doublet effective g values at g, = 4.0, g = 4.0, and = 2. For small values of EID, the two g = 4fresonances split symmetrically around g = 4, while gz drops slightly below 2.0. The EID values for the 2,3-PCD nitrosyl complexes were calculated from the observed effective g values. Mossbauer spectra can be analyzed with the same spin Hamiltonian augmented by nuclear hyperfine and Zeeman terms (see reference 41). Detailed analysis of Mossbauer spectra can yield values for all of the Hamiltonian parameters. In the current application, Mossbauer spectra are used only to determine the redox and spin state of the active site iron and to provide values for the characteristic absolute value of the quadrupole splitting and isomer shift parameters. This information can be obtained directly from the spectra with no further analysis. Quantitation of iron. Total iron was measured in a multielement analysis by inductively coupled plasma emission spectrometry (ICPES) in an Applied Research Laboratories model QA 137 spectrometer at the Soil Sciences Center, University of Minnesota. All samples were prepared from enzyme-NO complexes which were evacuated to remove the NO and then acidified with 2 N HC1 to release the iron. The total Fe2+ of the enzyme-NO complexes was determined by single integration of the S = 3/2 EPR spectrum of the complex for comparison with a standard solution of 1 mM Fe2+-EDTA-NO complex, which gives a similar spectrum and was standardized by ICPES. Corrections for g value anisotropy were made according to the method of Aasa and Vanngard (1). Specific removal of adventitiously bound ferric ion and manganese. Adventitiously bound Fe3' and Mn2+ were removed by two different methods. A 1-mg/ml solution of 2,3-PCD in 50 mM MOPS (pH 6.9) (buffer) was anaerobically incubated with 0.3 mM 4,5-dihydroxy-1,3-benzene disulfonic acid (Tiron) (Sigma) for 1 h at 4°C. The solution was made 1.5 M in NaCl and loaded onto a phenyl-Sepharose column (1 by 5 cm) preequilibrated with a buffer solution of 1.5 M NaCl and 2 mM cysteine. The Tiron adventitiously bound metal complex was removed by extensively washing the column with the same solution. The enzyme was eluted by buffer plus 2 mM cysteine. All column steps were performed at 4°C with N2 gas bubbling through degassed buffers. In the second method, a 1.8-mg/ml solution of 2,3-PCD in buffer was added to a slurry of Chelex resin (Bio-Rad) preequilibrated in buffer. This was centrifuged in a tabletop microcentrifuge for 2 min, and the enzyme-supernatant solution was removed for metal analysis. Anaerobic procedures and nitric oxide addition. Samples were made anaerobic by repeated cycles of evacuation and flushing with argon which had been passed over a column of BASF Inc. copper catalyst at 150°C. Transfers of anaerobic solutions were by gas-tight syringe. NO was added by slow bubbling (saturated solution, 3.3 mM) (Matheson Inc.). An Ar flush was maintained above the sample to remove excess NO and to protect the sample from 2. Approximately 3 min of bubbling (80 bubbles from 24-gauge tubing) was used and was sufficient to saturate the sample (4, 6). Enzyme-substrate-NO complexes could be made by initially adding either substrate or NO to the enzyme. The order of addition did not change the EPR spectrum of the resulting complex. NO is very toxic and should be handled in a well-ventilated hood. 4417 TABLE 2. Amino acid Amino acid CM.Cyse composition of B. Thrf Serf Glx Pro Gly Ala VaW Met Ileg Leu Tyr Phe His Lys Arg TrpO Total residues Total MW 2,3-PCDa Yield Subunit Total enzyme Physical properties and kinetic constants (residues/mol)C (residues/mol)d Physical properties MW Sedimentation velocity .................... 143,000 ± 5% Gel filtration ............................ 149,900 ± 10% Estimated from SDS-PAGE ............. 142,000 ± 10% Subunit MW ............................ 35,500 ± 10% 7.63 x 10-13 s ± 3% S.OOW D 20,w-............................................ 4.64 x 10-7cm2/s ± 3% Perrin shape factor (f/fo)...................1.3 ± 5% Subunit structure ... a4 Fe2+ content (mol/mol) . .................. 0 ± 0.15 5.8 pI .......................... 5.0 30.5 11.0 27.5 37.0 15.0 24.0 26.0 23.0 9.0 17.0 31.0 12.5 9.0 11.0 10.0 16.0 7.0 20 122 44 110 148 60 96 104 92 36 68 124 50 36 44 40 64 28 321.5 35,853.4 1,286 143,583' a N-terminal sequence: Ser-Leu-Glu-Met-Ala-Leu-Leu-Ala-Ala. b Values normalized to the yield of carboxymethylcysteine (absolute yield, 2.00 nmol). c On the basis of an estimated MW of 35,500. Values rounded to the nearest half-integer. d Assumes an a4 subunit structure. e Determined as carboxymethylcysteine. f Extrapolated to zero appears to be able to bind to the iron in the absence of 02 but vice versa (see below), an ordered bi-uni mechanism is likely to pertain to 2,3-PCD. Unfortunately, the appropriate product inhibition studies to confirm this conclusion cannot be completed in the case of the ring cleavage dioxygenases because the products spontaneously isomerize after they leave the active site. The kinetic constants and some rate constants for the 2,3-PCD-catalyzed reaction can be determined by analyzing the kinetic plots and replots shown in Fig. 3. Values of Km for 02 and PCA at saturating concentrations of the nonvaried substrate are given in Table 3. A limiting Vm~, cited in terms of a turnover number per iron at saturating concentrations of both substrates is also shown. For the case in which PCA was varied at several fixed 02 concentrations, the double reciprocal plots intersect very close to the horizontal axis, indicating that KmpCA is =KipcA. The value of KPCA obtained from the replots is given in Table 3. Because Ki = Kd for the first substrate to bind to the enzyme in any ordered mechanism (63), the Kd value for PCA (24 pM) can be determined. Solution of the net velocity equation for a bi-uni mechanism shows that KmpCA = KipCA if PCA and the product dissociate from the enzyme at approximately equal rates (see the following equations). ki Characteristics ............................ ....................... Steady-state kinetic constantsa'b 142 FM 24 PM KmPCAC .......................... 20 LM KiPCAC .. Turnover numbe( .......................... 210 s-1 active site-1 k1 (PCA association rate constant) .......8.75 x 106 M-1 S-1 k-, (PCA dissociation rate constant) .... 175 s-' k3 (product dissociation rate constant) ....210 s-1 KmO2- ........................... . ........................... a Kinetic and rate constants are determined assuming an ordered bi-uni reaction sequence (see text). b Approximate error in values is +10%. Rate constants k, and k-1 are determined as ratios of kinetic constants and therefore have a larger potential error. c Value determined for the condition in which both substrates are extrapolated to infinite concentration. time. g Extrapolated from a timed hydrolysis. h Determined from extinction coefficients at 280 and 288 nm. i Includes the weight of four iron atoms. not TABLE 3. Summary of characterization of 2,3-PCD (nmol)b 1.00 6.10 2.15 5.50 7.35 3.05 4.85 5.20 4.60 1.85 3.35 6.25 2.50 1.75 2.25 2.05 3.20 1.40 Asx macerans k2 E + PCA z E PCA + 02 t [E * PCA 02 k-E kP2 E + Product KmpCA k3 KiPCA ki =- k = ki Vmax Vmax. KiPCA Vmax KmPCA .Etotal KmPCA Etotal Etotal MW B 92,50066,000 45,00031,000 21,500 14,400FIG. 2. SDS-PAGE of 2,3-PCD. The 2,3-PCD was denatured in SDS and 2-mercaptoethanol, subjected to electrophoresis, and stained as described in Materials and Methods. (A) MW standards (Bio-Rad): phosphorylase B (92,500), bovine serum albumin (66,000), ovalbumin (45,000), carbonic anhydrase (31,000), soybean trypsin inhibitor (21,500), lysozyme (14,400). (B) Approximately 5 pg of 2,3-PCD. (C) Incubation buffer alone (same volume as used in lanes A and B). Minor bands appear in each lane because of an impurity in the incubation buffer. Downloaded from http://jb.asm.org/ on January 16, 2018 by guest EProduct] J. BACTERIOL. WOLGEL ET AL. 4418 VOL. 175, 1993 PROTOCATECHUATE 2,3-DIOXYGENASE FROM BACILLUS MACERANS 4419 TABLE 4. Comparison of substrate turnover by extradiol dioxygenases 2,3-PCD Substrate 0.03 3-Substituted catechols 3-Cl-catechol 3-CH3-catechol 0.01~~~~~~~~~~~~~~-50 40 -30 -20 0 2030 40 50 02 02 The product dissociation rate constant, k3, is defined as the turnover number (VmaiEtotal = 210 sol). The value for the second-order rate constant for PCA association can also be estimated [k1 = Vmjj(KmpcA Etoj) = 8.75 x 106 M s-]. Together with the Kd value for PCA, the calculated k1 value allows the dissociation rate constant for PCA to be estimated as k1 = 175 sSubstrate specificity. Sixteen compounds other than PCA were observed to be turned over (Table 4). Compounds that serve as substrates include 3-substituted and 4-substituted catechols, 2-substituted and 5-substituted PCAs, esters of PCA, 1-substituted PCAs with extended side chains, and a novel substrate analog, 4-NH2-3-OH benzoate (4-AHB). The absorbance maxima of most of the products were observed to be 350 to 382 nm, suggesting that the substrates are cleaved in a proximal extradiol manner between C-2 and C-3. Distal extradiol cleavage would be expected to yield products with A 4w (18, 19), whereas the products of intradiol cleavage would be expected to absorb in the UV range (25, 65). The cleavage of homo-PCA (HPCA) and 3,4-dihydroxyphenylpropionic acid (DHPP) each gave reaction products with two absorption maxima, potentially indicating two different ring cleavage products for these compounds. However, the optical spectra of the reaction products observed here are identical to published spectra of the reaction products of extradiol cleavage of HPCA and DHPP catalyzed by DHPP 2,3-dioxygenase from Pseudomonas ovalis (2, 54). This comparison strongly suggests that these two compounds are cleaved in an extradiol manner between the 2 and 3 carbons by 2,3-PCD. Thus, despite a very large substrate range, no evidence for cleavage other than proximal extradiol has been observed for 2,3-PCD. The common feature of all alternate substrates, with the 4,5-PCD 2,3-CTD turn- over' turnoverd 1 0.9 382 380 NIX Nof No Yes 100 5 2 1.5 0.5 <0.5 350 380 382 376 382 350 Yes Yes No No No No Yes Yes Yes Yes Yes Yes 10 3 372 372 Nof Nof Nof Nof 3 1 380,325 380,334 No No No 376 Nof Nof 30 265 Nof Yesf 3,4-(OH)2-5-CH3-benzoate 3,4,5-(OH)3-benzoate 10 1.2 375 368 Yes Yes Yesf 4-Substituted 3-OH-benzoic acid (3-OH-4-NH2-benzoate) <0.5 272 No NJ 4-Substituted catechols Simple substituents PCA 4-SO3-catechol 4-Cl-catechol Catechol 4-CH3-catechol Protocatechualdehyde Esters PCA methylester PCA ethylester Homoprotocatechuate analogs 3,4-(OH)2-phenylacetate 3,4-(OH)2-phenylpropionate 3,4-(OH)2-mandelic acid 2-Substituted PCA [2,3,4(OH)3-benzoic acid] 0.8 Nof 5-Substituted PCAs ND a The average of three trials by polarographic assay (see Materials and Methods). Values have an average error of approximately ±10%. Substrates were 1.5 mM. b Represents a single optical assay. Enzyme (0.1 pM), substrates (100 PM), 50 mM MOPS, pH 7.0. c From reference 3. d From reference 48. ¢ ND, not determined. f This work. exception of 4-AHB, is the presence of vicinal hydroxyl groups on the benzenoid ring. The unique aspects of the 4-AHB reaction will be examined in detail in a subsequent publication. EPR spectra of 2,3-PCD. The EPR spectra of 2,3-PCD are shown in Fig. 4. The enzyme as isolated (Fig. 4A) displayed signals from several distinct species: resonances near g = 4.3 reflect a high-spin (S = 5/2) Fe3+ in a rhombically distorted (EID = 1/3) electronic environment, resonances between g = 5.6 and 8.2 are characteristic of multiple, high-spin Fe31 species in more axial electronic environments, and the multiple resonances nearg = 2.0 result from Mn2". Quantification of the latter spectrum showed that the concentration of Mn2+ was only a few percent of the total subunit concentration (see Table 6 below). Direct quantitation of the iron by spectral integration was difficult because the spectra are very broad. Nevertheless, approximate quantitation indicated that the amount of EPR-active Fe3+ was much less than the amount of total iron in the sample, suggesting that much of the iron may have been in the Downloaded from http://jb.asm.org/ on January 16, 2018 by guest 1 / [PCA] (mM l) FIG. 3. Steady-state kinetics of 2,3-PCD turnover. Initial velocity of PCA turnover was measured at several fixed concentrations: 0, 290 PM; *, 150 PM; +, 90 PM; A, 35 p.M. Inset plot is a replot of the slope and intercept values from the main plot versus the reciprocal of the concentration. Solid lines are nonlinear regression fits to the original data converted to double reciprocal form. The observed initial velocities (measured in units of micromolar concentration per second) were divided by the enzyme active site concentration to convert the units of the y axis into those of the reciprocal turnover number (1/s-5). Conditions: enzyme, 0.016 P.M (0.064 pM sites); temperature, 250C; pH 7.0; MOPS buffer, 50 mM. Other conditions are given in Materials and Methods. Relative Product rate Xmp (nm)b (%)a J. BACTERIOL. WOLGEL ET AL. 4420 V 0 c'W1 -T ;127 20 4)~ ~ ~~41 ,0 + > X1 1450 Magnetic Field (Gauss) FIG. 4. EPR spectra of 2,3-PCD. X-band EPR spectra of 2,3PCD were recorded at 6.5 K. (A) Enzyme as isolated (70 A~M); (B) enzyme from A plus 0.5 mM Na ascorbate added anaerobically; (C) enzyme prepared as in B, treated with 8 mM Tiron for 5 min, and then chromatographed on phenyl-Sepharose to remove the Tironmetal chelates as described in Materials and Methods; (D) enzyme from B plus NO gas added anaerobically by bubbling with NO gas for 3 mmn Instrumental conditions: field center, 2,100 G; field sweep, 4,000 G; scan rate, 1,000 G/min; modulation amplitude, 10 G; modulation frequency, 100 kHz; microwave frequency, 9.22 GHz; microwave power, 0.2 mW. Relative scale factors are shown on the figure. (usually EPR-silent) Fe2+ state. Ascorbate diminished some of the Fe3+ signals, whereas the Mn2+ signal remained virtually unchanged (Fig. 4B). Concomitant with ascorbate reduction, the enzyme activity increased 5 to 10% depending upon the amount of Fe3+ in the preparation. We speculate that the Fe2+ of the enzyme is partially oxidized during purification and handling and that these factors determine the extent of EPR-active iron and ascorbate reactivation. EPR spectra of NO complexes of 2,3-PCD. Previous studies have shown that the EPR-silent Fe2+ of many enzymes can be made EPR active by forming the nitrosyl complex (5, 6, 29, 37, 61). Accordingly, upon anaerobic addition of NO to 2,3-PCD, a new and much more intense EPR spectrum with resonances at g = 4.11 and 3.95 was observed (Fig. 4D). Resonances with such features are characteristic of a species with an electronic ground state spin of S = 3/2. Because this type of signal has only been observed for the NO complex of bound in proteins and organic chelate complexes, this Fe2M constitutes an unambiguous demonstration that 2,3-PCD contains Fe2s. The spin concentration of this highly resolved spectrum could be accurately determined; it represents >95% of the iron in the ascorbate-reduced enzyme that had been treated with a chelator to remove adventitiously bound iron (see below). The low field portion of the spectrum is shown in a narrower magnetic field sweep range in Fig. 5A. Only one species is apparent. The positions of the 17550 1550 15 70 15 1650 1750 1850 Magnetic Field (Gauss) FIG. 5. EPR spectra of 2,3-PCD-nitric oxide complexes. Samples of 2,3-PCD were made anaerobic and exposed to NO gas (3.3 mM dissolved concentration) for 3 min prior to freezing. (A) 2,3-PCD (52 ,uM) plus 1 mM Na ascorbate. (B) Enzyme (26 ,uM) plus 1 mM Na ascorbate plus 2 mM PCA. (C) Enzyme (23 pM) plus 2 mM Na ascorbate plus 2 mM 2,3,4-(OH)3 benzoic acid. Instrumental conditions: field center, 1,660 G; field sweep, 400 G; scan rate, 100 G/min; modulation amplitude, 1 G; modulation frequency, 100 kHz; microwave frequency, 9.22 GHz; microwave power, 0.2 mW; temperature, 10.1 K. Relative scale factors are given on the figure after correction for enwyme concentration. resonances show that the iron was in a nearly axial electronic environment with an EID value of 0.014. Upon the addition of substrate, the splitting of the resonances around g = 4 increased, indicating that the electronic environment of the Fe2" had changed, becoming more rhombic, with an EID of 0.034 (Fig. 5B). A single species was observed, indicating that the complex was homogeneous. Quantitative analysis showed that nearly all of the enzyme had bound substrate. The weak resonances near g = 4.3 in the spectra from the NO-treated samples are due to the small amount of Fe3+, as was also seen in the spectra shown in Fig. 4A and B. Addition of alternate substrates such as 2,3,4-trihydroxybenzoate to the enzyme-NO complex elicited EPR spectra very similar to those of the enzyme-PCA-NO complex as shown in Fig. SC, suggesting that most substrates bind in a similar orientation in the active site. Substrates differed in their affinities for the enzyme-NO complex. In the case shown in Fig. 5C, the amount of 2,3,4-trihydroxybenzoate added was insufficient to saturate the enzyme, so a small amount of the g = 4.11, 3.95 species characteristic of substrate-free enzyme-NO complex was observed. The order of substrate and NO addition was not found to be critical. NO could be removed from the enzyme by evacuating the sample or exposing it to 02. After NO was removed, full catalytic activity was observed, and NO could again be bound to give the characteristic spectral features. In a separate experiment, PCA was anaerobically added to Downloaded from http://jb.asm.org/ on January 16, 2018 by guest 145 3.7 OH PROTOCATECHUATE 2,3-DIOXYGENASE FROM BACILLUS MACERANS VOL. 175, 1993 4421 TABLE 5. Mossbauer parameters of 57Fe-enriched 2,3-PCD' Enzyme Quadrupole splitting (AE0, mm/s) Isomer shift (8Fe, mm/s) Line width (I, mm/s) 2,3-PCDb 2,3-PCD + 10 mM PCA Typical Fe2" site (S = 2)d Typical Fe3" site (S = 5/2)d 2.57c 2.73 2-3.3 0.3-1.0 1.29 1.19 0.9-1.3 0.3-0.6 0.36 0.48 0.0k o.0k 5 A Spectra measured at 4.2 K in zero applied field. b Total iron concentration = 0.4 mM. Samples were prepared anaerobically. c The AEQ value was temperature dependent. At 150 K, AEQ = 2.30 mm/s. At 193 K, AEQ = 2.20 mm/s. d From reference 20. 0 CZoE~) o.1F~B k 0.3 0.4 p -8 -6 -4 -2 0 2 VELOCITY (mm/s) 4 6 8 FIG. 6. Mossbauer spectra of 57Fe-enriched 2,3-PCD at 4.2 K. (A) Uncomplexed enzyme as isolated (-500 ,uM). (B) Sample from A after anaerobic addition of 10 mM PCA. The solid lines are least-square fits of the data to one quadrupole doublet. the enzyme and the EPR spectrum was recorded (data not shown). The only change observed in the spectrum was a small decrease in the signals atg = 8.14, 5.6, and 4.3, similar to results seen in Fig. 4A versus B. No perturbation of the Mn2+ signal nearg = 2 was observed, making it unlikely that PCA binds to the Mn2+ (see reference 60). Mdssbauer spectroscopy of enzyme and the enzyme-substrate complex. While the NO complex provides a rapid and sensitive method to detect and quantitate Fe2+ in 2,3-PCD and other enzymes, a direct technique that does not involve binding of exogenous ligands to the metal is required to validate the method. Mossbauer spectroscopy is such a technique, but it requires substitution of "7Fe for 56Fe in the active site. This was achieved by growth of the organism on 57Fe-enriched media. The zero-field Mossbauer spectrum of 2,3-PCD shown in Fig. 6A consists of a single doublet with quadrupole splitting AE0 = 2.57 + 0.03 mm/s and isomer shift 8 = 1.29 + 0.02 mm/s at 4.2 K. These parameters unambiguously show that the iron site of the enzyme is in the high-spin ferrous state (Table 5). The value of AEQ is temperature dependent, suggesting that low-lying orbital states become thermally accessible at the temperature of the measurements; we found AEQ = 2.30 mm/s at 150 K and AEQ = substrate. The 4.2-K Mossbauer spectrum of the anaerobically prepared enzyme-PCA complex shown in Fig. 6B consists of one quadrupole doublet with AEQ = 2.73 mm/s and 8 = 1.19 mm/s. These parameters show that the iron site remains high-spin ferrous upon binding of substrate. The absorption lines of the enzyme-PCA complex are broader (-0.48 mm/s) than those observed for the uncomplexed enzyme (0.36 mm/s), suggesting some heterogeneity in the ligand environment of the Fe2+ site, i.e., AE0 is distributed. Iron and manganese content of 2,3-PCD and correlation with subunit MW and enzymatic activity. While the quantity of iron present clearly suggests that it is required for activity, the possible role of Mn~ +in catalysis must also be considered. The metal-chelating reagents Chelex and Tiron were used to remove any labile metals bound to the enzyme. Both reagents bind Mn2 , while Tiron has a high affinity for Fe3+ and Chelex preferentially binds Fe2+. The EPR spectrum of the enzyme after treatment with Tiron (Fig. 4C) shows that much of the EPR-active Fe3+ and Mn2+ was removed by this chelator. Samples of 2,3-PCD treated with Tiron and Chelex were complexed with NO to produce the S = 3/2 EPR signal. The height of the g = 4.11 peak was measured, and the ratio of the enzymatic activity to the peak height for each sample was determined (Table 6). Additionally, the ratios of total iron and manganese (as measured by ICPES) to subunit concentration and enzymatic activity were determined for each sample. If a given metal is required for activity, the ratio of the activity to the metal concentration should remain constant. This was not observed for Mn2" or the total iron. Only the amount of Fe2" that gives rise to the S = 3/2 signal maintained a constant relationship to activity. This iron could not be removed readily from the enzyme by these chelators. This finding, in addition to the fact that only Fe2+ is present in a stoichiometric amount relative to the subunit TABLE 6. Quantitation of metals bound to 2,3-PCD Sample 2.20 mm/s at 193 K. (These values suggest that the lowest orbital excited state is ca. 300 cm-1 above the ground state). Studies in an applied magnetic field of 6.0 T elicited magnetic patterns indicative of a site with a large and positive zero-field splitting parameter, D > 10 cm-1. The high-field spectra were, however, ill resolved, preventing a detailed analysis. Identical zero-field spectra were observed for aerobic and anaerobic samples, suggesting that the enzyme does not form a complex with 02 in the absence of Untreated Chelex treated Tiron treated TtlS=32Total iron/ Total iron/ subunit total activity signa/3/2 total Total Tta subunit manganese/ 0.1 0.014 0.07 0.04 0.01 0.006 toa activity (mo/mol)a,(MM/U)a Molaatvt,(o/mo0 (mol! (nM/Ur acttyb 1.4 1.6 1.02 0.21 0.24 0.14 1.06 1.06 1 Total metal concentrations determined by ICPES. Intensity of the S = 3/2 EPR signal of the NO complex of the enzyme contained in the sample divided by the total units of activity. Values are normalized to the Tiron-treated sample. It is argued in the text that the S = 3/2 signal is representative of the Fe2" in the active site of the enzyme. a b Downloaded from http://jb.asm.org/ on January 16, 2018 by guest 0.2 4422 WOLGEL ET AL. J. BACTERIOL. TABLE 7. Effect of metal chelators and oxidizing and reducing agents on enzymatic activity % Additiona Activity Activity None o-Phenanthroline (1 mM) a,a-Dipyridyl (1 mM) Tiron 8-OH quinoline (8 mM) EDTA (8 mM) Nitrilotriacetic acid (8 mM) H202 (1 mM)c K3Fe(CN)6 (1 mM) NaN3 (1 mM) 100 0 20 90 45 0 0 0 0 0 % Activity with ascorbate ~added` 110 0 20 95 45 0 0 101 98 92 concentration, clearly shows that it is Fe2' rather than Mn2+ that is responsible for activity. Effect of metal chelators and oxidizing and reducing agents on enzymatic activity. As shown in Table 7, strong Fe2+ chelators, such as o-phenanthroline and oc,a-dipyridyl, markedly inactivated 2,3-PCD at 1 mM concentrations. On the other hand, the Fe3" chelator Tiron showed little inactivation even at 8 mM. Other less-specific metal chelators such as EDTA and nitrilotriacetic acid inactivated the enzyme only at relatively high concentrations. A commonly used Mn21 chelator, 8-hydroxyquinoline, showed moderate inactivation, but this reagent will bind most divalent cations, including Fe2 . Substrate was observed to protect against inactivation only in the cases of N3-, o-phenanthroline, and a,a-dipyridyl (data not shown). The basis for N3- inactivation is not known, although it may act as an iron ligand, thereby inhibiting substrate binding as we have observed for 2,3-CTD (40). The oxidizing agents H202 and K3Fe(CN)6 completely inactivated 2,3-PCD. Full activity was restored if the samples were promptly reduced with sodium ascorbate. The EPR spectrum of the nitrosyl complex of the H202-treated sample showed nearly complete loss of the S = 3/2 signal. This signal was restored after the sample was exposed to ascorbate. All of these results are consistent with the proposed Fe2' requirement for enzyme activity and suggest that the active site is accessible to chelators and oxidation and/or reduction agents. DISCUSSION Through the use of altered growth conditions and the identification of stabilizing reagents, we have purified 2,3PCD to homogeneity for the first time. The enzyme has been characterized in terms of its physical, catalytic, and spectroscopic properties. The availability of the characterized enzyme permits comparisons with the other dioxygenases that catalyze cleavage of the ring of aromatic substrates. In particular, comparisons with 4,5-PCD and 3,4-PCD, which utilize the same substrate as 2,3-PCD, may allow recognition of features of particular relevance to the active site structure or the molecular mechanism. Also, because 2,3-PCD is one Downloaded from http://jb.asm.org/ on January 16, 2018 by guest a Enzyme (4 FLM) was incubated anaerobically in 50 mM MOPS (pH 7.0) for 5 min at 40C with each addition. Values shown are the average of three trials and have an error of approximately +10%. b Ascorbate was added at 1 mM and incubated for 5 min. c Inactivation of 30 pIM enzyme by treatment with 5 mM H202 resulted in comparable loss of the S = 3/2 EPR signal from the nitrosyl complex. Both the activity and the S = 3/2 signal were restored after treatment of the sample with 5 mM ascorbate. of the few dioxygenases purified from Bacillus species, it may be possible to determine whether there are speciesspecific characteristics that distinguish these dioxygenases from those of the better-studied genera such as Pseudomonas. Some of the characteristics of ring cleavage PCA and catechol dioxygenases are brought together in Table 8 to facilitate such comparisons in the following sections. Structure. The holoenzyme MW, the subunit MW, and the quaternary structure of 2,3-PCD from B. macerans are quite similar to those of most other extradiol catecholic dioxygenases. All of the proximal extradiol-cleaving dioxygenases have a single type of subunit with MWs of approximately 35,000. This group includes 2,3-CTD, which when presented with PCA catalyzes its cleavage to the same product as 2,3-PCD (6, 48). In contrast, the distal extradiol-cleaving dioxygenase 4,5-PCD is structurally quite distinct from 2,3-PCD, despite the fact that it utilizes the same substrate. The 4,5-PCD has two nonidentical subunits ("212), the larger of which is approximately the same size as the single subunit of the proximal extradiol-cleaving dioxygenases (5). This more complex structure for 4,5-PCD was controversial at the time it was first reported. However, the recent report of two structural genes for 4,5-PCD from Pseudomonas paucimobilis coding for subunits of the expected sizes shows that the subunit structure is, in fact, different from those of the proximal extradiol-cleaving enzymes (45). In marked contrast to the more typical extradiol dioxygenase, each of the many intradiol 3,4-PCD dioxygenases that have been characterized has two types of subunits. The fundamental unit of structure is a1-Fe +, but many different numbers of these protomer units are represented in the family of 3,4-PCD enzymes (for a review, see reference 37). Interestingly, the subunit structure of 1,2-CTD has recently been shown to be variable (43). The active enzyme is always composed of two subunits per iron, but any combination of a and I subunits is permitted. Ferrous iron site. The use in this study of EPR and Mossbauer spectroscopic techniques to show unequivocally that the active site metal of 2,3-PCD is Fe2+ validates numerous previous studies of other extradiol dioxygenases. In many of these studies (for example, see references 2, 47, and 49), the assignment of the metal as Fe2+ was based on either chelation studies, the observation of inactivation by oxidants, or the requirement for Fe2+ in reconstitution procedures. Unfortunately, each of these approaches is subject to ambiguities; for example, reconstitution of apo3,4-PCD requires Fe2' despite the fact that the iron is Fe3+ in the active enzyme (27). The spectroscopic approach also reveals differences in the active site structure that are more subtle than simply the redox state of the iron. For example, 2,3-CTD appears to be like all of the other proximal extradiol dioxygenases on the basis of structural criteria, but the Mossbauer AEQ value is distinct (67). Moreover, the Mossbauer spectrum of 2,3-CTD does not change when the substrate complex forms, unlike all other Fe2" dioxygenases thus far investigated. Nitrosyl complexes. The use of NO has proven to be a rapid and convenient method to detect, quantitate, and distinguish Fe2+ centers. It also offers information about the active site structure. For example, 4,5-PCD and 2,3-CTD each have more than one EPR-active species when complexed with NO (5, 6). In the case of 4,5-PCD, the enzyme-NO complex is homogeneous, but two species are observed when PCA binds, suggesting either two binding orientations for the substrate or two forms of the enzyme. Just the opposite is observed for 2,3-CTD in which the enzyme-NO complex is PROTOCATECHUATE 2,3-DIOXYGENASE FROM BACILLUS MACERANS VOL. 175, 1993 4423 TABLE 8. Comparison of the properties of intradiol and extradiol PCA and catechol dioxygenases Spectroscopic characteristics Structural characteristics Subunit Enzyme and source MW Extradiol dioxygenases 2,3-PCD B. macerans (ATCC 35889) plus PCA 4,5-PCD P. testosteroni (ATCC 49249) plus PCA 2,3-CTD Pseudomonas anvilla (putida) plus catechol Intradiol dioxygenases 3,4-PCD B. fuscum (ATCC 15993) plus PCA P. putida (aeruginosa) (ATCC 23975) plus PCA 1,2-CTD; P. arvilla (putida) (ATCC 23974) no. (sv) EPRE/D Mossbauer of NO Subunit structure 143,000 Fe2" (a-Fe2")4 Fe2+ a2122-Fe2+ 140,000 Fe2- (a-Fe2+)4 AEq 8 210 0.014 0.034 2.59 1.29 This work 2.73 1.21 a = 17,000, 13 = 33,000 360 0.015 0.064 0.037 2.23 1.28 5, 6 2.33 1.27 2.80 1.22 35,000 190 0.029, 0.021 0.005 3.28 1.31 5, 6, 35, 47, 67 0.40 0.44 57, 70 f 0.40 0.44 50, 51, 59 Fe2+ 315,000 Fe3+ (ap3-Fe3+)5C Fe3+ 587,000 Fe3` (a13-Fe3)12 complexa 35,500 Fe2+ 103,000 Fe2 Reference (mm/si a = 22,500, 13 = 40,000 416 0.055d 0.175w a = 22,300, P = 26,600 70 0.055d 3.28 1.31 f 0.175 FeF3e 63,000 Fe3+ a-Fe3+, aa-Fe3+, 1313-Fe3+ a = 30,000, 1 = 32,000 31 - 0.155c'h 0.45C 33, 34, 43 0.50 3 a EPR spectra recorded near 4 K of the anaerobic nitrosyl complex of the enzyme. b Mossbauer spectra recorded at 4.2 K in zero applied magnetic field. c Preliminary crystallographic studies indicate a hexameric structure (21a). d NO binds only to the reduced form of the enzyme (57). e Value for the enzyme-substrate-NO complex. f At least five species are present, presumably because of multiple substrate binding orientations. g Many species were detected for the nitrosyl complex of the enzyme. h Reference 55a. heterogeneous, but addition of catechol gives a species with a single sharp spectrum. The 2,3-PCD nitrosyl complexes with and without PCA both appear as single species, suggesting that the active site is homogeneous and that the substrate has a single binding orientation when a small 02-like molecule is bound to the iron. This unusual characteristic will facilitate more detailed spectroscopic studies designed to characterize the intermediates in the reaction cycle now in progress. Turnover characteristics. The turnover number of 2,3-PCD is comparable to those of 4,5-PCD and 2,3-CTD (5, 47). the enzyme is not Because of its relatively high Km for saturated with this substrate under typical assay conditions. It is generally thought that the Fe + enzymes have 2- to 10-fold higher turnover numbers than the Fe3" enzymes, but this is not the case as shown by the high turnover number of 3,4-PCD from Brevibacterium fuscum (70) (Table 8). For both 3,4-PCD and 2,3-PCD, the inherently high turnover number at saturating 02is not observed under normal assay conditions (or for that matter, in vivo). These results suggest that the Fe2+ and Fe3+ dioxygenases can have comparable catalytic capacities under favorable conditions and that differences in the observed rate derive primarily from their relative affinities for substrates or products. Substrate specificity. It is shown here that 2,3-PCD has the largest number of alternate substrates of any of the wellstudied intradiol and extradiol catecholic dioxygenases: 3,4PCD (25, 70), 4,5-PCD (5, 53), 1,2-CTD (34), and 2,3-CTD (48). The substrate range of 2,3-PCD is most closely approx02 imated by 2,3-CTD and chlorocatechol dioxygenase (9), but these enzymes do not readily accommodate bulky side chains in the position equivalent to C-1 of PCA. The 2,3C1D also does not accept halogens in the position immediately adjacent to the vicinal hydroxyl groups. Thus, the active site pocket of 2,3-PCD seems to be sufficiently open to bind substrate analogs with significant alterations in structure from PCA. Moreover, the ring cleavage chemistry does not appear to be as sensitive to inductive effects of the substituents as proposed for other dioxygenase enzymes (58). It seems likely that fundamentally the same molecular mechanism of ring cleavage applies to a wide variety of PCA analogs because the same site of ring cleavage is observed in every case. From previous studies, we have suggested that the site of ring cleavage depends strongly on the orientation of the substrate relative to the iron and the activated oxygen in the active site (4, 37). We have proposed that both the substrate and the oxygen bind to specific sites in the iron coordination field. Thus, 4,5-PCD and 2,3-PCD may use the same fundamental mechanism, while controlling the site of ring cleavage by altering the structure of the reactive complex. If proximity is the determining factor in the position of ring cleavage for the extradiol dioxygenases, then the current results suggest that the binding orientation of the substrate relative to the reactive oxygen species in the active site is primarily controlled by the enzyme structure rather than the functional groups of the substrates themselves. This specificity might be achieved by establishing chelation of the vicinal hydroxyl groups of the substrate to the iron as the Downloaded from http://jb.asm.org/ on January 16, 2018 by guest (ATCC 23973) Metal 4424 WOLGEL ET AL. primary binding determinant. Such chelate structures have been observed for other extradiol dioxygenase-substrate The observed turnover of catechol and the 3-substituted catechols indicates that the presence of a C-1 carboxylate group is not an absolute requirement for cleavage of catecholic compounds. Furthermore, a wide range of substitutions for the C-1 carboxylate is allowed, as evidenced by the turnover of the 4-substituted catechols and esters of PCA. However, the rate for the fastest substrate analog among any of these groups is only 1/10 that of PCA, indicative of the strong preference exhibited by the enzyme for the structure of its natural substrate. The role of Mn2". The results presented here show that Mn is not required for catalysis. This conclusion is supported by the observation that 2,3-PCD is inactivated by H202 and K3Fe(CN)6. It has been observed that these oxidants selectively inhibit Fe2+-containing enzymes while showing no effect against Mn2+-containing enzymes (38, 52, 60, 68). Moreover, the EPR spectrum due to Mn 2+ is not altered by anaerobic addition of substrate to enzyme, while large changes in the Fe2+ site are revealed by the changes that occur in the EPR spectra of the enzyme-NO complexes on addition of substrates. In a previous EPR study of an Mn 2-containing extradiol dioxygenase from Bacillus species, anaerobic addition of substrate was clearly shown to cause a perturbation of the Mn2+ spectrum (60), suggesting that a perturbation would have been observable in the present case had it occurred. Taken together, these results suggest that the Mn2+ present in 2,3-PCD is likely to be adventitiously bound to the surface of the enzyme. Studies of homo-2,3-PCD have revealed that the enzyme from Pseudomonas species utilized Fe2+ (2), while that from Bacillus species (60) utilized Mn2+, raising the possibility that Mn2+ is the metal of choice for Bacillus dioxygenases. The present results also show that it is incorrect to generalize that dioxygenases from Bacillus species specifically contain either iron or manganese, because enzymes containing both metals have now been characterized. Order of substrate binding. The kinetic data presented here are consistent with an ordered bi-uni mechanism in which PCA binding must precede 02 binding. This order is supported by the Mossbauer spectra, which show PCA binding to be independent of 02 but provide no evidence for an enzyme-02 complex in the absence of PCA. On the other hand, we have also observed that the 02 analog NO can bind either before or after PCA. Indeed, PCA and NO can bind simultaneously, as shown by the EPR spectra of the complex. In the case of other Fe2" dioxygenase enzymes, we have shown that the affinity for NO increases dramatically when the organic substrate is added (6, 30). Preliminary studies suggest that this will also be true for 2,3-PCD. Thus, the binding of organic substrates and NO is positively coupled for these enzymes. We have speculated that this coupling may result from a decrease in the redox potential of the iron due to the binding of a negatively charged catecholate or phenolate ligand to the iron (4, 30, 37). This would allow the iron to form a stronger bond with NO. The same phenomenon may apply to 02 binding, except that in the absence of the organic substrate, the 02 affinity may be too low to form a detectable amount of complex at physiological concentrations of 02. The putative coupling of 02 and organic substrate binding would assure that a substrate to oxidize is always present when 02 binds and becomes activated. The shift of electron density from the aromatic substrate to the 02 via the iron may also be a fundamental process in the oxygen activation mechanism (4, 30, 37). Conclusion. The purification of 2,3-PCD completes the isolation of all known types of enzymes of aerobic organisms that cleave PCA. The resulting family of protocatechuate dioxygenases represents all types of cleavage sites for catecholic dioxygenases as well as both of the mechanistically distinct subclasses of these enzymes. The studies regorted here strongly support the association of active site Fe + and Fe2+ with intradiol and extradiol ring cleavage, respectively. In contrast, the distinction between proximal and distal extradiol ring cleavage is not related to the redox state of the iron. Moreover, the spectroscopic studies reported here and in previous studies failed to reveal a substantial difference in the iron coordination sphere or the ability of the iron to interact with substrates and small molecule ligands. It therefore seems likely that the mechanisms of proximal and distal extradiol dioxygenase are fundamentally the same and that the site of ring cleavage is determined by the influence of the protein structure on the proximity of an activated oxygen species to the site of attack on the aromatic ring of catecholic substrates. ACKNOWLEDGMENTS This research was supported by National Institutes of Health grants GM24689 to J.D.L., GM22701 to E.M., and ES-A1-02184 to R.L.C. We thank Marcia A. Miller for measuring some of the data contained in Table 4. We also thank Gerald Bratt for assistance in use of the ultracentrifuge and Eric Eccleston for assistance in performing the amino acid analysis and N-terminal amino acid sequencing. REFERENCES 1. Aasa, R., and T. VanngArd. 1975. EPR signal intensity and powder shapes: a reexamination. J. Magn. Reson. 19:308-315. 2. Adachi, K., Y. Takeda, S. Senoh, and H. Kita. 1964. Metabolism of p-hydroxyphenylacetic acid in Pseudomonas ovalis. Biochim. Biophys. Acta 93:483-493. 3. Arciero, D. M. 1985. Ph.D. thesis. University of Minnesota, Minneapolis. 4. Arciero, D. M., and J. D. Lipscomb. 1986. Binding a 170-labeled substrate and inhibitors to protocatechuate 4,5 dioxygenase nitrosyl complex: evidence for direct substrate binding to the active site Fe2+ of extradiol dioxygenases. J. Biol. Chem. 260:2170-2178. 5. Arciero, D. M., J. D. Lipscomb, B. H. Huynh, T. A. Kent, and E. Munck. 1983. EPR and Mossbauer studies of protocatechuate 4,5 dioxygenase: characterization of a new Fe2' environ- Downloaded from http://jb.asm.org/ on January 16, 2018 by guest complexes (4, 30). The unusual dioxygenase substrate 4-AHB is turned over by 2,3-PCD, although at a very slow rate. This is the first report of cleavage of a nonorthodiol benzenoid ring by any extradiol dioxygenase that normally uses a catecholic substrate. Thus, these results indicate that the orthodiol arrangement on the benzenoid ring is not a strict requirement for turnover of substrates for 2,3-PCD and, by extension, the underlying Fe2' dioxygenase mechanism. It is interesting to note that 1,2-GTD can catalyze opening of the aromatic ring of 5-aminosalicylate as the only alternative substrate lacking p-hydroxyl groups (30). This enzyme also utilizes an active site Fe2' and appears mechanistically similar to the extradiol catecholic dioxygenases. The most prominent common feature of amino and hydroxyl groups is their capacity to tautomerize, and this may be the critical role played by these groups in the Fe2' dioxygenase mechanism. The PCA analogs 3-NH2-4-OH benzoate and 3,4-diaminobenzoate do not serve as substrates for 2,3-PCD, suggesting that only the hydroxyl group furthest from the site of ring cleavage can be replaced without loss of activity. J. BACTERIOL. VOL. 175, 1993 PROTOCATECHUATE 2,3-DIOXYGENASE FROM BACILLUS MACERANS tion of aromatic hydrocarbons, p. 181-252. In D. T. Gibson (ed.), Microbiology series, vol. 13. Microbial degradation of organic molecules. Marcel Dekker, New York. 29. Harpel, M. R., and J. D. Lipscomb. 1990. Gentisate 1,2dioxygenase from Pseudomonas: purification, characterization, and comparison of the enzymes from P. testosteroni and P. acidovorans. J. Biol. Chem. 265:6301-6311. 30. Harpel, M. R., and J. D. Lipscomb. 1990. Gentisate 1,2dioxygenase from Pseudononas: substrate coordination to active site Fe2" and mechanism of turnover. J. Biol. Chem. 265:22187-22196. 31. Hendrich, M. P., E. Mfinck, B. G. Fox, and J. D. Lipscomb. 1990. Integer spin EPR studies of the fully reduced methane monooxygenase hydroxylase component. J. Am. Chem. Soc. 112:5861-5865. 32. Hori, K., T. Hashimoto, and M. Nozaki. 1973. Kinetic studies on the reaction mechanism of dioxygenases. J. Biochem. (Tokyo) 74:375-384. 33. Kent, T. A., E. Munck, J. W. Pyrz, J. Widom, and L. Que, Jr. 1987. Mossbauer and EPR spectroscopy of catechol 1,2-dioxygenase. Inorg. Chem. 26:1402-1408. 34. Kojima, Y., H. Fujisawa, A. Nakazawa, T. Kakazawa, F. Kanetsuna, H. Taniuci, M. Nozaki, and 0. Hayaishi. 1967. Studies on pyrocatechase. I. Purification and spectral properties. J. Biol. Chem. 242:3270-3278. 35. Kojima, Y., N. Itada, and 0. Hayaishi. 1961. Metapyrocatechase: a new catechol-cleaving enzyme. J. Biol. Chem. 236: 2223-2228. 36. Laemmli, U. K. 1970. Cleavage of structural proteins during the assembly of the head of bacteriophage T4. Nature (London) 227:680-685. 37. Lipscomb, J. D., and A. M. Orville. 1992. Mechanistic aspects of dihydroxybenzoate dioxygenases, p. 243-298. In H. Sigel and A. Sigel (ed.), Metal ions in biological systems, vol. 28. Marcel Dekker, New York. 38. Lumsden, J., R. Cammack, and D. 0. Hall. 1976. Purification and physicochemical properties of superoxide dismutase from two photosynthetic microorganisms. Biocbim. Biophys. Acta 438:380-392. 39. Lundell, D. J., and J. B. Howard. 1977. Isolation and partial characterization of two different subunits from the molybdenum iron protein of Azotobacter vinelandii nitrogenase. J. Biol. Chem. 253:3422-3426. 40. Mabrouk, P. A., A. M. Orville, J. D. Lipscomb, and E. I. Solomon. 1991. Variable temperature variable field magnetic circular dichroism studies of the Fe(II) active site in metapyrocatechase: implications for the molecular mechanism of extradiol dioxygenases. J. Am. Chem. Soc. 113:4053-4061. 41. Munck, E. 1978. Mossbauer spectroscopy of proteins: electron carriers. Methods Enzymol. 54:346-379. 42. Munck, E., H. Rhodes, W. H. Orme-Johnson, L. C. Davis, W. J. Brill, and V. K. Shah. 1975. Nitrogenase. VIII. Mossbauer and EPR spectroscopy. The MoFe protein component from Azotobacter vinelandii OP. Biochim. Biophys. Acta 400:32-53. 43. Nakai, C., K. Horiike, S. Kuramitsu, H. Kagamiyama, and M. Nozaki. 1990. Three isozymes of catechol 1,2-dioxygenase (pyrocatechase), aa, ap, and OP1, from Pseudomonas arvilla C-1. J. Biol. Chem. 265:660-665. 44. Nakai, C., H. Kagamiyama, M. Nozaki, T. Nakazawa, S. Inouye, Y. Ebina, and A. Nakazawa. 1983. Complete nucleotide sequence of the metapyrocatechase gene on the TOL plasmid of Pseudomonas putida mt-2. J. Biol. Chem. 258:2923-2928. 45. Noda, Y., S. Nishikawa, K. Shiozuka, H. Kadokura, H. Nakajima, K. Yoda, Y. Katayama, N. Morohoshi, T. Haraguchi, and M. Yamasaki. 1990. Molecular cloning of the protocatechuate 4,5-dioxygenase genes of Pseudomonas paucimobilis. J. Bacteriol. 172:2704-2709. 46. Nozaki, M. 1974. Nonheme iron dioxygenase, p. 135-165. In 0. Hayaishi (ed.), Molecular mechanisms of oxygen activation. Academic Press, New York. 47. Nozaki, M., H. Kagamiyama, and 0. Hayaishi. 1963. Metapyrocatechase. I. Crystallization and some properties. Biochem. Z. 338:582-590. Downloaded from http://jb.asm.org/ on January 16, 2018 by guest ment. J. Biol. Chem. 258:14981-14991. 6. Arciero, D. M., A. M. Orville, and J. D. Lipscomb. 1985. 7O-water and nitric oxide binding by protocatechuate 4,5 dioxygenase and catechol 2,3 dioxygenase. J. Biol. Chem. 260:14035-14044. 7. Bevington, P. R. 1969. Data reduction and error analysis for the physical sciences, p. 237-239. McGraw-Hill, New York. 8. Bradford, M. M. 1976. A rapid and sensitive method for the quantitation of microgram quantities of protein utilizing the principle of protein-dye binding. Anal. Biochem. 72:248-254. 9. Broderik, J. B., and T. V. O'Halloran. 1991. Overproduction, purification, and characterization of chlorocatechol dioxygenase, a nonheme iron dioxygenase with broad substrate tolerance. Biochemistry 30:7349-7358. 10. Bull, C., and D. P. Ballou. 1981. Purification and properties of protocatechuate 3,4-dioxygenase from Pseudomonas putida: a new iron to subunit stoichiometry. J. Biol. Chem. 256:1267312680. 11. Cantor, C. R., and P. R. Schimmel. 1980. Biophysical chemistry, p. 539-590. W. H. Freeman, San Francisco. 12. Chapman, P. 1972. An outline of reaction sequences used for the bacterial degradation of phenolic compounds, p. 17-55. In Degradation of synthetic organic molecules in the biosphere, proceedings of a conference. National Academy of Science, Washington, D.C. 13. Chervenka, C. H. 1969. A manual of methods for the analytical ultracentrifuge, p. 23-33. Beckman Instruments, Inc., Palo Alto, Calif. 14. Crawford, R. L. 1975. Novel pathway for degradation of protocatechuic acid in Bacillus species. J. Bacteriol. 121:531-536. 15. Crawford, R. L. 1976. Pathways of 4-hydroxybenzoate degradation among species of Bacillus. J. Bacteriol. 127:204-210. 16. Crawford, R. L., J. W. Bromley, and P. E. Perkins-Olson. 1979. Catabolism of protocatechuate by Bacillus macerans. Appl. Environ. Microbiol. 37:614-618. 17. Dagley, S. 1978. Pathways for the utilization of organic growth substrates, p. 305-388. In L. N. Ornston and J. R. Sokatch, (ed.), The bacteria, vol. 6. Academic Press, New York. 18. Dagley, S., P. J. Geary, and J. M. Wood. 1968. The metabolism of protocatechuate by Pseudomonas testosteroni. Biochem. J. 109:559-568. 19. Dagley, S., and M. D. Patel. 1957. Oxidation of p-cresol and related compounds by a Pseudomonas. Biochem. J. 66:227-233. 20. Debrunner, P. G., E. Munck, L. Que, and C. E. Schulz. 1977. Recent Mossbauer of some iron-sulfur proteins and model complexes, p. 381-417. In W. Lovenberg (ed.), Iron sulfur proteins, vol. 3. Academic Press, New York. 21. Durham, D. R., L. A. Stirling, L. N. Ornston, and J. J. Perry. 1980. Intergeneric evolutionary homology revealed by the study of protocatechuate 3,4 dioxygenase fromAzotobacter vinelandii. Biochemistry 19:149-155. 21a.Earhart, C., D. H. Ohlendorf, and J. D. Lipscomb. Unpublished results. 22. Edelhoch, H. 1967. Spectroscopic determination of tryptophan and tyrosine in proteins. Biochemistry 6:1948-1954. 23. Fox, B. G., W. A. Froland, J. Dege, and J. D. Lipscomb. 1989. Methane monooxygenase from Methylosinus trichospoium OB3b: purification and properties of a 3 component system with high specific activity from a type II methanotroph. J. Biol. Chem. 264:10023-10033. 24. Freifelder, D. 1982. Physical biochemistry, 2nd ed., p. 455-469. W. H. Freeman, San Francisco. 25. Fujisawa, H., and 0. Hayaishi. 1968. Protocatechuate 3,4dioxygenase. I. Crystallization and characterization. J. Biol. Chem. 243:2673-2681. 26. Fujisawa, H., M. Uyeda, Y. Kojima, M. Nozaki, and 0. Hayaishi. 1972. Protocatechuate 3,4-dioxygenase. II. Electron spin resonance and spectral studies on interactions of substrates and enzyme. J. Biol. Chem. 247:4414-4421. 27. Fujiwara, H., and Nozaki, M. 1973. Protocatechuate 3,4 dioxygenase. IV. Preparation and properties of apo- and reconstituted enzymes. Biochim. Biophys. Acta 327:306-312. 28. Gibson, D. T., and V. Subramanian. 1984. Microbial degrada- 4425 4426 J. BACTERIOL. WOLGEL ET AL. 268:8596-8607. 58. Que, L., Jr., J. D. Lipscomb, E. Munck, and J. M. Wood. 1977. Protocatechuate 3,4 dioxygenase inhibition studies and mechanistic implications. Biochim. Biophys. Acta 485:60-74. 59. Que, L., Jr., J. D. Lipscomb, R. Zimmermann, E. Munck, N. R. 60. 61. 62. 63. 64. 65. 66. 67. 68. 69. 70. 71. Orme-Johnson, and W. H. Orme-Johnson. 1976. Mossbauer and EPR spectroscopy on protocatechuate 3,4 dioxygenase from Pseudomonas aeruginosa. Biochim. Biophys. Acta 452:320334. Que, L., Jr., J. Widom, and R. L. Crawford. 1981. 3,4Dihydroxyphenylacetate 2,3-dioxygenase: a manganese (II) dioxygenase from Bacillus brevis. J. Biol. Chem. 256:1094110944. Salerno, J. C., and J. N. Siedow. 1979. The nature of the nitric oxide complexes of lipoxygenase. Biochim. Biophys. Acta 579:246-251. Satyshur, K., S. T. Rao, J. D. Lipscomb, and J. M. Wood. 1980. Preliminary crystallographic study of protocatechuate 3,4-dioxygenase from Pseudomonas aeruginosa. J. Biol. Chem. 255: 10015-10016. Segel, I. H. 1975. Enzyme kinetics, p. 544-560. Wiley, New York. Smith, T. D., and J. R. Pilbrow. 1980. ESR of iron proteins, p. 85-168. In L. J. Berliner and J. Reuben (ed.), Biological magnetic resonance, vol. 2. Plenum Press, New York. Stanier, R. Y., and J. L. Ingraham. 1954. Protocatechuate acid oxidase. J. Biol. Chem. 210:799-808. Stanier, R. Y., N. J. Palleroni, and M. Doudoroff. 1966. The aerobic pseudomonads: a taxonomic study. J. Gen. Microbiol. 43:159-271. Tatsuno, Y., Y. Saeki, M. Nozaki, S. Otsuka, and Y. Maeda. 1980. Mossbauer spectra of metapyrocatechase. FEBS Lett. 112:83-85. Weisiger, R. A., and I. Fridovich. 1973. Superoxide dismutase: organelle specificity. J. Biol. Chem. 248:3582-3592. Whittaker, J. W., and J. D. Lipscomb. 1984. 70-water and cyanide ligation by the active site iron of protocatechuate 3,4 dioxygenase: evidence for displaceable ligands in the native enzyme and in complexes with inhibitors or transition state analogs. J. Biol. Chem. 259:4487-4495. Whittaker, J. W., J. D. Lipscomb, T. A. Kent, and E. Munck. 1984. Brevibacterium fuscum protocatechuate 3,4-dioxygenase: purification, crystallization and characterization. J. Biol. Chem. 259:4466-4475. Wolgel, S. A., and J. D. Lipscomb. 1990. Protocatechuate 2,3-dioxygenase. Methods Enzymol. 188:95-101. Downloaded from http://jb.asm.org/ on January 16, 2018 by guest 48. Nozaki, M., S. Kotani, K. Ono, and S. Senoh. 1970. Metapyrocatechase. III. Substrate specificity and mode of ring fission. Biochim. Biophys. Acta 220:213-223. 49. Nozaki, M., K. Ono, T. Nakazawa, S. Kotani, and 0. Hayaishi. 1968. Metapyrocatechase. II. The role of iron and sulfhydryl groups. J. Biol. Chem. 243:2682-2690. 50. Ohlendorf, D. H., J. D. Lipscomb, and P. C. Weber. 1988. Structure and assembly of protocatechuate 3,4-dioxygenase. Nature (London) 336:403-405. 51. Ohlendorf, D. H., P. C. Weber, and J. D. Lipscomb. 1987. Determination of the quaternary structure of protocatechuate 3,4-dioxygenase from Pseudomonas aeruginosa. J. Mol. Biol. 195:225-227. 52. Olson, P. E., B. Qi, L. Que, Jr., and L. P. Wackett. 1992. Immunological demonstration of a unique 3,4 dihydroxyphenylacetate 2,3-dioxygenase in soil Arthrobacter strains. Appl. Environ. Microbiol. 58:2820-2826. 53. Ono, K., M. Nozaki, and 0. Hayaishi. 1970. Purification and some properties of protocatechuate 4,5-dioxygenase. Biochim. Biophys. Acta 220:224-238. 54. Ono-Kaminoto, M. 1973. Studies on 3,4-dihydroxyphenylacetate 2,3-dioxygenase. I. Role of iron, substrate binding and some other properties. J. Biochem. (Tokyo). 74:1049-1059. 55. Orville, A. M., V. J. Chen, A. Kriauciunas, M. R. Harpel, B. G. Fox, E. Munck, and J. D. Lipscomb. 1992. Thiolate ligation of the active site Fe2+ of isopenicillin N synthase derives from the substrate rather than endogenous cysteine: spectroscopic studies of site specific cys-*ser mutated enzymes. Biochemistry 31:4602-4612. 55a.Orville, A. M., and J. D. Lipscomb. Unpublished observation. 56. Orville, A. M., and J. D. Lipscomb. 1989. Binding of isotopically labeled substrates, inhibitors and cyanide by protocatechuate 3,4-dioxygenase. J. Biol. Chem. 264:8791-8801. 57. Orville, A. M., and J. D. Lipscomb. 1993. Simultaneous binding of nitric oxide and isotopically labeled substrates or inhibitors by reduced protocatechuate 3,4-dioxygenase. J. Biol. Chem.