Rapid Reaction Studies on the Oxygenation Reactions of Catechol Dioxygenase*

The reaction of oxygen with catechol 1,2-dioxygen- ase from Pseudomonm arvilla ATCC 23974 in complex with catechol, 4-methylcatechol, and 4-flUOr0-catechol has been studied using single turnover stopped flow spectrophotometry. Two sequential enzyme intermediates have been resolved and their visible spectra characterized by computer-assisted methods. These intermediates are spectrally similar to those observed in a similar study with protocatechuate dioxygenase (Bull, C., Ballou, D. P., and Otsuka, S. J. Biol. Chem. 256, 12681-12686 (1981), although the first inter- mediate seen with the latter enzyme was not observed in this study. The rate of formation of intermediate I is oxygen-dependent and also accelerated by electron- donating substituents on the C-4 of the substrate. This is consistent with the proposed substrate reduction of dioxygen to form a hydroperoxide. Intermediate I is thus suggested to be a 6-hydroperoxycyclohexa-3,5- diene- 1-one. The decay of intermediate I is also accelerated by electron donors and is consistent with the rearrangement of intermediate hydroperoxide via an acyl migration mechanism. It is inconsistent with mechanisms involving nucleophilic attack at the carbonyl carbon. Intermediate I1 is proposed to be an enzyme-product complex based on the resemblance of its visible spectra to those of the benzoate complex of catechol 1,2-dioxygenase

opened products by non-heme iron containing catechol dioxygenases. Catechol 1,2-dioxygenase from Pseudomonas arvilla ATCC 23974 is one such enzyme, and is a member of the intradiol-cleaving group of catechol dioxygenases that typically contains high spin Fef3 in the active site (Nozaki, 1978;. These enzymes are burgundy-red in color due to a visible absorbance band a t -460 nm (t -3000-4000 M-' cm" Fe-') arising from tyrosinate to iron charge transfer interactions (Keyes et al., 1978;Tatsuno et al., 1978;Felton et al., 1978;Bull et al., 1979;Que and Heistand, 1979). The spectra of these enzymes can be characteristically perturbed by the anaerobic addition of substrates, or by substrate and product analogues Que and Epstein, 1981;Bull and Ballou, 1981;Walsh and Ballou, 1983).
Current ideas for the mechanism of catalysis by the Fe+3containing catechol dioxygenases focus around the observation that no Fe+2 species has yet been detected in the catalytic cycle. Thus, it has been proposed that, rather than the more classical activation of oxygen by Fe+', the iron acts essentially as a Lewis acid activating the substrate towards O2 attack and producing a peroxy adduct of the substrate (Hamilton, 1974;Que et al., 1977). This adduct may then rearrange to form product, conserving the two atoms of molecular oxygen in the resulting dicarboxylate. Rapid kinetic studies by Bull et al. (1981) on protocatechuate dioxygenase in which the enzyme-catalyzed oxygenation of the catechol, protocatechuate, was monitored, revealed two sequential enzyme intermediates in the reaction. It is not known to which chemical species in the proposed scheme these intermediates should be attributed, as the extremely transient nature of the intermediates has prevented any further characterization beyond that of their visible absorption spectra.
In this study, we have investigated the reaction of oxygen with catechol dioxygenase from P. arvilla in complex with the native substrate, catechol, and the two substrate analogues, 4-methylcatechol and 4-fluorocatechol. These analogues have relative turnover rates that are 76 and 19%, respectively, of that using catechol as substrate. Using stopped flow spectrophotometry, we have detected two sequential enzyme intermediates in single turnover reactions with each of these substrates. The rates of formation and decay of these intermediates vary according to the nature of the C-4 substituent of the substrate, enabling correlations to be made between the spectral forms and their possible chemical structures, in light of the current mechanistic proposals. Comparisons of these intermediates to those observed in the reaction of protocatechuate dioxygenase with protocatechuate and oxygen leads to tentative assignments of chemical structures to the species. The use of these substrates also enables a more reliable determination of the sequence in which these intermediates are formed.

Enzyme-Substrate Complexes-Anaerobic
titrations of catechol dioxygenase with either catechol, 4-methylcatechol, or 4-fluorocatechol produce similar changes in the spectrum of the enzyme that result in the formation of a broad absorbance band centered around 700 nm (Fig. 1, a and b; the spectrum of the enzyme in complex with 4-fluorocatechol is almost identical with that seen with the native substrate, and is not shown). These are typical of the spectral changes associated with the formation of ES2 in the non-heme, ferriccontaining dioxygenases, and imply that all three compounds bind to catechol dioxygenase in a similar way. The Kd values derived from these titrations are also similar and are shown in Table I. Small differences in the final ES spectrum using 4-methylcatechol to that using the native substrate and 4fluorocatechol can be noted in the 460-500 nm region where titration with 4-methylcatechol produces almost no change in the enzyme spectrum, whereas catechol and 4-fluorocatechol produce a decrease in absorbance and hence a distinct isosbestic point at 540 nm.
Attempts at measuring the rate of formation of the enzymecatechol complex were unsuccessful as the reaction was complete within the dead time of the stopped flow apparatus (-2.5 ms), even using concentrations of substrate approaching stoichiometry with that of enzyme ("40 p~) .
This implies that the second order rate constant for ES formation is greater than 10' M-' s-'. However, using 4-methylcatechol, the reaction was somewhat slower and hence more tractable. A plot of the observed rate of ES formation (as monitored at 700 or 440 nm) uersus concentration of 4-methylcatechol was linear and passed through the origin; the slope of the line yielded a second order rate constant of 1.1 X lo6 M" s-'. This lower rate could reflect some steric hindrance to binding due to the methyl group at the 4-position. It is interesting to note that both pyrogallol and 3-methyoxycatechol behave like catechol in that the rate of ES formation with these substrate analogues is too fast to monitor in the stopped flow apparatus. Thus, compared to the 4-position, the 3-position seems relatively unhindered.
Reaction of ES with Oxygen-Single enzyme turnover events using catechol as substrate were monitored at wavelengths between 320 and 720 nm in the stopped flow apparatus, as described under "Materials and Methods." A selection of reaction traces recorded at various wavelengths is shown in Fig. 2a. A total of three kinetic phases can be resolved, corresponding to the formation and decay of two intermediates in the reaction. This is best seen at 620 nm as at this wavelength, the absorbance changes due to each phase are opposed. The decay of the second intermediate, i.e. the final phase, results in the formation of resting enzyme and free product. The rates of the three phases were determined to be 240, 36, and 19 s" at an oxygen concentration of 0.96 mM ( Table I). The rate of the first phase was linearly dependent on the oxygen concentration, whereas the following two rates were independent of the oxygen concentration. A pseudo-first order plot of the observed rate of the first phase uersus oxygen concentration showed the reaction to be essentially irreversible, and yielded a second order rate constant of 2.5 X lo5 M"

S-1.
The turnover number of the enzyme with catechol as substrate can be calculated from the rates determined in the stopped flow experiments by the following relationship: llturnover number = 11121 + l/kz + l / k S At 0.24 mM oxygen, where the observed rate of kl is 60 s-', the turnover number from the rapid kinetics data can be calculated to be 10.3 s-'. This compares favorably with the value of 11.0 s" determined under the same conditions by steady state kinetics techniques using catalytic amounts of enzyme and saturating amounts of catechol.
Using the computer analysis techniques described under "Materials and Methods," it is possible to obtain the extinction coefficients of the two intermediates in the reaction from the kinetic time courses taken at 20-nm intervals in the 320-720 nm region. Computer-generated time courses are shown overlaying the data collected in the stopped flow experiments in Fig. 2. The fit is excellent at all wavelengths, indicating that the determined rates and extinctions are reasonable. In computer simulations of a given kinetic record, it was found that adjustment of more than 5 1 0 % of either the rate constants or the extinctions determined by the fitting routine gave poorer fits. Another useful test of both the kinetic scheme, and the extinction values and first order rate constants derived from the analysis of data obtained using 1 mM oxygen was the following. Experimental data recorded at different oxygen concentrations (0.3 and 0.45 mM) were simulated using the same scheme and values determined previously. The oxygen addition step, kl, was altered to account for the difference in oxygen concentration. Excellent fits were obtained, providing confidence that the scheme and the rate constants are reasonable. Fig. 3a shows the resulting spectra of the two intermediates. The spectrum of the first intermediate lacks the absorbance band centered at 700 nm typical of the ES and also shows a significant decrease in extinction at 460 nm, to produce a low shoulder at 480 nm. Note, however, the large increase in extinction a t lower wavelengths (340-380 nm). The second intermediate is characterized by a poorly resolved peak at 480 nm (c = 3000 M" cm") and an increased extinction (over that of resting enzyme) between 580 and 700 nm. However, this absorbance does not take the form of a distinct new band as in the ES.

E + S +ES
Dissociation constants were determined by anaerobic spectral titration of the enzyme with the appropriate compound. The values for k,, were determined by anaerobically mixing the enzyme with the substrate in the stopped flow apparatus. The values of k,, kz, and kS were obtained by analysis of kinetic records following the reaction of oxygen with the enzyme in complex with substrate in the stopped flow apparatus. Experimental details are in the text. The turnover numbers calculated from the rapid kinetic data were derived from the following relationship: l/turnover number = l / h + l / k 2 + l / k , The value for k, was taken to be that at 21% oxygen saturation. The turnover numbers obtained from steady state kinetic data were derived from Lineweaver-Burk plots using catalytic amounts of enzyme ( -1 PM) and at 21% oxygen saturation. Oxygen equilibration was performed at 25 "C.  The smooth curues represent computer-generated simulations of the data using the scheme, rate constants, and extinction values as described in the text. In many cases, the fit is so close that the experimental data and the simulated curves are almost indistinguishable. It should be noted that the experimental data consist of four sets of 100 points and that the time scale over which each of these sets of points is collected is user-determined so that in the records above, for example, the first set was collected over the first 40 ms of the reaction. Thus, accurate data are obtained for every phase of the reaction so that accurate determinations of the observed rate constants for each of the processes can be made from a single data record.
It was of interest to compare both the kinetics and spectra of the intermediates seen in the oxygenation of substrate analogs with those of the native substrate, in order to gauge the effect of ring substituents on the reaction. The reaction of oxygen with the 4-methylcatechol complex of the enzyme was qualitatively similar to that using catechol in that triphasic kinetics were again observed. However, the rates of the first two processes were significantly faster than those with the native substrate, while the final decay to resting enzyme and free product was marginally slower (Table I). Computer simulations of the reaction curves using extinctions for the two intermediates as shown in Fig. 36, and the appropriate rate constants, gave excellent fits (Fig. 26).
Intermediate I with 4-methylcatechol as substrate has a pronounced shoulder at 480 nm ( t = 2600 M" cm") and intermediate I1 has a distinct peak at 500 nm ( t = 2600 M" cm"). Thus, both intermediates have somewhat more resolved spec-ra than those found using catechol as substrate. The analysis of the data from the experiments using 4methylcatechol is somewhat more facile since the rates of the second and third kinetic phases are separated by a factor of 4.4, as compared to only 1.9 in the catechol case. However, in both cases, the closeness of fit of the simulated curves to the experimental data makes us confident that the spectra of the intermediates are accurate.
The kinetic course of the reaction of oxygen with the Wavelength, nrn Wavelength, nm Wavelength, nm enzyme in complex with 4-fluorocatechol was also triphasic (Fig. 2c). This is rather less apparent than with the other two substrates as there is no wavelength where all three phases are distinctly opposed. However, careful analysis of the reaction traces (in the 360-440 nm region especially) shows a biphasic return to resting enzyme and free product after the initial bimolecular reaction with oxygen. The rates of the phases are shown in Table I. With the halogenated compound, all three rates are slower than with the native substrate. The oxygen reaction is slowed by "25%, but the subsequent two reactions are very much more affected. Computer simulation of the reaction traces using the extinctions of the intermediates shown in Fig. 3c and the appropriate rate constants again gave excellent fits (Fig. 2c). Intermediate I in this case is almost identical to that found with catechol as substrate although intermediate I1 is rather different, particularly in having a similar extinction to resting enzyme in the 600 nm region and above (thus explaining the lack of wavelengths where all three phases are opposed).
Previous "0-labeling studies (Hayaishi et al., 1957) have shown that both atoms of molecular oxygen are incorporated into the product. This implies that little or no exchange of label occurs during catalysis. Since mechanistic interpretations rely heavily on this observation, we undertook careful "02-labeling experiments to determine if under some conditions any exchange of label with solvent could occur. It was expected that the conversion of I to I1 would be the most likely step in which oxygen exchange might occur. A larger difference in the rates of formation and decay of intermediate I was observed for 4-fluorocatechol than for catechol. Therefore, we reasoned that this halogenated substrate would afford a better opportunity for observing exchange of labeled oxygen with the solvent. However, "02-labeling experiments with both catechol and 4-fluorocatechol showed complete retention of both atoms of in the products (>99%). This is consistent with earlier studies using catechol (Hayaishi et al., 1957).

DISCUSSION
These studies show that catechol, 4-methylcatechol, and 4fluorocatechol are oxygenated by catechol 1,2-dioxygenase via a common mechanism. The ES for the substrates are similar, displaying a long wavelength absorption band arising from a catecholate to iron charge transfer interaction (Fig. 1) (Felton et al. 1978;Que and Heistand, 1979). Moreover, in rapid mixing experiments two comparable sequential intermediates are seen when each of the ES complexes is mixed with oxygen. The rapid oxygen-dependent disappearance of the catecholate charge transfer band concomitant with the formation of intermediate I strongly suggests that the substrate has become oxygenated in the process. Studies with the related enzyme, protocatechuate dioxygenase, have also shown a similar re-action scheme in which ES reacts with oxygen to form sequentially, ESO,, ES02*, and E + P (Bull et al., 1981).
However, the first intermediate seen with catechol 1,2-dioxygenase is spectrally more similar to the second intermediate observed with protocatechuate dioxygenase (ES02*) than to the first (ESO,).
We have no clear evidence that a species analogous to ESO, occurs in the present study although the mechanisms of the two enzymes appear to be very similar in all other respects. ESO, may not be observable in the scheme for catechol 1,2dioxygenase because ESO, is transformed into ES02* (intermediate I) as fast as it is formed. Given that intermediate I is formed at 240 s-', the conversion of ESO, to ES02* would have to be >lo00 s" to satisfy this hypothesis (This rate is 450 s-' for protocatechuate dioxygenase.) With 4-methylcat-echo1 as substrate, for which the rate of oxygen addition is significantly greater than for the other two substrates (making kinetic resolution more favorable), the spectrum of intermediate I has higher absorbance at "440 nm and lower absorbance at "550 nm (Fig. 2) than those for the other two substrates. This type of spectral difference could be accounted for by a slight contribution of an ESO, species as seen with protocatechuate dioxygenase. Alternatively, the variations in the spectra for intermediate I could arise from substituent effects rather than from unresolved kinetic processes.
The spectra of intermediate I1 obtained with the three substrates are characterized by a shift of the absorption maxima to "490 nm and show some variability with the different reactions. The spectrum of this intermediate very strongly resembles that of the benzoate complex of catechol dioxygenase . Also, protocatechuate dioxygenase forms complexes with &carboxylates, particularly at pH values below 7 (Que and Epstein, 1981;Ballou and Bull, 1978). Its natural product, P-carboxy-cis,cis-muconate, as well as the product analogs, 6-carboxy-cis,trans-muconate, terephthalate, and glutarate form complexes with spectra which are similarly red-shifted, but have variable extinctions and Amax values. Therefore, we believe that intermediate I1 is an EP. Unfortunately, catechol 1,2-dioxygenase has such a low affinity for product that we have been unsuccessful in generating an EP for comparison (even with concentrations of cis,&-muconate up to 5 mM).
No reduced iron species has as yet been detected in the catalytic cycle of any of the catechol dioxygenases. Therefore, current mechanistic proposals for the reaction catalyzed by the intradiol catechol dioxygenases involve Fe+3-assisted activation of the substrate to electrophilic attack by oxygen. Fraser and Hamilton (1982) and Jefford and Cadby (1981) have proposed that essentially any reasonable mechanism for this reaction has, as a first step, the formation of an ahydroperoxy ketone. It is proposed that the intermediate formed is the 6-hydroperoxycyclohexa-3,5-dien-l-one (species Our results show how substituents in the 4-position of catechols affect the rates of formation and decay of intermediate I. This allows us to examine postulated mechanisms for consistency with the results using the enzyme. We find that both the first and second processes in the reaction of oxygen with catechol 1,a-dioxygenase ES are accelerated by electron donation. The rate of the first process is dependent on oxygen concentration and thus would be expected to be enhanced by substituents which increase electron density on the substrate and make it more susceptible to electrophilic attack by oxygen. Indeed, Walsh and Ballou (1983) have shown that, in the reaction of oxygen with the protocatechuate dioxygenase-6chloroprotocatechuate complex, the bimolecular step involving the initial attack on oxygen is %-fold slower than that for the natural ES, again demonstrating that the presence of an electron-withdrawing group retards the oxygen reaction (Walsh and Ballou, 1983).
The assignnent of intermediate I to the hydroperoxy species described in Scheme 1 is an attractive hypothesis and its formation would be favored by electron donors. Considerable work on the mechanism of rearrangement of hydroperoxy species such as these has recently been reported. The work reported here is the first that directly bears on this mechanism for an enzyme-catalyzed process.
Three mechanisms for such rearrangements which are consistent with "0-labeling experiments are shown in Scheme 2. In pathway 1, the rearrangement is initiated by a nucleophilic attack on the carbonyl carbon by some enzyme-bound base or by an external nucleophile. Sawaki and Ogata (1975) have shown that this mechanism applies to the base-catalyzed rearrangement of hydroperoxy ketones in which the reactions are accelerated by electron-withdrawing groups. The cleavage of ribulose bisphosphate to phosphoglycolate and 3-phosphoglycerate catalyzed by ribulose bisphosphate carboxylase-oxygenase is suggested to proceed by this mechanism in which 1 atom of molecular oxygen is incorporated into the products while the other is released into the solvent water (Lorimer, 1981). Pathway 2 involves a dioxetane intermediate and has been shown to be unimportant in a wide variety of rearrangements of this type (Muto and Bruice, 1980;Jefford et al., 1978;Ogata, 1975, 1978;Fraser and Hamilton, 1982). Pathway 3 proceeds via acyl migration and has been shown to apply to the acid-catalyzed rearrangement of hydroperoxy ketones. These reactions are accelerated by electrondonating substituents, which stabilize the incipient carbonium ion (Sawaki and Ogata, 1978).
In our studies we have found that the decay of intermediate I is accelerated by electron donors. The data are inconsistent with either pathway 1 or 2, since both mechanisms involve nucleophilic attacks on the acyl carbon, which would be expected to be enhanced by electron-withdrawing substituents. It is possible, however, that a nucleophilic attack by an enzymic group is not a rate-limiting step in the rearrangement so that the appropriate substituent effect would not be observed. The data are consistent, however, with the rearrangement of an intermediate a-hydroperoxyketone via acyl migration (pathway 3). In the proposed mechanism, the Fe+3 in the active site may serve as a Lewis acid both in the activation of the substrate toward oxygen and in the subsequent Criegee (1948) rearrangement of the peroxide to form products. The ' ' Oz studies imply that these interesting acid-catalyzed rearrangements occur extremely efficiently since no exchange with solvent occurs.